- Email: [email protected]
Long-Term Results After the Rastelli Repair for Transposition of the Great Arteries
PEDIATRIC CARDIAC SURGERY:
Long-Term Results After the Rastelli Repair for Transposition of the Great Arteries Jürgen Hörer, MD,* Christian Schreiber, MD, PhD,* Eva Dworak, MD, Julie Cleuziou, MD, Zsolt Prodan, MD, Manfred Vogt, MD, PhD, Klaus Holper, MD, PhD, and Rüdiger Lange, MD, PhD Departments of Cardiovascular Surgery, and Pediatric Cardiology and Congenital Heart Disease, German Heart Center Munich at the Technical University, Munich, Germany
Background. This study sought to assess risk factors for late mortality after the Rastelli operation for patients with transposition of the great arteries, ventricular septal defect, and left ventricular outflow tract obstruction. Methods. Records of 39 patients who underwent the Rastelli operation between 1977 and 2004 were reviewed. Median age at the time of operation was 5.1 years (2.2 years within the last 5 years). Results. There were no early deaths. During a median follow-up of 8.9 years (range, 0 to 25 years), 2 patients died of sudden death, 1 of pneumonia, 1 during reoperation, and 2 received heart transplantation. Freedom from death or transplantation was 93.8% ⴞ 4.3% and 57.5% ⴞ 15.1% at 10 and 20 years, respectively. Freedom from conduit replacement was 48.8% ⴞ 10.3% and 32.5% ⴞ 10.3% at 10 and 20 years, respectively. Subvalvular and valvular left ventricular outflow tract obstruction (p ⴝ 0.012), stenosis of the peripheral pulmonary arteries (p < 0.001), enlargement of the ventricular septal defect
(p ⴝ 0.030), and longer ischemic time (p ⴝ 0.015) were predictive for death or transplantation. Patients younger than 4 years at the time of the Rastelli operation showed a trend toward a better freedom from death or transplantation (p ⴝ 0.068), but needed significantly more conduit replacements (p ⴝ 0.038) compared with patients 4 years or older. Conclusions. The Rastelli operation is a low-risk procedure with regard to early mortality. The status of the pulmonary arteries and ventricular septal defect enlargement are predictive for long-term survival. Patients 4 years of age or older at the time of the Rastelli operation require fewer reoperations for conduit exchange. Nevertheless, early Rastelli repair is recommended because patients 4 years or older are at risk for a higher long-term mortality.
A
successful operation for this special subset of patients, making the left ventricle as the subaortic ventricle. The Rastelli operation, in its original description, consisted of connecting the left ventricle to the aorta by suturing a polytetrafluoroethylene (Teflon) patch between the VSD and the aortic orifice, and of reconstructing the continuity from the right ventricle to the pulmonary artery with a homograft [4]. Although other surgical procedures, like the “réparation à l’étage ventriculaire” [5] and the Nikaidoh operation [6] were introduced, the Rastelli operation remained the preferred treatment in many congenital cardiac units for the past three decades. The disadvantages of the Rastelli operation, however, consist in the reconstruction of the left ventricular outflow tract with an intracardiac prosthetic baffle, and in the reconstruction of the right ventricular outflow tract with an extracardiac conduit, which may lead to subsequent reoperations for LVOTO and for right ventricular outflow tract obstruction (RVOTO). In the early experience, the Rastelli operation was often delayed to allow
pproximately 4% of neonates and infants who undergo surgery for transposition of the great arteries (TGA) also present with an additional ventricular septal defect (VSD) and a left ventricular outflow tract obstruction (LVOTO) [1]. Early survival of these patients without surgery is superior compared with children who present with a simple TGA without VSD, as the interventricular communication allows for better mixing of systemic and pulmonary venous return [2]. Moreover, the pulmonary vascular bed is protected by the LVOTO, which also keeps the left ventricle in training for supporting the systemic circulation after biventricular correction. In 1969 Rastelli and associates [3] introduced the first
Accepted for publication Jan 29, 2007. *Doctors Hörer and Schreiber contributed equally to the study. Address correspondence to Dr Hörer, Department of Cardiovascular Surgery, German Heart Center Munich at the Technical University, Lazarettstrasse 36, Munich, D-80636 Germany; e-mail: hoerer@ dhm.mhn.de.
© 2007 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2007;83:2169 –75) © 2007 by The Society of Thoracic Surgeons
0003-4975/07/$32.00 doi:10.1016/j.athoracsur.2007.01.061
CARDIOVASCULAR
The Annals of Thoracic Surgery CME Program is located online at http://cme.ctsnetjournals.org. To take the CME activity related to this article, you must have either an STS member or an individual nonmember subscription to the journal.
2170
HÖRER ET AL LONG-TERM OUTCOME AFTER RASTELLI REPAIR
Ann Thorac Surg 2007;83:2169 –75
Table 1. Palliative Procedures Before Rastelli Operation
CARDIOVASCULAR
Type of Procedure
n
mBTS mBTS ⫹ atrial septectomy cAPS cAPS ⫹ ligation of arterial duct cAPS ⫹ atrial septectomy cAPS ⫹ commissurotomy of the pulmonary valve cAPS ⫹ ligation of arterial duct ⫹ atrial septectomy cAPS ⫹ ligation of arterial duct ⫹ atrial septectomy, mitral valve repair Atrial septectomy Balloon atrial septostomy
13 3 9 1 1 2 1 1
cAPS ⫽ central aorta to pulmonary artery shunt; Blalock-Taussig shunt.
1 20
mBTS ⫽ modified
larger conduits to be placed. This might have resulted in poor long-term outcome owing to the deleterious effect of prolonged cyanosis and volume overload after palliative procedures [7]. The aim of the present study was to evaluate potential risk factors for late mortality and reoperations after the Rastelli operation for correction of TGA with VSD and LVOTO during a 25-year period.
Patients and Methods Patient Selection All patients who underwent a Rastelli operation for TGA with LVOTO and VSD between 1977 and 2004 at the German Heart Center Munich were included in a retrospective follow-up study. Patients who underwent the Rastelli repair for double-outlet right ventricle with malposition of the great arteries (n ⫽ 19) were excluded. Double-outlet right ventricle with malposition of the great arteries was differentiated from TGA with VSD if there was more than 50% overriding of the pulmonary artery relative to the ventricular septum, and if there was a discontinuity between the mitral valve and the pulmonary valve, and thus presence of a subaortic and subpulmonary conus. Furthermore, 3 patients from foreign countries were excluded because follow-up data could not be generated. This study was approved by the ethics committee of the Technical University Munich. Individual consent for the study was obtained from each patient.
Study Group The study group was made up of 21 male and 18 female patients. Thirty-two palliative operations were performed in 23 patients (Table 1). The VSD was located toward the outlet in all patients, with an inlet extension in 5 patients (12.8%). Five patients (12.8%) presented with pulmonary atresia. The remaining patients presented with LVOTO. Eight patients (20.5%) presented with stenosis of the pulmonary arteries (Table 2). Median age at the time of operation was 5.1 years for all patients (range, 1 to 23 years), and 2.2 years for the patients operated on
within the last 5 years of the study period. Median weight was 16.7 kg (range, 6 to 54 kg). Patients 4 years of age or older at the time of the Rastelli operation presented with significantly higher hemoglobin levels (19.5 ⫾ 2.8 g/dL), compared with patients younger than 4 years (17.7 ⫾ 2.4 g/dL; p ⫽ 0.040).
