The influence of coronary anatomy on the arterial switch operation in neonates

The influence of coronary anatomy on the arterial switch operation in neonates To determine whether coronary anatomy influences the outcome of the neonatal arterial switch operation, we examined the results in aU newborn infants (n = 70) with D-transposition of the great arteries who had a corrective operation at our institution between March 1987 and April 1991. The origin and distribution of coronary arteries were identified preoperatively by echocardiography, aortic root angiography, or selective coronary arteriography and intraoperatively by direct inspection. However, the arterial switch operation was performed independent of the coronary anatomy in aU but two candidates for the operation. Four early deaths occurred and five surviving patients had symptoms of impaired cardiac function. No late deaths have occurred in patients followed up for 2 to 50 months. Evidence of myocardial ischemia was present in three of the four deaths and in four of the five patients with cardiovascular symptoms. Patients with commissural or intramural coronary origins between the great arteries had significantly greater cardiovascular morbidity and mortality because of coronary ischemia than patients with the most common coronary pattern. Thus coronary anatomy may influence surgical management and the postoperative course of newborn infants with transposition. (J 'fHORAC CARDIOVASC SURG 1992;104:706-12)

Ronald W. Day, MD, Hillel Laks, MD, and Davis C. Drinkwater, MD, Los Angeles, Calif.

Anatomic correction in newborn infants with D-transposition of the great arteries can be performed by the arterial switch operation with low early mortality in some institutions. I, 2 The ability of the left ventricle to assume the systemic workload is one determinant of surgical outcome.v" However, D-transposition ofthe great arteries is associated with diverse coronary patterns that may also influence the risk of the arterial switch operation.v 6 At our institution, coronary anatomy was generally not used to exclude patients from surgical treatment. Therefore we examined the results of all arterial switch operations in neonates to determine whether selected coronary patterns were associated with increased cardiovascular morbidity and mortality. From the Divisions of Pediatric Cardiology and Cardiothoracic Surgery,University of California, LosAngeles Schoolof Medicine, Los Angeles, Calif. Ronald Day wassupported by the National Institutesof Health Grant HL-074I 2 and the Laubisch Fund for Cardiovascular Research Received for publication June II, 1991. Accepted for publication Oct. 28, 1991. Addressfor reprints: HillelLaks,MD, Division of Cardiothoracic Surgery,UCLA Schoolof Medicine, 10833 Le Conte Ave., LosAngeles, CA 90024. 12/1/35157

706

Patients and methods Patients. From March 1987 to April 1991, 70 patients with D-transposition of the great arteries had an arterial switch operation before 28 days of age. The operation was performed at a mean age of 5.6 days (range 2 to 26 days), and all but one patient had the operation before 14 days of age. Associated cardiovascular anomalies included a left superior vena cava (two patients), dextrocardia (two patients), juxtaposition of the atrial appendages (one patient), a single ventricular septal defect (five patients), and multiple ventricular septal defects (two patients). The coronary anatomy was imaged by echocardiography,? fixed-biplane aortic root angiography, or the venous technique of selective coronary arteriography before the operation.s Early in our series, one neonate was denied surgical treatment because of the preoperative description of a large and tortuous proximal left coronary artery with the usual coronary distribution. Another early patient had a neonatal Senning procedure because of the intraoperative finding of a coronary origin that was considered too distant from the native pulmonary trunk for successful reimplantation. Classification of coronary anatomy. The intraoperative description of the coronary anatomy was used to classify patients by the Leiden convention," However, variations in coronary anatomy were further distinguished by identifying (I) the epicardial course of each major coronary branch crossing the plane ofthe great arteries and (2) unusual coronary origins. For clarity, variations were descriptively stated rather than abbreviated with the symbols proposed by Yamaguchi," As summarized in Table I, this method of classification provided an accurate, yet concise, description of numerous coronary patterns.

