Ventricular function after anatomic repair in patients with atrioventricular discordance

Transcription

1 Ventricular function after anatomic repair in patients with atrioventricular discordance Since June 1989, 18 patients with atrioventricular discordant anomalies under 15 years of age underwent anatomic correction, or double switch operation, with 2 hospital deaths (11 %). Ages ranged from 1 year 4 months to 12 years (mean 7.2 years) and body weights ranged from 7.1 to 32.5 kg (mean 20.1 kg) at operation. Visceral situs was solitus in 12 and inversus in 6. Ventriculoarterial connection was discordant in 4 and double-outlet in 14. AU had associated congenital cardiac anomalies consisting of ventricular septal defect in 17, atrial septal defect in 9, pulmonary atresia in 10, pulmonary stenosis in 6, systemic tricuspid regurgitation in 9, mitral regurgitation in 6, bilateral superior venae cavae in 6, patent ductus arteriosus in 6, WoUf-Parkinson-White syndrome in 2, and congenital atrioventricular block in 1. Previous palliation had been performed 19 times in 11 patients including systemic-pulmonary shunts in 10 patients and pulmonary banding in 1 patient. Anatomic correction consisted in the Mustard procedure in 12 or the Senning procedure in 6 at the atrial level and in external conduit repair in 14, arterial switch operations in 3, or direct anastomosis between the pulmonary artery and right ventricle in 1 at the ventriculoarterial level. The arterial switch operation was indicated in patients with a normal pulmonary valve, and external conduit repair was indicated for patients with pulmonic stenosis or atresia. Anatomic right ventricular end-diastolic volume showed significant reduction from % ± 46.3 % of normal to 78.8 % ± 24.0 %, mainly owing to unloading of the ventricle after the operation (p < ). Right ventricular ejection fraction remained unchanged (56 % ± 8 % to 56 % ± 6 %). Anatomic left ventricular end-diastolic volume and left ventricular ejection fraction remained unchanged before and after the operation: volume % ± 50.0% of normal to 127.3% ± 32.8% and ejection fraction from 61 % ± 8% to 56% ± 8%. Cardiac index averaged 3.1 ± 0.5 L/min per square meter after the operation. The double switch operation is best indicated in patients with atrioventricular discordance associated with systeniic tricuspid regurgitation and/or with poor systemic right ventricular function. (J THoRAe CARDIOVASC SURG 1994;107: ) Yasuharu Imai, MD,a Kazuo Sawatari, MDa (by invitation), Shuichi Hoshino, MDa (by invitation), Kazuo Ishihara, MDa (by invitation), Makoto Nakazawa, MDb (by invitation), and Kazuo Momma, MDb (by invitation), Tokyo, Japan Atrioventricular (A V) discordant anomalies have unique features in the A V as well as ventriculoarterial From the Department of Pediatric " and Pediatric Cardiology,b The Heart Institute of Japan, Tokyo Women's Medical College, Tokyo, Japan. Read at the Seventy-third Annual Meeting of The American Association for Thoracic Surgery, Chicago, Ill., April 25-28, Address for reprints: Yasuharu Imai, MD, Department of Pediatric, The Heart Institute of Japan, Tokyo Women's Medical College, 8-1 Kawadacho, Shinjuku-ku, Tokyo 162, Japan by Mosby-Year Book, Inc /94 $ /6/52972 (V A) arrangements, in which the anatomic right ventricle (RV) serves as a systemic ventricle and the anatomic left ventricle (LV) as a pulmonic counterpart. Although conventional repair yielded excellent results on a shortterm basis, long-term systemic RV function is still a matter of concern. 1-6 Deterioration of systemic RV function associated with tricuspid regurgitation in patients with complete transposition of the great arteries after Mustard or Senning procedures 7-14 also suggests poor systemic R V function in A V discordance. The double switch operation for A V discordant anomalies has the advantage of creating nearly normal A V and V A relationships, in which the anatomic LV serves as a 1272

2 The Journal of Thoracic and Volume 107, Number 5 lmai et al Table I Age at Patient operation Visceral TR MR No. (yr) situs V A connection Associated anomalies (grade) (grade) I 2 Solitus {SLL} Discordant Congenital A V block 3/4 No 2 6 Solitus {SLL} Discordant VSD (perimembranous), SPS, B-SVC 3/4 1/4 (deft) 3 8 Solitus {SLL} DORV VSD (infundibular), PA, ASD, PAPVR No No 4 11 Inversus {IDD} DORV VSD (perimembranous), PA, ASD, PDA No 3/4 (deft) 5 9 Inversus {IDD} DORV VSD (perimembranous), PA, ASD, B-SVC 2/4 2/4 (deft) 6 5 Solitus {SLL} DORV VSD (perimembranous), PA, PFO, WPW 2/4 No 7 Solitus {SLL} Discordant VSD (perimembranous), PH, WPW, B-SVC 1/4 No 8 4 Solitus {SLL} DORV VSD (perimembranous), PS, ASD, PDA 1/4 1/4 9 8 Solitus {SLL} DORV VSD (infundibular), PA, ASD, PDA, B-SVC 1/4 No 10 7 Solitus {SLL} DORV VSD (perimembranous), PA, PFO No No 11 8 Solitus {SLL} DORV VSD (perimembranous), PA, ASD, B-SVC No No 12 8 Inversus {IDD} DORV VSD (perimembranous), PA, PFO 1/4 1/ Inversus {IDD} DORV VSD (perimembranous), PS, PFO No No Inversus {IDD} DORV VSD (perimembranous), PS, ASD, PDA Trivial No 15 9 Inversus {IDD} DORV VSD (perimembranous), PS, ASD, PDA No No 16 5 Solitus {SLL} DORV VSD (infundibular), PS, ASD, PSVT 1/4 1/ Solitus {SLL} DORV VSD (perimembranous), PFO, PDA No No 18 4 Solitus {SLL} Discordant VSD (perimembranous), PS, ASD, single RCA Trivial Trivial VA, Ventriculoarterial; TR. tricuspid regurgitation; MR, mitral regurgitation; AV, atrioventricular; VSD, ventricular septal defect; SPS, subpulmonary stenosis; B-SVC, bilateral superior vena cava; DORV, double-outlet right ventricle; PA, pulmonary atresia; ASD, atrial septal defect; PAPVR, partial anomalous pulmonary venous return; PDA, patent ductus arteriosus; PFO, patent foramen ovale; WPW, Wolff-Parkinson-White syndrome; PH, pulmonary hypertension; PS, pulmonary stenosis; PSVT, paroxysmal supraventricular tachycardia; RCA, right coronary artery; {SLL}, situs solitus, I-loop, and I-transposition; {IDD}, situs inversus, d-loop, and d-transposition. systemic ventricle and the anatomic R V functions as a pulmonic ventricle after the operation. The purpose of this report is to present our series of anatomic repair of A V discordance and to compare its postoperative ventricular function with that of the conventional external conduit repair in A V discordance and in complete transposition of the great arteries. Patients and methods From June 1989 to April 1993, 18 patients with A V discordance under 15 years of age underwent the double switch operation, consisting of a Mustard procedure in 12 patients and a Senning procedure in 6 at the atrial level and of an arterial switch operation in 3, an external conduit repair in 14, and direct anastomosis between the pulmonary artery and R V in 