Right ventricular outflow tract obstruction after arterial switch operation for the Taussig–Bing heart^☆

Abstract

Objective: Incidence of right ventricular outflow tract obstruction (RVOTO) may be suspected to be higher after arterial switch operation (ASO) for Taussig–Bing heart than after ASO for transposition of the great arteries (TGA), as Taussig–Bing anomaly is frequently associated with aortic arch obstruction and subvalvular aortic stenosis. We evaluated the risk to develop RVOTO after ASO for Taussig–Bing heart. Methods: The 34 Taussig–Bing cases who underwent ASO from 1984 to 2005 were reviewed. RVOTO was defined as peak echo-gradient ≥30 mmHg across right ventricular outflow tract. Kaplan–Meier method was used to estimate time-related events. Results: Subaortic stenosis was resected in 25 patients, 20 of whom (80%: 20/25) were discharged from hospital free from RVOTO. There was one early death: 2.9% mortality. Three patients died late. Actuarial survival was 85.1% ± 7.0% from 54 month onwards. Eleven survivors (36.7%: 11/30) experienced postoperative RVOTO. Obstruction was seen in 82% (9/11) of cases at subvalvular and/or valvular level. Surgery (n = 4) or percutaneous intervention (n = 2) was required in six patients. Patients discharged from hospital with RVOTO (n = 8) were more likely to undergo reintervention for RVOTO (p = 0.026). Freedom from reintervention for RVOTO decreased rapidly in the first two years to 86.5 ± 6.3%, slowly thereafter (80.4 ± 8.4% at year 7) and stabilized at 70.3 ± 11.9% from year 11 on. Risk for RVOTO occurrence was 23.5 ± 7.3% early after repair and progressively increased to level out at 53.6 ± 11% at year 11. Patients who underwent subaortic resection were more likely (p = 0.023) to be free from RVOTO occurrence or development. In the period under review, for patients who underwent ASO for simple (n = 355) and complex (n = 92) TGA, reoperation rate for neopulmonary stenosis was 0.3% (1/355) and 5.4% (5/92), respectively, to be compared to 11.8% (4/34) RVOTO rate of reoperation for Taussig–Bing heart in this study. Conclusions: Postoperative right-sided obstruction occurs more frequently after ASO repair of Taussig–Bing heart than after TGA arterial switching, leading to higher reintervention rate. Resection of the commonly associated subaortic stenosis often prevents RVOTO development.

1 Introduction

The success story of the arterial switch operation (ASO) used for correction of transposition of the great arteries (TGA) is still dimmed by complications, the most frequent one being right ventricular outflow tract obstruction (RVOTO) [1–4]. Commonly reported anatomic factors in occurrence of neopulmonary stenosis are associated aortic arch obstruction, side-by-side position of the great arteries, coronary artery anomalies and preoperative presence of organic subaortic RVOTO [1,5–7]. As these features are often seen in Taussig–Bing anomaly, RVOTO incidence might be expected to be higher after ASO for this lesion.

This study was conducted to evaluate the risk for patients with Taussig–Bing heart to develop right-sided outflow obstruction after arterial switch procedure.

2 Patients and methods

2.1 Definitions

Definition of Taussig–Bing malformation encompassed all sorts of double outlet right ventricle (DORV) lesions with subpulmonary ventricular septal defect (VSD) and without pulmonary stenosis, the aorta arising entirely and the pulmonary trunk at least 50% from right ventricle, that were amenable to ASO.

Postoperative RVOTO was defined as a peak echo-Doppler gradient at least 30 mmHg across the right ventricular outflow tract (RVOT). Level of stenosis was classified as subvalvular (right ventricular infundibulum), valvular (neopulmonary valve: leaflets or annulus) or supravalvular (reconstructed pulmonary trunk or right and left pulmonary artery).

Reintervention for RVOTO included any postoperative surgical or percutaneous interventional procedure to address any right-sided outflow obstruction.

2.2 Patients

All consecutive 34 patients who underwent ASO for Taussig–Bing anomaly from 1984 until the end of December 2005 were reviewed. These included 27 patients reported in a previous publication [8].

