| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2006;82:2059-2066
© 2006 The Society of Thoracic Surgeons
a Department of Cardiovascular Surgery, Children’s Hospital Boston—Harvard Medical School, Boston, Massachusetts
b Department of Cardiology, Children’s Hospital Boston—Harvard Medical School, Boston, Massachusetts
Accepted for publication June 9, 2006.
* Address correspondence to Dr del Nido, Department of Cardiovascular Surgery, Children’s Hospital Boston, 300 Longwood Ave, Bader 273, Boston, MA 02115 (Email: pedro.delnido{at}tch.harvard.edu).
Presented at the Forty-second Annual Meeting of The Society of Thoracic Surgeons, Chicago, IL, Jan 30–Feb 1, 2006. Winner of the J. Maxwell Chamberlain Memorial Award for Congenital Heart Surgery.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
METHODS: From August 1992 to July 2005, 44 patients (median age at surgery, 1.6 years; range, 0.6 to 39.6 years) with congenitally corrected transposition of the great arteries had anatomic repair. Left ventricular function and mitral regurgitation were evaluated by echocardiography at follow-up. Twenty-three patients had a Rastelli procedure, and 21 underwent an arterial switch. Twelve patients (27%) were pacemaker dependent at latest follow-up.
RESULTS: Early mortality was 4.5% (n = 2) with 1 late death as a result of leukemia. Median follow-up was 3.0 years (range, 7 days to 12.4 years). Left ventricular function remained unchanged (normal) in 35 patients, improved in 1 patient, and deteriorated in 8 patients (18%). Mitral regurgitation was unchanged in 30 patients, improved in 6 patients, and worsened in 8 patients (18%). Development of left ventricular dysfunction was significantly associated with pacemaker implantation (p = 0.005) and a widened QRS (>20% > 98% percentile of normal; p = 0.03).
CONCLUSIONS: Anatomic repair of congenitally corrected transposition can be performed with low operative mortality. However, late left ventricular dysfunction is not uncommon, with higher incidence in those requiring pacing and with a prolonged QRS. Resynchronization may be of value in patients requiring a pacemaker.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Decreased longevity in patients with ccTGA has been documented in natural history studies as well as multiinstitutional registries [2]. Classic or physiologic repair operations for ccTGA have been possible since the late 1950s [3]. This approach was directed at treating the associated defects such as VSD or subpulmonary obstruction. However, in patients undergoing a physiologic approach as well as in unoperated on patients, the morphologic right ventricle (RV) remains supporting the systemic circulation, potentially leading to late RV dysfunction and tricuspid regurgitation (TR) with disappointing late results. In fact, systemic ventricular failure may affect as many as 40% of patients within 3 years of physiologic repair [4, 5] and 56% of unoperated on patients older than 45 years [2].
Three key breakthroughs in cardiac surgery have resulted in anatomic repair in the late 1980s and early 1990s [6, 7]: first, the extensive experience accumulated with the Senning and Mustard procedures for atrial switch for d-transposition of the great arteries and the Rastelli operation in patients with associated pulmonary stenosis; second, the excellent surgical results obtained with the arterial switch procedure for d-transposition of the great arteries with or without VSD [8–11]; and third, the pioneering reports of pulmonary artery banding to retrain the left ventricle (LV) in preparation for an arterial switch [12]. The anatomic repair or so-called double-switch is a challenging operation that incorporates these three surgical advantages. The LV is connected to the aorta by an arterial switch or Rastelli operation, and AV concordance is established by a Senning or Mustard operation. In patients with preoperative low LV pressure (untrained LV), the pulmonary artery is banded to "prepare" the LV to become the systemic circuit. With increasing surgical experience, and choosing select physiology such as infants with VSD and pulmonary hypertension, or with significant LV outflow tract obstruction, the double-switch is now performed at a younger age not requiring pulmonary artery banding. Excellent early and mid-term results with good hemodynamics, normal RV and tricuspid valve function, and low morbidity and mortality have led to wide adoption of this approach. Nevertheless, postoperative LV dysfunction late after anatomic correction of ccTGA has been reported and remains a concern. Factors contributing to LV deterioration are poorly understood. In this report we review our experience with anatomic repair of ccTGA and evaluate its intermediate results, focusing on factors influencing LV function.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Preoperatively RV function, qualitatively assessed echocardiographically, was normal or mildly depressed in 40 patients (91%) whereas 4 patients (9%) had moderately depressed RV function. Tricuspid regurgitation, assessed qualitatively by echocardiography, was none, trivial, or mild in 30 patients (78%) and significant (moderate or severe) in 14 patients (31%), with 8 patients having Ebstein-like anomaly of the tricuspid valve with displaced or tethered septal leaflet. Left ventricular function was normal in 41 patients (93%) and mildly depressed in 2 patients (5%), with no patients having moderate LV dysfunction and 1 patient (2%) having severely depressed LV function before the double-switch procedure. No patient had more than trivial or mild mitral regurgitation before repair, with none or trivial in 33 patients (75%) and mild in 11 patients (25%). To assess accuracy of echocardiography, we also reviewed magnetic resonance imaging studies and catheterizations when available.
