European Heart Journal (1997) 18, 1470-1477

The double switch procedure for anatomical repair ofcongenially corrected transposition of the great

arteries in infants and children

V. M. Reddy, D. B. McElhinney, N. H. Silverman and F. L. Hanley

Divisions of Cardiothoracic Surgery and Pediatric Cardiology, University of California, San Francisco,California, U.S.A.

Aims To assess outcomes of anatomical repair (doubleswitch procedure) in infants and children with congenitallycorrected transposition of the great arteries.

Methods and results Between September 1993 andAugust 1996, 17 patients with congenitally corrected trans-position underwent surgery at UCSF. Anatomical repairwas performed in 11 of these patients, at ages ranging from4-8 months to 7-8 years (median 3-2 years). The remainingsix patients did not undergo anatomical repair due tounfavourable anatomy (n=2), prior conduit repair (n=2),biventricular dysfunction (n = l), and isolated completeatrioventricular block (n = l). The 11 patients who under-went anatomical repair make up the study group for thepresent report. All 11 patients had a malalignment ventricu-lar septal defect, while pulmonary outflow tract obstructionwas present in nine patients and significant tricuspid valvepathology or dysfunction was present in five. Anatomicalrepair was achieved with a Senning (n=7) or a Mustard(n=4) procedure combined with an arterial switch opera-tion plus ventricular septal defect closure (n=4), or aRastelli procedure with left ventricle to aortic baffle and

right ventricle to pulmonary artery conduit (n = 7). Therewas one early death and no patients developed surgicalcomplete atrioventricular block. At a median follow-up of22 months, there were no late deaths. Two patients requireda total of three late reoperations, and all patients wereasymptomatic on no cardiac medication. Follow-upechocardiography revealed normal biventricular function inall patients.

Conclusions Anatomical repair of corrected transpositioncan be achieved with low rates of early mortality andsurgical heart block, and favourable mid-term results.Long-term follow-up will be necessary to determine if thedouble switch approach improves the natural history ofcorrected transposition when compared to less aggressivesurgical approaches that leave the right ventricle in thesystemic circulation.(Eur Heart J 1997; 18: 1470-1477)

Key Words: Atrioventricular discordance, ventriculo-arterial discordance, Rastelli procedure, arterial switchoperation, Senning operation, Mustard operation.

Introduction

Until recently, the standard surgical approach to con-genitally corrected transposition of the great arteries(corrected transposition) has been simply to repair theassociated lesions, which are common and most ofteninclude ventricular septal defect, tricuspid valveanomalies or insufficiency, and pulmonary stenosis oratresia'1"71. In this approach, the right ventricle is

Revision submitted 4 February 1997, and accepted 24 February1997.

Presented in part at the XVIIIth Congress of the European Societyfor Cardiologists Birmingham, UK, 25-28 August 1996

Correspondence: V. Mohan Reddy, MD, 505 Parnassus Avenue,M593, San Francisco, CA 94143 0118 U.S.A.

allowed to remain as the systemic pumping chamber.Much of the data regarding ventricular function inisolated corrected transposition and corrected transpo-sition following intracardiac repair of associated defectsis inconclusive regarding the ability of the morphologi-cally right systemic ventricle to tolerate the pressureload of the systemic circulation. However, a number ofstudies have demonstrated depressed right ventricularfunction in corrected transposition, especially duringexercise'68^131. Similarly, as long-term follow-up hasbecome available for patients after atrial level repair ofsimple transposition of the great arteries, heightenedconcerns have arisen regarding the chronic performanceof the right ventricle in the systemic circulation'14"181.These concerns, along with the frequent dysfunction ofthe systemic tricuspid valve in corrected transposition,

0195-668X/97/091470+08 S18.00/0 & 1997 The European Society of Cardiology

Double switch procedure 1471

have prompted a number of surgeons to undertakeanatomical repair of corrected transposition, in whichan atrial switch procedure (Senning or Mustard) iscombined with either an arterial (arterial switch) orventricular (Rastelli) level repair'618"241. At UCSF, wehave been performing anatomical correction in selectedpatients with corrected transposition since September1993. In the present report, we review our experiencewith 11 patients who have undergone anatomical repairof corrected transposition since this time.

