Anatomic repair of transposition of great arteries with ventricular septal defect and aortic arch obstruction

Surgery for Congenital Heart Disease

Anatomic repair of transposition of great arteries with ventricular septal defect and aortic arch obstruction One-stage versus two-stage procedure Between September 1, 1982, and March 1, 1992, 40 patients underwent anatomic repair of transposition of the great arteries, ventricular septal defect, and aortic arch obstruction. In group I, 26 patients (65 %) underwent repair in a two-stage procedure, phases A and B. Phase A included repair of the aortic arch obstruction with (16 patients) or without (10 patients) pulmonary artery banding through a left thoracotomy (mean age 18.7 ± 23.4 days). There were three deaths and three reoperations. Phase B included an arterial switch operation with closure of the ventricular septal defect (mean age 95.5 ± 122 days). There were five early deaths and two late deaths. Eight patients required reoperation. Mean delay between phase A and phase B was 77.5 ± 109 days. In group I, there were eight early and two late deaths, and 11 patients required reoperation. The mean stay in the intensive care unit was 24.7 ± 20 days. Mean follow-up of 59.6 ± 21.4 months was completed in all survivors. All but one were in New York Heart Association class I without medication. Actuarial survival rate and rate of freedom from reoperation at 5 years were 57.5% and 49.9%, respectively. In group II, 14 patients (35 %) had a one-stage procedure through midsternotomy: an arterial switch operation with closure of the ventricular septal defect and repair of the aortic arch obstruction (mean age 10.2 ± 5.5 days). There were two early deaths (14.2%) and one late death after reoperation for overlooked multiple ventricular septal defects. Two patients required reoperation. The mean stay in the intensive care unit was 11.7 ± 2.5 days. Mean follow-up of 22.4 ± 16.7 months was achieved in all survivors. They were all in New York Heart Association class I without medication. Actuarial survival rate and rate of freedom from reoperation at 3 years were 78.5% and 81.5%, respectively. The one-stage procedure allowed complete repair in neonates without the need for multiple operations. We believe that it may decrease early mortality rates (14.2% versus 30.7%), reduce the reoperation rate and cumulative stay in the intensive care unit (11.7 days versus 24.7 days, p = Not significant), and significantly decrease the overall rate of morbidity (p < 0.01). However, requirements for surgical intervention with a one-stage or a two-stage procedure must include accurate assessments of intracardiac and aortic arch anatomy. (J THORAC CARDIOVASC SURG 1993;105:925-33)

Claude Planche, MD (by invitation), Alain Serraf, MD (by invitation), Juan V. Comas, MD (by invitation), Francois Lacour-Gayet, MD (by invitation), Jacqueline Bruniaux, MD (by invitation), and Anita Touchot, MD (by invitation), Paris, France Sponsored by John W. Kirlin, MD, Birmingham, Ala.

From Marie-Lannelongue Hospital, Universite Paris-Sud, France. Address for reprints: Claude Planche, MD, Department of Pediatric Surgery. Marie-Lannelongue Hospital, 133 Avenue de la Resistance, 92350 Le Plessis-Robinson, France.

Read at the Seventy-second Annual Meeting of The American Association for Thoracic Surgery, Los Angeles, Calif., April 26-29, 1992. Copyright

1993 by Mosby-Year Book, Inc.

0022-5223/93 $1.00+ .10

12/6/43346

925

9 2 6 Planche et al.

Transposition of the great arteries (TGA) with ventricular septal defect (VSD) and aortic arch obstruction (AAO) is a rare malformation; its prevalence, anatomy, and physiopathology have been widely described. 1,5 The surgical treatment of this highly complex form of TGA is still a therapeutic challenge. Symptoms generally appear soon after birth, and surgery must be performed in critically ill neonates. However, with the introduction of prostaglandin E 1 and the atrial septostomy maneuver, the infants are brought to the operating room in a more stable condition and surgical intervention can be achieved as a semielective procedure. The surgical treatment has not yet been well established. The two-stage procedure includes early repair of the AAO, with and without pulmonary artery banding, through a left thoracotomy followed, in the second stage, by an arterial switch operation with closure of the VSD. However, this approach involves risks during the first stage of the procedure and during the intermediary period before the complete repair. Morever, pulmonary artery banding may worsen the almost invariable anterior malalignment of the infundibular septum with subsequent subaortic obstruction. Complete repair in a one-stage operation through midsternotomy during the neonatal period seems to be an attractive alternative because it may avoid the disadvantages of a two-stage procedure. We report our total experience in anatomic repair of TGA, VSD, and AAO with these two approaches, with particular emphasis on preoperative, perioperative, and postoperative features.