Median age at ASO was 34 days (range 8–154 days) for patients who underwent primary complete repair (n = 27). It was 66 days (range 14–291 days) for those with two-stage correction (n = 7). Fifteen patients were neonates. Mean comprehensive Aristotle score was 16.8 ± 1.6, range 15–21 for primary repair cases and 18.1 ± 1.3, range 16–20 for the other patients. The main clinical and surgical characteristics of these patients are shown in Table 1 . Subvalvular aortic stenosis (primitive RVOTO) caused by conal muscle was preoperatively present in 28 patients, but was recognized only in 24 of them at the time of surgery. In an additional case, it was discovered perioperatively (overall 25 patients with operatively recognized RVOTO). Obstruction of aortic arch was common (n = 23), and the repair was staged in seven patients. The usual coronary artery pattern with left coronary artery (LCA) arising from Sinus 1 and right coronary artery (RCA) arising from Sinus 2 was found in only 32% (11/34) of cases.

2.3 Management

Our surgical management of the Taussig–Bing anomaly by arterial switch procedure has been reported previously [8,9]. Therefore, we only detail aspects pertaining to the reconstruction of the right ventricular outflow. Right and left pulmonary arteries were dissected beyond their point of branching to allow pulmonary anastomosis without straining. Subaortic obstructing (fibro)muscular tissue was resected either through the aortic valve or transatrially through the tricuspid valve in 74% (25/34) of all cases. Site of resection was usually the ventriculo-infundibular fold and the right anterior free wall as previously described [6]. The remaining nine patients (26%: 9/34) did not undergo any RVOT resection. A single large glutaraldehyde-treated autologous pericardial patch was used to reconstruct the neopulmonary artery. In 12 cases (35%: 12/34) who had side-by-side great arteries, we shifted the pulmonary anastomosis into the right pulmonary artery in order to minimize the risk of compression of the anteriorly transferred coronary artery, as well as to decrease tension on the pulmonary anastomosis. This was realized by closing the leftward end of the distal divided main pulmonary trunk and extending the opening into the right pulmonary artery. The pulmonary artery was not translocated anteriorly (Lecompte maneuver) in five patients (15%: 5/34) in order to prevent compression of the left coronary artery button.

2.4 Follow-up and evaluation of right ventricular outflow tract

To assess follow-up, all available clinical and surgical records, echocardiographic and postoperative catheterization data were collected. A thorough transthoracic echo-Doppler examination was realized before discharge from hospital. Particularly, the peak pressure gradient (calculated using the simplified Bernoulli equation: gradient = 4V², where V is the peak instantaneous transvalvular or transvascular Doppler velocity) was recorded. Afterwards, patient was usually reviewed by the referring cardiologist at interval 1, 3 and 6 months and at least once yearly thereafter, and we were informed about outcome. Assessment for this study took place from April to June 2006 by contacting attending physicians. The study put emphasis on variables related to right ventricular outflow tract. Our policy is to consider reintervention if peak echo-Doppler gradient through RVOT is ≥50 mmHg.

2.5 Statistical analysis

Kaplan–Meier curves for actuarial survival, freedom from reintervention for RVOTO and for RVOTO occurrence were calculated using the Graph Pad Prism (San Diego, CA, USA). The following parameters served as end points: time of death, first surgical or percutaneous reintervention to manage RVOTO, first demonstration of a peak echo-Doppler gradient at least 30 mmHg across RVOT. The probabilities are given as percentages (%) ± standard error of the mean (S.E.M.). Univariate analysis with Fisher’s exact test was used to compare variables. The level of statistical significance was set at a p-value of less than 0.05. Medians and means are given with range and standard deviation.

3 Results

3.1 Survival and functional status

There was one hospital death in a neonate who was reoperated upon on the ninth postoperative day because of residual aortic arch obstruction. He died 2 months later from multi-organ failure. Three patients died late. The first patient succumbed to septicemia unrelated to cardiac surgery after 8 months. The second late death was sudden, after 36 months, presumably from cardiac arrhythmia. Rhythm disturbances had already complicated postoperative course. LAD arose from Sinus 2 and peak echo-Doppler gradient through the neopulmonary valve was 50 mmHg at the day of discharge from hospital. The cause of the third late death (54 months) was attributed to an acute coronary insufficiency and myocardial infarction. The patient had a single coronary artery system, which previously came off Sinus 2.

