Long-term follow-up after simultaneous arterial switch operation and aortic arch repair

Abstract

Open in new tabDownload slide

OBJECTIVES

Taussig–Bing anomaly (TBA) and transposition of the great arteries (TGA) with hypoplastic or interrupted aortic arch (AA) are rare anomalies. Various operative techniques and a high incidence of reinterventions are described. The aim of this retrospective single-centre study was to evaluate operative data, mortality and reintervention rate with special regard to the AA.

METHODS

At the Children’s Heart Center Linz, 50 patients with the above-mentioned diagnosis have been corrected by a simultaneous repair between 2001 and 2022. Thirty-seven children had TBA, 13 had TGA and 5 of them had an interrupted AA. The median age at operation was 7 [interquartile range (IQR) 5–9] days, weight 3.38 (IQR 2.9–3.8) kg and follow-up 9.3 (IQR 3.1–14.5) years. The AA reconstruction was performed without patch material in 49 cases.

RESULTS

There was 1 in-hospital mortality in a TBA patient and 1 late mortality (7 years later, neuroblastoma). 14/49 patients needed at least 1 reoperation (28.6%, all TBA) and 3 further patients had catheter reintervention or radiofrequency ablation only (6.1%, 2 TBA). Seventy-five percent of these procedures affected the right heart/pulmonary arteries; there was 1 re-coarctation repair.

CONCLUSIONS

The simultaneous correction of TBA and TGA with AA obstruction or interruption is a safe operation with very low mortality. The AA reconstruction with minimized use of patch material resulted in a low restenosis rate.

INTRODUCTION

Historically, the surgical two-stage strategy of Taussig–Bing anomaly (TBA) and transposition of the great arteries (TGA) with associated aortic arch (AA) obstruction has been a frequent approach [1, 2]. In the current era, the one-stage repair of above-mentioned diagnoses seems to be preferred at most centres [1, 3–11]. Nevertheless, in special anatomic cases as those with interrupted AA as well as in low weight babies, the question of one- or two-stage strategy may still come up nowadays. Apart from timing of the corrective surgery, decisions regarding the operative technique as type of the AA reconstruction or performance of a Le Compte manoeuvre have to be made. This retrospective single-centre study should evaluate peri- and postoperative data, mortality and reintervention rate in a relatively large series of simultaneous arterial switch operation (ASO) and AA repair with minimized use of patch material for the AA reconstruction.

PATIENTS AND METHODS

Ethics statement

This study was approved by the ethics committee of the Medical Faculty at Johannes Kepler University Linz on 6 April 2022 (EK Nr: 1065/2020). Informed consent was waived because of the retrospective nature of the study and the use of anonymous clinical data for the analysis.

Patients

At the Children’s Heart Center Linz, 50 patients have undergone simultaneous repair of TBA or TGA with either hypoplastic or interrupted AA or coarctation between 2001 and 2022. Patient’s details are provided in Table 1. In patients born at our centre, the single-stage repair was performed during the neonatal period. Two of them were early preterm babies [1 TBA and 1 TGA with intact ventricular septum (IVS)] born during the 31st week of gestation. They were operated at their 7th and 9th day of life with a weight of 1.8 and 2.0 kg. The most important extracardiac pathological findings in patients born at our centre were an imperforate anus in 1 and a necrotizing enterocolitis and meningitis in another newborn. Six patients, who came from other centres, showed up late: 1 girl with TBA and interrupted AA was already provided with a stent in the arterial duct and a pulmonary artery (PA) banding, and she was operated at our hospital at the age of 55 days. Another boy with TBA, restrictive ventricular septal defect (VSD) and hypoplastic AA from abroad had a status post-PA banding, and the complete repair was performed at the age of 4.3 years. Although these 2 patients actually were operated before, they were included in our series because of their simultaneous performance of an ASO and AA reconstruction. The residual 4 patients (2 TBA, 1 TGA/VSD, 1 TGA/IVS) did not have any surgical procedure before and were provided with a single-stage repair on days 46, 63, 64 and 85 of life. In 2 of them, a diagnostic catheterization was performed preoperatively.