Patients and methods Forty patients underwent anatomic repair ofTGA, VSD, and AAO at Marie Lannelongue Hospital. According to the therapeutic option, two different groups were defined: Patients in group I underwent repair in a two-stage procedure, phases A and B; patients in group II underwent a one-stage procedure (Table I). Group I. Group I consisted of 26 patients (65%), most of whom underwent surgical intervention between 1982 and 1987. Only 3 patients were included in this protocol between 1987 and 1992.

In all patients, diagnosis was made on the basis of cardiac catheterization with cineangiographic studies and echocardiographic investigations. A balloon atrial septostomy was performed in 19 patients (73%), and 12 of those patients (46.1%) also underwent infusion of prostaglandin E, (0.0025 /-Lg/kg per minute) to improve oxygen saturation in the systemic circulation and to maintain stable hemodynamic balance. Six infants (23%) were supported preoperatively with artificial ventilation and inotropic drugs. Extracardiac anatomy. Twenty-three patients had D-TGA, 1 had L-TGA, and 2 had the Taussig-Bing malformation. The great vesselswere in a strictly anteroposterior position or there was a mild displacement of the aorta to the right in 24 patients

The Journal of Thoracic and Cardiovascular Surgery May 1993

(92.3%), and the great vessels were side by side in 2 patients (7.6%).

The diameters of the aorta and the pulmonary trunk were similar in 8 patients (30.75%, grade 1). The ratio of the pulmonary artery to the aorta was greater than 1.5:1 in 8 patients (30.75%, grade 2). The pulmonary artery was at least twicethe size of the aorta in the other 10 patients (38.5%, grade 3).6 Four different AAOs were encountered? (Fig. I). In each case, an isthmic coarctation was associated with either a hypoplasia of the distal transverse arch between the left carotid and the left subclavian arteries in 18 patients (69.2%) or a hypoplasia of the proximal and distal transverse arch, between the innominate and the left subclavian arteries in 6 patients (23.1 %). Also, examination of the AAO revealed a localized isthmic coarctation in 1 patient (3.8%) and an interrupted aortic arch (IAA)-type A of the Celoria and Patton" classification-in 1 patient (3.8%). The ductus arteriosus was widely patent in II patients (42.3%), of moderately patent in 12 patients (46.15%), and nearly closed in 2 patients (7.7%). According to the classification of Yacoub and RadleySmith.? coronary artery distribution was type A in 14 patients (53.8%), type D in 6 patients (23.1 %), and type E in 6 patients (23.1%). Intracardiac anatomy. The location of the VSD was perimembranous in 16 patients, trabecular in 2 patients (7.7%),

infundibular in 6 patients, and subpulmonary in 2 patients. One patient had multiple VSDs. The infundibular septum was anteriorly displaced in 12 cases (46.1%). In addition, 4 patients (15.3%) had an abnormal parietal band. A critical subaortic stenosis was present in 5 cases (19.2%). Four patients (15.4%) had moderate right ventricular hypoplasia,and 2 (7.7%) showed aberrant tricuspid chords. One patient (3.8%) had situs inversus, and I had transposed ventricles. Surgical treatment during phase A. There were 21 male and 5 female patients. Mean age at operation was 18.7 ± 23.4 days, mean weight was 3.4 ± 0.6 kg, and mean body surface area was 0.20 ± 0.02 m-, For the repair of AAO, an extended end-to-end anastomosis through a left thoracotomy was always performed, even for the patient with lAA. Associated pulmonary artery banding was performed in 15 patients (60%). Surgical treatment during the interval between phases A and B. The mean delay between the two phases was 77.5 ± 109 days. During this period three patients died and three patients required a reoperation. Surgical treatment during phase B. Twenty male and three female patients remained. Mean age was 95.5 ± 122 days, mean weight was 4.3 ± 0.8 kg, and mean body surface area was 0.24 ± 0.05 m', All of the patients had bicaval cannulation, continuous cardiopulmonary bypass at 20° C, and no circulatory arrest. Myocardial protection was achieved by blood cardioplegia, which was maintained by administering cardioplegic solution every 30 minutes. After closure of the atrial septal defect, the VSD was closed with a Dacron patch by various methods. We always attempted to use the approach that involved the right atrium. This approach allowed closure of the VSD in only 6 patients (26%), mostly involvingperimembranous VSD without malalignment. Most (10 cases, 43.4%) were closedvia the right ventricle and some (7 cases, 30.4%) were closed via the pulmonary artery. In most cases, after closure of the VSD the right ventricle was closed without the interposition of a patch. A parietal band resection was necessary in four cases that

The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5

Planche et at.