Surgical Technique All operations were performed by means of a midline sternotomy using conventional cardiopulmonary bypass with bicaval cannulation and crystalloid cardioplegic solution. The median cardiopulmonary bypass time was 136 minutes (range, 89 to 330 minutes). One patient (cardiopulmonary bypass time, 330 minutes) underwent aortic valve replacement with a xenograft conduit in addition to the Rastelli operation. The median ischemic
Table 2. Characteristics of 39 Patients Who Underwent the Rastelli Operationa Variable Morphology Inlet extension of VSD One or more additional muscular VSDs LVOTO subvalvular LVOTO valvular LVOTO subvalvular and valvular Pulmonary atresia Stenosis of the main pulmonary arteries within the hilar bifurcation Stenosis of the peripheral pulmonary arteries Hemodynamics at cardiac catheterization Right ventricular to left ventricular pressure ratio Pulmonary to systemic pressure ratio Left ventricular to pulmonary artery gradient (mm Hg) Right ventricular to aorta gradient (mm Hg) Pulmonary to systemic flow ratio Aortic oxygen saturation (%) Pulmonary artery oxygen saturation (%) Hemoglobin level (g/dL) Surgical Prior palliative operation Year of operation before 1995 Age at the time of Rastelli operation (y) Body weight at the time of Rastelli operation (kg) VSD enlargement Xenograft French maneuver Cardiopulmonary bypass time (min) Aortic cross-clamp time (min) Postoperative atrioventricular block
Value 5 (12.8%) 5 (12.8%) 5 (12.8%) 3 (7.8%) 26 (66.7%) 5 (12.8%) 8 (20.5%) 5 (12.8%) 1.00 ⫾ 0.06 0.23 ⫾ 0.11 70.8 ⫾ 21.3 2.7 2.1 79.7 85.9 18.7
⫾ ⫾ ⫾ ⫾ ⫾
4.2 1.9 5.6 7.1 2.8
23 (59.0%) 19 (48.7%) 5.8 ⫾ 5.1 18.0 ⫾ 11.0 34 (87.2%) 10 (25.6%) 4 (10.3%) 154 ⫾ 52 93 ⫾ 33 5 (12.8%)
Values are given as mean ⫾ standard deviation or as number with frequencies.
a
LVOTO ⫽ left ventricular outflow tract obstruction; lar septal defect.
VSD ⫽ ventricu-
Ann Thorac Surg 2007;83:2169 –75
HÖRER ET AL LONG-TERM OUTCOME AFTER RASTELLI REPAIR
2171
CARDIOVASCULAR
Fig 1. Estimated survival (A) and estimated survival as a function of age at the time of correction (B).
time was 83 minutes (range, 46 to 171 minutes). All operations were performed with the use of hypothermia with a mean rectal temperature of 23.7° ⫾ 1.5°C. Any systemic to pulmonary artery shunts were ligated. In 23 patients (59.0%), the VSD was enlarged at the anterosuperior margin of the defect. Cardiopulmonary bypass time (p ⫽ 0.76) and ischemic time (p ⫽ 0.13) were not significantly longer in patients who received enlargement of the VSD. The VSD was baffled to the aorta using a polyethylene terephthalate fiber (Dacron) patch in 34 patients (87.2%) and a polytetrafluoroethylene patch (Gore-Tex, W.L. Gore & Assoc, Flagstaff, AZ) in the remaining patients. The patch was fixed by a running suture in 19 patients (48.8%), by interrupted sutures in 5 patients (12.8%), and by a combination of both in the remaining patients. The right ventricular to pulmonary
artery connection was reconstructed with a Hancock porcine graft in 10 patients (25.6%) and with a homograft in 29 patients (74.4%). The mean size of the porcine graft was 17.4 ⫾ 4.8 mm (range, 12 to 25 mm), and the mean size of the homografts was 17.4 ⫾ 3.2 mm (range, 12 to 24 mm). The conduit was positioned to the left side of the aorta in all patients. The pulmonary bifurcation was placed anteriorly to the aorta (French maneuver [5]) in 4 patients (10.4%).
Follow-Up Final follow-up was obtained by a written questionnaire or by telephone and conducted between September 2004 and March 2006. Three patients were lost to follow-up at 0, 2, and 18 years after the Rastelli operation. The func-
Table 3. Data From Patients Who Died Late or Received Heart Transplantation Patient No.
Palliative Operations (number)
Age at Rastelli Operation (y)
1
11.5
3
(1) cAPS, and commissurotomy of the pulmonary valve (2) cAPS (1) mBTS (2) cAPS (1) mBTS
4
(0)
5
(1) mBTS, and atrial septectomy (2) mBTS (1) mBTS, and atrial septectomy
2
6
Reoperations (number)
Cause of Death or Heart Transplantation
Age at Death or Heart Transplantation (y)
(0)
Sudden death
12.9
4.6
(1) VSD closure, augmentation of pulmonary artery
Death of pneumonia
10.9
7.3
(1) Conduit exchange (2) Conduit exchange and tricuspid valve repair (1) Conduit exchange and aortic valve replacement (2) Conduit exchange (0)
Cardiac transplantation for biventricular heart failure
18.9
At reoperation for conduit exchange
22.2
Sudden death
39.6
Cardiac transplantation for biventricular heart failure
27.4
8.7
23.0
9.0
cAPS ⫽ central aorta to pulmonary artery shunt; shunt; VSD ⫽ ventricular septal defect.
(1) Residual VSD closure and relief of LVOTO (2) Conduit exchange and aortic valve replacement
LVOTO ⫽ left ventricular outflow tract obstruction;
mBTS ⫽ modified Blalock-Taussig
2172
HÖRER ET AL LONG-TERM OUTCOME AFTER RASTELLI REPAIR
CARDIOVASCULAR
tional status was determined according to the New York Heart Association class and ability index [8]. In case of death, the relatives and the general physicians of the patients were contacted to determine the cause of death. On follow-up examinations, patients underwent echocardiography and electrocardiography. Data of echocardiographic examinations from the outpatients’ clinics were evaluated according to written notes by one of the authors. Ventricular function was graded as normal or as mildly, moderately, or severely impaired. Insufficiency of the atrioventricular valves, the aortic valve, and of the homograft or xenograft valves in the right ventricular outflow tract was graded as none, mild, moderate, or severe. Left ventricular outflow tract obstruction and RVOTO at final follow-up were graded as mild stenosis when the maximal instantaneous pressure gradient was less than 30 mm Hg, as moderate stenosis when the gradient was 30 to 60 mm Hg, and as severe stenosis when the gradient was greater than 60 mm Hg.
Data Analysis Descriptive data for continuous variables are presented as mean ⫾ standard deviation or as medians with ranges; categorical variables are presented as relative frequencies. The outcome variables were defined as time from the Rastelli operation to death or cardiac transplantation, and time to reoperation for RVOTO, LVOTO, VSD, and pacemaker implantation. Reoperation for RVOTO was defined as any operation that included relief of an RVOTO, with or without additional surgical procedures. Reoperations for LVOTO, VSD, and pacemaker implantation were defined similarly. The probability of freedom from events was estimated according to the Kaplan– Meier method. Freedom-from-event curves were compared using the log-rank test. Probability values less than 0.1 were considered as a statistical trend, and probability values less than 0.05, as significant. Analyses were performed with SPSS 14.0.2 for Windows (SPSS Inc, Chicago, IL).
Results Postoperative Hospital Data There was no in-hospital mortality. Median hospital stay was 22 days (range, 7 to 98 days), median time at the intensive care unit was 5 days (range, 2 to 80 days), and median time on ventilation was 2 days (range, 1 to 70 days). There were five early reoperations, three for bleeding, and two for closure of a residual VSD. Five patients (12.8%) showed transient or permanent atrioventricular block postoperatively; 3 of them (7.8%) required permanent pacemaker implantation. The incidence of postoperative transient or permanent atrioventricular block was not significantly higher in patients who received enlargement of the VSD.
Late Deaths Within a median follow-up of 8.9 years (range, 0 to 25 years), there were 4 late deaths and 2 cardiac transplan-
Ann Thorac Surg 2007;83:2169 –75
Table 4. Predictors for Death and Transplantation Variable LVOTO subvalvular and valvular Stenosis of main pulmonary arteries within hilar bifurcation Stenosis of peripheral pulmonary arteries Hemoglobin level ⬎18.5 g/dL Age at time of Rastelli operation ⬎4 years Body weight at time of Rastelli operation ⬎17 kg VSD enlargement Cardiopulmonary bypass time ⬎136 minutes Aortic cross-clamp time ⬎83 minutes
Numbers
Events
26
6
0.012
8
2
0.018
5
2
⬍0.001
19 22
6 6
0.094 0.068
23
6
0.071
23 19
5 5
0.030 0.029
19
6
0.015
LVOTO ⫽ left ventricular outflow tract obstruction; lar septal defect.
