Influence of Bovine Arch Anatomy on Surgical Outcomes of Coarctation of the Aorta

This study aimed to evaluate the outcome of coarctation of the aorta (CoA) repair with a special interest in bovine arch anatomy. Fifty-six patients who underwent CoA repair between 2010 and 2021 were included in this retrospective study. Of these, 11 patients had bovine arch anatomy. Surgical outcomes were reviewed. Computed tomography was used to analyze aortic arch geometry for all cases preoperatively. The gap between anastomotic sites was calculated at the linear region of the lesser curvature of the aortic arch between the distal ascending aorta and the proximal descending aorta. CoA repair was performed at a median age of 39 days (median body weight 3.3 kg). Thirty-two patients underwent extended direct anastomosis, 22 patients underwent direct anastomosis, and two patients underwent lesser curvature patch augmentation. The median follow-up period was 47 months. There were no early deaths. In patients who underwent direct and extended direct anastomosis, nine recoarctation and one left pulmonary venous obstruction events occurred. Moreover, freedom from these adverse events was 81% in normal arch and 50% in bovine arch patients at 10 years (P = 0.04). Two patients with a bovine arch anatomy who underwent lesser curvature patch augmentation had good outcomes. The distal arch was narrower and longer, and the gap between anastomotic sites was longer in patients with a bovine arch anatomy than with a normal arch (P < 0.01). In CoA with a bovine arch anatomy, the gap between anastomotic sites was long. This adversely influenced the outcomes of the CoA repairs.


Introduction
Bovine arch anatomy is characterized by a "bovine trunk," wherein the innominate artery and left common carotid artery arise as a single trunk from the ascending aorta. Both imaging and autopsy studies indicate that the prevalence of bovine arch is 15-37% [1][2][3]. Moreover, some reports indicate that the frequency of complications with coarctation of the aorta (CoA) is as low as 3-5% [4][5][6]. Direct anastomosis (DA) and extended direct anastomosis (EDA) are commonly used techniques for CoA repair. However, DA and EDA can cause bronchial compression and/or recoarctation. Factors including repair in the neonatal period, lower weight, hypoplastic aortic arch, and ipsilateral descending aorta are all associated with a high incidence of bronchial compression and/or recoarctation [7,8]. However, these reports have not commented on arch branching, such as bovine arch, being a risk factor for these complications. In a recent report, Turek et al. demonstrated that arch anatomy often goes undocumented on preoperative imaging for coarctation. Moreover, they suggested that the presence of a bovine arch increases the risk of recoarctation in infants and neonates undergoing EDA [9]. Thus, the strategy for CoA with a bovine arch should be discussed. At our institution, depending on the degree of aortic arch hypoplasia and the distance between the bovine trunk and descending aorta, DA, EDA, or glutaraldehyde-treated autologous pericardial patch reconstruction is performed. This study aimed to evaluate the midterm outcomes of arch reconstruction for CoA with a bovine arch.

Ethical Considerations
Ethical approval for this study was obtained from the Ethics Committee and Institutional Review Board Committee of Osaka Women's and Children's Hospital (Approval no. 1536). Clinical data were obtained via a retrospective review of medical records. The requirement for written informed consent was waived because of the retrospective design of the study.

Patient Population and Data Collection
Between January 2010 and December 2021, 63 patients underwent aortic arch reconstruction for CoA at a single institution. Four patients with left arch and right descending aorta anomaly and three patients who underwent subclavian artery flap were excluded. In total, 11 patients had a bovine arch anomaly (17%). The remaining 56 patients were retrospectively reviewed by their clinical records, echocardiograms, preoperative computed tomography (CT) scans, operative findings, and surgical outcomes. According to the surgical outcomes and after referring to analyzed aortic arch geometry with a CT scan before surgery, two patients who underwent patch augmentation were excluded.