Methods

DefinitionsAnatomical structures are referred to in terms oftheir morphological characteristics, unless otherwisespecified. Thus, for example, the morphologically rightfunctionally systemic atrioventricular valve in patientswith unrepaired congenitally corrected transposition isreferred to in the present report as the tricuspid valve.Similarly, semilunar valves and great vessels followingthe arterial switch procedure are designated as theneo-aorta and neo-pulmonary artery.

Patients and operative techniquesDuring the 3 year period between September 1993 andAugust 1996, 17 patients with corrected transpositionunderwent surgery at UCSF. Patients with correctedtransposition who had other abnormalities that pre-cluded biventricular repair were excluded. In six of thesepatients, anatomical repair was not performed becauseof unfavourable anatomy for an arterial switch orRastelli procedure (n = 2), previous left ventricle topulmonary artery conduit repair (n = 2), biventriculardysfunction (n=l), and complete heart block with noassociated defects (n=l). Details of these patients andoperative procedures are summarized in Table 1. Adouble switch procedure was originally planned in twoof these patients, pending surgical exploration. Oninspection, one of these patients (patient 16, who hadpulmonary hypertension) was found to have chordalattachments of both atrioventricular valves to the rimof the septal defect, which was small and posteriorlylocated. In order to perform a Rastelli baffle, it wouldhave been necessary to enlarge the septal defect to theextent that atrioventricular valve integrity would havebeen compromised and complete heart block wouldhave been a likely consequence. In addition, the pulmo-nary valve was bicuspid and mildly regurgitant, rulingout an arterial switch. In the other patient (patient 17),attachments of the tricuspid valve to the septal crest andmuscular left ventricular outflow tract obstruction pre-cluded a Rastelli tunnel, while an arterial switch was outof the question due to a severely dysplastic and stenoticpulmonary valve.

The other 11 patients, aged 4-8 months to7-8 years (median 3-2 years), underwent anatomical

correction with a double switch procedure (Table 2). All11 patients had a malalignment ventricular septal defect,nine had either pulmonary stenosis (n = 5) or atresia(n=4), five had significant tricuspid valve pathology ordysfunction, and five had juxtaposition of the cardiacapex and venae cavae (apicocaval juxtaposition). Sixpatients had undergone one or more previous pro-cedures, including systemic to pulmonary shunts inthe four patients with pulmonary atresia and one patientwith severe pulmonary stenosis, pulmonary arterybanding (patient 6), aortic coarctation repair (patient 8),and bidirectional cavopulmonary shunt (patients 10and 11).

In these 11 patients, atrial switch was performedwith either a Senning (n = 7) or a Mustard (n = 4) pro-cedure. The internal or external Senning baffle wasaugmented with a patch of fresh or glutaraldehyde fixedautologous pericardium in four patients (patients 1, 2, 6,7), and with in situ autologous pericardium in twopatients (patients 8 and 9). The right atrial free wall wasaugmented in two patients who underwent a Mustardatrial switch (patients 3 and 10). In two patients, onewith situs solitus (patient 11) and one with situs inversus(patient 10), the Mustard procedure was modified be-cause of the presence of a previously performed bidirec-tional cavopulmonary shunt, so that only the inferiorvena cava was baffled to the tricuspid valve (Fig. 1).

The atrial switch was combined with an arterialswitch procedure in four patients, with patch closure ofthe ventricular septal defect. In three of these patients,standard coronary transfer techniques were performed.In the other patient (patient 6), the commisures of thepulmonary and aortic valves were not aligned, the leftcoronary artery arose directly opposite the neo-aorta,and the right coronary arose anteriorly, a long distancefrom the neo-aorta. Thus, Aubert's technique'251 wasapplied, with construction of a baffle from the neo-aortato the coronary orifices through a surgically createdaortopulmonary window. In the seven patients who didnot undergo an arterial switch procedure, ventriculo-arterial discordance was corrected with a Rastelli pro-cedure, in which the left ventricle was baffled to theaorta through the ventricular septal defect and the rightventricle was connected to the pulmonary artery with acryopreserved aortic allograft conduit (n = 6) or a rightventricular outflow tract patch with subvalvar muscleresection (n= 1). Conduit placement varied according tothe individual patient's anatomy, but was generally at alateral angle in order to minimize the chance of sternalcompression. The ventricular septal defect was enlargedin two Rastelli patients by removing a wedge of musclefrom the superior-anterior aspect of the defect. In allpatients, the ventricular septal patch or left ventricle toaorta baffle was sutured to the right side of the ventricu-lar septum in order to avoid the conduction axis'1926"27'.