Volume 104 Number 3 September 1992

Fig. 1 illustrates some examples of the patterns and unusual coronary origins that were identified in our patients. In all patients, the aortic root was located within a portion of the quadrant rightward and anterior to the pulmonary trunk. Surgical technique. Patients were generally treated with prostaglandin E j to maintain a patent ductus arteriosus until the operation. The procedures were performed with cardiopulmonary bypass plus deep hypothermia. (ultimate perfusate temperature of 180 C) with low flow. In most cases, a single right atrial cannula was used. When bypass was begun, the ductus arteriosus was divided and the ends oversewn. The aorta was crossclamped after the perfusate had been cooled to 24 0 C. Warm blood cardioplegic solution was administered into the root of the aorta followed by cold blood cardioplegic solution, and the perfusate was cooled further to 18 0 C. The aorta was then transected just above the aortic valve. The coronary buttons were excised. The pulmonary artery was transected just proximal to the bifurcation. The distal aorta was brought posterior to the pulmonary artery as described by LeCompte and colleagues.!'' The coronary buttons were then implanted in the proximal native pulmonary artery. The technique has varied with individual anatomy and has evolved with time. Recently, the site of the valve commissure has initially been defined with a marking suture. The distal aorta was then anastomosed to the proximal native pulmonary artery with absorbable PDS suture material (Ethicon, Inc., Somerville, N.J.). The crossclamp was transiently released to completely distend the proximal portion of the new aorta so that appropriate sites for the coronary anastomoses could be selected. At these sites, Ll-shaped segments of the new aorta were excised and each coronary button was implanted with a running 7-0 Prolene suture (Ethicon). Cold blood cardioplegic solution was then given into the aorta below the crossclamp. Thereafter, perfusate was warmed to 24 0 C. A W-shaped pericardial patch, treated with glutaraldehyde, was used to reconstruct the proximal portion of the new pulmonary artery as described by Planche and associates. I I Bypass was then discontinued and the atrial septal defect was closed via a small right atrial incision. The atrial incision was then closed during 5 to 6 minutes of circulatory arrest, and bypass was again instituted and rewarming completed. The left side of the heart was vented via the left atrial appendage. The posterior part of the new pulmonary artery anastomosis was performed with absorbable PDS suture material. Warm blood cardioplegic solution was administered into the aortic root, and the crossclamp was released while the ascending aorta was being vented. The pulmonary artery anastomosis was completed and the patient was weaned from bypass as rewarming continued. In two patients with commissural coronary origins, an attempt was made to preserve the commissure, which resulted in an inadequate coronary button. Subsequently, an appropriately sized button has been excised by including the superior half of the commissure. An intramural coronary segment between the great arteries was not recognized in one patient and a pericardial graft was required to reconstruct and reimplant the left main coronary artery. In a second patient with an intramural coronary course, the intramural segment was incorporated in a larger than usual coronary button. This was left facing anteriorly and a pericardial patch was used to create a conduit between the new aorta and the coronary button. Morbidity and mortality. The number of patients with cardiovascular morbidity and mortality associated with coronary ischemia was determined for each coronary pattern. Cardio-

Neonatal arterial switch operation

707

Table I. Classification of coronary anatomy in

D-transposition of the great arteries Leiden classification of coronary origin Sinus

Using the perspective of an individual looking from the aorta to the pulmonary artery: Sinus I-The sinus adjacent to the pulmonaryartery on the right-hand side of the observer Sinus 2-The sinusadjacent to the pulmonaryartery on the left-handside of the observer Coronary arteries

Right coronaryartery (R) Anteriordescending artery (AD) Circumflex artery (Cx) A comma is used to indicate that major branchesoriginate from a common vessel, whereas a semicolon denotesseparate origins

Supplemental descriptive classification Epicardial course of major coronary branches

Anterior-A coronarybranch passing anterior to the aorta Posterior-A coronarybranch passing posterior to the pulmonary artery Between-A coronarybranch passing between the great arteries (usuallyintramural) Unusual origins

Commissural-A coronaryorigin near an aortic commissure Separate--Separate origins of two coronary branches from the same aortic sinus Remote or distal-Origin of the circumflex artery and the posterior descending artery as a distal bifurcation of the right coronaryartery Aortic position relative to the pulmonary artery