p Value
VSD ⫽ ventricu-
tations. Freedom from death or cardiac transplantation was 93.8% ⫾ 4.3% and 57.5% ⫾ 15.1% at 10 and 20 years, respectively (Fig 1A). Patient 1 (Table 3) presented with stenosis of the peripheral and central pulmonary arteries. She died of ventricular fibrillation during school sport 16 months after correction. Two months before, sinus rhythm with right bundle-branch block and normal left ventricular function without LVOTO was documented. Patient 2 also presented with multiple stenoses of the peripheral and central pulmonary arteries. At the time of the Rastelli operation, the patient received patch plasty of both pulmonary arteries; however, 1 year after correction, she required reoperation for pulmonary artery stenosis and residual VSD. She died of pneumonia 6 years after the Rastelli operation. Patient 3 underwent two reoperations for conduit exchange and tricuspid valve repair after the Rastelli operation. Before cardiac transplantation he presented with severe tricuspid regurgitation and severely impaired right and left ventricular function. Patient 4 died during the second reoperation for conduit exchange at another hospital 13 years after Rastelli repair. Patient 5 died of sudden death 16 years after the Rastelli operation. Three months before, sinus rhythm with right bundle-branch block and moderately impaired left ventricular function with mild aortic insufficiency was documented. The maximum pressure gradient in the right ventricular outflow tract was 60 mm Hg. The postoperative course of patient 6 was complicated by severe low cardiac output syndrome and pneumonia. He received mechanical ventilation for 70 days. He underwent reoperation for residual LVOTO 5 months after Rastelli repair, and for conduit exchange and aortic valve replacement 11 years after Rastelli repair. Before cardiac transplantation, 18 years after correction, severely impaired left ventricular function and stenosis and insufficiency of the degenerated aortic valve prostheses were observed.
HÖRER ET AL LONG-TERM OUTCOME AFTER RASTELLI REPAIR
2173
Fig 2. Estimated freedom from conduit reoperation, left ventricular outflow tract obstruction (LVOTO), and residual ventricular septal defect (VSD; A), and estimated freedom from conduit reoperation as a function of age at the time of correction (B).
Among the morphologic characteristics presented in Table 2, subvalvular and valvular LVOTO (p ⫽ 0.012) and stenosis of the peripheral pulmonary arteries (p ⬍ 0.001) emerged as risk factors for late death or transplantation (Table 4). There was no significant influence for late death or transplantation to be derived from any of the hemodynamic variables as listed in Table 2. Among operative variables, enlargement of the VSD at the time of the Rastelli repair (p ⫽ 0.030), longer cardiopulmonary bypass time (p ⫽ 0.029), and longer ischemic time (p ⫽ 0.015) were predictive for death or transplantation. There was a trend toward an increased risk for late death or transplantation in patients who presented at an older age (p ⫽ 0.068; Fig 1B) or with a higher body weight (p ⫽ 0.071) or an elevated hemoglobin level (p ⫽ 0.094) at the time of the Rastelli repair.
Late Reoperations During follow-up, 18 patients underwent a total of 30 reoperations. Seventeen patients required conduit exchange for RVOTO. Freedom from conduit reoperation, from reoperation for LVOTO, and from reoperation for residual VSD at 20 years was 32.5% ⫾ 10.3%, 93.3% ⫾ 4.7%, and 82.0% ⫾ 8.0%, respectively (Fig 2A). Freedom from conduit reoperation at 10 years after correction was 36.3% ⫾ 15.4% in patients younger than 4 years of age at the time of the Rastelli operation compared with 57.6% ⫾ 12.9% in patients 4 years of age or older (p ⫽ 0.038; Fig 2B). There was no significant influence on freedom from conduit exchange to be derived from the presence of pulmonary atresia or pulmonary stenosis at the time of the Rastelli operation, from the type of conduit used for outflow tract reconstruction, or whether a French maneuver was performed.
Final Follow-Up Mean age of the patients at final follow-up was 15.0 ⫾ 9.4 years. Data of echocardiographic and electrocardiographic examinations at final follow-up could be obtained from 30 patients. Left ventricular function was normal in 87% of the patients, and 83% presented without LVOTO, 73% without aortic insufficiency, and 67% without mitral insufficiency. Right ventricular function was normal in 77% of the patients, and 27% presented
with no RVOTO, 50% with no pulmonary insufficiency, and 50% with no tricuspid insufficiency. Three patients presented with severe stenosis of the conduit (gradient ⬎ 60 mm Hg) and are scheduled for conduit exchange. At final follow-up, 25 patients presented with sinus rhythm and complete right bundle-branch block. Two of the 3 patients who had received pacemaker implantation during hospital stay for the Rastelli operation presented with pacemaker rhythm at final follow-up. In 2 patients, permanent pacemaker implantation had been performed after hospital discharge for complete atrioventricular block, and 1 patient received implantation of an implantable cardioverter defibrillator for ventricular tachycardia. Freedom from pacemaker implantation was 89.6% ⫾ 4.9% and 74.7% ⫾ 10.8%, at 5 and 20 years, respectively. Enlargement of the VSD at the time of the Rastelli operation had no significant influence on freedom from pacemaker implantation. Information concerning the functional status at the time of final follow-up could be obtained from 30 patients. Accordingly, 53% of the patients were in New York Heart Association class I, 43% in class II, and 3% in class III. The majority (90%) of the patients led a normal life with full-time work or school (ability index class I), 7% were able to work part-time (class II), and 3% experienced noticeable limitations on activities (class III). Only 7% of the patients were taking medications for heart failure (angiotensin-converting enzyme inhibitors, diuretics, or digitalis), and 7% were taking antiarrhythmics.
Comment Early Mortality and Reoperations During Hospital Stay In the present study population we could demonstrate that the Rastelli repair for TGA with VSD and LVOTO can be performed without operative mortality. Lee and colleagues [9] also reported a series of 10 patients who had undergone the Rastelli operation without operative mortality. In the series reported by Kreutzer and associates [10], early mortality rate was 7% without early deaths within the last 7 years of the study. Hence, nowadays the Rastelli operation can be applied to this special subset of patients with very low operative risk. However, because
CARDIOVASCULAR
Ann Thorac Surg 2007;83:2169 –75
2174
HÖRER ET AL LONG-TERM OUTCOME AFTER RASTELLI REPAIR
CARDIOVASCULAR
of the complexity of the operation, early complications may occur during hospital stay. In the literature, 10% to 14% of the patients require an early reoperation during the same hospital stay for such complications as complete atrioventricular block, postoperative bleeding, residual VSD, or residual LVOTO [9, 11]. In the present cohort, the prevalence was 18%, mainly as a result of prolonged bleeding and residual VSD.
Reoperations Despite the advantage of incorporating the left ventricle into the systemic circulation, the Rastelli operation is by far not an anatomic correction for at least two reasons. First, the left ventricular outflow tract is reconstructed with an intracardiac prosthetic baffle. Second, the right ventricular outflow tract is reconstructed with an extracardiac conduit. This leads to inevitable reoperations for conduit exchange. Reoperations are required predominantly for stenosis and insufficiency of the extracardiac conduit. In the present cohort, freedom from right ventricular outflow tract reintervention was 49% at 10 years after the Rastelli operation, and thus compares favorably to the range of 21% to 27% reported in the literature [9, 10]. Also according to these previously published series, the time to right ventricular outflow tract reintervention was not affected by the type of the conduit used for right ventricular outflow tract reconstruction. A similar conduit exchange rate of homografts compared with porcine xenografts for conduit diameters less than 15 mm has already been reported from our institution [11]. Age at the Rastelli operation, younger than 1 year in the series reported by Kreutzer and coworkers [10], and younger than 4 years in the present series, was the only reported risk factor for shorter time to conduit reoperation. Freedom from right ventricular outflow tract reoperation was reported to be significantly higher after the réparation à l’étage ventriculaire procedure, compared with the Rastelli operation [9]. The difference between the two techniques consists in a direct anastomosis of the pulmonary trunk to the superior margin of the right ventriculotomy, insertion of a monocusp valve, and closure of the right ventricular outflow tract with an anterior patch in the réparation à l’étage ventriculaire procedure, instead of implantation of a prosthesis, as performed in the Rastelli operation. The disadvantage of the réparation à l’étage ventriculaire procedure, however, is that it cannot be applied to all patients, especially not to those presenting with pulmonary atresia. In addition, during the réparation à l’étage ventriculaire procedure, according to its original description, the pulmonary bifurcation is placed anteriorly to the aorta, which is also known as the French maneuver. Especially in this subset of patients, who often present with a large ascending aorta that is not relocated posteriorly in contrast to an arterial switch operation [12], the French maneuver may lead to tension of the branch pulmonary arteries and consequently to pressure load of the right ventricle. In the present cohort, the French maneuver was applied in 4 patients. In one of them, both
Ann Thorac Surg 2007;83:2169 –75
branch pulmonary arteries had to be augmented 1 year after the initial operation. Left ventricular outflow tract obstruction may persist owing to a restrictive VSD that was not sufficiently enlarged. Or it may develop as a result of muscular hypertrophy of the outflow tract. Kreutzer and colleagues [10] reported a freedom from LVOTO reintervention to be 84% at 15 years in contrast to 93% at 15 years in the present cohort. The lower reintervention rate in our series may be related to the fact that we enlarged the VSD more frequently (59% versus 47%). By resecting the conal septum or the anterior superior margin of the defect, even a small, restrictive VSD can be enlarged to the size of the aortic annulus, and a straighter left ventricular outflow tract can be created. The higher rate of freedom from LVOTO reoperation may also be explained by the higher median age at the time of Rastelli operation in the present cohort compared with the series reported by Kreutzer and associates [10] (5.1 versus 3.1 years), as the authors reported age younger than 1 year at the initial operation to be a risk factor for shorter time to LVOTO reintervention.