Operative technique

All operations included an ASO, a VSD closure by a pericardial patch, if necessary, an ASD closure and an AA repair. At the Children’s Heart Center Linz, every AA reconstruction since 2003 has been performed using whole body perfusion [12], which was the case in 44 patients of this cohort. Antegrade cerebral perfusion was used in 6 patients, who were operated before 2003. The AA repair was performed by resection and extended end-to-end anastomosis (REEEA) in 33 children, by ascending descending aortic anastomosis in 12 children. In 2 patients with interrupted AA type A, an isolated end-to-side anastomosis was possible (Fig. 1a); in the third patient, an additional AA enlargement using a curved polytetrafluorethylene patch on the inner curvature was carried out [13] (Fig. 1b). In the 2 cases with interrupted AA type B, the left subclavian artery was ligated first and then an end-to-side anastomosis was performed (Fig. 1c). The VSD could be closed via the right atrium in 46 cases, partially via the right atrium plus transpulmonary in 3 cases and once through an incision in the right ventricular outflow tract (RVOT). The decision for a trap-door plasty out of a pericardial patch was made directly before the reimplantation of the coronary button into the neoaortic root, when there seems to be too much tension or kinking despite the extensive mobilization. This was the case in 18 children. The size mismatch between neoaortic root and ascending aorta was regarded as significant in 22 cases; therefore, in these patients, a pericardial patch for ascending aorta enlargement was used. A Le Compte manoeuvre was performed in all children. The pulmonary bifurcation was transferred towards the right PA in 18 patients and towards the left PA in 2 patients. For the neo-pulmonary anastomosis, a nonabsorbable suture was used in the first 5 cases, and then, we switched to absorbable material.

Statistics

All data of continuous variables were checked for normal distribution (test of normality: Kolmogorov–Smirnov with Lilliefors significance correction, type I error = 10%) and in the case of normal distribution also for variance heteroscedasticity (Levene test, type I error = 5%). In the case of normality and variance homogeneity, the independent two-sample t-test was used for subgroup comparisons. In the case of normality but no variance homogeneity, Welch’s t-test was applied. For variables without normally distributed data and for variables measured on ordinal scales, the exact Mann–Whitney U-test was used. Dichotomous variables were compared by the Fisher’s exact test, and the other categorical variables by the exact chi-square test. For the comparison of the occurrence of reinterventions, depicted by Kaplan–Meier plots, the log-rank test was used. Two-sided 95% confidence intervals of incidences were calculated according to Clopper–Pearson.

The influence of covariables on the occurrence of reinterventions on the pulmonary arteries was investigated by Cox regression analyses. The type I error was not adjusted for multiple testing. Therefore, the results of inferential statistics are descriptive only. Statistical analyses were performed using the open-source R statistical software package, version 4.1.2 (The R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Peri- and postoperative data and mortality

The median bypass and aortic cross-clamp times are presented in Table 1. Both were significantly longer in TBA than in TGA patients (P = 0.003 and <0.001), which is also owed to the included cases without VSD in the TGA group. In total, 18 children (36.0%) required a trap-door plasty and 20 children (40.0%) a pulmonary bifurcation transfer. Both were significantly more frequent in TBA than in TGA patients (48.6% vs 0%, P = 0.002 and 51.4% vs 7.7%, P = 0.008). A patch enlargement of the ascending aorta was performed in 19 TBA and 3 TGA cases (P = 0.108). The median stay on intensive care unit was 12 [interquartile range (IQR), 7–15] days, and the median hospital stay was 20 (IQR, 15–25) days; these results were similar if only looking at the 44 newborn patients: 12 (IQR, 8–15) and 20 (IQR, 15–25) days. No significant difference between TBA and TGA cases was seen (P = 0.673 and 0.427). There was 1 in-hospital mortality of a boy with TBA. He was operated on his 7th day of life (3 kg) and had a hypoplastic AA, side-by-side position of the great arteries, a complex coronary situation and required extracorporeal membrane oxygenation (ECMO) therapy postoperatively. Unfortunately, he died on postoperative day 9 because of multiorgan failure. During the complete follow-up period of 9.3 (IQR, 3.1–14.5) years, 1 further patient with TBA died 7 years postoperatively due to a neuroblastoma.