927

Table I. Patients Procedure

No,

Two-stage Phase A Phase B One-stage

26

23 14

Age (days)

Weight

21/5

18.7 ± 23

20/3

95.5 ± 122 10.2 ± 5.5

3.4 ± 0.6 4.3 ± 0.8 3.1 ± 0.3

M/F

10/4

(kg)

0.2 ± 0.02 0.24 ± 0.05 0.19 ± 0.01

M/F, Male/female ratio: BSA, body surface area.

Table II. Comparison of overall morbidity between the two groups Procedure

Early mortality

Late mortality

Two-stage Phase A Phase B One-stage

3 (11.5%) 5 (21.7%) 2 (14.2%)

o 2

1

Reoperation 3 8*

2*

Overall morbidity

20/26 (76.9%) 4/14 (28.6%)t

"One late death. tp < 0.01.

involved the right ventricle (17.3%). The arterial switch was then performed according to the standard technique.!" The Lecompte maneuver was performed in all cases, including those with Taussig-Bing malformation. For completion of the aortic anastomosis, the anterior wall of the distal aorta was vertically incised if the diameter of the pulmonary artery was much larger than the aortic diameter. Mean aortic crossclamp time was 96.4 ± 16.5 minutes, and the mean period of cardiopulmonary bypass was 198 ± 28.7 minutes. Delayed sternal closure was necessary in three patients, with an average delay time of 2.3 ± 0.5 days. Cumulative (for phases A and B) mean intubation time and time in an intensive care unit were 10.1 ± 5.7 days and 23.7 ± 20.2 days, respectively. Group II. Group II consisted of 14 patients (35%) who, from 1987 to 1992, underwent surgical treatment with a one-stage procedure through a median sternotomy: arterial switch operation with closure of VSD and repair of AAO. All patients underwent diagnostic procedures in the previously mentioned explorations and received prostaglandin E[ infusion. Eleven (78.6%) required a balloon atrial septostomy. Two infants (14.3%) were supported with artificial ventilation. Extracardiac anatomy. All patients had D-TGA with the vessels in a strictly anteroposterior position. The diameters of the aorta and pulmonary trunks were grade 1 in 2 patients (14.3%), grade 2 in 8 patients (57.1 %), and grade 3 in the other 4 patients (28.6%). Pathologic study of the AAO revealed coarctation and hypoplasia of the distal transverse arch in 6 patients (42%). Six patients (42.9%) also had coarctation and hypoplasia of the proximal and distal transverse arch. One (7.1 %) had a localized isthmic stenosis and one (7.1%) had IAA type A. The ductus arteriosus was systemic in 10 patients (71.4%) and of moderate size in 4 patients (28.6%). According to the classification of Yacoub and Radley-Smith, coronary artery distribution was type A in 10 patients (71.9%), type C in I patient (7.1%), and type D in 3 patients (21.4%). Intracardiac anatomy. The VSD was perimembranous in 8 patients (57.1%), infundibular in 2 patients (14.2%), subpulmonary in 3 patients (21.4%), and multiple in I patient (7.1 %). An anterior malalignment of the infundibular septum was

present in II cases (78.5%). Two patients (14.2%) had an abnormal parietal band, and a significant subaortic stenosis was present in 5 cases (35.7%). One patient (7.1%) had aberrant tricuspid chords, and another (7.1 %) had a left superior vena cava. Surgical treatment. There were 10 male and 4 female patients. Mean age was 10.2 ± 5.5 days, mean weight was 3.1 ± 0.3 kg, and mean body surface area was 0.19 ± 0.01 m 2 . Before cardiopulmonary bypass was begun, right and left pulmonary arteries were controlled with small tourniquets. Cardiopulmonary bypass was then instituted between the ascending aorta and the right atrium. One patient with IAA had a double aortic and pulmonary artery cannulation. I I During cardiopulmonary bypass, with the heart still beating, both pulmonary arteries were snared, and the lower part of the body was perfused via the systemic ductus arteriosus. When ventricular fibrillation occurred, both pulmonary arteries were unclamped and the ductus was ligated; the lower hemibody was perfused through the coarctation. When central body temperature reached 20° C, the cardiopulmonary bypass was stopped and the aortic arch was reconstructed according to three different techniques: (I) extended end-to-end anastomosis in 12 patients, as was described for repair of IAA by Bailey and associatesl-; (2) patch angioplasty in two patients and; (3) lateral anastomosis of the proximal and distal aortic arches in the last patient. I 3 The additional advantage of the last two techniques was reduction of the tremendous mismatch in size between the aorta and the pulmonary artery. Afterward hypothermic perfusion was reinstituted and the aorta was unclamped. During this time of cold myocardial reperfusion, the atrial septal defect was closed. The aorta was clamped again, and a blood cardioplegic solution was injected into the aortic root; the arterial switch and VSD closure were then done. The VSD was closed through the right ventricle in 4 patients, through the pulmonary artery in 6, or through both in 4. In addition, a parietal band was resected in 3 patients (21.4%). Two patients with angiographic evidence of subaortic stenosis did not show abnormal stenotic subaortic anatomy after the right ventriculotomy. However, in both of these patients the right

928

Planche et al.