Additional procedures performed are listed inTable 2. Total perfusion time ranged from 196 to356 min (median 269 min), and aortic cross-clamptime ranged from 136 to 249 min (median 188 min).Circulatory arrest was not employed.

Eur Heart J, Vol. 18, September 1997

tn

o

ia

Table 1 Diagnostic and surgical details of patients undergoing procedures other than double switch for corrected transposition

P',t,Vnt9 A S e Segmental v q n Pulmonary Tricuspid(years) anatomy outflow valve

Surgery Reason, double switchnot performed

Complications

12

13

14

15

16

17

26-8

22-7

1-8

39 2

5-4

4-2

(I,D,D)*

(S,L,L)

(S,L,L)

(S.L.L)

(S,L,L)*

(S,L,L)

Yes

Yes

No

No

Yes

Yes

PA, s/p LV-PArt conduit

PA, s/p LV-PArt conduit

Normal

Normal

subPS, small conus

severe PS, thickeneddysplastic valve

Mild TR

Normal

Normal

Severe TR

Attached tocrest, trace '

Attached toseptal crest

Conduit removal, LV-PArt conduit patchaugmentation, VSD and ASD closure

Conduit replacement (24 mm Ao)t,LPA augmentation, ASD closure

Unipolar epicardial pacemaker placement

Tricuspid valve replacement (33 mm St. Jude)

VSD closure, LV-PArt homograft (21 mm Ao)f,pulmonary artery ligation and division

Previous conduit repair

Previous conduit repair

CAVB only — no defects

Bioventricular dysfunction

Unfavourable anatomy

None

None

None

None

None

Unfavourable anatomy None

"Indicates patients with apicocaval juxtaposition, t'ndicates size and type of allograft conduit used.(I,D,D), (S,L,L) = segmental anatomy; Ao = aorta; ASD4=5atrial septal defect; CAVB=complete atrioventncular block; JET=junctional ectopic tachycardia; LV-PArt=left ventricle topulmonary artery; PA = pulmonary atresia; PAB = pulmonary artery band; PS = pulmonary stenosis; T/D=take- down; TR = tricuspid regurgitation; TS = tricuspid stenosis; VSD = ventricularseptal defect (malalignment).

Table 2 Diagnostic and surgical details of patients undergoing anatomical repair of corrected transposition

Patient , A g e . SeS™'l ta l VSD(years) anatomy outflow

Tricuspidvalve Surgery Complications

<o

Senning+Rastelli (18 mm Ao)f, VSD enlargement, resect supravalvar tricuspid ring Effusions

Senning+ASO, VSD closure JET

Mustard+ Rastelli (23 mm Ao)t, VSD enlargement, T/D shunt Cortical watershed

infarcts

Mustard+ Rastelli, resect subpulmonary muscle, VSD closure JET

Attached to septal crest Senning+ASO, VSD closure JET, effusions

;re TR Senning+ASO (Aubert procedure), VSD closure, T/D PAB None

Senning+Rastelli (23 mm Ao)|, T/D shunts Effusions

Senning+ASO, VSD closure Effusions

Senning + Rastelli (21 mm Ao)|), resect subpulmonary muscle, T/D shunts None

Modified Mustard+ Rastelli (20 mm Ao)t, resect subpulmonary muscle None

Modified Mustard + Rastelli (23 mm Ao)t, T/D shunts Effusions

'Indicates patients with apicocaval juxtaposition, tindicates size and type of allograft conduit used.(I,D,D), (S,L,L), (A,L,L:) = segmental anatomy; Ao = aorta; ASO= arterial switch operation; JET=junctional ectopic tachycardia; PA = pulmonary atresia; PAB = pulmonary artery band;PS = pulmonary stenosis; T/D = take- down; TR = tricuspid regurgitation; TS = tricuspid stenosis; VSD=ventricular septal defect (malalignment).

