Primary Arterial Switch Operation as a Strategy for Total Correction of Taussig–Bing Anomaly
A 21-Year Experience
Background—Studies of the arterial switch operation for Taussig–Bing anomaly demonstrate significant rates of reintervention and mortality, particularly after initial palliation to delay complete repair. We aimed to describe the long-term outcomes of our 21-year practice of single-stage arterial switch operation for all patients with Taussig–Bing anomaly.
Methods and Results—A retrospective study was performed, and 43 patients with Taussig–Bing anomaly were identified between 1990 and 2011. Median age at arterial switch operation was 7 (range, 2–192) days, and median operative weight was 3.2 (1.4–6.2) kg. Aortic arch obstruction was present in 30 patients (70%). Hospital mortality was 7% (n=3). Follow-up was available for 37 hospital survivors at a mean of 8.1 (±6.3) years. Late mortality was 2% (n=1). At follow-up, all patients were in New York Heart Association functional class I. Freedom from transcatheter or surgical reintervention was 73% at 1 year, 64% at 5 years, and 60% at 10 years. Eleven patients underwent 13 catheter reinterventions on the pulmonary arteries (n=8) or aortic arch (n=5). Seven patients underwent 11 reoperations, including relief of right ventricular outflow tract obstruction (n=5), pulmonary arterioplasty (n=3), recoarctation repair (n=2), and tricuspid valve repair (n=1). By multivariate analysis, a preoperative aortic valve annulus z score of ≤−2.5 was associated with reintervention (hazard ratio, 7.66 [95% confidence interval, 1.29–45.6], P=0.03).
Conclusions—Although reintervention is common, primary correction of Taussig–Bing anomaly with arterial switch operation can be achieved in all patients with low mortality and good long-term outcomes.
- arterial switch operation
- congenital heart disease
- double outlet right ventricle
- long-term outcome
- Taussig–Bing anomaly
- thoracic surgery
- transposition of great vessels
Taussig–Bing anomaly is a rare congenital cardiac malformation, first described in 1949 by Helen B. Taussig and Richard J. Bing.1,2 The second most common variant of double outlet right ventricle, the anomaly was originally distinguished by a transposed aorta arising entirely from the right ventricle, a pulmonary artery overriding a ventricular septal defect (VSD), side-by-side great vessels, and bilateral subarterial conus with absence of pulmonary–mitral fibrous continuity.1,2 Most surgical series have expanded the definition to include any double outlet right ventricle (with or without pulmonary–mitral continuity) with a subpulmonary VSD, in which left ventricular output flows preferentially to the pulmonary artery.2–4
Despite >30 years of experience with various management strategies, Taussig–Bing anomaly continues to present a considerable surgical challenge. This is largely related to associated anomalies, the most significant of which may be the varying degrees of right ventricular outflow tract and aortic arch obstruction caused by a malaligned ventricular septum.5 Today, primary arterial switch operation (ASO) with VSD closure is the gold standard for Taussig–Bing anomaly.2,6,7 Significant rates of reintervention, mortality, and neoaortic complications have been demonstrated in studies of this lesion, particularly in patients who had undergone palliative procedures such as pulmonary artery banding before definitive repair.6,7
Since 1990, single-stage correction with primary ASO has been our institution’s practice for all children with Taussig–Bing anomaly. We aimed to describe the long-term outcomes and functional status of patients managed with this surgical approach and to identify factors that predict mortality, reintervention, neoaortic insufficiency (AI), and neoaortic root dilation.
After obtaining approval from the Institutional Review Board of Columbia University with a waiver of consent, we retrospectively reviewed our hospital and surgical databases for all admissions diagnosed with transposition of the great arteries or double outlet right ventricle. Patients with Taussig–Bing anomaly were identified as those with a transposed aorta and a pulmonary artery arising in greater part from the right ventricle, overriding the VSD. Anatomy was confirmed on the preoperative echocardiogram, with clarification from the operative note as needed. Perioperative information was collected from the inpatient medical record, and follow-up was obtained from the primary cardiologist.
Data were analyzed using IBM SPSS version 19. Categorical variables are expressed as frequency (percentage), and continuous variables are expressed as mean (±SD) for values with normal distribution or median (range) for skewed data. Log transformation was performed for all data without a normal distribution. To test comparisons between groups, χ2 and Fisher exact tests were used for categorical variables and 2-tailed Student t test for continuous variables. Possible predictors of outcomes with P<0.1 by univariate analysis were then investigated using a multivariate Cox proportional hazards model, and P<0.05 were considered statistically significant.