Electrocardiograms were examined preoperatively and at the latest follow-up visit for morphologic changes, and QRS duration was measured using electronic calipers. The value of QRS duration was adjusted for age by calculating the percentage above or below the 98th percentile for normal individuals of the same age. Significantly prolonged QRS was defined as 20% above the 98th percentile for patient age.
Surgical Technique
Before anatomic repair, 14 patients (32%) had a modified Blalock-Taussig shunt and 5 (11%) had pulmonary artery banding to retrain the LV. Atrial switch was accomplished by a Senning procedure in 22 patients and a Mustard operation in 22. Twenty-three patients had a Rastelli procedure for LV outflow reconstruction, and 21 underwent an arterial switch. Ten patients with subpulmonary stenosis had extensive resection of accessory valve tissue and underwent an arterial switch. The technical details of the procedures performed to restore the AV concordance (Mustard or Senning) as well as the ventriculoarterial concordance (Rastelli or arterial switch) have been previously described [13]. If during rewarming or before surgery there was evidence of high-grade AV block, permanent dual-chamber epicardial leads were implanted.
Data Collection and Statistical Analysis
Data were obtained from retrospective review of hospital charts and medical records preoperatively and postoperatively. Ventricular function was calculated by qualitative assessment. When significant dysfunction was present we correlated the data with echocardiographic end-systolic wall stress to fractional shortening and velocity of circumferential fiber shortening [14], and magnetic resonance imaging studies when available. Mitral regurgitation and TR were assessed by means of the vena contracta jet, with less than 4 mm being mild regurgitation; from 4 to 6, moderate; and more than 6 mm, severe regurgitation. We considered significant mitral regurgitation or TR when the insufficiency was moderate or higher. We compared RV and LV function and mitral regurgitation and TR before repair with the most recent echocardiogram to determine progression.
Relationships between patient and surgical characteristics and the presence of significant LV dysfunction at most recent follow-up were examined using Fisher’s exact test; continuous variables such as age at surgery were categorized before the analyses were performed.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
One patient in our series required heart transplantation 7 days after repair. This was a 3-year-old child with severe preoperative TR caused by an Ebstein-like abnormality, straddling tricuspid valve and a small RV. The TR persisted after anatomic repair, and this was the only patient in our series having severe TR postoperatively. No attempt was made to repair the tricuspid valve at the time of corrective surgery. This patient had a cardiac arrest within hours of surgery and was placed on extracorporeal membrane oxygenation while awaiting transplantation. Preoperative QRS duration was 80 milliseconds.
Long-term outcomes were analyzed in the 41 patients who were successfully discharged from the hospital and did not require cardiac transplantation. There was only 1 late death attributable to leukemia 15 months after surgical repair. Postoperatively, 6 patients required reintervention: 2 patients for pulmonary venous pathway stenosis (1 Mustard and 1 Senning), 2 in the Rastelli group requiring conduit replacement, 1 closure of residual VSD, and 1 patient with preoperative pulmonary artery banding who exhibited severe regurgitation requiring aortic valve replacement.
Ten patients (23%) with LV outflow tract obstruction had extensive resection of subpulmonary stenosis with the aim of performing an arterial switch procedure rather than a Rastelli operation to restore the ventriculoarterial concordance. Two patients in this group had LV dysfunction during the follow-up, one moderate and one severe; one of them died 15 months after repair as a result of leukemia. Both patients with LV dysfunction had prolonged QRS on electrocardiogram. No patient in our series experienced LV outflow tract obstruction during the follow-up period.
At most recent evaluation, at median follow-up of 3.0 years (range, 7 days to 12.4 years), RV function was moderately depressed in 1 patient (2%) whereas 40 patients (98%) had normal RV function after repair. Tricuspid regurgitation was significant in 4 patients (10%) after repair, with 2 of these having an Ebstein-like anomaly of the tricuspid valve. Right ventricular function (p = 0.015) and TR (p = 0.048) improved after anatomic repair. Mitral regurgitation was moderate in 3 patients (7%) on late follow-up with no patient having severe mitral regurgitation at most recent echocardiography.