Fig. 1. Aortogram of hypoplastic arch and coarctation.

ventriculotomywas reconstructedwith a fresh autologous pericardial patch to enlarge this region. Mean aortic crossclamptime, without inclusion of circulatory arrest time, was 79.2 ± 13.4 minutes, mean duration of circulatory arrest was 26.1 ± 12.6 minutes, and mean period of cardiopulmonary bypass, without circulatory arrest time, was 187.1 ± 17.1 minutes. Delayedsternal closurewas necessaryin 7 patients (50%), with an average delay time of 3.6 ± 1.9days. Mean intubation period was 7 ± 2.2 days. Mean length of stay in an intensivecare unit was 11.5 ± 2.2 days. Statistical analysis was performed by standard techniques, and time-related events were examined by actuarial methods. Continuousdata were presented as mean ± standard deviation and crude ratio with confidence limits (Cl.) of 70%. Results Group I

Mortality. During phase A, there were three perioperative deaths (12% CL; 5.1% to 21.8%). Two were related to reactive pulmonary vascular disease. Neither of these patients received a pulmonary artery band at the time of aortic repair. In both cases, a pulmonary biopsy sample corroborated a hypertensive pulmonary disease grade of III to IV of the Heath-Edwards!" classification. Both patients were operated on beyond the neonatal period because of early misdiagnosis. They were critically ill when they arrived in the operating room, and they died soon after the operation.

The Journal of Thoracic and Cardiovascular Surgery May 1993

The third death occurred in a critically ill neonate with associated preoperative neurologic disorders and anuria. He died of multiple organ failure the first day after the operation. During phase B, there were five early (21.7% CL; 12.4% to 34.1%) and two late deaths. Early mortality was related to coronary disease in three cases, supravalvular aortic stenosis in one case, and right ventricular failure in the remainder of the patients, who had a moderate underdevelopment of the right ventricle. One late death was related to left ventricular dysfunction in a patient who underwent a heart transplantation in another institution and died of acute graft rejection soon afterward. The other late death was probably caused by arrhythmias 2 years after reconstruction of the superior vena cava for correction of stenosis. In this particular case, enlargement of the superior vena cava may certainly have injured the sinus node region. The overall mortality rate for patients with TGA, VSD, and coarctation who underwent a two-stage repair program was 38.46%. Reoperations. During phase A, there were three reoperations (12% CL; 5.1% to 21.8%). Two patients required emergency pulmonary artery banding because of persistent heart failure. The third patient underwent a ligation of the thoracic duct because of recurrent chylothorax. During phase B, there were eight reoperations (34.8% CL; 27.4% to 47.7%). They were related in five cases to right ventricular outflow tract obstruction. Characteristically, all of them had a moderate anterior displacement of the infundibular septum; in four of these patients, this malalignment was probably exacerbated by a pulmonary artery banding.!" The mean right ventricular-pulmonary artery gradient before and after the reoperations was 99 ± 31.7 mm Hg and 34 ± 5.4 mm Hg, respectively. Another infant had a thrombosis of the superior vena cava. In this case an enlargement was achieved with a heterologous pericardial patch. One patient had a recurrent coarctation and underwent a left subclavian flap angioplasty. The last reoperation, which took place after a heart transplantation, was mentioned previousl y. The cumulative overall reoperation rate for phases A and B was 42.3% (II patients). Moreover, morbidity occurred in 20 of 26 cases that involved early and late death and reoperation (76.9% CL; 69.7% to 85.9%) (Table II). Functional status. Mean follow-up for survivors was 67.1 ± 27 months (range 2 to III months). All but one were in New York Heart Association class I without medications. All the survivors were in normal sinus rhythm, with no Q waves or ST-T changes. Each had right bundle branch block.

The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5

Planche et al.