Late Mortality The 58% freedom from death or cardiac transplantation at 20 years in the present cohort is a sobering statistic. However, it does not stand on its own. Kreutzer [10] reported a 52% freedom from death or cardiac transplantation at 20 years. Hence, survival of patients presenting with TGA with VSD and LVOTO is worse compared with that of patients who had undergone an arterial switch operation [13, 14] or even an atrial redirection [15]. The developmental status of the peripheral pulmonary artery plays an important role for the intermediate and long-term outcome after the Rastelli operation. In the present study group, stenosis of the peripheral pulmonary arteries was a strong predictor for late death or transplantation. Hence, decisions related to the planning of the Rastelli operation must be strongly influenced by the status of the peripheral pulmonary arteries. Patients who received enlargement of the VSD exhibited a significantly higher risk for late death. This might be related to the damage of the conduction tissue or the injury of the septal perforating arteries. In the present study group, postoperative atrioventricular block was not significantly more frequent in patients who received enlargement of the VSD during Rastelli repair. Longer cardiopulmonary bypass time and longer ischemic time also emerged as risk factors for late death. Cardiopulmonary bypass time and cross-clamp time were not significantly longer in patients who received enlargement of the VSD; hence potential deleterious effects of prolonged cardiac arrest time may affect long-term survival.
Timing of the Rastelli Operation On the one hand, lower age at the time of correction predicts smaller conduits for right ventricular outflow tract reconstruction, and consequently a higher replacement rate. On the other hand, older children sustain a
longer time of palliation, which may lead to myocardial dysfunction as a consequence of prolonged cyanosis. Because the present cohort represents in part a historical collective, the median age of 5 years at the time of correction was rather high. Freedom from conduit reoperation was significantly higher in patients younger than 4 years of age at the time of operation. In the early experience, the surgeons tended to postpone the correction to make VSD baffling easier, and to allow for a larger conduit to be used. With growing experience of early correction in our department, the Rastelli operation was performed in younger patients. Hence, the median age at correction decreased to 2.2 years within the last 5 years of the study period. In the present cohort, there was a trend toward an increased risk for death or transplantation in patients 4 years of age or older at the time of the Rastelli operation. The survival curves of patients younger and older than 4 years of age diverge clearly from 10 years of follow-up onward, indicating that the benefit of early correction might emerge clearly in the longer-term. Interestingly, patients who were 4 years of age or older at the time of the Rastelli operation presented with a significantly higher hemoglobin level. In addition, there was a trend toward an increased risk for late death or transplantation in patients presenting with a higher hemoglobin level at the time of the Rastelli operation. The higher hemoglobin level in patients presenting with older age at the time of the Rastelli operation indicates that longer palliation leads to chronic cyanosis with the consequence of an erythrocytosis. Presumably, the sequelae of prolonged cyanosis, such as myocardial fibrosis, may have a negative effect on survival in the long term.
Conclusions The Rastelli operation is a low-risk procedure with regard to early mortality. The status of the pulmonary arteries and VSD enlargement are predictive for longterm survival. Patients who are 4 years of age or older at the time of the Rastelli operation require fewer reoperations for conduit exchange. However, there is a significant mortality in the long term in patients 4 years of age or older. Therefore, early Rastelli repair is recommended because the initial operation in early infancy bears no higher operative risk and reoperations for conduit exchange can be performed with a low mortality [16].
Study Limitations The study design was a retrospective follow-up study covering a long period of patient inclusion. Changes in preoperative, operative, and postoperative management may have affected the outcome variables in a way not covered by our analysis. According to Concato and coworkers [17], the ratio of events per variable concerning
HÖRER ET AL LONG-TERM OUTCOME AFTER RASTELLI REPAIR
2175
the end point death or transplantation was too small to perform a multivariate analysis.
References 1. Williams WG, McCrindle BW, Ashburn DA, Jonas RA, Mavroudis C, Blackstone EH. Outcomes of 829 neonates with complete transposition of the great arteries 12–17 years after repair. Eur J Cardiothorac Surg 2003;24:1–10. 2. Hanlon CR, Blalock A. Complete transposition of the aorta and the pulmonary artery: experimental observations on venous shunts as corrective procedures. Ann Surg 1948;127: 385–97. 3. Rastelli GC, Wallace RB, Ongley PA. Complete repair of transposition of the great arteries with pulmonary stenosis: a review and report of a case corrected by using a new surgical technique. Circulation 1969;39:83–95. 4. Rastelli GC. A new approach to “anatomic” repair of transposition of the great arteries. Mayo Clin Proc 1969;44:1–12. 5. Lecompte Y, Zannini L, Hazan E, et al. Anatomic correction of transposition of the great arteries. J Thorac Cardiovasc Surg 1981;82:629 –31. 6. Nikaidoh H. Aortic translocation and biventricular outflow tract reconstruction. A new surgical repair for transposition of the great arteries associated with ventricular septal defect and pulmonary stenosis. J Thorac Cardiovasc Surg 1984;88: 365–72. 7. Perloff JK. Systemic complications of cyanosis in adults with congenital heart disease. Hematologic derangements, renal function, and urate metabolism. Cardiol Clin 1993;11:689 –99. 8. Warnes CA, Somerville J. Tricuspid atresia in adolescents and adults: current state and late complications. Br Heart J 1986;56:535– 43. 9. Lee JR, Lim HG, Kim YJ, et al. Repair of transposition of the great arteries, ventricular septal defect and left ventricular outflow tract obstruction. Eur J Cardiothorac Surg 2004;25: 735– 41. 10. Kreutzer C, De Vive J, Oppido G, et al. Twenty-five-year experience with Rastelli repair for transposition of the great arteries. J Thorac Cardiovasc Surg 2000;120:211–23. 11. Lange R, Weipert J, Homann M, et al. Performance of allografts and xenografts for right ventricular outflow tract reconstruction. Ann Thorac Surg 2001;71(Suppl):S365–7. 12. Jatene AD, Fontes VF, Paulista PP, et al. Anatomic correction of transposition of the great vessels. J Thorac Cardiovasc Surg 1976;72:364 –70. 13. Hutter PA, Kreb DL, Mantel SF, Hitchcock JF, Meijboom EJ, Bennink GB. Twenty-five years’ experience with the arterial switch operation. J Thorac Cardiovasc Surg 2002;124:790 –7. 14. Pretre R, Tamisier D, Bonhoeffer P, et al. Results of the arterial switch operation in neonates with transposed great arteries. Lancet 2001;357:1826 –30. 15. Moons P, Gewillig M, Sluysmans T, et al. Long term outcome up to 30 years after the Mustard or Senning operation: a nationwide multicentre study in Belgium. Heart 2004;90: 307–13. 16. Mohammadi S, Belli E, Martinovic I, et al. Surgery for right ventricle to pulmonary artery conduit obstruction: risk factors for further reoperation. Eur J Cardiothorac Surg 2005; 28:217–22. 17. Concato JPP, Holford TR, Feinstein AR. Importance of events per independent variable in proportional hazards analysis. I. Background, goals, and general strategy. J Clin Epidemiol 1995;48:1495–501.