Twelve patients required chronic ventricular pacing for nodal AV dysfunction. Of these 12 patients, 4 had preoperative ventricular pacing at a mean of 4.0 years before undergoing repair, 4 had preoperative AV block, and 4 experienced AV block perioperatively and had pacemaker implantation before discharge. All the pacemakers were implanted before or soon after the operation with no late pacemaker implants. All patients received a dual-chamber system. In 18 patients (44%) the follow-up electrocardiogram showed a QRS duration that was higher than 20% above the 98th percentile for normal individuals of the same age.
Left ventricular function on late follow-up was normal in 35 patients (85%), mildly depressed in 2 (5%), moderately depressed in 2 (5%), and severely depressed in 2 (5%). Cardiac catheterization revealed no coronary artery abnormalities in patients with LV dysfunction. The 4 patients with moderate to severe LV dysfunction also had AV block requiring ventricular pacing. Therefore, 33% of the 12 patients requiring long-term pacing also had moderate to severe LV dysfunction. Tables 1 and 2
display the preoperative and postoperative determinants of late LV dysfunction. The only significant determinants were the need for chronic ventricular pacing (p = 0.005) and prolonged QRS duration (p = 0.03). Other related variables with borderline significance were presence of arrhythmia in the electrocardiogram (p = 0.011), QRS duration greater than 80 milliseconds (p = 0.12), electrical dyssynchrony (p = 0.11), and percentage of QRS greater than the 98th percentile (p = 0.076).
|
|
|
Comment |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Left ventricular dysfunction appearing late has not been a commonly described complication late after anatomic repair for ccTGA. However, several reports have reported deterioration of LV contractility after double-switch operation. In 1994, Yagihara and colleagues [19] reported a series of 10 patients with ccTGA who had anatomic repair. Eight patients received a Rastelli procedure and 2 an arterial switch for restoration of the AV concordance. There was 1 early and 2 late deaths in this group. Four patients experienced LV dysfunction (ejection fraction < 0.55), and 2 patients had rhythm disturbances. Interestingly, the 2 late deaths reported occurred in patients with both impaired LV function and complete heart block [19]. Brawn and Barron [20] reported a series of 54 patients with an early and late mortality of 3 and 2 patients, respectively. New signs of LV dysfunction were observed in 6 patients, 2 of whom had neoaortic regurgitation. All patients in the double-switch group had some degree of regurgitation. In 4 patients it was moderately severe, requiring valve replacement in 2 [20]. In a recent report describing the same 54 patients with ccTGA, Langley and coworkers [21] found 6 patients (11%) with acquired impaired systolic ventricular function after anatomic repair. One required heart transplantation and another was awaiting transplantation for LV dysfunction. Of the 49 survivors, 46 (94%) were in New York Heart Association class I, even though 3 of these patients had impaired LV performance. Furthermore, the 2 late deaths observed were attributable to LV dysfunction. No information was provided regarding pacemaker dependency or QRS duration and LV dysfunction. Debaney and associates [22] reported a series of 22 patients undergoing arterial switch operation and Senning procedure with no early or late mortality. One patient required cardiac transplantation for progressive LV failure after a preliminary banding [22]. In a series of 22 patients reported by Imamura and colleagues [23], 1 patient had moderate LV dysfunction 5 months after a Rastelli plus Senning operation. Interestingly, for these authors mild pulmonary stenosis or dynamic subpulmonary stenosis did not preclude the performance of an arterial switch and an atrial switch. In fact, their findings indicate that this may prove to be advantageous in that the LV is likely preconditioned by this obstruction [23]. Acquired mild LV impairment was also observed by Karl and associates [16] in a series of 14 patients undergoing a double-switch procedure.