Right or anterior, left, side by side,or posterior

vascular morbidity was determined by symptoms of decreased ventricular function or an arrhythmia associated with acute ST segment changes greater than 2 mm from baseline or impaired coronary filling confirmed by angiography. Fisher's exact test was used to determine the probability of a statistically significant difference (p < 0.05) in morbidity and mortality, comparing patients with the usual (lAD,Cx;2R) coronary anatomy and patients with coronary patterns characterized by (1) a course posterior to the pulmonary artery, (2) an intramural course between the great arteries, (3) a course anterior to the aorta, or (4) an unusual coronary origin. Long-term assessment of patients after the operation was not uniform. Heart catheterization with aortic root or selective coronary angiography was performed approximately I year after the operation in most of the patients followed up at the V niversity of California at Los Angeles. However, the majority of patients (83%) were followed up outside our institution without an invasive evaluation of coronary perfusion. Thus a correlation of coronary anatomy with asymptomatic coronary occlusion could not be determined. Results

Morbidity and mortality. The early results in all 70 patients were available, yet intermediate follow-up (2 to

708

The Journal of Thoracic and Cardiovascular Surgery

Day, Laks, Drinkwater

(IAD,Cx;2R) The Usual Coronary Anatomy

(IAD;2R,Cx) Posterior Circumflex Course

(2R,AD,Cx) Posterior Left Course

(2R;2AD,Cx) Intramural Coronary Course Between the Great Arteries

(IR;2AD,Cx) Anterior Right Course Posterior Left Course

(IR,AD;2Cx) Anterior Right Course Posterior Circumflex Course

(IAD,Cx;2R) Commissural Origins Between the Great Arteries

(IAD;2R;2Cx) Separate Origins from the Same Aortic Sinus

( IAD;2R,Cx) Remote Circumflex Origin from the Distal Right Coronary

Fig. 1. Illustrations and descriptiveclassification of the coronary artery patterns identifiedin our neonates having the arterial switch operation.

50 months, mean 27 months) could be obtained in only 64 of the surviving 66 patients. Four deaths occurred in the immediate postoperative period for an overall mortality of 5.7%(4/70). There have been no late deaths. Five surviving patients had cardiovascular symptoms. The deaths of three patients and the cardiovascular symptoms of four patients were associated with coronary ischemia. The fourth death occurred in a patient with multiple ventricular septal defects, and one patient had early left ventricular dysfunction without evidence of coronary ischemia. Two symptom-free patients have required surgical repair of supravalvular or branch pulmonary stenosis that

was not amenable to balloon angioplasty. Of note, coronary position was revised or adjusted in seven patients during repeated periods of cardiopulmonary bypass. Three of these patients died and three had cardiovascular symptoms. Table II lists the morbidity and mortality associated with ischemia for each coronary pattern. The patient's age at the time of operation and the presence of associated cardiovascular anomalies were not statistically significant determinants of morbidity and mortality. Of the 43 patients with straightforward (IAD,Cx;2R) coronary anatomy, only one had early ischemic electrocardio-

Volume 104 Number 3 September 1992

Neonatal arterial switch operation

709

Table II. Coronary pattern and the morbidity and mortality related to coronary ischemia Coronary pattern

Total

Morbidity

Mortality

Usual coronary pattern (I AD,Cx,2R) 43 I Epicardial course of the anterior descending and/or the circumflex posterior to the pulmonary artery Total 16 0 (lAD;2R,Cx) II 0 (2R,AD,Cx) Single coronary origin 3 0 (I AD;2R;2Cx) Separate origins 2 0 Epicardial course between the great arteries (intramural) (2AD,Cx;R) Intramural 2 2 Epicardial course of the right coronary anterior to the aorta Total 3 0 (I R;2AD,Cx) I 0 (I R,AD;2Cx) 2 0 Commissural origins between the great arteries Total 6 Commissural left coronary (lAD,Cx;2R) 2 Commissural right coronary (IAD,Cx;2R) 3 0 (IAD,2R,Cx) I 0 Remote origin of the circumflex from the right coronary artery (lAD;2R,Cx) Remote I 0