1 patient at the V A level. The Mustard procedure with patch augmentation of the functionally systemic atrium was selected for small right atrium, and arterial switch was done in patients with an intact pulmonary valve (Tables I and II). The patients comprised 10 boys and 8 girls. Ages at operation ranged from 1 year 4 months to 12 years (mean 7.2 ± 2.9 years), and body weights ranged from 7.1 to 32.5 kg (mean 20.1 ± 7.1 kg). Before the anatomic correction, palliative procedures were performed 19 times in 11 patients. These operations consisted of 16 shunt procedures in 10 patients, formalin fixation of a patent ductus in 1 patient, and pulmonary artery banding for preparation of an anatomic LV in 1 patient. A 2-year-old boy who had grade 3 systemic tricuspid regurgitation associated with a small ventricular septal defect (VSD) underwent two successive pulmonary arterial bandings for preparation of the LV 7 months before the anatomic correction. Pulmonary artery banding resulted in spontaneous closure of the VSD and reduction in tricuspid regurgitation at the time of the double switch operation. Associated anomalies were seen in all cases: VSD in 17, atrial septal defect in 9, pulmonary atresia in 10, pulmonary stenosis in 6, bilateral superior vena cava in 6, patent ductus arteriosus in 6, Wolff-Parkinson-White syndrome in 2, congenital A V block in 1, and single coronary artery in 1. Systemic tricuspid regurgitation was seen in 9, mitral regurgitation in 6, and aortic regurgitation in 4 patients. Situs solitus {SDD} was seen in 12 and situs inversus {ILL} in 6 patients. Risk factors regarding application of conventional repair by using the systemic R V were observed in 15 of 18 patients (83%), including tricuspid regurgitation, poor ejection fraction, small R V volume, or large R V volume (Table III). LV and R V function was evaluated in the cineangiogram before and 1 to 2 months after operation. The volume of the morphologically LV was calculated with the area-length method and that of the morphologically R V with Simpson's rule by the method of Graham and associates. 15 Volume occupied by papillary muscles for the LV was adjusted according to the method of Graham, Jarmakani, and Canent l6 as follows: When the measured LV volume (V) was less than 15 ml, adjusted LV volume was measured as follows: V' = V For LV volume more than 15 ml, the following equation was applied: V' = 0.733V. Similarly, volume for the morphologically RV, calculated by Simpson's method, was adjusted by the following formula: V' = 0.649V. Both adjusted ventricular volumes were then expressed as a percentage of the expected normal volume of the respective ventricle per body surface area (BSA) according to the formula by Nakazawa and associates!7 as follows:

3 Imai et al. The Journal of Thoracic and May 1994 Table II Patient Previous Atrial No. palliation switch V A switch and other procedures Date of operation Results PAB, re-pab Senning 2 No Mustard 3 R-BT, L-BT (modified) Mustard 4 R-BT, L-BT, central shunt Senning 5 No Senning 6 No Mustard 7 No Senning 8 No Senning 9 L-BT (modified) Mustard \0 R-BT, PDA fixation Mustard II L-BT, R-BT (modified) Mustard 12 L-BT, central shunt Senning 13 L-BT (modified) Mustard 14 L-BT Mustard 15 L-BT, R-BT (modified) Mustard 16 No Mustard 17 R-BT Mustard 18 No Mustard Arterial switch, pacemaker implantation Arterial switch, MVR with 23 mm SJM Rastelli, P APVR repair Rastelli, MVP (cleft closure), VSD enlarged Rastelli, MVP (cleft closure) + MAP Rastelli, VSD enlarged, Sealy Arterial switch, VSD closure, Sealy Rastelli, LP A angioplasty Rastelli Rastelli Rastelli, VSD enlarged, LP A angioplasty Rastelli Rastelli, VSD enlarged Rastelli, VSD enlarged Rastelli, VSD enlarged Rastelli, VSD enlarged, A VR, Sealy Rastelli, RP A angioplasty RVOTR June 1989 September 1989 May 1990 July 1990 July 1990 September 1990 January 1991 March 1991 June 1991 June 1991 September 1991 November 1991 November 1991 December 1991 January 1992 April 1992 May 1992 November 1992 HD (3 days) HD (33 days) PAB, Pulmonary arterial banding; MVR, mitral valve replacement; SJM, St. Jude Medical valve; HD, hospital death; R-BT, right Blalock-Taussig shunt; L-BT, left Blalock-Taussig shunt; PAPVR, partial anomalous pulmonary venous return; MVP, mitral valvuloplasty; VSD, ventricular septal defect; MAP, mitral annuloplasty; LPA, left pulmonary artery; PDA, patent ductus arteriosus; AVR, aortic valve replacement; RPA, right pulmonary artery; RVOTR, right ventricular outflow reconstruction. Normal LVEDV = 72.5 X BSA1.43 Normal RVEDV = 75.1 X BSA1. 43 where L VEDP and RVEDP are LV and R Vend-diastolic volume, respectively. Postoperatively, pulmonic ventricular volume and its ejection fraction were calculated by excluding the subvalvular portion of the external conduit in all types of external conduit repair both in conventional repair and in the double switch operation. Regarding stroke volume index, commonly expressed as milliliters per square meter of body surface area, ventricular end-diastolic volume (expressed as percent of normal value) multiplied by ejection fraction was arbitrarily termed as percent stroke volume index in this article for the sake of standardization and of easy comparison among patients with different body surface areas. All mean values were expressed as a mean ± 1 standard deviation, and the paired or unpaired t test was used for statistical analysis. Preoperative RV and LV pressures were equal in all patients; cardiothoracic ratio averaged 50.9% ± 6.4%, ranging from 40% to 63%; pulmonary artery index 1 8 ranged from 182 to 593 (mean ± mm 2 /BSA); systemic arterial partial pressure of oxygen ranged from 30 to 84 mm Hg with a mean of 43.6 ± 13.2 mm Hg; systemic arterial oxygen saturation ranged from 56% to 97% (mean 60.9% ± 7.8%); and pulmonary/systemic flow ratio ranged from 0.33 to 2.24 (mean 1.1 ± 0.6). Preoperative systemic R Vend-diastolic volume