Actuarial survival was 85.1 ± 7.0% at 54 months and remained constant thereafter (Fig. 1 ). Follow-up for the late 30 survivors is complete. Median time after ASO repair is 70 months, range 4–224 months. Ten patients underwent, after first hospital discharge, a total of 15 reinterventions to manage RVOTO (n = 6), aortic arch stenosis (n = 2), neoaortic valve regurgitation (n = 1) and complete atrio-ventricular block (n = 1). Twenty-two patients (73%: 22/30) were in functional NYHA class I and eight in class II. There was no case of significant (neo) pulmonary regurgitation. The (neo) aortic valve functioned normally or showed trivial or Grade 1 regurgitation in 26 cases. It was Grade 2 incompetent in two patients. Aortic valve was replaced by a mechanical valve in one patient and was planned to be repaired or replaced in the remaining patient with Grade 3 regurgitation.

3.2 Right ventricular outflow tract obstruction

Outcome and levels of peak echo-Doppler gradients across RVOT at the time of discharge from hospital and at the last follow-up examination are schematically displayed in Fig. 2 . Early echo-Doppler results are missing for two patients operated upon in 1989. They did not undergo any reintervention and their actual pulmonary valve function is normal. They have been excluded from risk evaluation of RVOTO occurrence. Thus, a total of 31 (33 − 2) early survivors and 28 (30 − 2) late survivors have been evaluated.

Eleven late survivors (39%: 11/28) presented with a peak Doppler gradient of at least 30 mmHg, either early or late after ASO. Obstruction was located in 82% (9/11) of cases at subvalvular and/or valvular level. It is to be noted that 20 of the 25 patients (80%: 20/25) who also underwent resection of subaortic stenosis left hospital free from RVOTO.

Six patients had to undergo surgery (n = 4) or percutaneous intervention (n = 2). Table 2 details these patients lesions that were associated with Taussig–Bing anomaly, the type of procedure and outcome. Patient number 1 was reoperated twice because of recurrent RVOTO 25 months after ASO. The procedure consisted of transpulmonary resection of the hypertrophied infundibulum and removal of the previously implanted PTFE (polytetrafluoroethylene) conduit. A dilating stent was placed in the left pulmonary artery 50 months after initial operation. At recent follow-up, 3 years after the last intervention, echo-Doppler still exhibited supravalvular 98 mmHg gradient and thus, patient was scheduled for reintervention. It is to be noted that patients number 5 and 6 had left hospital after ASO repair with low RVOT gradients: 0 and 10 mmHg. Probability to undergo reintervention for neopulmonary stenosis was 50% (4/8) for patients discharged from hospital with RVOTO and 8.7% (2/23) for those leaving hospital without RVOTO: a significant difference of p = 0.026. Freedom from first reintervention for RVOTO (Fig. 3 ) decreased rapidly in the first 2 years to 86.5 ± 6.3%, slowly thereafter 80.4 ± 8.4% at year 7 and stabilized at 70.3 ± 11.9% from year 11 on.

RVOT pressures evolved over time. As shown in Fig. 4 , gradients above 40 mmHg increased in the follow-up. Overall gradient increased in 11 (39%: 11/28) patients, decreased in 14 (50%: 14/28) patients and did not change in three. Nine patients (30% of cases) are currently showing RVOTO, with severity (gradient ≥ 50 mmHg) in three. Gradient is under 25 mmHg in 20 patients (67% of cases) as shown in Fig. 4.

The risk to develop RVOTO (Fig. 5 ) early after ASO was estimated to be 23.5 ± 7.3%. Cumulative risk increased up to 36.1% ± 9.0% at year 4 and slowed down to 53.6 ± 11.0% level at year 11. Univariate analysis (Table 3 ) of variables presumably involved in RVOTO occurrence did not incriminate any individual associated cardiovascular anomaly. Particularly, patients with aortic arch obstruction or subvalvular aortic stenosis did not experience more RVOTO than those without these lesions. But two-stage repair as factor reached almost the significance level: p = 0.07. Furthermore, two variables related to surgical techniques applied to avoid compression of the transferred coronary arteries were statistically significant: the no use of Lecompte maneuver and the shifting of the pulmonary anastomosis into the right pulmonary artery. Most importantly, patients who underwent subaortic resection were more likely (p = 0.023) to be free from RVOTO occurrence or development.

In the period under review, for patients who underwent ASO for simple (n = 355) and complex (n = 92) TGA in our institution, reoperation rate for neopulmonary stenosis was 0.3% (1/355) and 5.4% (5/92), respectively. The value is comparable to the 11.8% (4/34) rate of reoperation for Taussig–Bing heart in this study. This gives a p-value of 0.0002 and 0.25 versus simple and complex TGA respectively.