1

2

3

4

5

6

7

8

9

10

m 11c

11

0-4

7-0

0-4

0-4

3-2

4-2

11

5-8

7-8

7-4

(S.L.L)

(S,L,L)

(1,D,D)*

(A,L,L)*

(S,L,L)

(S,L,L)

(S,L,L)*

(S,L,L)

(S,L,L)

(I,D,D)*

(S,L,L)*

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

PS

Normal

Severe PS

subPS

PS

PAB

PA

subPS

PA

PA

PA

TR/TS

Mild TR

Normal

Normal

Attached

Ebstein's

Normal

Mild TR

Normal

Normal

Normal

I

1474 V. M. Reddy et al.

Figure 1. Diagram of double switch repair in patient 10,whose anatomy is situs inversus (I,D,D) with apicocavaljuxtaposition. This patient had previously undergone bi-directional cavopulmonary shunt, with anastomosis of thesuperior vena cava (SVC) to the pulmonary artery (PA).The modified Mustard intraatrial baffle (MM) directsinferior vena caval (IVC) return to the tricuspid valve andmorphological right ventricle (RV), and does not encom-pass the SVC orifice. An allograft valved conduit (C) isused to connect the RV to the PA. The modified Mustardbaffle allows ample room for drainage of pulmonaryvenous blood to the morphological left ventricle (LV) viathe mitral valve. The Rastelli baffle (RB), which is suturedto the right ventricular aspect of the septum, directs LVblood through the ventricular septal defect to the aorta(Ao).

Results

There was one early death among the 11 double switchpatients. This was patient 6, who had undergone Sen-ning repair and an arterial switch, with Aubert's pro-cedure instead of simple coronary transfer. The cause ofdeath was multi-organ failure and brain death secondaryto mediastinal haemorrhage and cardiac tamponade(Table 2). Other postoperative complications are sum-marized in Table 2. In six patients, the chest was leftopen because of extensive myocardial oedema or persist-ent bleeding following prolonged cardiopulmonary by-pass and cardioplegic arrest. Sternal closure wasperformed in the intensive care unit between 1 and 7days postoperatively. Duration of ventilator support

ranged from 1 to 9 days (median: 5 days), and post-operative hospital stay ranged from 8 to 49 days(median: 14 days).

At follow-up ranging from 1 to 34 months(median: 22 months), there were no deaths. Two patientshad undergone a total of three reoperations. One patientunderwent reoperation for (1) Rastelli baffle obstruc-tion 8 months post-repair, with patch augmentation ofthe intraventricular baffle and resection of obstructingmuscle bundles, and for (2) replacement of the rightventricle to pulmonary artery conduit 24 months post-operatively. Another patient underwent reoperation 11months after double switch to relieve right ventricularoutflow tract obstruction and right ventricular aneurysm(with a peak right ventricular pressure of 80 mmHg anda peak left ventricular pressure of 98 mmHg), with rightventricular myectomy, transannular patch, and pulmo-nary artery augmentation. On most recent follow-upechocardiography, all 10 patients had normal biven-tricular function, one patient had mild bilateral atrio-ventricular valvar insufficiency, one patient had mildmitral insufficiency, and one patient mild aortic insuffi-ciency. No patients were taking any cardiac medicationsand all were in New York Heart Association class I.