Between 1990 and 2011, 43 patients met our definition of Taussig–Bing anomaly, and all underwent primary ASO. Patient characteristics are demonstrated in Table 1. We included 1 patient who, at an outside hospital, had undergone aortic coarctation repair via lateral thoracotomy for the purpose of hemodynamic stabilization only 6 days prior to ASO. No patient had undergone palliation to delay the definitive repair.
Median age at ASO was 7 (2–192) days, and median operative weight was 3.2 (1.4–6.2) kg. Three infants died in the hospital. Follow-up was available in 37 of 40 hospital survivors (93%), with a mean duration of 8.1 years (±6.3).
Anatomic characteristics are demonstrated in Table 1. The position of the aortic valve in relation to the pulmonary valve was anterior and rightward in 20 patients (47%), side by side in 20 (47%), and directly anterior in 3 (7%). A usual coronary artery arrangement, with the left anterior descending and circumflex coronaries arising from sinus 1 and the right coronary arising from sinus 2 (classified as 1LCx-2R by the Leiden convention) was noted in 17 patients (40%), whereas the majority had various atypical branching patterns.
Preoperatively, aortic arch obstruction was present in 30 patients (70%), including 24 with arch hypoplasia, 5 with interrupted aortic arch, and 1 with isolated coarctation. Subaortic right ventricular outflow tract obstruction (RVOTO) requiring relief at the time of ASO was present in 12 (28%) patients. Patients with aortic arch obstruction had significantly smaller z scores of the aortic valve annulus (mean, −1.3±1.3) than those with an unobstructed arch (mean, −0.1±1.3; P=0.02). The native pulmonary valve annulus, which would function postoperatively as the aortic outflow, was often dilated, and assigning z scores according to aortic valve criteria yielded a median of 2.5 (−3.4 to 5.4).
Operative data are demonstrated in Table 2. A standard ASO was performed under cardiac arrest on cardiopulmonary bypass, typically using deep hypothermia. The ascending aorta was transected at its midportion, and the pulmonary artery was transected just before bifurcation. In most patients, the VSD was closed via the pulmonary artery using a Gore-Tex patch and continuous running suture. Coronary artery buttons were excised from the aorta and reimplanted using the trap-door technique, if indicated. The neopulmonary root was repaired with 2 separate patches of fresh autologous pericardium.
In those with aortic arch obstruction, the aortic isthmus, coarctation, and all visible ductal tissue were excised, and partial anastomosis was performed between the descending aorta and the distal aortic arch. The proximal arch and often the distal ascending aorta were augmented with cryopreserved pericardium or pulmonary allograft. Relief of RVOTO was achieved by transaortic resection of any infundibular extension of the subaortic conus or by patch augmentation. To avoid compression of an aberrant coronary artery, 1 patient required a right ventricle–pulmonary artery conduit. Defects in the atrial septum were closed. The pulmonary bifurcation was then mobilized and anastomosed to the neopulmonary artery, typically using the Lecompte maneuver if the native aorta was anterior to the pulmonary artery.
Perioperative Complications and Early Morbidity
Eighteen patients (42%) experienced significant postoperative morbidity, including delayed sternal closure or reopening in 8 (19%) patients and permanent pacemaker insertion for complete heart block in 2 (5%) patients. Additional complications are listed in Table 2. Median postoperative hospital stay for survivors was 9 (5–311) days.
Hospital mortality, defined as death before discharge, occurred in 3 patients (7%) during the years 1999, 2001, and 2007. All were the result of a cardiac arrest within 24 hours of surgery. The first death occurred in a 32-day-old, full-term infant with severe intrauterine growth retardation (operative weight, 1.4 kg), intestinal malrotation, and a repaired omphalocele. Additional cardiovascular anomalies included hypoplastic aortic arch with severe coarctation and a single coronary artery from sinus 2. Death occurred after the development of acute hypotension and complete heart block. Autopsy did not reveal a definitive cause of death.