929

100

90

78.5%

80

-G5 INGLE STAGE -- TWO-STAGE

70 63.9%

57.5%

60

50

+-----r--~----.----,r---...,...--___r--...---__,

o

20

40

60

80 MONTHS

Fig. 2. Actuarial survival in the two groups. At the last echocardiographic Doppler study, the mean shortening fraction of the left ventricle was 38.9% ± 4.1 %. The mean gradient between the right ventricle and the pulmonary artery was 22.4 ± 18.1 mm Hg, with a range of 8 to 60 mm Hg. The mean gradient between the left ventricle and the descending thoracic aorta was 6 ± 7.1 mm Hg. One patient had moderate aortic insufficiency. Actuarial survival rate and rate of freedom from reoperation at 5 years were 57.5% (CL; 45.7% to 64.3%) and 49.9% (CL; 41.4% to 62.5%), respectively (Figs. 2 and 3). Group II Mortality. There were two early deaths (14.2% CL; 4.7% to 30.6%) and one late death (8.3% CL; I % to 25.5%). Early deaths were related to coronary disease in one case and to supra valvular aortic stenosis in the other. The late death occurred at reoperation in a patient with overlooked multiple VSDs. This patient died postoperatively of acute right ventricular failure. Reoperations. There were two reoperations (14.2% CL; 4.7% to 30.6%). One was reported, for overlooked multiple VSDs, while this study was in progress. The second reoperation was performed in an infant who had recurrent coarctation with a 55 mm Hg gradient. It was diagnosed I year after the first operation. The surgical treatment involved subclavian flap angioplasty. The cumulative morbidity rate for patients with TGA, VSD, and coarctation who underwent a one-stage procedure

program was 28.6% (CL; 20.8% to 45.6%) (Table II). This rate was significantly lower than that for the two-stage procedure (p < 0.01). Functional status. Mean follow-up of 24.3 ± 18.7 months (range I to 52 months) was achieved in all survivors. They were all in New York Heart Association class I without medications. Ten patients were in normal sinus rhythm, with right bundle branch block and no Q waves or ST-T changes. One infant needed a permanent pacemaker because of postoperative sick sinus syndrome with a second-degree atrioventricular block. At the last echocardiographic Doppler examination the mean shortening fraction of the left ventricle was 38.8% ± 3.3%. The mean gradient between the right ventricle and the pulmonary artery was 13.5 ± 6.9 mm Hg, with a range of 8 to 30 mm Hg. The mean gradient between the left ventricle and AAO repair was 4.6 ± 5.5 mm Hg. The aortic valve was competent in all patients. Actuarial survival rate and rate of freedom from reoperation at 3 years were 78.5% (CL; 61.5% to 90.3%) and 81.5% (CL; 69.3% to 95.2%), respectively (Figs. 2 and 3). Comparison of the actuarial results by the log-rank test did not show any statistical significancebetween the two groups. Discussion Several studies have drawn attention to the nearly invariable anatomic associations in TGA with VSD and

The Journal of Thoracic an"d

9 30

Planche et al.

Cardiovascular Surgery

May1993

%

100 90

------

81.5%

--

80 70

51 NGLE -5T AGE -- TWO-STAGE ~

58.9%

60

49" 9% 50 40 0

20

40

60

80 MONTHS

Fig. 3. Freedom from reoperation in the two groups.

AAO. As suggested by Milanesi and associates.' TGA should be considered a nosographic entity that includes a YSD, underdevelopment of the right ventricle, and a right ventricular outflow and/or inflow obstruction. To understand and treat this complex anomaly, one must understand the anterior displacement of the infundibular septal components'< 17 with a prominence of the ventriculoinfundibular fold. This lesion may cause a subaortic obstruction18. 19 with a left posterior overriding of the pulmonary valve and malaligned YSD. It has been postulated that the malalignment of right ventricular outlet structures causes aortic arch anomalies by diverting flow from the aorta to the pulmonary artery.20,21 Kurosawa and Van Mierop!" reported that 88% of the hearts with TGA, YSD, and an anteriorly displaced infundibular septum had an aortic coarctation or an interrupted aortic arch. In the present clinical series, there features were confirmed with a somewhat lower percentage (57.5%). Anomalies of the right ventricular inflow, such as mild hypoplasia of the tricuspid valve or aberrant tricuspid chords, have a lesser influence on the physiopathology but may be determinants of the right ventricular size. 3, 4, 17,22,23 From a clinical standpoint cyanosis, congestive heart failure, decreased amplitude of femoral pulses, and acidosis characterize most of these infants. The ductus arteriosus patency, which was maintained by infusion of

prostaglandins E 1, allows these patients to survive the critical postdelivery period and to be quickly operated on in a more stable hemodynamic condition. The high morbidity rate reported in the two-stage procedure group of the present series is probably related to the absence of prostaglandins E 1 at the beginning of the experience, at which time the infants were brought to the operating room in a very poor clinical condition. Repair of the AAO is imperative in the neonatal period, whereas repair of TGA and YSD may be delayed to a second procedure if the condition is well palliated. However, this approach carries risks during repair of the coarctation and during the intermediary period. Therefore one may consider the feasibility of complete repair during a one-stage procedure through midsternotomy. Review of the literature supports this approach.f 5,13,16.18 However, a one-stage procedure may represent a serious surgical challenge. Theoretically, the treatment in a two-stage procedure has advantages and disadvantages. If the first phase involves a pulmonary artery banding, some deleterious effects are possible. (1) A subaortic stenosis caused by a concentric hypertrophy of the subaortic conal myocardium can transform an insignificant gradient between the right ventricle and aortic root into a significant one; (2) pulmonary branch stenosis may occur as a result of distal migration of the band; (3) neoaortic valve insufficiency,