Identification of Arch Anatomy
All patients underwent preoperative CT scans. Bovine arch anatomy was defined as a common aortic arch origin of the left common carotid artery and innominate artery or the left common carotid artery originating as a branch of the innominate artery [1-3, 9, 10]. The gap between the anastomotic sites was calculated at the linear of the lesser curvature of the aortic arch, between the distal ascending aorta and the proximal descending aorta. The distance was indexed to the diameter of the descending aorta at the diaphragm level. The diameter of the distal arch was calculated just distal to the bovine trunk or left common carotid artery by Z score [11]. The length of the distal arch was calculated between the left common carotid artery and the left subclavian artery. Hypoplastic aortic arch was defined by a distal arch diameter Z score less than − 3. A long gap between anastomotic sites was defined as more than 2.5 mm/mm indexed to the diameter of the descending aorta.

Operation Technique
Three different operative techniques were used for aortic arch reconstruction, including DA, EDA, and patch augmentation. DA was chosen for patients without a hypoplastic arch and EDA for patients with a hypoplastic arch. From 2018, patch augmentation was chosen for patients with bovine arch, especially in cases of long gap between anastomotic sites. DA was performed via posterolateral left thoracotomy; the left chest was entered through the 3rd intercostal space. Dissection was performed to mobilize the entire aortic arch and descending aorta. Heparin, at a dosage of 100-150 U/kg, was administered. The ductus arteriosus was ligated. Next, the aortic arch was clamped with a vascular clamp incorporating the left carotid and left subclavian arteries. The distal clamp was then placed across the descending aorta, and the ductal tissue was completely resected. The incision was extended to the undersurface of the aortic arch onto the distal arch. The anastomosis between the aortic arch and descending aorta was formed using a continuous suture technique with fine polypropylene suture (6-0 or 7-0 prolene; Ethicon, Somerville, NJ) [12]. EDA was then performed through a median sternotomy. The procedure was performed using bicaval cannulation and arterial inflow at the ascending aorta with antegrade cerebral perfusion through the innominate artery. The vascular clamp was placed partially on the brachiocephalic artery to maximize the area of extended anastomosis. Both the isthmus and ductus arteriosus were ligated and divided. Then, all the ductal tissue was excised from the descending aorta. A longitudinal incision was then made in the distal ascending aorta/bovine trunk, corresponding to the size of the descending aorta. The descending aorta was then anastomosed to the aortic arch incision by the same method as DA [7]. Patch augmentation was performed through a median sternotomy. The procedure was performed in almost the same way as EDA; however, patch augmentation was added to the inner curvature of the arch. The patch was glutaraldehyde-treated autologous pericardium and ellipse-shaped [13].

Follow-Up
Follow-up included clinical examination, transthoracic echocardiography (TTE), and CT scan before discharge. The evaluation of recoarctation was determined by TTE reports with a flow velocity of 2.5 m/s or greater across the repair site.

Statistical Analysis
Statistical analysis was performed using Statistical Package for Social Sciences (SPSS) software, version 18.0 (SPSS Inc., Chicago, IL). Frequencies are presented as absolute numbers and percentages. Continuous data are presented as mean ± SD or median values with ranges, as appropriate. Median values were compared using Student's t-test. Kaplan-Meier analysis was used to estimate the proportion of patients who did not encounter postoperative adverse events, including recoarctation, bronchial compression, and/or left pulmonary venous stenosis. For all tests, P values of < 0.05 were considered significant.