The second death occurred in a 10-day-old, full-term, 3.2 kg infant with interrupted aortic arch, tricuspid valve attachments to the subaortic conus, and 1LCx-2R coronary pattern. Death occurred because of refractory postoperative ventricular arrhythmias, and autopsy demonstrated complete obstruction of the reimplanted right coronary artery.
The third death occurred in a 7-day-old, full-term, 3.3 kg neonate with the unique association of accessory mitral valve tissue resulting in left ventricular outflow tract obstruction (which required resection) and a hypoplastic, bicuspid pulmonary valve. Coronary configuration was 1RL-2Cx. Death occurred after sudden pulmonary hemorrhage and cardiac arrest. Autopsy was unrevealing.
Late mortality occurred in 1 patient (2%). Six months after an uncomplicated neonatal ASO with aortic arch reconstruction, the patient succumbed to severe pneumonia complicated by sepsis, pneumothoraces, and pneumoperitoneum. Echocardiogram had confirmed normal biventricular systolic function before death. Autopsy revealed a moderate residual coarctation. In this cohort, a low mortality rate prevented identification of significant risk factors.
Surgical or transcatheter reintervention was required in 13 hospital survivors, with a median time of 0.8 (0.1–5.1) years to the first reintervention (Table 3). Eleven patients underwent 13 catheter reinterventions, including balloon angioplasty or stenting of the pulmonary arteries (n=8) and balloon angioplasty of the aortic arch (n=5). Seven patients underwent 11 reoperations, including relief of RVOTO (n=5), pulmonary arterioplasty (n=3), recoarctation repair (n=2), and repair of iatrogenic tricuspid valve injury sustained during a previous reoperation (n=1). Of the 7 patients who required reoperation, 2 had undergone unsuccessful transcatheter intervention aimed at the residual lesions (including 1 recoarctation and 1 RVOTO with branch pulmonary artery stenosis).
Freedom from any transcatheter or surgical reintervention was 73% at 1 year, 64% at 5 years, and 60% at 10 years (Figure). After controlling for the effects of preoperative aortic arch obstruction, younger age at ASO, operative weight <3 kg, and longer circulatory arrest times (all of which yielded P<0.10 in univariate analysis), an aortic valve annulus z score ≤−2.5 emerged as a significant predictor of any reintervention (Table 4) with a hazard ratio of 7.66 (95% confidence interval, 1.29–45.6; P=0.03).
Freedom from reoperation was 88% at 1 year and 78% at 5 and 10 years. Longer circulatory arrest time (P=0.01) was associated with reoperation by univariate analysis, but no predictors were identified in multivariate analysis.
The most recent available echocardiogram demonstrated normal biventricular function in all hospital survivors. Moderate or more AI was present in 2 patients. One had moderate-to-severe AI at 2.2 years postoperatively, and another had moderate AI at 5.9 years. Neither had undergone reintervention at latest follow-up. Predictors of significant AI could not be identified because of the rarity of this outcome.
Severe neoaortic root dilation, defined as an aortic sinus z score of >5, was noted in 6 patients (16%). No specific factors were associated with severe dilation, including the preoperative z score (according to aortic valve parameters) of the often dilated native pulmonary valve annulus (P=0.55). No patients had been referred for root intervention at latest follow-up.
Other Long-Term Outcomes
At follow-up, all patients were in New York Heart Association functional class I. All but 3 patients maintained sinus rhythm (2 were paced and 1 was in an atrial rhythm).
Taussig–Bing anomaly has presented a surgical challenge for >30 years. The complexity of this malformation is due to a combination of anatomic features, including subaortic RVOTO,6–8 aortic arch obstruction,2,3,5–7,9,10 and unusual coronary patterns.3,5,6,9,11 These and other patient-related factors may contribute to higher complication, mortality, and reintervention rates compared with other forms of transposition of the great arteries.12 In the past, the outcomes of patients who underwent initial palliation followed by atrial switch were less than satisfactory,13 and palliation to delay the arterial switch resulted in 5-year event-free survival rates as low as 35%.7 This naturally led to single-stage ASO as a primary method of repair in all patients with Taussig–Bing anomaly, and improved outcomes have been achieved as the surgical approach has evolved.7 Benefits of single-stage repair include avoiding multiple surgeries, averting the effects of prolonged cyanosis, and preventing the development of AI, congestive heart failure, ventricular hypertrophy, and pulmonary vascular disease.7,11
To our knowledge, we have identified one of the largest reported cohorts of consecutive patients with Taussig–Bing anomaly to undergo single-stage repair with primary ASO. Our 21-year practice of the currently preferred treatment strategy, as well as our long duration of near-complete follow-up, makes our outcomes data highly applicable to the counseling of today’s patients.