The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5

expected after a proximal movement of the band, may develop;(4) there may be pericardial adhesions with a risk of damaging coronary arteries during the arterial switch; (5) late left ventricular functional abnormalities, most likely caused by aortic regurgitation, may occur.i" If, in the first phase, the pulmonary banding is absent, delaying repair of TGA and VSD for more than 2 to 3 months may not be possible because of severe congestive heart failure and early development of pulmonary vascular disease, which has been described by Clarkson and colleagues" as occurring in as many as 25% of children by the age of 3 months. Nevertheless, a two-stage procedure can be useful in the patients whose condition has deteriorated substantially. In such patients an AAO repair with ductus arteriosus ligation can improve and balance the previously impaired situation. Several reports, n, 26, 27 have shown the feasibility of repair of AAO by means of an anterior approach. Also, neonatal anatomic repair of TGA with VSD has a very promisingoutcome." It was therefore attractive to try the same techniques in patients who have TGA with VSD and AAO. Complete repair has to address four types of difficulty: brain protection, myocardial protection, reconstruction of the aortic arch, and the arterial switch. The former difficulties are commonly addressed by deep hypothermia with circulatory arrest and by cardioplegia. The surgical procedures have been described previously. However, some aspects have to be emphasized. An extended end-to-end anastomosis was always the first choice for repair of the AAO. Direct aortic anastomosiswas always possible if adequate mobilization of the ascending and descending aorta was achieved.i?: 30 Only two patients (5%), one in each group, had a recurrent coarctation. These were repaired by a subclavian flap angioplasty technique, which avoided any prosthetic material. At a distant follow-up, the mean gradient was very low. In some cases, however, even with a wide dissection of the aorta, the aortic reconstruction is not feasible without any excessivetension, and therefore a patch angioplasty should be used. Another important factor in these patients was the size discrepancy between the two great vessels. For culmination of the aortic anastomosis, a vertical split in the anterior wall of the distal aorta was not always sufficient to reduce this size mismatch. For this purpose, a patch augmentation of the distal aorta was used in three cases. Although, at distant follow-up, supravalvular aortic stenosis was not an obvious residual anomaly, two patients died early after operation for this complication. For closure of the VSD, the nearly invariable mala-

Planche et at.

93 I

lignment has necessitated an approach through the right ventricle, the pulmonary artery, or both (83.7%). The pulmonary approach is possible in this complex because of the unusually large diameter of the pulmonary artery. There were no more rhythm disturbances, and damage in the future aortic valvewas related to this treatment in only one case (5.8%). A major difference between the two groups was the right atrial approach, used in only 6 patients (26%) for a two-stage procedure. The approach was used most probably because of a larger tricuspid anulus. Arterial switch was performed according to the standard technique. Assisted mainly by trapdoor flaps, the coronary transfer has been performed in all forms of coronary pattern. However, for type C, the translocation should be carefully performed according to atypical principles. The Lecompte maneuver was accomplished in all the patients and was not responsible for late pulmonary stenosis. Attentive inspection of the results of these procedures outlined some useful inferences. Overall morbidity rate was significantly lower in one-stage repair than in twostage repair. Also, right ventricular outflow tract obstruction occurs at a lower rate than in two-stage procedure. This may have occurred because the patients were specifically selected for the one- or two-stage group, because more attention was paid to the particular features and their surgical implications for patients in the onestage group, or because the previous pulmonary artery banding fixed the infundibular hypertrophy, 19.31 which led to post-operative significant right ventricular outflow tract obstruction. At present, the neonatal repair considerably reduces the risk of right ventricular outflow tract obstruction. The influence of familiarity and experience with the arterial switch operation is likely to be a major factor in the difference between the outcomes of the two groups. To validate the one-stage repair of TGA, VSD, and AAO, the results of neonatal anatomic repair ofTGA and VSD were recently reviewed and compared with the results previously mentioned.F The mortality and reoperation rates were not significantly different in these two groups of patients. Finally, there are the socioeconomic advantages of the one-stage procedure; mean time spent in intensive care unit and, logically, its high cost were reduced by more than 50%. In conclusion, the one-stage repair of TGA, VSD, and AAO through a midsternotomy provides a better outcome than a two-stage repair. However, surgical indications require an accurate assessment of intracardiac and aortic arch anatomy. In cases of multiple VSDs or those

9 32

Planche et al.