4 Discussion

Supravalvular neopulmonary artery stenosis is the most common complication and cause of reintervention after ASO for TGA, with a reported incidence varying from 1 to 42%, depending on obstruction criteria and length of follow-up [4], reoperation being observed up to 9–10 years after arterial switching [1,2]. Apart from the anatomic variables cited in Section 1, technical factors have been incriminated for RVOTO development: insufficient mobilization of the pulmonary arteries and inadequate size or form of the pericardial patch used to reconstruct the neopulmonary trunk. Some are controversial such as the Lecompte maneuver and the use of glutaraldehyde-treated pericardial patch. Whatever the cause, with experience, the reoperation rate for pulmonary stenosis in most series is reported to be decreasing in the last 10 years.

As for the Taussig–Bing lesion, little has been published about RVOTO complication after ASO correction. The multi-institutional study of Williams et al. [1] with a considerable number of patients does not include Taussig–Bing cases. The largest series from one institution with 79 Taussig–Bing patients [2] only reveals that the incidence of all reoperation was higher in complex TGA, without describing further details. The first study of RVOTO problem after ASO repair of these complex DORV lesions was published by Wetter et al. [8] from our Unit.

Peak echo-Doppler gradient ≥30 mmHg adopted in this study to define RVOTO could be challenged. The incidence of truly significant pulmonary stenosis may have been overestimated. Gradients 25–49 mmHg are commonly defined as mild stenosis. In the series of Ullmann et al. [4], 24.2% of cases had such postoperative gradients. Nevertheless, the authors claimed a low incidence of late supravalvular pulmonary stenosis after ASO for TGA. Like other investigators [3,10], we defined the threshold to be 30 mmHg, considering that such gradient is not insignificant, even if a good number would indicate an intervention at 40 mmHg. Indeed, on one hand, this study shows that patients discharged from hospital with gradient ≥30 mmHg are more likely to undergo reintervention for RVOTO than those leaving hospital with gradient under 30 mmHg (p = 0.026). It also indicates that RVOTO rarely develops in patients discharged from hospital with gradient under 30 mmHg. On the other hand, according to Krovetz and Goldbloom [11], normal systolic pressure in the right ventricle (RV) is 32.9 ± 14.7 mmHg. If a RVOT stenosis causes a peak echo gradient of at least 30 mmHg, corresponding systolic pressure in RV could be assumed to be at least 62.9 mmHg. As 50% of infants are reported to have 90 mmHg systolic systemic blood pressure on average [12], RV pressure may then be estimated to be at least 70% (62.9/90) of the systemic arterial pressure. This level of RV pressure, if it persists, can certainly be detrimental to RV function.

RVOTO evolved over time. Valvular and subvalvular obstruction appeared early and its occurrence approached 0% in later follow-up, which corresponds to higher reintervention rate in the first 2 years. This evolution was similar to the behavior of postoperative proximal pulmonary stenosis observed after ASO for TGA [1]. As shown in Fig. 4, half of pressure gradients subsided. But obstruction increased in all three children discharged from hospital with gradients above 40 mmHg. It is noteworthy that this group of patients keep requiring further management.

As described by Boyadjev et al. [5] and Akiba et al. [13], there is an anatomic substrate for subvalvular RVOTO after ASO TGA repair. In this series, the Taussig–Bing malformation was associated with stenosing subaortic conal muscle in at least 74% (25/34) of cases. Resection was effective in relieving this obstruction. Indeed, only 3 out of 25 patients (12%) with confirmed preoperative subaortic stenosis left hospital with RVOTO. Patients who underwent subaortic resection were more likely (p = 0.023) to be free from occurrence or development of RVOTO. Bulging myocardial tissue under the aortic valve must be identified and removed to avoid the problem of subpulmonary obstruction after arterial switching. Nevertheless, we cannot recommend infundibulotomy and transannular enlargement, considering that early survival does depend on a competent neopulmonary valve. Moreover, infundibulotomy could not be feasible in those patients with RCA crossing infundibulum, as seen in 29% (10/34) of our cohort.