Complications

Three further newborns with TBA required ECMO therapy postoperatively. Twice this was caused by a junctional ectopic tachycardia, and these patients could be weaned off ECMO without problems. One other girl required ECMO therapy because of low saturation and could be weaned successfully after receiving a PA patch plasty 5 days later.

In a subgroup comparison of ECMO versus non-ECMO patients, a significant difference was only shown regarding the incidence of a trap-door plasty (100% in ECMO patients vs 30.4% in non-ECMO patients, P = 0.013) and, as assumed, duration of ICU stay (median 27 vs 12 days, P = 0.039). No significant difference was seen regarding TBA/TGA, coronary anatomy, AA pathology, weight or age at operation, cross-clamp or cardiopulmonary bypass times, type of AA reconstruction, VSD, hospital stay or other possible influencing factors.

There was 1 patient with postoperative aortic valve (AV) block III, who required a permanent pacemaker implantation. In total, 13 patients showed arrhythmias peri- or postoperatively. Apart from the described cases, there were only temporary arrhythmias that could be treated conservatively (2× temporary AV block, 8× tachycardia).

We had 4 newborns with peri- or postoperative intracerebral complications. There was 1 intracerebral haemorrhage in the preterm boy with TGA/IVS, who was operated at a weight of 2.0 kg. At his last follow-up, he was 13 years old, went to school and showed mild development disability. Three further children had minor cerebral events peri- or postoperatively with no clinical consequences after 2, 8 and 12 years of follow-up. The first patient of the whole cohort (TGA/VSD/hypoplastic AA) developed a postoperative peritonitis requiring diagnostic laparotomy, drainage and antibiotics. After 18 years of follow-up, the same patient was diagnosed with a combined development disability without any evidence for a neurological event during the hospital stay. In 4 children, at least a temporary unilateral paresis of the diaphragm was described without clinical problems during the follow-up period. There was no case with a postoperative left bronchial stenosis.

Reinterventions

14/49 patients (28.6%) needed at least 1 surgical reintervention, in 8 of them transcatheter reinterventions were performed additionally. All of the re-operated patients had a TBA (TBA versus TGA, P = 0.011). An overview regarding the type of surgical reintervention is shown in Table 2. Two further patients (1 TBA and 1 TGA/VSD) only had catheter reinterventions, and 1 boy (TBA) 2 radiofrequency ablation procedures (6.1%). The Kaplan–Meier curve regarding time to surgical reintervention is provided as Fig. 2, regarding time to surgical or transcatheter reintervention as Fig. 3.

One girl with TBA had multiple VSDs, which could not be closed completely during her first operation. Therefore, 198 and 200 days later 2 further operations were necessary. The same patient was the case with an atrioventricular block III after the first procedure. During the follow-up, she also developed a pacemaker infection and needed in total 5 electrophysiological reoperations.

Reinterventions on the right ventricular outflow tract and pulmonary arteries

In total, 18 surgical procedures in 11 patients affected the RVOT, pulmonary valve or arteries. In 4 patients, 2–4 operations regarding right ventricle-to-pulmonary artery (RVPA) conduit as well as 1–4 transcatheter reinterventions at the same location were needed. Three of them had the problem of a right coronary artery crossing the RVOT (Yacoub type D) and an RVPA conduit had to be implanted 31, 101 and 221 days after the primary repair. One of these children initially had an interrupted AA, and this was the only patient with interrupted AA who needed any reintervention during the whole follow-up period. One child with Yacoub type B required RVPA conduit implantation on postoperative day 510. Two of these 4 RVPA conduit patients developed an infective endocarditis during the follow-up period, once on a transcatheter implanted valve and once on a pulmonary homograft. Both infected valves were replaced by homografts [11]. In total, 6 patients needed a patch plasty on the main PA each because of a supravalvular stenosis after 216–5170 days. In 2 children, only transcatheter dilatations of the main and once also on the right PA were necessary. We did not find a statistical difference regarding type of AA reconstruction and need for surgical and transcatheter reintervention on the pulmonary arteries: the reintervention rate after ascending descending aortic anastomosis was 25.0% vs REEEA 21.2% (P = 0.965). The Kaplan–Meier curve is provided as Fig. 4. In the Cox regression analyses none of the following covariables showed an influence on the occurrence of reintervention on the pulmonary arteries: type of AA repair (ascending descending anastomosis versus REEEA), diagnosis (TBA versus TGA), age at operation (days), time since single-stage repair (days), pulmonary anastomosis (with versus without absorbable material), patch ascending aorta (no versus yes), position of the aorta (side by side versus anterior), transfer of pulmonary bifurcation (no versus yes) and interrupted AA (no versus yes).