in which there is another complex intracardiac anomaly that is not amenable to early repair, the two-stage procedure should be done. REFERENCES 1. Bowers DE, Schiebler GL, Krovetz J. Interruption of the aortic arch with complete transposition of the great vessels: hemodynamic and angiocardiographic data of a case diagnosed during life. Am J Cardiol 1965;16:442-8. 2. Liebman J, Cullum L, Belloc NB. Natural history of transposition of the great arteries: anatomy and birth and death characteristics. Circulation 1969;40:237-62. 3. Milanesi 0, Thiene G, Bini RM, Pellegrino PA. Complete transposition of the great arteries with coarctation of aorta. Br Heart J 1982;48:566-71. 4. Moene RJ, Ottenkamp J, Oppenheimer-Dekker A, Bartelings MM. Transposition of the great arteries and narrowing of the aortic arch. emphasis on right ventricular characteristics. Br Heart J 1985;53:58-63. 5. Vogel M, Freedom RM, Small horn JF, Williams WG, Trusler GA, Rowe RD. Complete transposition ofthe great arteries and coarctation of the aorta. Am J Cardiol 1984;53:1627-32. 6. Serraf A, Bruniaux J, Lacour-Gayet F, et al. Anatomic correction of transposition of the great arteries with ventricular septal defect: experience with 118 cases. J THORAC CARDIOVASC SURG 1991;102:140-7. 7. Moulaert AJ, Bruins CC, Oppenheimer-Dekker A. Anomalies of the aortic arch and ventricular septal defects. Circulation 1976;53:1011-5. 8. Celoria GC, Patton RB. Congenital absence of the aortic arch. Am Heart J 1959;58:407-15. 9. Yacoub MH, Radley-Smith R. Anatomy of the coronary arteries in transposition of the great arteries and methods for their transfer in anatomical correction. Thorax 1978; 33:418-24. 10. Planche C, Bruniaux J, Lacour-Gayet F, et al. Switch operation for transposition of the great arteries in neonates: a study of 120 patients. J THORAC CARDIOVASC SURG 1988;96:354-63. II. Sell JE, Jonas RA, Mayer JE, Blackstone EH, Kirklin JW, Castaneda AR. The results of a surgical program for interrupted aortic arch. J THORAC CARDlOVASC SURG 1988; 96:864-77. 12. Bailey LL, Jacobson JG, Vymeister E, et al. Interrupted aortic arch complex: successful total correction in neonate. Ann Thorac Surg 1978;25:66-70. 13. Pigott JD, Chin AJ, Weinberg PM, Wagner HR, Norwoord WI. Transposition of the great arteries with aortic arch obstruction. J THORAC CARDIOVASC SURG 1987; 94:82-6. 14. Heath D, Edwards JE. The pathology of hypertensive pulmonary vascular disease: a description of six grades of structural changes in the pulmonary arteries with special reference to congenital cardiac septal defects. Circulation 1958;18:533.

The Journal of Thoracic and Cardiovascular Surgery May 1993

15. Robertson MA, Penkoske PA, Duncan NF. Right pulmonary artery obstruction after pulmonary artery banding. Ann Thorac Surg 1991;51:73-5. 16. Kurosawa H, Van Mierop LHS. Surgical anatomy of the infundibular septum in transposition of the great arteries with ventricular septal defect. J THORAC CARDIOVASC SURG 1986;91:123-32. 17. Moene RJ, Oppenheimer-Dekker A, Bartelings MM. Anatomic obstruction of the right ventricular outflow tract in transposition of the great arteries. Am J Cardiol 1983;51:1701-4. 18. Schneeweiss A, Motro M, Shern-Tov A, Neufeld HN. Subaortic stenosis: an unrecognized problem in transposition of the great arteries. Am J Cardiol 1981;48:336-9. 19. Waldman JD, Schneeweiss A, Edwards WD, Lamberti JJ, Shem-Tov A, Neufeld HN. The obstructive subaortic conus. Circulation 1984;70:339-44. 20. Rudolph AM, Heymann MA, Spitznas U. Hemodynamic considerations in the development of narrowing of the aorta. Am J Cardiol 1972;30:514-25. 21. Shinebourne ES, Elseed AM. Relation between fetal flow patterns, coarctation of the aorta, and pulmonary blood flow. Br Heart J 1974;36:492-8. 22. Huhta JC, Edwards WD, Danielson GK, Feldt RH. Anomalies of the tricuspid valve in complete transposition of the great arteries with ventricular septal defect. J THORAC CARDIOVASC SURG 1982;83:569-76. 23. Riemenschneider TA, Vincent WR, Ruttenberg HD, Desilets DT. Transposition of the great vessels with hypoplasia of the right ventricle. Circulation 1968;38:386402. 24. Gibbs JL, Qureshi SA, Wilson N, Radley-Smith R, Yacoub MH. Doppler echocardiographic comparison of haemodynamic results of one- and two-stage anatomic correction of complete transposition. Int J Cardiol1987; 18:8592. 25. Clarkson PM, Neutze JM, Wardhill JC, Barratt-Boyes BG. The pulmonary vascular bed in patients with complete transposition of the great arteries. Circulation 1976;53:53943. 26. Norwood WI, Lang P, Castaneda AR, Hougen TJ. Reparative operations for interrupted aortic arch with ventricular septal defect. J THORAC CARDIOVASC SURG 1983;86:8327. 27. Turley K, Yee ES, Ebert PA. The total repair of interrupted arch complex in infants: the anterior approach. Circulation 1984;70(Pt. 2):116-20. 28. Di Donato RM, Wernovsky G, Walsh EP, et al. Results of the arterial switch operation for transposition of the great arteries with ventricular septal defect: surgical considerations and midterm follow-up data. Circulation 1989; 80:1689-705. 29. Lacour-Gayet F, Bruniaux J, Serraf A, et al. Hypoplastic transverse arch and coarctation in neonates. J THoRAc CARDIOVASC SURG 1990;100:808-16. 30. Amato JJ, Rheinlander HF, Cleveland RJ. A method of enlarging the distal transverse arch in infants with hypo-