Inasmuch as patients with ccTGA rarely manifest symptoms requiring correction during infancy, palliation with a systemic-to-pulmonary shunt to treat cyanosis is often performed when necessary and repair is frequently delayed beyond the first few months and occasionally years of age. In our series, 22 patients (50%) had corrective surgery between 3 months and 12 months of age. Because of the frequent existence of subvalvar or valvar pulmonary obstruction and of a concomitant large VSD in our study group, pulmonary artery banding was rarely required. The largest experience with retraining the LV in preparation for a double-switch procedure was reported by Poirier and Mee [24] in 2000. In this study, 84 patients (45 with ccTGA), underwent pulmonary artery banding before the arterial switch operation. The overall mortality was 15.4%, and 91% of the survivors showed normal LV function at follow-up echocardiography. Mean follow-up for patients with ccTGA was 21 months. The results in prepubescent patients were good, but the response of older patients was less predictable and was associated with a higher early and late mortality [24]. Nevertheless, pulmonary artery banding is not an innocuous procedure. Various authors have reported significant morbidity and mortality after the operation [25, 26]. Furthermore, we have recently identified pulmonary artery banding before arterial switch operation as a risk factor for the development of aortic root dilatation and aortic regurgitation in the long-term [27], and neoaortic regurgitation has been recently reported as a factor contributing to late LV dysfunction in patients having atrial plus arterial switch procedure for ccTGA [28]. In our study, 5 patients had pulmonary artery banding in preparation for the double-switch procedure at a mean age of 5.8 years (range, 0.4 to 11.5 years). All patients in this group but one, with mild depression, have good LV function at follow-up. However, 3 of the patients have mild to moderate aortic regurgitation, one requiring aortic valve replacement. Thus, our experience is similar to that reported by other centers with respect to neoaortic regurgitation in banded patients but not with respect to LV dysfunction. This may be related to the age at anatomic repair, with our patients being on average younger than those reported in other series [21, 24].
The most frequent early complication after surgical correction of ccTGA [21, 29, 30] is complete heart block. Patients with ccTGA are at risk of complete heart block even without surgery owing to the location and characteristics of the conduction system in ccTGA. The estimated incidence of complete heart block is 2% of patients per patient-year after diagnosis, and 40% of the patients develop complete heart block during a 20-year period [31]. Loss of AV synchrony can contribute to LV dysfunction because of ineffective contraction. In fact, various reports have identified the presence of arrhythmias or pacemaker implantation as risk factors for systemic ventricle dysfunction in unoperated on patients with ccTGA [2]. In our study, 14 patients had dual-chamber pacemakers inserted, and 12 of these were pacemaker-dependent for AV conduction, thus had ventricular pacing. These 12 patients were more likely to exhibit postoperative LV dysfunction than those without a pacemaker (p = 0.005). In fact, all patients who exhibited LV dysfunction at follow-up were pacemaker-dependent. Furthermore, increased duration of QRS was also related to LV dysfunction. The most likely scenario is that QT prolongation and dyssynchrony (paced or intraventricular conduction delay or bundle-branch block) in conjunction with some baseline ventricular dysfunction leads to progressive dysfunction. In fact, the biggest change in QRS was in the immediate postoperative period with minimal long-term change.
Cardiac resynchronization therapy has been proposed for the treatment of heart failure in adults with decreased ventricular function and interventricular conduction delay (prolonged QRS complex) who remain symptomatic despite optimal medical therapy. Recent data indicate that cardiac resynchronization therapy is an effective therapy for symptomatic heart failure in children as well, in the acute and the chronic setting and after heart surgery [32–37]. Although some groups have proposed clinical indication for resynchronization therapy in children, there is a widely accepted approach to pediatric patients with decrease ventricular function and interventricular conduction delay [38]. Because either ventricular pacing (as part of dual-chamber pacing) or prolonged QRS or both were present in all patients with moderate or severe LV dysfunction in our study, we speculate that resynchronization therapy may be effective in improving LV function in this subgroup of patients by means of avoiding the asynchronous contraction of the LV because of the free wall and the septum contraction at different times. In fact, in our series 1 patient with congenital complete AV block with a QRS of 130 to 140 milliseconds, dual-chamber pacemaker, and severe LV dysfunction improved ventricular size and function after upgrading to a biventricular system.
In conclusion, LV dysfunction after anatomic repair of ccTGA is a potentially serious late complication. We found that although most patients who experienced LV dysfunction had a mild to moderate decrease in LV performance during follow-up, the concern exists as to whether this problem will be progressive. We found an association with late-onset LV dysfunction and pacemaker implantation or interventricular conduction delay, which leads us to speculate that resynchronization therapy may be effective in this patient cohort.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
In their series of 44 patients, about half received a Rastelli and the other half a switch in combination with atrial redirection by either Senning or Mustard technique. Survival was excellent for this complex group of patients, with 3 deaths, but the need for permanent pacemaker therapy was high, 27% requiring a permanent pacer implantation.
Using predominantly qualitative echocardiographic techniques, their findings demonstrated that patients requiring a pacemaker and those with a prolonged QRS complex were more likely to have left ventricular dysfunction, and the authors concluded that resynchronization may be of value in these patients. Importantly, they did not identify left ventricular retraining as a predictor of decreased late function, but the number of patients in that group was small.