4 The Journal of Thoracic and Volume 107, Number 5 Imai et al. I 275 ranged from 62% to 241%, with a mean of 122.2% ± 43.6%; RV ejection fraction ranged from 41% to 73% (mean 56% ± 8%); pulmonic LV end-diastolic volume ranged from 82% to 269%, averaging 134.6% ± 50.0%; and LV ejection fraction ranged from 40% to 75% (mean 61% ± 8%). Postoperative catheterization and angiography were performed in 15 patients from 18 to 67 days (mean 32.4 ± 12.9 days) after operation and in the other 2 patients 11h years later. Systemic ventricular function after the double switch operation was compared with that after conventional external conduit repair, bridging between the LV apex and pulmonary artery in 20 patients with A V discordance in our series. Also, ventricular function after the double switch operation, consisting of external conduit repair combined with the Mustard or Senning procedure, was compared with that in 23 patients with complete transposition with VSD and with reduced pulmonary blood flow who underwent external conduit repair approximately in the same period in our institution. Catheterization was performed 22 to 56 days (mean 30.8 ± 8.5 days) after the operation in patients with A V discordance who underwent conventional conduit repair and 22 to 95 days (mean 38.1 ± 19.6 days) after the operation in patients with transposition who underwent conduit repair. Operative techniques. Intracardiac repair was performed through a median sternotomy with the use of moderately hypothermic cardiopulmonary bypass (25 0 to 27 0 C) via ascending aortic and direct bicaval or tricaval cannulation. Glucose-insulin-potassium solution (20 ml/kg) was used to induce cardioplegia and was repeated in half the dose every 40 minutes thereafter. The cardioplegic solution contained 20 meq of potassium and 20 units of regular insulin in a liter of 5% dextrose solution. As an atrial switch procedure, the Senning repair was preferred in patients with a normal or large right atrium, and the Mustard procedure with liberal patch augmentation of the functionalleft atrium was chosen for patients with a small right atrium because of an overhanging anatomic LV in mesocardia or dextrocardia in situs solitus. At the V A level, an arterial switch operation was performed in patients with a normal pulmonary valve, and external conduit repair with intraventricular rerouting was preferred for those with pulmonary atresia or pulmonary valvular stenosis. In patient 18, RV outflow reconstruction with a monocusp patch was performed after direct anastomosis between the R V and distal pulmonary artery in an attempt to eliminate an external conduit. The combination of double switch procedures was as follows: Mustard and external conduit repair in 10 patients, Mustard and R V outflow reconstruction in I, Senning and external conduit repair in 4, Mustard and arterial switch operation in 1, and Senning with arterial switch in 2 patients. Atrial switch procedures. The Mustard procedure was performed through a vertical incision in the right atrium, and an additional transverse incision was made toward the confluence of the right pulmonary veins to crea te at-shaped opening. After liberal atrial septectomy, a systemic venous channel was created with a glutaraldehyde-treated equine pericardial baffle (Xenomedica AG, Luzern, Switzerland). The ostium of the coronary sinus was always incorporated in the systemic venous channel, because the anterior conduction system was the rule in A V discordance. In patients with bilateral superior venae cavae, the coronary sinus was cut back deep into the left atrium and its cranial wall was used as a pedicled flap to partially create a ceiling of the pulmonary venous channel in patients 2, 9, and 11. Table III Patient No. Risk factors of systemic RV against conventional repair TR grade 3/4, large RVEDV (213% of normal) 2 TR grade 3/ 4, Ebstein's anomaly, poor RVEF (41%) 3 Small RVEDV (74% of normal) 4 None 5 TR grade 2/4 6 TR grade 2/4 7 TR grade 1/ 4, large RVEDV (241% of normal) 8 TR grade 1/4, small RVEDV (74% of normal) 9 TR grade 1/4 io Poor RVEF (42%) II Small RVEDV (62% of normal) 12 TR grade 1/4 13 None 14 Trivial TR by Doppler echocardiogram IS Poor RVEF (42%) 16 TR grade 1/4 17 None 18 Trivial TR by Doppler echocardiogram RV, Right ventricle; TR, tricuspid regurgitation; RVEDV, right ventricular enddiastolic volume; RVEF, right ventricular ejection fraction. Similarly, in patient 3 partial pulmonary venous return into the right superior vena cava was corrected by diverting the inferior vena cava and persistent left superior vena cava into the anatomic left atrium by a partial Mustard baffle, in which the ostium of the right superior vena cava into the anatomic right atrium was kept and the right superior vena cava distal to the anomalous pulmonary venous connection was ligated. In all Mustard cases, the functional left atrium was deliberately enlarged with Xenomedica membrane or an autologous pericardial patch down to the confluence of the right pulmonary veins. In patients having the Senning operation, a standard Senning procedure was performed in patients 1,4,8, and 12, and cutback of the coronary sinus was added in patients 5 and 7, who had a persistent contralateral superior vena cava. Switch procedures at VA level. An arterial switch operation, done by the French maneuver in the presence of a normal pulmonary valve, was performed in patients 1,2, and 7. Mirror image of the normal coronary pattern was seen in all cases and round coronary buttons were implanted into punched-out holes in the facing wall of the pulmonary artery above the level of the sinus of Valsalva. The VSD was closed through the right atrial approach. Pledget-supported mattress sutures were used exclusively in all patients. Those sutures were placed on the anatomic R V side along the cranial and anterior margin of the defect and on the margin of LV aspect of the ventricular septum along its posterior margin. In external conduit repair, ventricular rerouting was performed with a double-layered Xenomedica membrane. The intraventricular conduit was also fixed in place with pledgetsupported mattress sutures placed exclusively on the RV side of the ventricular septum and then onto the aortic valve ring mainly with a running suture. Along the posterior margin of the defect, sutures were placed through the tricuspid valve ring and then directly on the margin of the VSD, disregarding the possibility of the posterior conduction system regardless of the vis-

5 Imai et al. The Journal of Thoracic and May 1994 ceral situs and of the V A connection. An extracardiac conduit was placed from the left to the ascending aorta in situs solitus and from the right to the ascending aorta in situs inversus. A VSD extending from the membranous septum to the aortic valve ring without a recognizable infundibular septum (total conus defect in our classification) was the best indication for rerouting, and this favorable type of VSD was seen in 3 of 15 patients undergoing intraventricular rerouting (patients 3, 9, and 16). In the remaining II patients, a VSD was located in the membranous portion of the septum with various degrees of infundibular extension, as seen in patients with ordinary tetralogy of Fallot. In patients 13 and 16, the defect was located in the middle of the infundibular septum (mid-conus defect according to our classification, or infundibular muscular defect). Provided the anterior conduction system was present in all cases, its protrusion on the R V side could be carved