The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5

plasia and coarctation of the aorta. Ann Thorac Surg 1977;23:261-3. 31. Shrivastava S, Tadavarthy SM, Fukuda T, Edwards JE. Anatomic causes of pulmonary stenosis in complete transposition. Circulation 1976;54:154-60. 32. Serraf A, Comas JV, Lacour-Gayet F, Bruniaux J, Bouchart F, Planche C. Neonatal anatomic repair of transposition of the great arteries and ventricular septal defect. Eur J Cardiothorac Surg 1992;6:630-4.

Discussion Dr. Constantine Mavroudis (Chicago, Ill.). Can you speculate why survival has improved? Is it due to young age, anatomy, or increasing experience? Dr. Planche. I think the improved results are due to improvement in surgical technique, surgical treatment of the aortic arch, surgical treatment by means of the anterior approach, and, of course, experience in the switch operation. Dr. Christo I. Tchervenkov (Montreal, Quebec, Canada). I would like to thank Dr. Planche for demonstrating through his achievements that which may not always be obvious to some of us or to our cardiology colleagues. A complex operation will lead to better survival, provided a complete and accurate repair is achieved. We are in full agreement with the single-stage approach for this difficult group of patients. At the Montreal Children's Hospital, we have dealt with four patients with AAO; three had a double-outlet right ventricle with subpulmonary VSD (the Taussig-Bing anomaly), and one had transposition with an intact ventricular septum. In all four patients, the intent was to do a single-stage repair. In one of them, soon after induction of anesthesia, because of hypotension and ST changes, the arch was repaired with extended end-toend anastomosis and pulmonary artery banding via a left thoracotomy. Because of persistent heart failure, the patient was

Planche et al. 9 3 3

returned to the operating room 2 weeks later and an arterial switch operation, VSD closure to the pulmonary artery, was successfully carried out. The other three patients underwent a single-stage procedure by extensive reconstruction of the aortic arch with a pulmonary homograft patch from the upper descending thoracic aorta all the way down to the proximal anastomosis plus an arterial switch operation. The VSD was closed into the pulmonary artery for the Taussig-Bing malformation. All four patients survived the operation, and none has required reoperation. They are all doing well between 15 months and 3 years later. It was not very clear to me whether these were two separate series or whether the two different approaches overlapped. Also, I would like to know whether Dr. Planche believes that the improved results with the single-stage operation are due to operating on less sick babies that have not had palliation for 2 or 3 months or whether there are anatomic reasons for the improved survival. Perhaps there is more residual AAO after the extended end-to-end anastomosis via left thoracotomy at the proximal aortic arch. The discrepancy between the proximal and distal parts of the aorta and the necessity for extensive tailoring may have led to distortion of the aortic valve and perhaps distortion to coronary arteries. Finally, in the patient with subaortic stenosis that was caused by anterior malalignment VSD, how was the right ventricular outflow tract dealt with at the initial operation? Were any transannular patch procedures done? Dr. Planche. There are two differents points in the first question. At the beginning of our experience in the early 1980s, we used a two-stage approach, and since 1987 we have used a single primary repair. For subaortic stenosis, we usually open the right ventricle and resect the parietal band and abnormal muscular band. We did not use a transannular patch as the primary repair. Another difficulty is related to the uncommon situation in which the anterior coronary artery runs in front of the aortic root, so that it is not possible to enlarge the aortic root. In this case it is necessary to insert a conduit.