Discussion

Selection of patients for anatomical repair

Over the past several years a number of centres havereported favourable results with anatomical repair forcongenitally corrected transposition'18"241. Most sur-geons who perform the double switch procedure do notrecommend this approach in all patients with correctedtransposition, but primarily in patients with systemicright ventricular or tricuspid valvar dysfunction'22231.Clearly, it is these patients in whom anatomical repairoffers the greatest immediate benefit, in the form ofimproved systemic ventricular and atrioventricularvalvar function. However, the double switch proceduremay prove to be the treatment of choice in most patientswith corrected transposition. While there is no evidenceyet that anatomical repair will improve chronic ventricu-lar function in patients without a compromised rightventricle, the data of Imai et al. suggest that mid-termsystemic ventricular performance in patients with adouble switch may be somewhat better than in patientswho have undergone conventional repair'221. Similarly,though long-term data are not available, a double switchmay be effective not only in the presence of tricuspidregurgitation, but also in preventing regurgitation inpatients with tricuspid valvar anomalies that are likelyto lead to dysfunction.

In general, we prefer to perform anatomicalrepair in all patients with corrected transposition and nocontraindications, including those with intact systemicright ventricular and tricuspid valve function. Thisapproach is based largely on the literature regarding

Eur Heart J, Vol. 18, September 1997

Double switch procedure 1475

right ventricular failure and tricuspid dysfunction lateafter atrial repair of simple transposition1'4"181 and ourown experience with such patients (unpublished data).However, we realize that the double switch procedure isnot the optimal approach in all patients with correctedtransposition and two adequate ventricles. The relativelyhigh incidence of native pulmonary outflow tract pathol-ogy in corrected transposition'28' leaves many patientsunsuitable for an arterial switch procedure. Likewise,construction of a left ventricle to aorta baffle can becomplicated by an unfavourably sized or located ven-tricular septal defect, which may be difficult to enlargedue to the anterior position of the conduction axis incorrected transposition when viewed from the left-sidedright ventricle'261. Similarly, intraventricular baffle tech-niques may be difficult in the presence of anomalousatrioventricular valves, especially with chordae thatattach to the septal crest. A combination of these factorsprevented us from performing a double switch pro-cedure in two of our recent patients (patients 16 and 17).

Selection of surgical components foranatomical repair

Even when there is no contraindication to anatomicalrepair, it is important to select the proper componentsfor the atrial and ventricular/arterial switch procedures.The one death in our series of anatomical repair was thepatient in whom an arterial switch with the Aubertprocedure was performed. Although the coronaries andthe Aubert repair could not be clearly implicated in thepatient's demise, the poor outcome highlights the poten-tial importance of coronary artery anatomy in selectingpatients for double switch repair of corrected transposi-tion. As recently reported by McKay et alP9\ thesemilunar valve commissures are frequently malalignedin corrected transposition, resulting in coronary arteriesthat are sometimes poorly situated for transfer. Becausethe native pulmonary artery is usually deeply seatedbetween the atrioventricular valves, it is generally poss-ible to implant the coronaries above the sinotubularjunction and avoid any problem with commissures at thepoint of implantation. However, this was not the case inour patient, who had non-aligned semilunar valve com-missures and a left coronary artery originating directlyopposite the neo-aorta (native pulmonary artery). In thispatient, the right coronary also arose anteriorly and itwould have been necessary to turn this coronary back onitself in order to make the transfer. Anomalous coronaryanatomy is not necessarily a contraindication to ana-tomical repair, but in retrospect this patient may havebeen better served with a Rastelli procedure.

Probably the major shortcoming of the doubleswitch procedure is that it includes a Senning orMustard atrial switch, which may be complicated byvenoatrial obstruction'30'31' or supraventricular arrhyth-mias'32'33'. This might be particularly problematic inpatients with corrected transposition and apicocaval