CARDIOVASCULAR
Ann Thorac Surg 2007;83:2169 –75
Long-Term Results After the Rastelli Repair for Transposition of the Great Arteries Jürgen Hörer, MD,* Christian Schreiber, MD, PhD,* Eva Dworak, MD, Julie Cleuziou, MD, Zsolt Prodan, MD, Manfred Vogt, MD, PhD, Klaus Holper, MD, PhD, and Rüdiger Lange, MD, PhD Departments of Cardiovascular Surgery, and Pediatric Cardiology and Congenital Heart Disease, German Heart Center Munich at the Technical University, Munich, Germany
Background. This study sought to assess risk factors for late mortality after the Rastelli operation for patients with transposition of the great arteries, ventricular septal defect, and left ventricular outflow tract obstruction. Methods. Records of 39 patients who underwent the Rastelli operation between 1977 and 2004 were reviewed. Median age at the time of operation was 5.1 years (2.2 years within the last 5 years). Results. There were no early deaths. During a median follow-up of 8.9 years (range, 0 to 25 years), 2 patients died of sudden death, 1 of pneumonia, 1 during reoperation, and 2 received heart transplantation. Freedom from death or transplantation was 93.8% ⴞ 4.3% and 57.5% ⴞ 15.1% at 10 and 20 years, respectively. Freedom from conduit replacement was 48.8% ⴞ 10.3% and 32.5% ⴞ 10.3% at 10 and 20 years, respectively. Subvalvular and valvular left ventricular outflow tract obstruction (p ⴝ 0.012), stenosis of the peripheral pulmonary arteries (p < 0.001), enlargement of the ventricular septal defect
(p ⴝ 0.030), and longer ischemic time (p ⴝ 0.015) were predictive for death or transplantation. Patients younger than 4 years at the time of the Rastelli operation showed a trend toward a better freedom from death or transplantation (p ⴝ 0.068), but needed significantly more conduit replacements (p ⴝ 0.038) compared with patients 4 years or older. Conclusions. The Rastelli operation is a low-risk procedure with regard to early mortality. The status of the pulmonary arteries and ventricular septal defect enlargement are predictive for long-term survival. Patients 4 years of age or older at the time of the Rastelli operation require fewer reoperations for conduit exchange. Nevertheless, early Rastelli repair is recommended because patients 4 years or older are at risk for a higher long-term mortality.
A
successful operation for this special subset of patients, making the left ventricle as the subaortic ventricle. The Rastelli operation, in its original description, consisted of connecting the left ventricle to the aorta by suturing a polytetrafluoroethylene (Teflon) patch between the VSD and the aortic orifice, and of reconstructing the continuity from the right ventricle to the pulmonary artery with a homograft [4]. Although other surgical procedures, like the “réparation à l’étage ventriculaire” [5] and the Nikaidoh operation [6] were introduced, the Rastelli operation remained the preferred treatment in many congenital cardiac units for the past three decades. The disadvantages of the Rastelli operation, however, consist in the reconstruction of the left ventricular outflow tract with an intracardiac prosthetic baffle, and in the reconstruction of the right ventricular outflow tract with an extracardiac conduit, which may lead to subsequent reoperations for LVOTO and for right ventricular outflow tract obstruction (RVOTO). In the early experience, the Rastelli operation was often delayed to allow
pproximately 4% of neonates and infants who undergo surgery for transposition of the great arteries (TGA) also present with an additional ventricular septal defect (VSD) and a left ventricular outflow tract obstruction (LVOTO) [1]. Early survival of these patients without surgery is superior compared with children who present with a simple TGA without VSD, as the interventricular communication allows for better mixing of systemic and pulmonary venous return [2]. Moreover, the pulmonary vascular bed is protected by the LVOTO, which also keeps the left ventricle in training for supporting the systemic circulation after biventricular correction. In 1969 Rastelli and associates [3] introduced the first
Accepted for publication Jan 29, 2007. *Doctors Hörer and Schreiber contributed equally to the study. Address correspondence to Dr Hörer, Department of Cardiovascular Surgery, German Heart Center Munich at the Technical University, Lazarettstrasse 36, Munich, D-80636 Germany; e-mail: hoerer@ dhm.mhn.de.
© 2007 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2007;83:2169 –75) © 2007 by The Society of Thoracic Surgeons
0003-4975/07/$32.00 doi:10.1016/j.athoracsur.2007.01.061
CARDIOVASCULAR
The Annals of Thoracic Surgery CME Program is located online at http://cme.ctsnetjournals.org. To take the CME activity related to this article, you must have either an STS member or an individual nonmember subscription to the journal.
2170
HÖRER ET AL LONG-TERM OUTCOME AFTER RASTELLI REPAIR
Ann Thorac Surg 2007;83:2169 –75
Table 1. Palliative Procedures Before Rastelli Operation
CARDIOVASCULAR
Type of Procedure
n
mBTS mBTS ⫹ atrial septectomy cAPS cAPS ⫹ ligation of arterial duct cAPS ⫹ atrial septectomy cAPS ⫹ commissurotomy of the pulmonary valve cAPS ⫹ ligation of arterial duct ⫹ atrial septectomy cAPS ⫹ ligation of arterial duct ⫹ atrial septectomy, mitral valve repair Atrial septectomy Balloon atrial septostomy
13 3 9 1 1 2 1 1
cAPS ⫽ central aorta to pulmonary artery shunt; Blalock-Taussig shunt.
1 20
mBTS ⫽ modified
larger conduits to be placed. This might have resulted in poor long-term outcome owing to the deleterious effect of prolonged cyanosis and volume overload after palliative procedures [7]. The aim of the present study was to evaluate potential risk factors for late mortality and reoperations after the Rastelli operation for correction of TGA with VSD and LVOTO during a 25-year period.
Patients and Methods Patient Selection All patients who underwent a Rastelli operation for TGA with LVOTO and VSD between 1977 and 2004 at the German Heart Center Munich were included in a retrospective follow-up study. Patients who underwent the Rastelli repair for double-outlet right ventricle with malposition of the great arteries (n ⫽ 19) were excluded. Double-outlet right ventricle with malposition of the great arteries was differentiated from TGA with VSD if there was more than 50% overriding of the pulmonary artery relative to the ventricular septum, and if there was a discontinuity between the mitral valve and the pulmonary valve, and thus presence of a subaortic and subpulmonary conus. Furthermore, 3 patients from foreign countries were excluded because follow-up data could not be generated. This study was approved by the ethics committee of the Technical University Munich. Individual consent for the study was obtained from each patient.
Study Group The study group was made up of 21 male and 18 female patients. Thirty-two palliative operations were performed in 23 patients (Table 1). The VSD was located toward the outlet in all patients, with an inlet extension in 5 patients (12.8%). Five patients (12.8%) presented with pulmonary atresia. The remaining patients presented with LVOTO. Eight patients (20.5%) presented with stenosis of the pulmonary arteries (Table 2). Median age at the time of operation was 5.1 years for all patients (range, 1 to 23 years), and 2.2 years for the patients operated on
within the last 5 years of the study period. Median weight was 16.7 kg (range, 6 to 54 kg). Patients 4 years of age or older at the time of the Rastelli operation presented with significantly higher hemoglobin levels (19.5 ⫾ 2.8 g/dL), compared with patients younger than 4 years (17.7 ⫾ 2.4 g/dL; p ⫽ 0.040).