safely. The outlet portion of the septum generally was composed of infundibular septum and the original ventricular septum, in which the former covered the cranial aspect of VSD on the R V side. Because the infundibular septum should not contain conduction system, its resection should be done safely. It was our practice to evaluate the anatomic composition of the VSD in the LV aspect through a right atrial approach. Along the anterior margin of the VSD, the lower margin of the infundibular septum could occasionally be recognized as a second margin on the opposite side of the septum restricting the VSD. This portion of the infundibular septum could be resected. Also, a wedge resection of the ventricular septum proper at the posterocaudal corner of the VSD was naturally possible. In patients with doubleoutlet RV with AV discordance with or without pulmonary atresia, resection of the infundibular septum also could be performed after the structure on the LV surface was analyzed. Enlargement of the VSD was necessary in 4 patients (patients 4,6, II, and 15) and carving of the infundibular septum was performed in 2 (patients 13 and 14). An extracardiac conduit bearing three cusps was hand made with Xenomedica membrane on the operating table. 19 Diameters of the conduit ranged from 20 to 26 mm. In patient 18, the distal end of the pulmonary artery was brought down to the cranial margin of the right ventriculotomy to the left of the aorta and a direct anastomosis was performed to create an autologous posterior wall of the outflow tract, which then was enlarged with an autologous pericardial patch bearing a monocusp. Repair of AV valve. Mitral regurgitation resulting from a cleft anterior leaflet was repaired by suture closure and annuloplasty in patients 2, 3, and 4. Mitral valve replacement with a 23 mm St. Jude Medical valve (St. Jude Medical, Inc., St. Paul, Minn.) had to be done in patient 2 after unsuccessful valve repair. Mild to severe systemic tricuspid regurgitation was seen in 9 patients before the operation, and Reed's tricuspid annuloplasty was performed in 5 patients at the commissure between the septal and posterior leaflets. Other concomitant procedures. An accessory pathway was confirmed between the left atrium and anatomic RV by epicardial mapping, and Sealy's operation combined with cryoablation for type A Wolff-Parkinson-White syndrome was performed in patients 5 and 6. In patient 15, repeated episodes of paroxysmal supraventricular tachycardia were controlled by Sealy's operation and cryoablation, and prosthetic aortic valve replacement with a 19 mm St. Jude Medical valve was also per- formed. In patient 1, a pacemaker was implanted for congenital A V block. Results All but 2 patients survived the operation, and the hospital mortality rate was 11 %. One (patient 17) died of pulmonary hypertension and septicemia 33 days after the operation and another (patient 2) died of low cardiac output because of massive regurgitation of the neoaortic valve 3 days after the operation. In patient 17, who had associated pulmonary atresia, preoperative measurement of pulmonary vascular resistance index was erroneous. Pulmonary mean arterial pressure measured by insertion of a catheter through a Blalock-Taussig shunt yielded a low pressure of 8 mm Hg and the resultant pulmonary vascular resistance index was reported to be 1.3 Wood units. However, postoperative catheterization disclosed a prohibitive pulmonary vascular resistance index of 9.3 Wood units. In retrospect, preoperative catheterization data were reevaluated. The left pulmonary venous wedge pressure was found to be elevated (33 mm Hg) and pulmonary vascular resistance index was corrected to 9.4 Wood units, a figure that should have contraindicated total correction. Follow-up in 16 survivors ranged from 4 to 46 months (mean 22.9 ± 10.4 months). Currently 14 are in New York Heart Association functional class I and 2 are in class II. Reoperation for repair of superior vena caval obstruction 1 month after the operation and closure of residual shunts at both the atrial and ventricular levels 31 months later were performed in patients 10 and 4, respectively. None had permanent surgical A V block after the double switch operation. Regarding residual regurgitation of A V valves, postoperative tricuspid regurgitation was grade 1/4 in 3 patients by cineangiogram and was judged trivial by Doppler echocardiogram in the other 7 patients. Systemic mitral regurgitation was grade 1/4 in 4 patients on cineangiogram, and an additional 4 patients were found to have trivial regurgitation by Doppler echocardiogram. Postoperative parameters. After the double switch operation, the cardiothoracic ratio increased from 50.9% ± 6.4% to 58.9% ± 4.7% (p = ). Anatomic R Vend-diastolic volume showed a significant reduction from 122.2% ± 46.3% of normal to 78.4% ± 24.3% because of unloading of the ventricle and diversion of its outflow tract into the systemic ventricular outlet by intraventricular rerouting after the operation (p = ), and RV ejection fraction remained unchanged (56% ± 8% to 56% ± 6%). Postoperatively, systolic pressure in the R V showed a marked decrease to 44.7 ± 17.4 mm Hg and RV /LV systolic pressure ratio

6 The Journal of ThoraCic and Volume 107, Number 5 Imai et al LVEF RVEF ±.08 NS.56± ±.08 NS.57±.O I I I I Before After 0.31 Before After Fig. 1. Ejection fraction (EF) before and after double switch operation. LVEF showed insignificant decrease after operation, but remained within normal limits in the majority of cases. NS, Not significant. decreased to 0.49 ± 0.2. The anatomic LV end-diastolic volume and LV ejection fraction remained unchanged before and after operation: end-diastolic volume 134.6% ± 50.0% of normal to 127.3% ± 32.8% (p = ) and ejection fraction from 61% ± 8% to 56% ± 8% (p = ). Cardiac index averaged 3.1 ± 0.5 Ljm 2 after the operation, ranging from 2.0 to 4.4 Ljm 2 Changes in LV ejection fraction before and after the double switch operation showed an increase in 7 patients and a decrease in 8 patients (Fig. O. In 3 patients (patients 6, 11, and 12), medium-term follow-up catheterization was performed 1 to 11/2 years after the operation. In patient 6, LV ejection fraction increased from 0.42 at 1 month after operation to 0.50 at 1 year, and cardiac index also increased from 2.6 to 3.4 Ljmin per square meter. In the other 2 patients, exercise testing by supine bicycle ergometer was performed during catheterization. Marked increases in cardiac output compared with the resting state were observed. Cardiac index increased from 3.3 and 3.0 Ljmin per square meter at rest to 5.5 and 4.6 Ljmin per square meter after exercise in patients 11 and 12, respectively. Comparison with