juxtaposition, as was found to be the case in the seriesreported by Yagihara et a/.'2'1. However, there areoptions that might help reduce the frequency of suchcomplications. In our two most recent double switchprocedures, performed in patients who had previouslyundergone a bidirectional cavopulmonary shunt, theatrial switch component was a modified Mustard pro-cedure in which the intra-atrial baffle was not extendedto incorporate the orifice of the superior vena cava. As aresult, the atrial suture line was significantly less thanwith a standard Mustard or a Senning, and the region ofthe sinoatrial node was completely avoided. These modi-fications will probably reduce the likelihood of some ofthe more common electrophysiological complicationsof the Mustard operation'331. In addition, not extendingthe patch to the superior vena cava allowed more roomfor the pulmonary venous pathway. This is likely tominimize the potential for venoatrial obstruction'301,especially in patients with apicocaval juxtaposition,which was the case with the two patients in our serieswho underwent this modification. We (unpublisheddata) and others'341 have found the bidirectional cavo-pulmonary shunt to function well as an auxilliaryprocedure in the biventricular repair of a variety ofcongenital heart defects. Though the cavopulmonaryshunt had been previously performed in our twopatients, this procedure may be a useful addition to thedouble switch operation in certain circumstances, suchas patients with double-outlet right ventricle and a smallright ventricle or pulmonary outflow tract, patients withapicocaval juxtaposition, or as a means of avoidingpotential atrial complications. On the other hand, thereare also possible drawbacks to the use of a bidirectionalcavopulmonary shunt in patients with corrected trans-position. Because transection of the superior vena cavaeliminates the standard access pathway for transvenouspacemaker placement, the management of completeatrioventricular block in patients with corrected trans-position and bidirectional cavopulmonary shunt will bemore complicated than when a cavopulmonary shunt isnot present. We decided to leave the cavopulmonaryshunt in both of our patients in whom this procedurehad previously been performed, but it is too early todetermine whether this is preferable to reconstructingthe superior cavoatrial connection.

Anatomical repair vs conventional repair

The optimal means of comparing anatomical and con-ventional repair of corrected transposition would be arandomized clinical trial. However, given the individu-alized nature of these complex repairs, such a trial isunlikely to be performed. In comparing the literature onthe two approaches, it seems clear that outcomes arecomparable, and perhaps even better with anatomicalrepair. There have been five hospital deaths (9%) amongthe 56 patients reported in the literature to have under-gone double switch repair, including the presentreport'618"24'. These results are not noticeably different

Eur Heart J, Vol. 18, September 1997

7476 V. M. Reddy et al.

from outcomes in patients undergoing less aggressiverepair in which the right ventricle is left in the systemiccirculation (4-18%). One of the major technicaladvantages of anatomical repair is that, because themorphological right ventricle is switched to its naturalrole in the pulmonary circulation, the ventricular septaldefect may be closed either through a right ven-triculotomy (in cases of Rastelli repair) or through theneo-pulmonary valve (in cases of arterial switch). Thisallows more ready accessibility to the right ventricularaspect of the septum than do traditional techniques,which is important from the point of view of avoidingthe conduction tissues'19'26'271. Though it has becomestandard procedure to place the patch on the right sideof the septum with conventional repair as well, surgicalheart block remains a significant problem with such anapproach14"7'35'. In contrast, only two cases of surgicalheart block have been reported among the 56 patients inthe double switch literature16'18"241, including no cases inour 11 patients. Aside from surgical heart block, how-ever, it might be expected that there will be greatersurgical morbidity after the double switch procedure,which generally requires extended periods of cardiopul-monary bypass and cardioplegic arrest. Five of our 11patients required chest tube drainage of effusions forlonger than 7 days, three had a transient junctionalrhythm, and one suffered bilateral cortical watershedinfarcts. All of these problems had resolved by the timeof discharge, and there were no demonstrable differencesin post-hospital condition between double switch andnon-double switch patients.

Conclusions

We and others have described a variety of approaches toanatomical repair of corrected transposition, in patientswith the full range of anatomical variations, includingsitus solitus, situs inversus, situs ambiguous, and withthe usual and juxtaposed apicocaval relationships'18"241.The natural history of the right ventricle in patients withcorrected transposition is not fully understood, largelybecause of the unknown number of patients who areasymptomatic and do not have associated defects thatrequire intervention. As a result, the denominator isunclear. Though the evidence is strong that the rightventricle does not hold up well in patients with simpletransposition and a Senning or Mustard repair'14"181, it isdebatable whether this accurately reflects the situationwith corrected transposition'"1. As more data becomeavailable regarding outcomes following double switchprocedures in patients with various anatomical diag-noses, it may be possible to determine subsets of patientsin whom anatomical repair is indicated or contraindi-cated, or circumstances in which particular modifica-tions, such as the addition of a cavopulmonary shunt,may be more appropriate. Regardless, long-termfollow-up will be necessary to determine if the doubleswitch approach favourably influences the natural his-tory of corrected transposition when compared to less

aggressive surgical approaches in which the rightventricle and tricuspid valve remain in the systemiccirculation.