Results
The patients' demographic characteristics and intraoperative outcomes are described in Table 1. There was no difference between patients with a normal arch and patients with a bovine arch except for the number of patch augmentation techniques. The median descending aorta clamping time was The median follow-up period was 42 months (range 5 months to 12 years). There was no difference between the normal arch and bovine arch groups in the median follow-up period [normal arch group; 47 months (range 1 month to 10 years), bovine arch group; 32 months (range 1-12 years)] (P = 0.66). There were two deaths in the normal arch group-one was due to an acute myocardial infarction post a Ross-Konno operation and the other due to right heart failure post complete atrioventricular septal defect repair; however, there were no deaths during the hospital stay which were associated with CoA repair. Freedom from adverse events was 81% in patients with a normal arch and 50% in patients with a bovine arch who underwent DA or EDA (P = 0.04) (Fig. 1). Recoarctation occurred in six patients with a normal arch, five of whom required balloon angioplasty, and three patients with a bovine arch, one of whom required balloon angioplasty. No patient needed surgical reintervention in either group. The last follow-up TTE in these cases with recoarctation shows the median flow velocity across the repair site 1.7 m/s (range 1.0-2.3 m/s). Left pulmonary veinous stenosis occurred in one patient with a bovine arch but no bronchial stenosis occurred in either group. These adverse events occurred over a median period of 10 months postoperatively (range 1-30 months). Of the 10 patients with these adverse events, six patients had undergone EDA and four had undergone DA. Two patients who underwent the patch augmentation technique had good outcomes during a follow-up period of 2.3 years and 2.0 years.
The preoperative CT scan data are shown in Table 2. The distal arch was narrower and longer and the gap between the anastomotic sites was longer in patients with a bovine arch than in those with a normal arch (P < 0.01). In the bovine arch group, the relationship of preoperative CT scan data and postoperative complication is described in Fig. 2. Patients with adverse events had a smaller median distal arch diameter (Z score, − 6.7 vs − 3.0) (P < 0.01) and longer gap between the anastomotic sites (2.1 vs 1.7 mm/ mm) (P = 0.04) For the two patients with a bovine arch who