graphic changes and acute heart failure that resolved by opening the chest in the intensive care unit. The sternum was subsequently closed and this patient completely recovered, with normal coronary filling by selective arteriography 3 months after the operation. The course of one or more major coronary branches was posterior to the pulmonary artery in 16 patients. Two patients in this group had separate origins of the circumflex and right coronary branches from sinus 2 ( IAD;2R;2Cx). Although none of these 16 patients had evidence of coronary ischemia, global left ventricular hypokinesia was observed immediately after the operation in one of the patients with separate origins of the circumflex and right coronary branches. The ventricular function of this patient improved over 2 days with increased inotropic support and afterload reduction. One ofthe three patients with a coronary course anterior to the aorta had an elevated right atrial pressure and electromechanical dissociation shortly after closure of the sternum. Right ventricular infarction was evident when this patient's heart was directly visualized during attempted resuscitation. When compared with patients having the usual coronary anatomy, there was not a statistically significant difference in morbidity and mortality resulting from ischemia for patients with an epicardial coronary course posterior or anterior to the great arteries. Our first patient with an intramural coronary segment, which had not been recognized preoperatively and which had been repaired with a pericardial graft, was initially

p Value

0 0 0 0 0

p> 0.05

0

p z; 0.005

I 0 I

p> 0.05

2

p

::s 0.005

I 0 0

p> 0.05

free of symptoms. However, 2 years after the operation, syncope developed as a result of transient periods of complete heart block. Selective coronary arteriography showed that the left coronary artery was completely occluded, and extensive collaterals had developed from the right coronary artery to the areas supplied by the anterior descending and circumflex branches. The second patient with an intramural coronary segment had early ischemic electrocardiographic changes and left ventricular dysfunction that prolonged his initial hospitalization and resulted in a second hospitalization for intravenous medical management. Among the six patients with at least one commissural coronary origin between the aorta and pulmonary artery, two early deaths occurred and one patient had several early postoperative episodes of acute hemodynamic compromise that were resolved by opening the sternum. Overall, the morbidity and mortality related to ischemia were significantly greater for patients with posterior commissural origins or an intramural coronary segment than for patients with the usual coronary anatomy. Complications of coronary ischemia were not limited to the early patients in our series. Of the initial 20 patients, five had commissural coronary origins between the great arteries or an anterior coronary arterial course. In this group, one patient with the usual coronary anatomy experienced morbidity and two patients with commissural origins died as a result of coronary ischemia. In the following 50 patients, six additional patients had com-

7I0

Day, Laks, Drinkwater

The Journal of Thoracic and Cardiovascular Surgery

Fig. 2. Technique for reirnplantation of an intramural coronary segment between the great arteries. A, An intramural coronary course between the aorta and pulmonaryartery is occasionally associated with D-transposition of the great arteries. B, Separate coronarybuttons may be excised whenthere is an adequate distance between the left and right ostia. The entire intramural segmentand the superiorportion of the aortic valve commissure are included in the large coronarybutton. C, The button of the intramural segmentis left facinganteriorly, and the superiormargin is sewnto the native pulmonarytrunk to avoid potentialtorsion or kinking of proximal coronarybranches. The anastomosis of the ascending aorta and the pulmonarytrunk is performedwith an aortic window adjacent to the intramural button. D, A pericardialpatch is then sewn from the aortic window to the lateral and inferior margins of the button to complete the conduit.

missural ongms, an intramural coronary segment between the great arteries, or an anterior coronary arterial course. In this group, two patients with intramural coronary segments experienced morbidity and one patient with an anterior coronary course died as a result of coronary ischemia.

Discussion Over the past decade, several institutions have reported their experience with diverse coronary patterns and outcome of the arterial switch operation for transposition of the great arteries.Ui'? Despite an increased mortality for some coronary patterns, many reports suggest that no coronary pattern is a contraindication for anatomic correction. Quaegebeur and associates':' found no correlation of mortality with the coronary anatomy; however, the epicardial course of the coronary branches was not defined in their report. Similarly, mortality was not increased in patients with atypical coronary patterns in the study of Kurosawa, Imai, and Kawada.!" In these two series, however, 9% to 13% of patients with the usual coronary pattern died, and the cause of each death was not explained. Coronary patterns with commissural origins or an intramural segment between the great arteries

increased the risk of the arterial switch operation in a multicenter study reported by Yamaguchi and coworkers. IS In a neonatal study, Planche and colleagues 1 I also reported an increased mortality associated with commissural coronary origins between the great arteries. Recently, a large series of 290 patients having the arterial switch operation was reported by Mayer and colleagues. 16 They found an increased risk for the arterial switch operation when the entire left coronary circulation passed posterior to the pulmonary artery, whether all vessels originated from sinus 2 as a single coronary trunk or whether the right coronary originated separately from sinus 1 with an epicardial course anterior to the aorta. Unfortunately, patients with commissural coronary origins between the great arteries were not identified in this series, and approximately one third of the patients with intramural coronary segments or major coronary branches passing anterior to the great arteries had a Senning operation. With experience, fewer patients were deferred to Senning operations and fewer deaths resulted from coronary ischemia; however, the morbidity and mortality of the neonatal Senning procedures were not reported. Our entire experience with arterial switch in neonates is included in this study. From the onset, coronary anat-