References

[1] Allwork SP, Bentall HH, Becker AE el al. Congenitallycorrected transposition of the great arteries: Morphologicstudy of 32 Cases. Am J Cardiol 1976; 38: 910-23.

[2] Fox LS, Kirklin JW, Pacifico AD, Waldo AL, Bargeron LM.Intracardiac repair of cardiac malformations with atrio-ventricular discordance. Circulation 1976; 54: 123-7.

[3] Marcelletti C, Maloney JD, Ritter DG, Danielson GK,McGoon DC, Wallace RB. Corrected transposition and ven-tricular septal defect: Surgical experience. Ann Surg 1980; 191:751-9.

[4] Williams WG, Suri R, Shindo G, Freedom RM. Morch JE,Trusler GA. Repair of major intracardiac anomalies associ-ated with atrioventricular discordance. Ann Thorac Surg1981; 31: 527-31

[5] Westerman GR, Lang P, Castaneda AR, Norwood WI.Corrected transposition and repair of associated intracardiacdefects. Circulation 1982; 66 (Suppl I): 1-197-202.

[6] Sano T, Riesenfeld T, Karl TR, Wilkinson JL. Intermediate-term outcome after intracardiac repair of associated cardiacdefects in patients with atrioventricular and ventriculoarterialdiscordance. Circulation 1995; (Suppl II): II-272-8.

[7] Termignon JL, Leca F, Vouhe PR et al. 'Classic' repair ofcongenitally corrected transposition and ventricular septaldefect. Ann Thorac Surg 1996; 62: 199-206.

[8] Masden RR, Franch RH. Isolated congenitally correctedtransposition of the great arteries. In: Hurst JW, ed. UpdateIII: The Heart. New York: McGraw-Hill, 1980; 59-83.

[9] Graham TP, Parrish MD, Boucek RJ et al. Assessment ofventricular size and function in congenitally corrected trans-position of the great arteries. Am J Cardiol 1983; 51: 244-51.

[10] Benson LN, Burns R, Schwaiger M et al. Radionuclideangiographic evaluation of ventricular function in isolatedcongenitally corrected transposition of the great arteries. Am JCardiol 1986; 58: 319-24.

[11] Peterson RJ, Franch RH, Fajman WA, Jones RH. Compari-son of cardiac function in surgically corrected and congeni-tally corrected transposition of the great arteries. J ThoracCardiovasc Surg 1988; 96: 227-36.

[12] Dimas AP, Moodie DS, Sterba R, Gill CC. Long-termfunction of the morphologic right ventricle in adult patientswith corrected transposition of the great arteries. Am Heart J1989; 118: 526-30.

[13] Kirklin JW, Barratt-Boyes BG. Cardiac Surgery. New York:Churchill Livingstone, 1992: 1511-33.

[14] Redington AN, Rigby ML, Odershaw P, Gibson DG,Shinebourne EA. Right ventricular function 10 years after theMustard operation for transposition of the great arteries:Analysis of size, shape, and wall motion. Br Heart J 1989; 62:455-61.

[15] Bender HW, Stewart JR, Merrill WH, Hamon JW, GrahamTP. Ten years' experience with the Senning operation fortransposition of the great arteries. Physiological results andlate follow-up. Ann Thorac Surg 1989; 47: 218-23.

[16] Martin RP, Quereshi SA, Ettedgui JA et al. An evaluation ofright and left ventricular function after anatomical correctionand intra-atrial repair operations for complete transpositionof the great arteries. Circulation 1990; 82: 808-16.

[17] Chang AC, Wernovsky G, Wessel DLet al. Surgical manage-ment of late right ventricular failure after Mustard or Senningrepair. Circulation 1992; 86 (Suppl II): H-140-9.