The authors’ experience is virtually identical to our own. In Ann Arbor we have performed 44 anatomic repairs for congenitally corrected transposition, 24 patients undergoing a switch and 20 a Rastelli. There were 3 early and late deaths, and 1 went on to transplantation. The results in the switch group in our series have generally been more favorable than those in the Rastelli group, and I note in the authors’ abstract that they identified patients with a Rastelli operation as having worse left ventricular function but did not reach the same conclusion in the presentation or the manuscript. Perhaps they could explain that.
Our experience, as well as that reported this morning, prompts a number of questions. It would not seem likely that patients with congenitally corrected transposition and a large VSD would have LV dysfunction following either type of anatomic repair. Thus the key question perhaps relates to those undergoing LV retraining. How old were the patients who underwent retraining of the anatomic left ventricle in the authors’ series? Do the authors feel that retraining perhaps is limited by the development of left ventricular hypertrophy but without associated hyperplasia and neovascularization, which can result in subendocardial fibrosis and diastolic dysfunction, something we have identified particularly in our older patients? In light of their findings, what do they currently recommend regarding retraining for patients to undergo a double switch?
When the authors refer to resynchronization, it wasn’t immediately obvious what they meant. Is that the establishment of AV synchrony, LV pacing, or biventricular pacing? It is now well recognized that pacing the right ventricle results in asynchronous contraction of the left ventricle as the free wall and the septum contract at different times and that LV pacing is an effective mode to overcome this. How was permanent pacing established in the patients who already had LV dysfunction? Were the leads on the anatomic left or right ventricle? Were all patients paced with AV synchrony? Do they now currently recommend routine biventricular pacing for all patients?
I congratulate the authors for undertaking this important study and providing crucial data to guide us in the treatment of this challenging condition. Thank you.
DR BAUTISTA-HERNANDEZ: Thank you for your comments Dr Bove. Regarding the question about the high incidence of pacemakers in our series, I would like to point out that we have only 11% of new-onset AV block after operation. As I had shown before in my presentation, 4 patients had perioperative AV block requiring a pacemaker. So 11% of new complete AV block requiring a pacemaker is just about the incidence reported by other authors.
With regard to pulmonary artery banding, I also would like to highlight that for us that operation is not an innocuous procedure. We think that it is associated with morbidity and mortality, and furthermore, we have recently reported the association of pulmonary artery banding with aortic root dilation and aortic insufficiency. With increasing surgical experience and improved timing of surgery, and due to the fact that many patients have subvalvular or valvular pulmonary obstruction and a large VSD, we think that the double switch can be performed at a younger age, avoiding pulmonary artery banding in the vast majority of the cases.
With respect to the Rastelli question, in the abstract we reported that Rastelli was associated with LV dysfunction, but this was when we considered the whole population; that means the 44 patients. After extensive discussions among the authors, we decided to exclude the 2 early deaths and the 1 early transplantation for the analysis of long-term outcomes. So in this case, the Rastelli operation doesn’t come out like a risk factor for LV dysfunction.
And the last question about recommendations for resynchronization, we believe that any patient with a pacemaker or a prolonged QRS on electrocardiogram should be considered for resynchronization therapy, that is, biventricular pacing.
DR MUHAMMAD MUMTAZ (Cleveland, OH): I had the privilege of working with Dr Roger Mee. He presented his data at the World Congress of Cardiology last year, and I helped him with the data. Our data were very similar, and not to just repeat it, but we had 75 patients, out of which 30 were Senning-Rastelli and 45 were Senning-switch, and one of the risk factors in our series was LV retraining at an older age, like Dr Bove mentioned.
My question is related to the atrial switch. I noticed there were 22 patients with Mustard, and in our series all of the patients had a Senning, none of them had a Mustard. I am just curious to know what was the reason or on what basis did you decide to do a Mustard versus a Senning procedure?
DR BAUTISTA-HERNANDEZ: Thank you. The atrial switch procedure depended on surgeon’s preferences, and we found in our study that the Mustard or the Senning operation at the atrial level didn’t have any relation to LV function or late outcome.
DR MUMTAZ: The concern there is not LV function but late baffle problems and pulmonary venous obstruction, and those usually occur years later and not necessarily early on.
DR BAUTISTA-HERNANDEZ: We didn’t have any baffle obstruction in the late follow-up.
DR MUMTAZ: Thank you.
DR BAUTISTA-HERNANDEZ: You are welcome.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  V. Bautista-Hernandez, F. Serrano, J. M. Palacios, and J. M. Caffarena Successful Neonatal Double Switch in Symptomatic Patients With Congenitally Corrected Transposition of the Great Arteries Ann. Thorac. Surg., February 1, 2008; 85(2): e1 - e2. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