conventional repair in A YD. Systemic ventricular function after the double switch operation was compared with that of conventional external conduit repair between the LV apex and pulmonary artery in 20 patients with A V discordance who were younger than 15 years of age. Although changes in systemic ventricular end-diastolic pressure (11.9 ± 5.5 versus 12.9 ± 3.5 mm Hg), systemic ventricular end-diastolicvolume(127.3% ± 32.8% versus 104.5% ± 24.9%), systemic ventricular ejection fraction (0.56% ± 10% versus 0.53% ± 0.06%), and systemic percent stroke volume index (66.5% ± 13.3% versus 55.5% ± 15.0%) were statistically insignificant, all parameters of systemic ventricular function suggested better performance after the double switch operation (Fig. 2). Comparison with transposition of the great arteries. Ventricular function after the double switch operation, consisting of external conduit repair combined with the Mustard or Senning procedure, was compared with that in the 23 patients with complete transposition with VSD and with reduced pulmonary blood flow who had external conduit repair. Differences in right atrial pressure (10.0 ± 5.5 versus 8.4 ± 3.7 mm Hg), LV end-diastolic pressure (12.9 ± 4.8 versus 10.6 ± 4.0 mm Hg), pressure gradient between the LV and aorta (2.5 ± 5.3 versus 3.3 ± 4.2 mm Hg), pressure gradient between the RV and distal pulmonary artery (13.6 ± 13.5 versus 16.9 ± 11.3 mm Hg), RVjLV peak systolic pressure ratio (0.48 ± 0.21 versus 0.51 ± 0.14), RV ejection fraction (0.57 ± 0.06 versus 0.55 ± 0.09), and LV ejection fraction (0.56 ± 0.1 versus 0.56 ± 0.09) were statistically insignificant. However, cardiac index (3.2 ± 0.5 versus 3.7 ± 0.8 Ljm 2, p = ), RV end-diastolic volume (72.7% ± 19.4% versus 115.7% ± 47.0%, p = ), RV percent stroke volume index (41.1 % ± 11.2% versus 62.3% ± 23.3%, p = ), LV end-diastolic volume (122.5% ± 28.5% versus 170.5% ± 45.1%, p = ), and LV percent stroke volume index (66.5% ± 13.3% versus 94.1 ± 22.5%, p = ) were significantly larger in the transposition group (Fig. 3). These significant differences in ventricular volume in the two groups were merely reflections of increased volumes before the operation in the transposition group. Preoperatively, RV end-diastolic volume, RV percent stroke volume index, LV end-diastolic volume, and LV percent stroke volume index in patients with

7 1278 Imai et al. The Journal of Thoracic and May 1994 EDP EDV EF troke Volume mmhg Fig. 2. Systemic ventricular function after double switch operation was not significantly different from that after conventional Rastelli procedure in patients with A V discordance. However, all parameters were better in the double switch group. EDP, End-diastolic pressure; ED V, end-diastolic volume; EF, ejection fraction; NS, not significant; D-S, double switch; C-R, conventional Rastelli. mmllg 20 LVEDP LVEF % 250 LVEDV P<.OO5 Cardiac Index Wmm o P<.05 Fig. 3. Systemic LV function after double switch operation, consisting of external conduit repair and atrial switch in A V discordance, (AVD), was compared with that of conventional Rastelli repair in complete transposition of the great arteries (TGA). LV ejection fraction (LVEF) and end-diastolic pressure (LVEDP) were similar. LV end-diastolic volume (LVEDV) and cardiac index were significantly smaller in A V discordance but remained within normal limits. transposition were already larger than those in patients with atrioventricular discordance: R Vend-diastolic volume, 140.5% ± 41.7% versus 106.7% ± 26.6% (p = ); RV percent stroke volume index, 80.0% ± 21.2% versus 60.1% ± 15.2% (p = ); LV end-diastolic volume, 164.3% ± 62.3% versus 119.9% ± 35.2% (p = ); and LV percent stroke volume index, 99.6% ± 36.6% versus 73.2% ± 25.5% (p = ). These significant differences in ventricular volumes in the two groups before the operation were due to younger age at operation and to higher pulmonary / systemic flow ratio in the transposition group. A postoperative increase in LV volume after the Rastelli-type repair was observed in both the transposition and double switch groups. Ratios of LV end-diastolic volume before and after the operation were 113% ± 26% of the preoperative value in the transposition group and 108% ± 28% in the double switch group. These increases in LV end-diastolic volume were accompanied by corresponding reductions in R Vend-diastolic volume after the operation. Ratios of RV end-diastolic volume before and after the operation were 86% ± 28% in the transposition group and 67% ± 18% in the double switch group. Because the volume occupied by the intraventricular conduit, which was located in the outflow portion of the anatomic RV, was inevitably added to the postoperative LV end-diastolic volume and the same amount had to be subtracted from the RV end-diastolic volume, at least part of the changes in ventricular volumes would be caused by this mechanism, besides the unloading effect on R Vend-diastolic volume and diminution in tricuspid regurgitation. Also, addition of a noncontractile intraventricular conduit to LV volume accounted for postoperative diminution in LV ejection fraction in both the double switch and transposition groups. Arterial switch operation. In 2 patients who survived

8 The Journal of Thoracic and Volume 107, Number 5 Long-term fate of the anatomic RV as a systemic ventricle still remains controversial in patients with A V discordant anomalies. 1-6 Systemic R V failure usually associated with tricuspid regurgitation after the Mustard or Senning procedure in complete transposition of the great arteries was experienced with increasing frequency,7-14 and function of the failing R V can be normalized as the pulmonary ventricle by conversion from an atrial switch to an arterial switch operation. Also, our clinical experience in conversion from atrial switch to arterial switch after systemic R V failure in 6 patients with transposition, 5 of whom survived and were doing well, showed that the systemi~ LV greatly improved the quality of life in these patients. In view of the facts that systemic RV function rapidly deteriorates in association with tricuspid regurgitation and that tricuspid regurgitation of the systemic R V is one of the major risk factors in conventional surgical correction of associated anomalies in atrioventricular discordance,2,6 the double switch operation or anatomic correction carries the best hope for longevity. Anatomic correction in atrioventricular discordance as a combined atrial switch and Rastelli repair has been reported by Ilbawi and associates 20 in 1990 and by Di Donato and colleagues 21 in Combined arterial switch and atrial repair as an anatomic repair for A V discordance has never been reported to our best knowledge except for our preliminary report. 