[18] Cochrane AD, Karl TR, Mee RBB. Staged conversion toarterial switch for late failure of the systemic right ventricle.Ann Thorac Surg 1993; 56: 854-62.

Eur Heart J, Vol. 18, September 1997

Double switch procedure 1477

[19] Ilbawi MN, DeLeon SY. Backer CL et al. An alternativeapproach to the surgical management of physiologically cor-rected transposition with ventricular septal defect and pulmo-nary stenosis or atresia. J Thorac Cardiovasc Surg 1990; 100;410-5.

[20] Di Donato RM, Troconis CJ, Marino B et al. CombinedMustard and Rastelli operations: An alternative approach forrepair of associated anomalies in congenitally corrected trans-position in situs inversus (I,D,D). J Thorac Cardiovasc Surg1992; 104: 1246-8.

[21] Yagihara T, Kishimoto H, Isobe F et al. Double switchoperation in cardiac anomalies with atrioventricular andventriculoarterial discordance. J Thorac Cardiovasc Surg1994; 107: 351-8.

[22] Imai Y, Sawatari K, Hoshino S, Ishihara K, Nakazawa M,Momma K. Ventricular function after anatomic repair inpatients with atrioventricular discordance. J Thorac Cardio-vasc Surg 1994; 107: 1272-83.

[23] Bove EL. Congenitally corrected transposition of the greatarteries: Ventricle to pulmonary artery connection strategies.Semin Thorac Cardiovasc Surg 1995; 7: 139-44.

[24] Stumper O, Wright JGC, De Giovanni JV, Silove ED, SethiaB, Brawn WJ. Combined atrial and arterial switch procedurefor congenital corrected transposition with ventricular septaldefect. Br Heart J 1995; 73: 479-82.

[25] Aubert J, Pannetier A, Couvelly JP, Unal D, Rouault F,Delarue A. Transposition of the great arteries: New techniquefor anatomical correction. Br Heart J 1978; 40: 204-8.

[26] Anderson RH, Becker AE, Arnold R, Wilkinson JL. Theconducting tissues in congenitally corrected transposition.Circulation 1974; 50: 911-23.

[27] de Leval MR, Bastos P, Stark J, Taylor JFN, Macartney FJ,Anderson RH. Surgical technique to reduce the risks of heart

block following closure of ventricular septal defect in atrio-ventricular discordance. J Thorac Cardiovasc Surg 1979; 78:515-26.

[28] Anderson RH, Becker AE, Gerlis LM. The pulmonary out-flow tract in classically corrected transposition. J ThoracCardiovasc Surg 1975; 69: 747-57.

[29] McKay R, Anderson RH, Smith A. The coronary arteries inhearts with discordant atrioventricular connections. J ThoracCardiovasc Surg 1996; 111: 988-97.

[30] Williams WG, Trusler GA, Kirklin JW et al. Early and lateresults of a protocol for simple transposition leading to anatrial switch (Mustard) repair. J Thorac Cardiovasc Surg1988; 95: 717-26.

[31] Satomi G, Nakamura K, Takao A, Imai Y. Two-dimensionalechocardiographic detection of pulmonary venous chennelstenosis after Senning's operation. Circulation 1983; 68:545-9.

[32] Turley K, Ebert PA. Total correction of transposition of thegreat arteries: Conduction disturbances in infants youngerthan three months of age. J Thorac Cardiovasc Surg 1978; 76:312-20.

[33] Vetter VL, Tanner CS, Horowitz LN. Electrophysiologicconsequences of the Mustard repair of d-transposition of thegreat arteries. J Am Coll Cardiol 1987; 10: 1265-73.

[34] VanArsdell GS, Williams WG, Maser CM et al. Superior venacava to pulmonary artery anastomosis: An adjunct to bi-ventricular repair. J Thorac Cardiovasc Surg 1996; 112:1143-8.

[35] McGrath LB, Kirklin JW, Blackstone EH, Pacifico AD,Kirklin JK, Bargeron LM. Death and other events aftercardiac repair in discordant atrioventricular connection.J Thorac Cardiovasc Surg 1985; 90: 711-28.

Eur Heart J, Vol. 18, September 1997