Volume 104 Number 3 September 1992

amy and specific complications related to coronary ischemia were recorded. Only two patients were deferred to a Senning operation on the basis of the coronary pattern; thus the impact of coronary anatomy was probably not biased by the exclusion of some patients. One of the neonates deferred to a Senning procedure died shortly after the operation, and the other had an uncomplicated repair during infancy. Only one of our patients with the usual coronary pattern had symptomatic myocardial ischemia. In contrast to the experience of Mayer and colleagues, 16 there were no complications resulting from ischemia with coronary branches passing posterior to the pulmonary artery, unless the right coronary artery also originated from a contralateral sinus and passed anterior to the aorta. There was, however, significantly greater risk for anatomic correction in our patients with commissural origins or intramural segments between the great arteries. The increased morbidity and mortality associated with some coronary patterns may reflect distinct learning curves for these patterns and the need for improved techniques to prevent coronary occlusion. Two of our recent patients with commissural coronary origins have been successfully managed by including the superior half of the adjacent commissure in the coronary button. An intramural coronary course can be managed by modifications of the Aubert technique. I?, 18 Alternatively, a large coronary button that contains the intramural segment can be left facing anteriorly and can be patched to the new aorta with a pericardial conduit, as shown in Fig. 2. As illustrated for this case, the distance between the left and right ostia was large enough to allow the excision of separate, yet adequately sized buttons. However, when separate coronary branches originate in close proximity from the same sinus, a large single button that encircles both ostia is usually preferred. Finally, delayed closure of the sternum may decrease the risk of early coronary compression and ischemia independent of the coronary pattern or method of reimplantation. The early and intermediate results of the neonatal arterial switch operation are good for transposition of the great arteries, with or without a ventricular septal defect. I, 2, 19, 20 Good results have also been reported with atrial switch procedures during infancy for transposition with an intact ventricular septum; however, outcome is less favorable for transposition with a ventricular septal defect. 21. 23 Further, significant arrhythmias and systemic right ventricular failure may occur after atrial switch procedures.P: 25 Assuming that the long-term outcome of the arterial switch was considerably better than atrial switch alternatives, Danford/" suggested that the arterial switch operation is the procedure of choice unless early mortality exceeds 20% for simple transposition and 39%

Neonatal arterial switch operation 7 1 1

for transposition with a ventricular septal defect. We believe that a considerably lower mortality than this can be achieved by the arterial switch operation; however, an atrial switch may be preferred in a small subset of patients if the outcome of anatomic correction continues to be influenced by some coronary patterns. Further, the relative risk of neonatal and infant atrial switch operations must be determined. If neonatal atrial repair is not a comparable alternative, the method and timing of the operation may be influenced by foreknowledge of the coronary anatomy. The long-term results of the arterial switch operation are unknown, yet some patients have had coronary occlusions early in life. 16, 19 Longitudinal assessment of ventricular function and coronary perfusion may be required after the operation to confirm adequate growth of the coronary anastomoses and to identify coronary patterns with increased risk of late obstruction. At the same time, the efficacy of improved surgical techniques for more difficult coronary patterns will be determined. We thank Steven L. Day for preparing the illustrations. REFERENCES I. Castaneda AR, Trusler GA, Paul MH, Blackstone EH, Kirklin JW. The early resultsof treatment of simpletransposition in the current era. J THORAC CARDIOVASC SURG 1988;95: 14-28, 2. Norwood WI, DobellAR, Freed MD, KirklinJW, BlackstoneEH. Intermediate resultsofthe arterial switchrepair. J THORAC CARDIOVASC SURG 1988;96:854-63. 3. Yacoub MH, Radley-Smith R, MacLaurin R, Two-stage operation for anatomical correction of transpositionof the great arteries with intact ventricular septum. Lancet 1977;1:1275-8. 4. Jonas RA, GigliaTM, Sanders SP, et al. Rapid, two-stage arterial switch for transposition of the great arteries and intact ventricularseptum beyondthe neonatal period. Circulation I989;80(Pt 2):1203-8. 5. Shaher RM, Puddu Gc. Coronary arterial anatomy in completetransposition of the great arteries. Am J Cardiol 1966; 17:355-61. 6. Gittenberger-de Groot AC, Sauer U, Oppenheimer-Dekker A, Quaegebeur J. Coronary arterial anatomy in transposition of the great arteries: a morphologic study. Pediatr Cardiol 1983;4(Pt 2):1 15-24. 7. Pasquini L, Sanders SP, Parness lA, Colan SO. Diagnosis of coronaryartery anatomy by two-dimensional echocardiography in patients with transposition of the great arteries. Circulation 1987;75:557-64. 8. Day RW, Isabel-Jones JB, Wetzel GT, Oku GS, Jarmakani JM. Description of a venous technique for selective coronary arteriography in newborns with transposition of the great arteries. J Am Call Cardiol 1989;14:1308-11.