Hospital mortality rates vary widely in the literature from 4% to 21%.5–9 Prior studies have concluded that staged repair,6 higher operative weight,6 and certain coronary artery patterns10 are significant risk factors for mortality after ASO for Taussig–Bing anomaly. The need for aortic arch augmentation during ASO has also been associated with mortality in transposition of the great arteries and its variants,12 but this is not always the case.5,7,10 Although we noted a significant risk of perioperative morbidity in this complex population, 93% of our patients survived to hospital discharge.
Reintervention is common after ASO for Taussig–Bing anomaly, with reported rates of 22% to 40%.3,6,11 In our population, the time to first reintervention was short, and 9 of the 13 patients requiring any reintervention did so within the first postoperative year. Similar to previous studies,7–10,14 branch pulmonary stenosis was the most common reason for catheter reintervention in our cohort, and RVOTO was the most common indication for reoperation.
In our series, transcatheter or surgical reintervention was not predicted by complexities of Taussig–Bing anomaly, such as great vessel relationship, RVOTO, or coronary anomalies. Patients with aortic valve z scores of ≤−2.5, however, were significantly more likely to require reintervention than those with z scores of >−2.5, even after controlling for factors such as aortic arch obstruction.
AI is a time-dependent risk in any patient who has undergone ASO.15 Patients with Taussig–Bing anomaly and other forms of double outlet right ventricle have a higher risk of developing significant AI and of requiring aortic valve replacement compared with patients with transposition of the great arteries.16,17 Prior pulmonary artery banding (which may distort and increase flow velocity across the native pulmonary valve), older operative age, and higher operative weights have previously been associated with the development of significant AI in Taussig–Bing anomaly,6 which further supports single-stage repair as a possible protective strategy. In this cohort, we demonstrated a low rate of significant AI after primary ASO.
Neoaortic Root Dilation
Dilation and stiffening of the aortic root are common after ASO15,18 and may be related to several different factors. These include distortion of the neoaorta by prior pulmonary artery band, impairment of flow to the vasa vasorum of the neoaorta, complex suture lines at sites of coronary reimplantation, and inherent properties of the pulmonary arterial root functioning in a systemic position.18 During the ASO for Taussig–Bing anomaly, the aorta is anastomosed to a pulmonary artery that is frequently quite dilated, and neoaortic root z scores above normal can be expected to persist as the child grows. Despite this, few of our patients developed severe neoaortic root dilation, and no patient had required root intervention at latest follow-up.
The current report assumes the typical limitations of a retrospective study. Changes in perioperative management during the study period may have affected our results. Finally, details about the progression of neoaortic root dilation and AI over time were unavailable and would be beneficial to this study.
In one of the largest series to date, it is shown that although reintervention is common, primary correction of Taussig–Bing anomaly with ASO can be achieved in all patients with low mortality and good long-term outcomes. The need for reintervention begins most often during the first postoperative year; however, close lifetime follow-up is warranted to assess for development of neoaortic valve insufficiency and root dilation. A preoperative aortic valve annulus z score of ≤−2.5 is associated with the need for reintervention.
- © 2013 American Heart Association, Inc.
- Comas JV,
- Mignosa C,
- Cochrane AD,
- Wilkinson JL,
- Karl TR
- Alsoufi B,
- Cai S,
- Williams WG,
- Coles JG,
- Caldarone CA,
- Redington AM,
- Van Arsdell GS
- Griselli M,
- McGuirk SP,
- Ko CS,
- Clarke AJ,
- Barron DJ,
- Brawn WJ
- Huber C,
- Mimic B,
- Oswal N,
- Sullivan I,
- Kostolny M,
- Elliott M,
- de Leval M,
- Tsang V
- Masuda M,
- Kado H,
- Shiokawa Y,
- Fukae K,
- Kanegae Y,
- Kawachi Y,
- Morita S,
- Yasui H
- Lange R,
- Cleuziou J,
- Hörer J,
- Holper K,
- Vogt M,
- Tassani-Prell P,
- Schreiber C