Discussion
In this study, of the 63 patients who underwent CoA in our institute, 11 patients (17%) were cases of concomitant bovine arch. Surgical outcomes of included 56 patients were reviewed, and aortic arch geometry was analyzed routinely before surgery with CT. In the normal arch and bovine arch groups, among patients who underwent DA or EDA, nine recoarctations and one left pulmonary venous stenosis occurred. Moreover, freedom from these adverse events was 81% in the normal arch group, and 50% in the bovine arch group at 10 years (P = 0.04). There were two patients with a bovine arch anatomy who underwent lesser curvature patch augmentation with good outcomes. Preoperative CT showed the distal arch was narrower and longer, and that the gap between anastomotic sites was significantly longer in patients with a bovine arch than in those with a normal arch (P < 0.01).
Turek et al. reported that 14 of the 49 patients (29%) who underwent CoA repair had concurrent bovine arch anatomy [9]. In this study, 11 of the 63 patients (17%) had bovine arch anatomy. These results indicate that CoA with bovine arch anatomy is not uncommon. Moreover, they reported that only around 6% of cases with concomitant bovine arch anatomy were preoperatively diagnosed and therefore concluded that if assessment of aortic morphology is uncertain on postnatal TTE, a CT scan should be performed [9]. In our institute, all patients diagnosed with CoA on fetal echocardiogram or postnatal TTE undergo CT. We should be careful and not overlook bovine arch when evaluating patients with CoA. Regarding the necessity of a preoperative CT scan, Langley et al. insisted that an accurate and complete imaging is an essential starting point of the CoA repair. As the available evidence on which these decisions are made increases, it is hoped that the management of CoA will become more standardized and that this will translate into improved outcomes [13].
Furthermore, Turek et al. showed that patients with a bovine arch who underwent EDA are at a significantly increased risk of recurrent arch obstruction. This could be due to proximal displacement of the left common carotid artery and distal displacement of the innominate artery in bovine arch anatomy [9]. It is suggested that the distal arch emerges as a direct leftward lateral branch in the bovine arch and that it may be long with a hypoplastic aortic arch. As such, modification of the surgical technique may be necessary to mitigate arch hypoplasia, especially if the distal arch is particularly long [12]. In our result, patients with both long gap between anastomotic sites and the hypoplastic aortic arch were observed to be at risk of recoarctation and pulmonary venous obstruction (Fig. 2).
Langley et al. reported an increased risk of recoarctation with EDA in cases with no proximal arch. In such cases, both the innominate artery and left common carotid artery are very close together, the distal arch is often long and hypoplastic, there is a relatively short isthmus, and a Fig. 2 The relationship between preoperative CT scan data and postoperative complications. * CoA with the bovine arch group included only patients who underwent direct anastomosis or extended direct anastomosis. CoA coarctation of the aorta; CT computed tomography tight coarctation. Moreover, they reported that the patch augmentation technique was effective in such cases. This technique involves EDA and glutaraldehyde-treated autologous pericardium patch augmentation of the inner curvature of the arch [13]. Roussin et al. also introduced the concept of the patch augmentation technique. They used pulmonary autograft patches in 23 patients with aortic arch obstruction; in their comparative midterm analysis, they noted that this technique yielded more superior results than other techniques, such as DA [14]. Lee et al. reported that augmentation of the lesser curvature with an autologous vascular patch has the following advantages: (1) reduction in the tension on the anastomosis because of elongation of the lesser curvature for arch reconstruction; (2) allowance for complete resection of the ductal tissue without fear of tension; and (3) avoidance of recurrent coarctation with the autologous tissue with growth potential [15]. Recently, Hasegawa et al. reported a study including 68 patients with a significantly longer gap between anastomotic sites. Compared to EDA, the patch augmentation technique created a significantly greater arch angle and arch/descending diameter ratio and preserved the arch width, which lead to satisfactory surgical outcomes [16]. Regarding the patch materials, Li et al. used glutaraldehyde-treated autologous pericardium for patch augmentation of the inner curvature of the arch with EDA and reported satisfactory outcomes. They insisted that the additional surgical procedure, which means a pulmonary autograft patch resected from the main pulmonary artery, may prolong the operative time and the patch may be limited by the length and width of the pulmonary artery [17]. In recent years, we have also used this technique for the reconstructed aortic arch in CoA cases with bovine arch anatomy, especially in patients with a longer gap (more than 2.5 mm/mm) between the anastomosis sites (Fig. 3). Thus far, we have performed this technique in two patients. The descending clamping time was comparable to other surgical methods, and the postoperative course was good with no adverse events. We used glutaraldehyde-treated autologous pericardium patch for the two patients because of insufficient length and width of the pulmonary artery. However, we think it is desirable to use a pulmonary autograft patch for patients who have sufficiently dilated main pulmonary artery, for example, the status post main pulmonary artery banding.
Finally, in terms of evaluating the total cost of recoarctation in patients with a bovine arch over their lifetime, Turek et al. report that although only 4 out of 14 patients (28.5%) had postoperative recoarctation, only balloon angioplasty treatment was required, and no patients required stenting or a repeat operation [9]. In our results, there were 3 out of 11 (27%) patients who had postoperative recoarctation, of which, only one patient required balloon angioplasty, and no other surgical reintervention was needed. One patient with postoperative pulmonary venous stenosis caused by descending aortic compression from the dorsal side was not critical, and no intervention was needed till date. However, we must continue with careful follow-up of patients with these adverse events to determine if surgical intervention will be necessary. This will also enable evaluation of whether routine patch augmentation procedures for aortic arch reconstruction of CoA with bovine arch anatomy are necessary.
This study had some limitations. It was a retrospective review of patients who underwent CoA repair in a single center. Moreover, our small sample size limited our ability to perform statistical adjustments.

Conclusion
In CoA with a bovine arch, the distal arch was narrower and longer than CoA with a normal arch. The gap between anastomotic sites was long in these patients, which influenced the postoperative adverse events. The lesser curvature patch However, further large-scale future studies, assessing the outcomes of CoA with a bovine arch anatomy, are needed to determine whether routine patch augmentation procedures for aortic arch reconstruction of CoA with a bovine arch anatomy are necessary.