22 In our series, the first double switch operation, consisting of a Senning and arterial switch operation, was performed on a 2-year-old boy with atrioventricular discordance associated with grade 3 tricuspid regurgitation on June 12, 1989, with success; a combined Mustard and Rastelli-type repair was then performed on May 25, 1990, also with success (Figs. 4 and 5). In 1992, direct anastomosis between the distal pulmonary artery and cranial margin of the right ventriculotomy was also successfully performed, and this modification, similar to REV (reparation a l'etage ventriculaire) in TGA reported by Lecompte and coworkers,23 would further reduce long-term morbidity and need for reoperation by eliminating a tubular prosthetic matelmai et al the combined Senning and arterial switch operation (patients 1 and 7), R Vend-diastolic volume showed a marked decrease from 213% to 115% of normal in patient 1 and from 241 % to 125% of normal in patient 7 because of the unloading effect after the operation. Postoperative systemic LV function revealed excellent results. LV enddiastolic volume decreased from 189% to 154% of normal in patient 1 and 267% to 194% of normal in patient 7. Postoperative LV ejection fractions were 0.53 and 0.55 and cardiac indexes were 3.2 and 3.1 in patients 1 and 7, respectively. Discussion rial in the outflow reconstruction. Two types of anatomic repair, combined arterial switch and the Senning procedure and combined atrial switch and direct anastomosis between the R V and the distal pulmonary artery, seemed to be new combinations performed for atrioventricular discordance. Hemodynamics. Because hemodynamic parameters and ventricular function after the double switch operation in patients with atrioventricular discordance are similar to those in the group with complete transposition of the great arteries after conventional Rastelli-type repair in terms of right atrial pressure, LV end-diastolic pressure, pressure gradient between the LV and aorta, RV jl V peak systolic pressure ratio, and RV and LV ejection fraction, further application of anatomic repair in atrioventricular discordance seems to be justified. LV ejection fraction at 1 month after operation was lower than normal in 5 of 16 cases studied. Our normal value for LV ejection fraction is 0.63 ± 0.05 (mean ± standard deviation), and the normal range (mean ± 2 standard deviations) is set between 0.53 and However, the fact that the medium-term results in 3 patients showed improved LV function and a favorable increase in cardiac output and LV ejection fraction after exercise would indicate that temporary depression of LV function in the immediate postoperative period could improve. In comparison with results of the conventional Rastelli procedure in patients with transposition, LV ejection fraction in both groups was exactly the same and decrease in L V ejection fraction after conduit repair was partly due to the addition of a noncontractile intraventricular conduit to the LV volume. LV function in the double switch operation proved to be similar to that in the conventional Rastelli procedure. Also, the fact that all the patients are currently leading a normal life in New York Heart Association Class I or II encourages us to continue using anatomic correction in atrioventricular discordance. Arterial switch operation. Combined atrial and arterial switch operation is preferable to conventional repair in patients with tricuspid regurgitation, because conversion of the systemic R V to a pulmonic ventricle even without direct surgical intervention on the tricuspid valve provides an excellent solution to a hitherto complicated problem. In 2 surviving patients after the combined Senning and arterial switch operation (patients 1 and 7), RV end-diastolic volume showed a marked decrease because of the unloading effect after anatomic correction, and new systemic LV function also was excellent. LV end-diastolic volume decreased from 189% to 154% of normal in patient 1 and from 267% to 194% of normal in patient 7. Postoperative LV ejection fraction was 0.53 and 0.55 and cardiac index was 3.2 and 3.1 in patients 1 and 7, respectively. However, massive neoaortic regurgitation seen in

9 I mai et al. The Journal of Thoracic and May 1994 Fig. 4. Combined arterial switch and Senning procedure in patient I. Preoperative systemic right ventriculogram. (left upper panel) showed massive tricuspid regurgitation, and pulmonic left ventriculogram (left lower panel) showed pulmonary band for preparation of anatomic left ventricle. After double switch operation, right ventricle served as pulmonary ventricle (right upper panel), and left ventricle was connected to aorta (right lower panel). Permanent pacemaker was implanted for congenital atrioventricular block. AO, Aorta; TR, tricuspid regurgitation; RV, right ventricle; P A, pulmonary artery; P AB, pulmonary artery band; L V, left ventricle. patient 2 with subvalvular pulmonary stenosis contraindicates conversion of the untrained low-pressure pulmonic valve into a neoaortic valve. Because combination ofthe arterial switch and Senning procedures revealed excellent systemic LV function after the operation, universal application of this type of anatomic repair should be considered in cases of atrioventricular discordance associated with VSD and pulmonary hypertension in the future. Tricuspid regurgitation. Systemic tricuspid regurgitation was considered to be a major problem in the surgical treatment of patients with atrioventricular discordance, and Ebstein's anomaly of the systemic A V valve was not amenable to conventional valvuloplasty. 2, It is also worth mentioning that postoperative tricuspid regurgitation was clinically insignificant in all patients including one with Ebstein-like anomaly despite simple application of tricuspid annuloplasty in 5 of 9 cases with preoperative tricuspid regurgitation. Significant reduction in both the end-diastolic volume and peak systolic pressure after the double switch operation made the incompetent tricuspid valve competent. Although longer follow-up is warranted, the double switch operation should be best indicated for atrioventricular discordance especially in the presence of systemic tricuspid regurgitation or in patients with poor systemic R V function. VSD. Location and size of the VSD was important in this type of repair. In isolated VSD in atrioventricular discordance, we had often seen the typical perimembranous VSD with a long infundibular septum. However, in cyanotic patients with atrioventricular discordance, some degree of infundibular extension of the VSD seemed to be the rule, partly by deviation of the infundibular septum toward the pulmonary artery in patients with pulmonic stenosis or by maldevelopment of the infundibulum in