Surgical Technique All operations were performed by means of a midline sternotomy using conventional cardiopulmonary bypass with bicaval cannulation and crystalloid cardioplegic solution. The median cardiopulmonary bypass time was 136 minutes (range, 89 to 330 minutes). One patient (cardiopulmonary bypass time, 330 minutes) underwent aortic valve replacement with a xenograft conduit in addition to the Rastelli operation. The median ischemic
Table 2. Characteristics of 39 Patients Who Underwent the Rastelli Operationa Variable Morphology Inlet extension of VSD One or more additional muscular VSDs LVOTO subvalvular LVOTO valvular LVOTO subvalvular and valvular Pulmonary atresia Stenosis of the main pulmonary arteries within the hilar bifurcation Stenosis of the peripheral pulmonary arteries Hemodynamics at cardiac catheterization Right ventricular to left ventricular pressure ratio Pulmonary to systemic pressure ratio Left ventricular to pulmonary artery gradient (mm Hg) Right ventricular to aorta gradient (mm Hg) Pulmonary to systemic flow ratio Aortic oxygen saturation (%) Pulmonary artery oxygen saturation (%) Hemoglobin level (g/dL) Surgical Prior palliative operation Year of operation before 1995 Age at the time of Rastelli operation (y) Body weight at the time of Rastelli operation (kg) VSD enlargement Xenograft French maneuver Cardiopulmonary bypass time (min) Aortic cross-clamp time (min) Postoperative atrioventricular block
Value 5 (12.8%) 5 (12.8%) 5 (12.8%) 3 (7.8%) 26 (66.7%) 5 (12.8%) 8 (20.5%) 5 (12.8%) 1.00 ⫾ 0.06 0.23 ⫾ 0.11 70.8 ⫾ 21.3 2.7 2.1 79.7 85.9 18.7
⫾ ⫾ ⫾ ⫾ ⫾
4.2 1.9 5.6 7.1 2.8
23 (59.0%) 19 (48.7%) 5.8 ⫾ 5.1 18.0 ⫾ 11.0 34 (87.2%) 10 (25.6%) 4 (10.3%) 154 ⫾ 52 93 ⫾ 33 5 (12.8%)
Values are given as mean ⫾ standard deviation or as number with frequencies.
a
LVOTO ⫽ left ventricular outflow tract obstruction; lar septal defect.
VSD ⫽ ventricu-
Ann Thorac Surg 2007;83:2169 –75
HÖRER ET AL LONG-TERM OUTCOME AFTER RASTELLI REPAIR
2171
CARDIOVASCULAR
Fig 1. Estimated survival (A) and estimated survival as a function of age at the time of correction (B).
time was 83 minutes (range, 46 to 171 minutes). All operations were performed with the use of hypothermia with a mean rectal temperature of 23.7° ⫾ 1.5°C. Any systemic to pulmonary artery shunts were ligated. In 23 patients (59.0%), the VSD was enlarged at the anterosuperior margin of the defect. Cardiopulmonary bypass time (p ⫽ 0.76) and ischemic time (p ⫽ 0.13) were not significantly longer in patients who received enlargement of the VSD. The VSD was baffled to the aorta using a polyethylene terephthalate fiber (Dacron) patch in 34 patients (87.2%) and a polytetrafluoroethylene patch (Gore-Tex, W.L. Gore & Assoc, Flagstaff, AZ) in the remaining patients. The patch was fixed by a running suture in 19 patients (48.8%), by interrupted sutures in 5 patients (12.8%), and by a combination of both in the remaining patients. The right ventricular to pulmonary
artery connection was reconstructed with a Hancock porcine graft in 10 patients (25.6%) and with a homograft in 29 patients (74.4%). The mean size of the porcine graft was 17.4 ⫾ 4.8 mm (range, 12 to 25 mm), and the mean size of the homografts was 17.4 ⫾ 3.2 mm (range, 12 to 24 mm). The conduit was positioned to the left side of the aorta in all patients. The pulmonary bifurcation was placed anteriorly to the aorta (French maneuver [5]) in 4 patients (10.4%).
Follow-Up Final follow-up was obtained by a written questionnaire or by telephone and conducted between September 2004 and March 2006. Three patients were lost to follow-up at 0, 2, and 18 years after the Rastelli operation. The func-
Table 3. Data From Patients Who Died Late or Received Heart Transplantation Patient No.
Palliative Operations (number)
Age at Rastelli Operation (y)
1
11.5
3
(1) cAPS, and commissurotomy of the pulmonary valve (2) cAPS (1) mBTS (2) cAPS (1) mBTS
4
(0)
5
(1) mBTS, and atrial septectomy (2) mBTS (1) mBTS, and atrial septectomy
2
6
Reoperations (number)
Cause of Death or Heart Transplantation
Age at Death or Heart Transplantation (y)
(0)
Sudden death
12.9
4.6
(1) VSD closure, augmentation of pulmonary artery
Death of pneumonia
10.9
7.3
(1) Conduit exchange (2) Conduit exchange and tricuspid valve repair (1) Conduit exchange and aortic valve replacement (2) Conduit exchange (0)
Cardiac transplantation for biventricular heart failure
18.9
At reoperation for conduit exchange
22.2
Sudden death
39.6
Cardiac transplantation for biventricular heart failure
27.4
8.7
23.0
9.0
cAPS ⫽ central aorta to pulmonary artery shunt; shunt; VSD ⫽ ventricular septal defect.
(1) Residual VSD closure and relief of LVOTO (2) Conduit exchange and aortic valve replacement
LVOTO ⫽ left ventricular outflow tract obstruction;
mBTS ⫽ modified Blalock-Taussig
2172
HÖRER ET AL LONG-TERM OUTCOME AFTER RASTELLI REPAIR
CARDIOVASCULAR
tional status was determined according to the New York Heart Association class and ability index [8]. In case of death, the relatives and the general physicians of the patients were contacted to determine the cause of death. On follow-up examinations, patients underwent echocardiography and electrocardiography. Data of echocardiographic examinations from the outpatients’ clinics were evaluated according to written notes by one of the authors. Ventricular function was graded as normal or as mildly, moderately, or severely impaired. Insufficiency of the atrioventricular valves, the aortic valve, and of the homograft or xenograft valves in the right ventricular outflow tract was graded as none, mild, moderate, or severe. Left ventricular outflow tract obstruction and RVOTO at final follow-up were graded as mild stenosis when the maximal instantaneous pressure gradient was less than 30 mm Hg, as moderate stenosis when the gradient was 30 to 60 mm Hg, and as severe stenosis when the gradient was greater than 60 mm Hg.
Data Analysis Descriptive data for continuous variables are presented as mean ⫾ standard deviation or as medians with ranges; categorical variables are presented as relative frequencies. The outcome variables were defined as time from the Rastelli operation to death or cardiac transplantation, and time to reoperation for RVOTO, LVOTO, VSD, and pacemaker implantation. Reoperation for RVOTO was defined as any operation that included relief of an RVOTO, with or without additional surgical procedures. Reoperations for LVOTO, VSD, and pacemaker implantation were defined similarly. The probability of freedom from events was estimated according to the Kaplan– Meier method. Freedom-from-event curves were compared using the log-rank test. Probability values less than 0.1 were considered as a statistical trend, and probability values less than 0.05, as significant. Analyses were performed with SPSS 14.0.2 for Windows (SPSS Inc, Chicago, IL).
Results Postoperative Hospital Data There was no in-hospital mortality. Median hospital stay was 22 days (range, 7 to 98 days), median time at the intensive care unit was 5 days (range, 2 to 80 days), and median time on ventilation was 2 days (range, 1 to 70 days). There were five early reoperations, three for bleeding, and two for closure of a residual VSD. Five patients (12.8%) showed transient or permanent atrioventricular block postoperatively; 3 of them (7.8%) required permanent pacemaker implantation. The incidence of postoperative transient or permanent atrioventricular block was not significantly higher in patients who received enlargement of the VSD.
Late Deaths Within a median follow-up of 8.9 years (range, 0 to 25 years), there were 4 late deaths and 2 cardiac transplan-
Ann Thorac Surg 2007;83:2169 –75
Table 4. Predictors for Death and Transplantation Variable LVOTO subvalvular and valvular Stenosis of main pulmonary arteries within hilar bifurcation Stenosis of peripheral pulmonary arteries Hemoglobin level ⬎18.5 g/dL Age at time of Rastelli operation ⬎4 years Body weight at time of Rastelli operation ⬎17 kg VSD enlargement Cardiopulmonary bypass time ⬎136 minutes Aortic cross-clamp time ⬎83 minutes
Numbers
Events
26
6
0.012
8
2
0.018
5
2
⬍0.001
19 22
6 6
0.094 0.068
23
6
0.071
23 19
5 5
0.030 0.029
19
6
0.015
LVOTO ⫽ left ventricular outflow tract obstruction; lar septal defect.