Reinterventions on the left ventricular outflow tract and aorta

Three patients developed a subaortic stenosis during the follow-up period, 2 of them were provided with a modified Konno procedure after 253 and 440 days and 1 with a myectomy 5170 days after TBA correction. There was 1 case of re-coarctation in a boy with TBA, large VSD reaching into the inlet septum and hypoplastic AA. His primary repair was a very complex operation with dissection of the innominate artery and the descending aorta during arterial cannulation, and equal problems showed up during the REEEA. A first stent into the isthmus region was implanted 79 days after the first operation, a second ‘stent in stent’ followed 2 years and stent dilatation 4 years later. At the age of 12 years all stents and a recurrent stenosis were removed surgically through a lateral thoracotomy. As the boy was of the weight and size of an adult a 20-mm interposition graft was performed. A second case of mild recurrent coarctation affected the 4th patient of our series after an ascending descending aortic anastomosis. He had in total 5 transcatheter reinterventions (primary because of PA and valve stenosis and later due to an iatrogenic aortic pulmonary window), as well as 1 surgical reintervention (PA patch plasty and valvuloplasty). The mild aortic coarctation was dilated during the catheterization sessions regarding the pulmonary arteries, and the last examination still showed mild coarctation with no need for reintervention.

At the last follow-up, we found 8 patients with significant dilation of the neoaortic sinus (35–51 mm); the largest is shown in Fig. 5. This pathology affected one 14-year-old boy and 7 patients between 16 and 19.9 years of age. Since in total 8 patients of our whole series were 16 years or older at the last follow-up, 87.5% of this oldest cohort showed a dilated neoaortic sinus. All these patients are checked by magnetic resonance imaging regularly, and following the ESC guidelines [14], no intervention was needed until now (all diameters <51 mm, stable, no or mild regurgitation). In 47/49 patients (95.9%) we did not find more than mild neoaortic regurgitation at their last follow-up, only in 2 patients aged 9 and 1 years mild to moderate insuffiency was described.

DISCUSSION

The procedure-related mortality after the correction of TBA or TGA and hypoplastic/interrupted AA was 2% at our centre, although a large variety of patients like preterm and low birth weight babies were included. This confirms the approach of the simultaneous repair, which is pursued by the majority of centres nowadays [1–8]. However, the reintervention rate of our TBA cohort was undoubtedly high and is comparable to other published series [1, 3, 6, 7]. The heart defect itself can be charged for unavoidable reinterventions on the pulmonary valve, which is too small in a certain percentage of cases. Interestingly, 3 of our 4 patients with the need for RVPA conduit implantation later on showed coronary anatomy Yacoub type D. The small valves were already recognized during the primary repair, but the impossibility of performing a transannular patch because of the crossing coronary artery was the reason for the preservation of the valve and a delay of RVPA conduit reoperations. These had to take place very early (after 31, 101 and 221 days) in the Yacoub type D patients. However, after implanting valved conduits in newborns reoperations during the first year of life can also be necessary [15]. These 4 children were the patients with the highest number of total surgical (2–4) and also transcatheter reinterventions per person. Two of them additionally developed endocarditis, once even in a pulmonary homograft, which is known to be the safest option to prevent valve infection [16, 17].

The number of reinterventions on other locations as on the AA, left ventricular outflow tract or on the pulmonary arteries seems to vary in different publications [4, 9–11]. Therefore, these problems might be the consequence of the complexity of the procedures and can probably be addressed by reconsidering technical issues. The restenosis rate of the AA was comparably low in our series, which suggests that the techniques of REEEA and ascending descending anastomosis in all but 1 cases seem to be appropriate. The rare need for adding material when reconstructing the AA in these malformations can be explained easily, as the distance between ascending and descending aorta becomes shorter after the ASO in the majority of malformations.