10 The Journal of ThoraCic and Volume 107, Number 5 Imai et al Fig. 5. Combined external conduit repair and Mustard procedure in patient 6. Preoperative left and right ventriculogram (left upper and lower panels) showed dextrocardia, double-outlet right ventricle, pulmonic stenosis, and tricuspid regurgitation. Postoperative cineangiograms (right upper and lower panels) indicated anatomic ventriculoarterial connections and unobstructed outflow tracts of both ventricles. AO, Aorta; RV, right ventricle; P A, pulmonary artery; LV, left ventricle. patients with double-outlet RV and pulmonary atresia. 27 This infundibular extension of varying degree is not necessarily unique to Asian populations and is also clearly shown on the angiogram reported by Ilbawi and his associates. 2o In cyanotic patients, ejection by the anatomic LV into left-sided anteriorly originating aorta, predisposing straight alignment of the ventricle to the aorta, resulted in dextrocardia or mesocardia in its cardiac position in {SLL} and in levocardia or mesocardia in {IDD}. Conduction system and postoperative arrhythmia. General agreement has been reached concerning the anterior conduction system in {SLL}.28 On the other hand, the posterior conduction system has been reported in atrioventricular discordance in situs inversus. 29, 30 In our series of atrioventricular discordance including conventional Rastelli repair, there were 11 cases of {IDD} type, in which the VSD was either closed or rerouted. We merely preserved the anterior conduction system and disregarded the posterior counterpart, but sinus rhythm could be maintained in all cases after the operation. Because there was no difference in A V septal alignment either in {SLL} or in {IDD}, an anatomic explanation seemed lacking to support the different conduction pattern in all cases of {IDD}: However, there was a possibility of the dual pathway forming a sling in atrioventricular discordance. The exact prevalence of posterior conduction in a large series of patients, especially with a good A V septal alignment, should be studied in detail. Regarding patients with Wolff-Parkinson-White syndrome or with preoperative paroxysmal supraventricular tachycardia, recurrence of tachyarrythmia was seen in patient 7 and was controlled with antiarrhythmic medication. In the remaining 2 patients, arrhythmia surgery combined with cryoablation successfully prevented recurrence. In summary, although the double switch operation is a

11 1282 Imai et al. The Journal of Thoracic and May 1994 time-consuming procedure and longer follow-up is warranted, it is best indicated for patients with atrioventricular discordance, especially in the presence of systemic tricuspid regurgitation or in patients with poor systemic R V function. Preoperative tricuspid regurgitation improved spontaneously without surgical intervention on the valve, and mere unloading of the anatomic R V was sufficient to control the regurgitation in our series. Postoperative ventricular function can be favorably compared with that of the conventional Rastelli procedure in complete transposition of the great arteries. Combined atrial switch and arterial switch procedures were performed in 3 patients, and 2 surviving patients are currently doing well in New York Heart Association functional class I. This combination anatomic repair should be considered for patients with atrioventricular discordance associated with VSD and pulmonary hypertension. Two types of anatomic repair, combined arterial switch and Senning procedure and combined atrial switch and direct anastomosis between the R V and the distal pulmonary artery, are new combinations reported for the treatment of patients with atri?ventricular discordance. REFERENCES 1. Graham TP, Parrish MD, Bonchek Rl, et al. Assessment of ventricular size and function in congenitally corrected transposition of the great arteries. Am 1 Cardiol 1983; 54: McGrath LB, Kirklin lw, Blackstone EH, Pacifico AD, Kirklin lk, Bargeron LM. Death and other events after cardiac repair in discordant atrioventricular connection. 1 THORAC CARDIOVASC SURG 1985;90: Benson LN, Burns R, Schwaiger M, et al. Radionuclide angiographic evaluation of ventricular function in isolated congenitally corrected transposition of the great arteries. Am 1 CardioI1986;58: Peterson Rl, Franch RH, Fajman W A, 10nes RH. Comparison of cardiac function in surgically corrected and congenitally corrected transposition of the great arteries. 1 THORAC CARDIOVASC SURG 1988;96: Dimas AP, Moodie DS, Sterba R, Gill CC. Long-term function of the morphologic right ventricle in adult patients with corrected transposition of the great arteries. Am Heart ;118: Lundstrom U, Bull C, Wyse RKH, Somerville 1. The natural and "unnatural" history of congenitally corrected transposition. Am 1 CardioI1990;65: Trusler GA, The Mustard procedure: still a valid approach. In: Congenital heart surgery: current technique and controversies. Pasadena: Appleton Davies, 1984: Mee RBB. Severe right ventricular failure after Mustard or Senning operation: two-stage repair. 1 THORAC CARDIO V ASC SURG 1986;92: Graham TP, Atwood GF, Boucek Rl, Beoerth RC, Bender HW. Abnormalities of right ventricular function follow- ing Mustard's operation for transposition of the great arteries. Circulation 1975;52: Godman Ml, Friedli B, Pasternac A, Kidd BSL, Trusler GA, Mustard WT. Hemodynamic studies in children four to ten years after the Mustard operation for transposition of the great arteries. Circulation 1976;53: Hagler Dl, Ritter DG, Mair DD, et al. Right and left ventricular function after the Mustard procedure in transposition of the great arteries. Am 1 CardioI1979;44: Parrish MD, Graham TP, Bender HW, lones lp, Patton 1, Partain CL. Radionuclide angiographic evaluation of right and left ventricular function during exercise after repair of transposition of the great arteries: comparison with normal subjects and patients with congenitally corrected transposition. Circulation 1983;67: Benson LN, Bonet 1, McLaughlin P, et al. Assessment of right ventricular function during supine bicycle exercise after Mustard's operation. Circulation 1982;65: Ramsay 1M, VenablesAW, KellyMl, KalffV. Right and left ventricular function at rest and with exercise after the Mustard operation for transposition ofthe great arteries. Br Heart ;51: Graham TP, larmakani 1M, Atwood GF, Canent RV. Right ventricular volume determinations in children: normal values and observations with volume or pressure load. Circulation 1973;47: Graham TP, larmakani 1M, Canent RV. Left heart volume estimation in infancy and children: reevaluation of methodology and normal values. Circulation 1971 ;43: Nakazawa M, Marks RA, Isabel-lones 1, larmakani 1M. Right and left ventricular volume characteristics in children with pulmonary stenosis and intact ventricular septum. Circulation 1976;53: Nakata S, Imai Y, Takanashi Y, et al. A new method for the quantitative standardization of cross-sectional areas of the pulmonary arteries in congenital heart disease with decreased pulmonary blood flow. 1 THORAC CARDIOVASC SURG 1984;88: Harada Y, Kawada M, Ishihara K, Higashidate M, Kurosawa H, Imai Y. A new valved conduit with commissures using a glutaraldehyde preserved equine pericardium. Kyobu Geka 1989;42:457-9 (in lapanese). 