p Value
VSD ⫽ ventricu-
tations. Freedom from death or cardiac transplantation was 93.8% ⫾ 4.3% and 57.5% ⫾ 15.1% at 10 and 20 years, respectively (Fig 1A). Patient 1 (Table 3) presented with stenosis of the peripheral and central pulmonary arteries. She died of ventricular fibrillation during school sport 16 months after correction. Two months before, sinus rhythm with right bundle-branch block and normal left ventricular function without LVOTO was documented. Patient 2 also presented with multiple stenoses of the peripheral and central pulmonary arteries. At the time of the Rastelli operation, the patient received patch plasty of both pulmonary arteries; however, 1 year after correction, she required reoperation for pulmonary artery stenosis and residual VSD. She died of pneumonia 6 years after the Rastelli operation. Patient 3 underwent two reoperations for conduit exchange and tricuspid valve repair after the Rastelli operation. Before cardiac transplantation he presented with severe tricuspid regurgitation and severely impaired right and left ventricular function. Patient 4 died during the second reoperation for conduit exchange at another hospital 13 years after Rastelli repair. Patient 5 died of sudden death 16 years after the Rastelli operation. Three months before, sinus rhythm with right bundle-branch block and moderately impaired left ventricular function with mild aortic insufficiency was documented. The maximum pressure gradient in the right ventricular outflow tract was 60 mm Hg. The postoperative course of patient 6 was complicated by severe low cardiac output syndrome and pneumonia. He received mechanical ventilation for 70 days. He underwent reoperation for residual LVOTO 5 months after Rastelli repair, and for conduit exchange and aortic valve replacement 11 years after Rastelli repair. Before cardiac transplantation, 18 years after correction, severely impaired left ventricular function and stenosis and insufficiency of the degenerated aortic valve prostheses were observed.
HÖRER ET AL LONG-TERM OUTCOME AFTER RASTELLI REPAIR
2173
Fig 2. Estimated freedom from conduit reoperation, left ventricular outflow tract obstruction (LVOTO), and residual ventricular septal defect (VSD; A), and estimated freedom from conduit reoperation as a function of age at the time of correction (B).
Among the morphologic characteristics presented in Table 2, subvalvular and valvular LVOTO (p ⫽ 0.012) and stenosis of the peripheral pulmonary arteries (p ⬍ 0.001) emerged as risk factors for late death or transplantation (Table 4). There was no significant influence for late death or transplantation to be derived from any of the hemodynamic variables as listed in Table 2. Among operative variables, enlargement of the VSD at the time of the Rastelli repair (p ⫽ 0.030), longer cardiopulmonary bypass time (p ⫽ 0.029), and longer ischemic time (p ⫽ 0.015) were predictive for death or transplantation. There was a trend toward an increased risk for late death or transplantation in patients who presented at an older age (p ⫽ 0.068; Fig 1B) or with a higher body weight (p ⫽ 0.071) or an elevated hemoglobin level (p ⫽ 0.094) at the time of the Rastelli repair.
Late Reoperations During follow-up, 18 patients underwent a total of 30 reoperations. Seventeen patients required conduit exchange for RVOTO. Freedom from conduit reoperation, from reoperation for LVOTO, and from reoperation for residual VSD at 20 years was 32.5% ⫾ 10.3%, 93.3% ⫾ 4.7%, and 82.0% ⫾ 8.0%, respectively (Fig 2A). Freedom from conduit reoperation at 10 years after correction was 36.3% ⫾ 15.4% in patients younger than 4 years of age at the time of the Rastelli operation compared with 57.6% ⫾ 12.9% in patients 4 years of age or older (p ⫽ 0.038; Fig 2B). There was no significant influence on freedom from conduit exchange to be derived from the presence of pulmonary atresia or pulmonary stenosis at the time of the Rastelli operation, from the type of conduit used for outflow tract reconstruction, or whether a French maneuver was performed.
Final Follow-Up Mean age of the patients at final follow-up was 15.0 ⫾ 9.4 years. Data of echocardiographic and electrocardiographic examinations at final follow-up could be obtained from 30 patients. Left ventricular function was normal in 87% of the patients, and 83% presented without LVOTO, 73% without aortic insufficiency, and 67% without mitral insufficiency. Right ventricular function was normal in 77% of the patients, and 27% presented
with no RVOTO, 50% with no pulmonary insufficiency, and 50% with no tricuspid insufficiency. Three patients presented with severe stenosis of the conduit (gradient ⬎ 60 mm Hg) and are scheduled for conduit exchange. At final follow-up, 25 patients presented with sinus rhythm and complete right bundle-branch block. Two of the 3 patients who had received pacemaker implantation during hospital stay for the Rastelli operation presented with pacemaker rhythm at final follow-up. In 2 patients, permanent pacemaker implantation had been performed after hospital discharge for complete atrioventricular block, and 1 patient received implantation of an implantable cardioverter defibrillator for ventricular tachycardia. Freedom from pacemaker implantation was 89.6% ⫾ 4.9% and 74.7% ⫾ 10.8%, at 5 and 20 years, respectively. Enlargement of the VSD at the time of the Rastelli operation had no significant influence on freedom from pacemaker implantation. Information concerning the functional status at the time of final follow-up could be obtained from 30 patients. Accordingly, 53% of the patients were in New York Heart Association class I, 43% in class II, and 3% in class III. The majority (90%) of the patients led a normal life with full-time work or school (ability index class I), 7% were able to work part-time (class II), and 3% experienced noticeable limitations on activities (class III). Only 7% of the patients were taking medications for heart failure (angiotensin-converting enzyme inhibitors, diuretics, or digitalis), and 7% were taking antiarrhythmics.
Comment Early Mortality and Reoperations During Hospital Stay In the present study population we could demonstrate that the Rastelli repair for TGA with VSD and LVOTO can be performed without operative mortality. Lee and colleagues [9] also reported a series of 10 patients who had undergone the Rastelli operation without operative mortality. In the series reported by Kreutzer and associates [10], early mortality rate was 7% without early deaths within the last 7 years of the study. Hence, nowadays the Rastelli operation can be applied to this special subset of patients with very low operative risk. However, because
CARDIOVASCULAR
Ann Thorac Surg 2007;83:2169 –75
2174
HÖRER ET AL LONG-TERM OUTCOME AFTER RASTELLI REPAIR
CARDIOVASCULAR
of the complexity of the operation, early complications may occur during hospital stay. In the literature, 10% to 14% of the patients require an early reoperation during the same hospital stay for such complications as complete atrioventricular block, postoperative bleeding, residual VSD, or residual LVOTO [9, 11]. In the present cohort, the prevalence was 18%, mainly as a result of prolonged bleeding and residual VSD.
Reoperations Despite the advantage of incorporating the left ventricle into the systemic circulation, the Rastelli operation is by far not an anatomic correction for at least two reasons. First, the left ventricular outflow tract is reconstructed with an intracardiac prosthetic baffle. Second, the right ventricular outflow tract is reconstructed with an extracardiac conduit. This leads to inevitable reoperations for conduit exchange. Reoperations are required predominantly for stenosis and insufficiency of the extracardiac conduit. In the present cohort, freedom from right ventricular outflow tract reintervention was 49% at 10 years after the Rastelli operation, and thus compares favorably to the range of 21% to 27% reported in the literature [9, 10]. Also according to these previously published series, the time to right ventricular outflow tract reintervention was not affected by the type of the conduit used for right ventricular outflow tract reconstruction. A similar conduit exchange rate of homografts compared with porcine xenografts for conduit diameters less than 15 mm has already been reported from our institution [11]. Age at the Rastelli operation, younger than 1 year in the series reported by Kreutzer and coworkers [10], and younger than 4 years in the present series, was the only reported risk factor for shorter time to conduit reoperation. Freedom from right ventricular outflow tract reoperation was reported to be significantly higher after the réparation à l’étage ventriculaire procedure, compared with the Rastelli operation [9]. The difference between the two techniques consists in a direct anastomosis of the pulmonary trunk to the superior margin of the right ventriculotomy, insertion of a monocusp valve, and closure of the right ventricular outflow tract with an anterior patch in the réparation à l’étage ventriculaire procedure, instead of implantation of a prosthesis, as performed in the Rastelli operation. The disadvantage of the réparation à l’étage ventriculaire procedure, however, is that it cannot be applied to all patients, especially not to those presenting with pulmonary atresia. In addition, during the réparation à l’étage ventriculaire procedure, according to its original description, the pulmonary bifurcation is placed anteriorly to the aorta, which is also known as the French maneuver. Especially in this subset of patients, who often present with a large ascending aorta that is not relocated posteriorly in contrast to an arterial switch operation [12], the French maneuver may lead to tension of the branch pulmonary arteries and consequently to pressure load of the right ventricle. In the present cohort, the French maneuver was applied in 4 patients. In one of them, both
Ann Thorac Surg 2007;83:2169 –75
branch pulmonary arteries had to be augmented 1 year after the initial operation. Left ventricular outflow tract obstruction may persist owing to a restrictive VSD that was not sufficiently enlarged. Or it may develop as a result of muscular hypertrophy of the outflow tract. Kreutzer and colleagues [10] reported a freedom from LVOTO reintervention to be 84% at 15 years in contrast to 93% at 15 years in the present cohort. The lower reintervention rate in our series may be related to the fact that we enlarged the VSD more frequently (59% versus 47%). By resecting the conal septum or the anterior superior margin of the defect, even a small, restrictive VSD can be enlarged to the size of the aortic annulus, and a straighter left ventricular outflow tract can be created. The higher rate of freedom from LVOTO reoperation may also be explained by the higher median age at the time of Rastelli operation in the present cohort compared with the series reported by Kreutzer and associates [10] (5.1 versus 3.1 years), as the authors reported age younger than 1 year at the initial operation to be a risk factor for shorter time to LVOTO reintervention.