The numbers of surgical or transcatheter reinterventions on the left, right and especially main pulmonary arteries were relatively high in our series (22.4%). Eight patients had 1 surgical reoperation each at this site, and they were provided with O- or Y-shaped patches. Although these procedures are of rather low complexity in comparison to other possible reoperations, the aim should be to prevent any kinds of reoperations. We performed a patch enlargement of the ascending aorta in 22 cases with a significant size mismatch between the neoaortic root and the ascending aorta, and this could be suspected of contributing to stretched pulmonary arteries. Four of these patients (18.2%) required reinterventions on the pulmonary arteries, whereas 21.4% of the patients without ascending aortic enlargement needed at least 1 reintervention at the PAs. Furthermore, the Cox regression analyses did not show an influence of a patch in the ascending aorta on the occurrence of reintervention on the PAs. That leads to the question if performing a Le Compte manoeuvre in all patients, which is not common at other centres [7], is the right approach, and this will be considered in our future cases. However, as in all cases, the stenosis could be relieved by 1 patch plasty procedure each and the small neo-pulmonary root is probably one of the influencing factors, we provided our latest TBA patient with a Y-shaped patch immediately during his first operation.

Although no reoperation because of a dilated neoaortic sinus was necessary until now, the fact that 87.5% of all patients 16 years or older had a significant dilation is alarming. This is owed to the large native pulmonary sinus of all these cardiac malformations on the one hand and potentially to the fact of pulmonary sinus tissue on the other hand. On the contrary, none of these showed more than mild neoaortic regurgitation, and the diameter stayed stable in the oldest patients. Fricke et al. [10] described the operative technique of reducing the sinotubular junction instead of enlarging the ascending aorta during the initial operation, and this technique will at least be considered in our future cases. No late neoaortic root dilation was described in this publication. However, 5 patients required neoaortic valve repair, 2 an ascending aorta stenosis repair and 1 a valve-sparing root replacement. We conclude that the neoaortic aneurysm or other pathologies of the neoaortic root seem to be concomitant with these patients after long-term follow-up, hence requiring close surveillance.

The most important limitations of this study are its retrospective nature, the small number of patients and the fact of using results from different examiners for the follow-up data.

ACKNOWLEDGEMENTS

The authors thank Dr. Wolfgang Schimetta and the Working Group for Systemic Optimization of Clinical Research Projects (ASOKLIF) for data analysis and statistics as well as Peter Kreuzer for the graphical design.

Conflict of interest: none declared.

DATA AVAILABILITY

The data underlying this article will be shared on reasonable request to the corresponding author.

Author contributions

Michaela Kreuzer: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Visualization; Writing—original draft. Eva Sames-Dolzer: Formal analysis; Supervision; Validation; Writing—review & editing. Andreas Tulzer: Data curation; Formal analysis; Writing—review & editing. Gregor Gierlinger: Data curation; Visualization. Roland Mair: Software; Visualization; Writing—original draft. Mohammad-Paimann Nawrozi: Data curation; Resources. Gernot Grangl: Data curation. Rudolf Mair: Conceptualization; Formal analysis; Methodology; Supervision; Validation; Writing—review & editing.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Yoshihisa Tanoue and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

Presented at the 36th Annual Meeting of the European Association for Cardio-Thoracic Surgery, Milano, Italy, 5–8 October 2022.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• AA

  Aortic arch

 
• ASO

  Arterial switch operation

 
• ECMO

  Extracorporeal membrane oxygenation

 
• IQR

  Interquartile range

 
• IVS

  Intact ventricular septum

 
• PA

  Pulmonary artery

 
• REEEA

  Resection and extended end-to-end anastomosis

 
• RVOT

  Right ventricular outflow tract

 
• RVPA

  Right ventricle to pulmonary artery

 
• TBA

  Taussig–Bing anomaly

 
• TGA

  Transposition of the great arteries

 
• VSD

  Ventricular septal defect