20. Ilbawi MN, DeLeon SY, Backer CL, et al. An alternative approach to the surgical management for physiologically corrected transposition with ventricular septal defect and pulmonary stenosis or atresia. 1 THORAC CARDIOV ASC SURG 1990;100: Di Donato R, Troconis C, Marino B, et al. Combined Mustard and Rastelli operations: an alternative approach for repair of associated anomalies in congenitally corrected transposition in situs inversus (IDD). 1 THORAC CARDIO VASC SURG 1992;104: Yamagishi Y, Imai Y, Hoshino S, et al. Anatomic correction of atrioventricular discordance. 1 THORAC CARDIO VASC SURG 1993;105: Lecompte Y, Zannini L, Hanzan E, et al. Anatomic

12 The Journal of Thoracic and Volume 107, Number 5 [rnai et al correction of transposition of the great arteries: a new technique without use of a prosthetic conduit. J THORAC CAR. DIOVASC SURG 1981;82: Anderson KR, Danielson GK, McGoon D, Lie JT. Ebstein's anomaly of the left-sided tricuspid valve: pathological anatomy of the valvular malformation. Circulation 1987;58(Pt 2): Allwork SP, Bentall HH, Cameron H, Gerlis LM, Wilkinson JL, Anderson RH. Congenitally corrected transposition of the great arteries: morphologic study of 32 cases. Am J CardioI1976;38:91O Westerman GR, Lang P, Castaneda AR, Norwood WI. Corrected transposition and repair of associated intracardiac defects. Circulation 1982;66(Pt 2): Okamura K, Konno S. Two types of ventricular septal defect in corrected transposition of the great arteries: reference to surgical approaches. Am Heart J 193;85: Anderson RH, Becker AE, Arnold R, Wilkinson JL. The conducting tissue in congenitally corrected transposition. Circulation 1974;50: Dick M, Van Praagh R, Rudd M, Folkerth T, Castaneda AR. Electrophysiologic delineation of the specialized A V conduction system in two patients with corrected transposition of the great arteries in.situs inversus {I,D,D}. Circulation 1977;55: Thiene G, Nava A, Rossi L. The conduction system in corrected transposition with situs inversus. Eur J CardioI1977; 6:57-'70. Discussion Mr. Jaroslav F. Stark (London, England). We have recently reviewed our experience at Great Ormond Street with 90 consecutive patients with A V discordance operated on between 1974 and Two incremental risk factors for death had high significance-anatomically abnormal tricuspid valve and tricuspid valve incompetence. These observations would certainly support your idea that the so-called double switch operation is preferable to the standard repair. Also, as you are closing the VSD from the RV side, I presume the risk of complete heart block would also be reduced. You have reported one pacemaker implantation. Was this the result of surgical heart block or was the patient already in complete heart block before the operation? Mr. Marc R. de Leval (London, England). Some years ago I looked at a number of hearts with A V discordance to define the relationship between the aortic valve and the VSD. My intention was to try to decide preoperatively which VSDs could be closed through the aorta. I noticed a great anatomic variety of relationships. Quite a few hearts had a large infundibular septum between the aorta and the VSD. Furthermore, in a number of specimens there was a dysplastic tricuspid valve with chordal attachments on the margin of the VSD and/or the infundibular septum. Sometimes the defect was partially obstructed by the tricuspid valve. My question is this: Do you think that you can do a Rastelli procedure in all patients or do you think that some of them should be excluded? Furthermore, if you want to enlarge a VSD, where do you make the incision, knowing that the conduction system runs along the anterosuperior aspect of the defect? Dr. Serafin Y. DeLeon (Maywood, Ill.). This just goes to show that surgeons really think alike. When we reported our experience in 1990, we did not call it a new approach because of the possibility that somebody had done it before. Your experience is nice. Combining the arterial switch and Mustard procedure is the ideal thing to do. In certain cases, however, a Rastelli procedure using a conduit is needed. In these cases, instead of using a conduit, we have uprooted the main pulmonary artery and valve, swinging it to the RV. We do this with the hope that these patients will never require further operations. Conduits will be outgrown or degenerate and will have to be replaced. Dr. Imai. First, I would like to thank Mr. Stark for his kind comments. Regarding heart block, one patient with congenital heart block required pacemaker implantation. However, we had no case in this series in which surgical block was induced. As far as tricuspid regurgitation is concerned, I completely agree with your view. We did have the same dismal follow-up study in patients with the systemic tricuspid regurgitation, whether it was repaired by the valvuloplasty or valve replacement. That is one reason why we started this kind of approach in this subset. In response to Mr. de Leval, enlargement of the VSD is sometimes difficult. And the types of VSDs found in the A V discordant anomalies varies a great deal. However, if you select patients with a VSD and pulmonic stenosis or atresia, most of the patients do have a so-called malaligned-type VSD, in that the length of the infundibular septum is not really long. Then, as you know, the ventricular septum is composed of ventricular septum proper overlapped by the infundibular septum on its R V side. As long as you are resecting the infundibular septum, not the ventricular septum proper, it is safe. In patients with the normal visceroatrial situs, the posterior margin of the VSD can be enlarged at your will, provided anterior conduction is the rule. Dr. DeLeon, I quite agree with your view. It is our custom now, so as to avoid prosthetic material while doing a conventional Rastelli repair in the complete transposition of the great arteries, to pull the distal part of the pulmonary artery directly on the RV, as suggested by Lecompte. In the last patient of our series, usage of an external conduit was avoided by direct anastomosis.