Late Mortality The 58% freedom from death or cardiac transplantation at 20 years in the present cohort is a sobering statistic. However, it does not stand on its own. Kreutzer [10] reported a 52% freedom from death or cardiac transplantation at 20 years. Hence, survival of patients presenting with TGA with VSD and LVOTO is worse compared with that of patients who had undergone an arterial switch operation [13, 14] or even an atrial redirection [15]. The developmental status of the peripheral pulmonary artery plays an important role for the intermediate and long-term outcome after the Rastelli operation. In the present study group, stenosis of the peripheral pulmonary arteries was a strong predictor for late death or transplantation. Hence, decisions related to the planning of the Rastelli operation must be strongly influenced by the status of the peripheral pulmonary arteries. Patients who received enlargement of the VSD exhibited a significantly higher risk for late death. This might be related to the damage of the conduction tissue or the injury of the septal perforating arteries. In the present study group, postoperative atrioventricular block was not significantly more frequent in patients who received enlargement of the VSD during Rastelli repair. Longer cardiopulmonary bypass time and longer ischemic time also emerged as risk factors for late death. Cardiopulmonary bypass time and cross-clamp time were not significantly longer in patients who received enlargement of the VSD; hence potential deleterious effects of prolonged cardiac arrest time may affect long-term survival.
Timing of the Rastelli Operation On the one hand, lower age at the time of correction predicts smaller conduits for right ventricular outflow tract reconstruction, and consequently a higher replacement rate. On the other hand, older children sustain a
longer time of palliation, which may lead to myocardial dysfunction as a consequence of prolonged cyanosis. Because the present cohort represents in part a historical collective, the median age of 5 years at the time of correction was rather high. Freedom from conduit reoperation was significantly higher in patients younger than 4 years of age at the time of operation. In the early experience, the surgeons tended to postpone the correction to make VSD baffling easier, and to allow for a larger conduit to be used. With growing experience of early correction in our department, the Rastelli operation was performed in younger patients. Hence, the median age at correction decreased to 2.2 years within the last 5 years of the study period. In the present cohort, there was a trend toward an increased risk for death or transplantation in patients 4 years of age or older at the time of the Rastelli operation. The survival curves of patients younger and older than 4 years of age diverge clearly from 10 years of follow-up onward, indicating that the benefit of early correction might emerge clearly in the longer-term. Interestingly, patients who were 4 years of age or older at the time of the Rastelli operation presented with a significantly higher hemoglobin level. In addition, there was a trend toward an increased risk for late death or transplantation in patients presenting with a higher hemoglobin level at the time of the Rastelli operation. The higher hemoglobin level in patients presenting with older age at the time of the Rastelli operation indicates that longer palliation leads to chronic cyanosis with the consequence of an erythrocytosis. Presumably, the sequelae of prolonged cyanosis, such as myocardial fibrosis, may have a negative effect on survival in the long term.
Conclusions The Rastelli operation is a low-risk procedure with regard to early mortality. The status of the pulmonary arteries and VSD enlargement are predictive for longterm survival. Patients who are 4 years of age or older at the time of the Rastelli operation require fewer reoperations for conduit exchange. However, there is a significant mortality in the long term in patients 4 years of age or older. Therefore, early Rastelli repair is recommended because the initial operation in early infancy bears no higher operative risk and reoperations for conduit exchange can be performed with a low mortality [16].
Study Limitations The study design was a retrospective follow-up study covering a long period of patient inclusion. Changes in preoperative, operative, and postoperative management may have affected the outcome variables in a way not covered by our analysis. According to Concato and coworkers [17], the ratio of events per variable concerning
HÖRER ET AL LONG-TERM OUTCOME AFTER RASTELLI REPAIR
2175
the end point death or transplantation was too small to perform a multivariate analysis.
References 1. Williams WG, McCrindle BW, Ashburn DA, Jonas RA, Mavroudis C, Blackstone EH. Outcomes of 829 neonates with complete transposition of the great arteries 12–17 years after repair. Eur J Cardiothorac Surg 2003;24:1–10. 2. Hanlon CR, Blalock A. Complete transposition of the aorta and the pulmonary artery: experimental observations on venous shunts as corrective procedures. Ann Surg 1948;127: 385–97. 3. Rastelli GC, Wallace RB, Ongley PA. Complete repair of transposition of the great arteries with pulmonary stenosis: a review and report of a case corrected by using a new surgical technique. Circulation 1969;39:83–95. 4. Rastelli GC. A new approach to “anatomic” repair of transposition of the great arteries. Mayo Clin Proc 1969;44:1–12. 5. Lecompte Y, Zannini L, Hazan E, et al. Anatomic correction of transposition of the great arteries. J Thorac Cardiovasc Surg 1981;82:629 –31. 6. Nikaidoh H. Aortic translocation and biventricular outflow tract reconstruction. A new surgical repair for transposition of the great arteries associated with ventricular septal defect and pulmonary stenosis. J Thorac Cardiovasc Surg 1984;88: 365–72. 7. Perloff JK. Systemic complications of cyanosis in adults with congenital heart disease. Hematologic derangements, renal function, and urate metabolism. Cardiol Clin 1993;11:689 –99. 8. Warnes CA, Somerville J. Tricuspid atresia in adolescents and adults: current state and late complications. Br Heart J 1986;56:535– 43. 9. Lee JR, Lim HG, Kim YJ, et al. Repair of transposition of the great arteries, ventricular septal defect and left ventricular outflow tract obstruction. Eur J Cardiothorac Surg 2004;25: 735– 41. 10. Kreutzer C, De Vive J, Oppido G, et al. Twenty-five-year experience with Rastelli repair for transposition of the great arteries. J Thorac Cardiovasc Surg 2000;120:211–23. 11. Lange R, Weipert J, Homann M, et al. Performance of allografts and xenografts for right ventricular outflow tract reconstruction. Ann Thorac Surg 2001;71(Suppl):S365–7. 12. Jatene AD, Fontes VF, Paulista PP, et al. Anatomic correction of transposition of the great vessels. J Thorac Cardiovasc Surg 1976;72:364 –70. 13. Hutter PA, Kreb DL, Mantel SF, Hitchcock JF, Meijboom EJ, Bennink GB. Twenty-five years’ experience with the arterial switch operation. J Thorac Cardiovasc Surg 2002;124:790 –7. 14. Pretre R, Tamisier D, Bonhoeffer P, et al. Results of the arterial switch operation in neonates with transposed great arteries. Lancet 2001;357:1826 –30. 15. Moons P, Gewillig M, Sluysmans T, et al. Long term outcome up to 30 years after the Mustard or Senning operation: a nationwide multicentre study in Belgium. Heart 2004;90: 307–13. 16. Mohammadi S, Belli E, Martinovic I, et al. Surgery for right ventricle to pulmonary artery conduit obstruction: risk factors for further reoperation. Eur J Cardiothorac Surg 2005; 28:217–22. 17. Concato JPP, Holford TR, Feinstein AR. Importance of events per independent variable in proportional hazards analysis. I. Background, goals, and general strategy. J Clin Epidemiol 1995;48:1495–501.
CARDIOVASCULAR
Ann Thorac Surg 2007;83:2169 –75