In a great majority of cases (82%), postoperative RVOTO was located at subvalvular and valvular level. The supravalvular site is more affected after ASO for TGA [1–4]. It is therefore unlikely that technical factors in reconstruction of the neopulmonary trunk are involved in the genesis and development of this obstruction. Analysis of the usually evoked anatomic factors did not point at any specific lesion. It may be explained by a small number of patients or by the fact that most patients had combined associated malformations. Nevertheless, with a p of 0.07, two-stage repair should be accepted as a risk factor. This finding added to the fact that staged correction compounded with higher comprehensive Aristotle score constitutes additional argument to further recommend primary complete early repair of the Taussig–Bing lesion.

In five instances, position of the great arteries and coronary anomalies did not allow Lecompte maneuver and the pulmonary anostomosis had to be shifted to the right pulmonary artery not to compromise coronary circulation. All five patients survived but had postoperative RVOTO and three needed reintervention. Thus, abnormal coronary artery pattern played a role in RVOTO development. It is also clear that native aortic annulus was small: size ratio of pulmonary artery and aorta was ≥2:1 in 53% (18/34) of cases. Detachment of coronary buttons left limited tissue for later growth of the neopulmonary root. This might be the main explication of obstruction at the valvular region, which triggers increasing adaptive infundibular hypertrophy and subvalvular stenosis.

We conclude that survival after ASO repair of the Taussig–Bing heart can be high despite concomitant major anomalies such as aortic arch obstruction, unusual coronary patterns and subaortic stenosis. In our experience, postoperative right-sided obstruction occurs or develops more frequently after arterial switching for Taussig–Bing heart than after ASO TGA correction, leading to higher reintervention rate. It rarely develops in patients discharged from hospital with gradient under 30 mmHg. It can be prevented by resection of subaortic stenosis in great majority of cases.

Appendix A

Conference discussion

Dr B. Maruszewski (Warsaw, Poland): What technical precautions do you use during repair of Taussig–Bing anomaly nowadays to prevent recurrent or residual right ventricular outflow tract stenosis? And did you change your surgical technique over the years to minimize this percentage of recurrent stenosis?

Dr Sinzobahamvya: It is very important to resect the subaortic muscle that causes obstruction. You go either through the native aortic valve or transatrially. There is ventriculo-infundibular fold and the free anterior wall of right ventricle where you have really to be aggressive. In our series of 25 who had preoperative subvalvular aortic stenosis, only 3 of them had, at the time of discharge from hospital, gradient of 30 mmHg or more through right ventricular outflow tract.

But I must also say that even if you do it, even in patients who have no gradient at time of discharge from hospital, or gradient less than 30 mmHg, right-sided obstruction can develop, albeit less frequently. Because the native aortic annulus is very small. And whatever you do under the annulus, the annulus remains unchanged. We think this is the main reason for postoperative pulmonary stenosis. Sometimes it is better to accept some gradient than to be too aggressive and to have, after the operation, a free outflow tract but bad contractility and thus, to jeopardize early survival.

Dr D. Barron (Birmingham, United Kingdom): Can I ask, you had some patients who had a one-stage repair, some who had a two-stage repair. What criteria did you use to decide on a one- or two-stage repair? And did that have an effect on your incidence of right outflow tract obstruction?

Dr Sinzobahamvya: We favor one-stage repair. There are seven patients who underwent two-stage repair. Among these seven patients, three came from other units, two were operated upon in the year 1987. We decided to do two-stage repair in the remaining two patients, as these two neonates presented in functional NYHA class V, with a comprehensive Aristotle score, which would have been more than 20. We decided to first stabilize them. The aortic arch was first repaired and we went back 2 weeks later to do the intracardiac repair and arterial switching. But fundamentally, one should perform one-stage complete repair.

Patients who underwent two-stage repair have a tendency to have more frequently right-sided obstruction. We did do an analysis. The probability was 0.07. If we had more cases, the difference would become significantly different. But we are unlikely to have more cases, as we know that we have to go after one-stage repair.

Dr C. Kreutzer (Buenos Aires, Argentina): I think what we must emphasize here is that subaortic resection means resection of the infundibular septum, because this is actually a fundamental part of the complex. The anterior deviation of the infundibular septum is creating subaortic stenosis, and this is the anatomic substrate for the aortic arch obstructions usually seen in Taussig–Bing hearts. So there is a clear indication that in all patients with a Taussig–Bing, even though you are not performing Kawashima type of repair, you must resect the infundibular septum.

Dr Sinzobahamvya: I agree with you.

References