7 12

Day. Laks, Drinkwater

9. Yamaguchi M. Arterial switch operation [reply to the Editor]. J THORAC CARDIOVASC SURG 1990;100:314. 10. LeCompte Y, Zannini L, Hazan E, et al. Anatomic correction of transposition of the great arteries: a new technique without use of a prosthetic conduit. J THORAC CARDIOVASC SURG 1981;82:629-31. II. 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. 12. 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. 13. Quaegebeur JM, Rohmer J, Ottenkamp J, et al. The arterial switch operation: an eight-year experience. J THORAC CARDIOVASC SURG 1986;92:361-84. 14. Kurosawa H, Imai Y, Kawada M. Coronary arterial anatomy in regard to the arterial switch procedure. Cardiol Young 1991;1:54-62. 15. Yamaguchi M, Hosokawa Y, Imai Y, et al. Early and midterm results of the arterial switch operation for transposition of the great arteries in Japan. J THORAC CARDIOVASC SURG 1990;100:261-9. 16. Mayer JE, Sanders SP, Jonas RA, Castaneda AR, Wernovsky G. Coronary artery pattern and outcome of arterial switch operation for transposition of the great arteries. Circulation I990;82(Pt 2):lV 139-45. 17. Aubert J, Pannetier A, Couvelly JP, Unal D, Rouault F, Delarue A. Transposition of the great arteries: new technique for anatomical correction. Br Heart J 1978;40:204-

8. 18. Takeuchi S, Katogi T. New technique for arterial switch operation in difficult situations. Ann Thorac Surg 1990;50: 1000-1.

The Journal of Thoracic and Cardiovascular Surgery

19. Wernovsky G, Hougen TJ, Walsh EP, et al. Midterm results after the arterial switch operation for transposition of the great arteries with intact ventricular septum: clinical, hemodynamic, echocardiographic, and electrophysiologic data. Circulation 1988;77:1333-44. 20. 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: 1689-1705. 21. Penkoske PA, Westerman GR, Marx GR, et al. Transposition of the great arteries and ventricular septal defect: results of the Senning operation and closure of the ventricular septal defect in infants. Ann Thorac Surg 1983;36:2818. 22. George BL, Laks H, Klitzner TS, Friedman WF, Williams RG. Results of the Senning procedure in infants with simple and complex transposition of the great arteries. Am J Cardiol 1987;59:426-30. 23. Turley K, Hanley FL, Verrier ED, Merrick SH, Ebert PA. The Mustard procedure in infants (less than 100 days of age). J THORAC CARDIOVASC SURG 1988;96:849-53. 24. Deanfield J, Camm J, Macartney F, et al. Arrhythmia and late mortality after the Mustard and Senning operation for transposition of the great arteries: an eight-year prospective study. J THORAC CARDIOVASC SURG 1988;96:569-76. 25. Graham TP, Atwood GF, Boucek RJ, Boerth RC, Bender HW. Abnormalities ofright ventricular function following Mustard's operation for transposition of the great arteries. Circulation 1975;52:678-84. 26. Danford DA. Factors influencing choice of procedure in transposition of the great arteries: a decision analysis approach. J Am Coli CardioI1990;16:471-5.