Left subclavian artery revascularization in thoracic endovascular aortic repair: single center’s clinical experiences from 171 patients

Background Left subclavian artery revascularization (LSA) is frequently performed in the setting of thoracic endovascular repair (TEVAR). The purpose of this study was to compare different techniques for LSA revascularization during TEVAR. Methods We performed a single center’s retrospective cohort study from 2016 to 2019. Patients were categorized by LSA revascularization methods, including direct coverage without revascularization (Unrevascularized), carotid-subclavian bypass (CSB), fenestrated TEVAR (F-TEVAR). Indications, demographics, operation details, and outcomes were analyzed using standard statistical analysis. Results 171 patients underwent TEVAR with LSA coverage, 16.4% (n = 28) were unrevascularized and the remaining patients underwent CSB (n = 100 [58.5%]) or F-TEVAR (n = 43 [25.1%]). Demographics were similar between the unrevascularized and revascularized groups, except for procedure urgent status (p = 0.005). The incidence of postoperative spinal cord ischemia was significantly higher between unrevascularized and revascularized group (10.7% vs. 1.4%; p = 0.032). There was no difference in 30-day and mid-term rates of mortality, stroke, and left upper extremity ischemia. CSB was more likely time-consuming than F-TEVAR [3.25 (2.83–4) vs. 2 (1.67–2.67) hours, p = 0], but there were no statistically significant differences in 30-day or midterm outcomes for CSB versus F-TEVAR. During a mean follow-up time of 24.8 months, estimates survival rates had no difference. Conclusions LSA revascularization in zone 2 TEVAR is necessary which is associated with a low 30-day rate of spinal cord ischemia. When LSA revascularization is required during TEVAR, CSB and F-TEVAR are all safe and effective methods, and F-TEVAR appears to offer equivalent clinical outcomes as a less time-consuming and minimally invasive alternative.


Background
Thoracic endovascular aortic repair (TEVAR) has gained widespread acceptance and serves as the predominant treatment approach for patients with thoracic aortic diseases. Durable outcomes of TEVAR require an adequate proximal seal zone. It is estimated that 40% of TEVAR will require coverage of zone 2 to create a proximal seal [1]. The coverage of the left subclavian artery during TEVAR to achieve a proximal seal is associated with increased risk of stroke, spinal cord ischemia, and upper extremity ischemia [2]. In 2009, the Society for Vascular Surgery consensus statement recommended routine revascularization of the left subclavian artery when covered by TEVAR for proximal sealing in Open Access *Correspondence: dongjinwang_gl@163.com 1 Department of Cardiothoracic Surgery, Nanjing Drum Tower Hospital, The Affiliated Hospital of Nanjing University Medical School, 321 Zhongshan Road, Nanjing 210000, Jiangsu, People's Republic of China Full list of author information is available at the end of the article elective cases. However, the statement noted that their recommendations were based on low-quality evidence [3]. Besides, the recommendations did not address the most effective technique for LSA revascularization. The purpose of this study was to compare 30-day and midterm outcomes of LSA coverage during TEVAR and evaluate revascularization techniques for the LSA including carotid-subclavian bypass (CSB) and fenestrated thoracic endovascular aortic repair (F-TEVAR).
All patients with thoracic aortic pathology aged 18-90 years were confirmed by CTA. Only those undergoing TEVAR for pathology requiring intentional coverage of the left subclavian artery to obtain a proximal seal were included. Both elective and emergent cases were included in the sample. The comparison groups were designated by exposure criteria. According to LSA revascularization methods, patients were divided into 3 groups, including Unrevascularized group, CSB group, F-TEVAR group. The first comparison group was unrevascularized LSA versus revascularized LSA. The second group was based on revascularization types, CSB versus F-TEVAR.

Preoperative imaging and revascularization decision
General principles of LSA revascularization include dominant left vertebral artery, presence of a patent left internal mammary artery to coronary artery bypass graft, occluded right vertebral artery, functioning left upper extremity arteriovenous access, long-segment thoracic aortic coverage. Thoracic aortic pathology was diagnosed by CT angiography. The proximal vertebral arteries were all imaged on CT angiography. However, preoperative imaging of cerebral anatomy was not performed routinely. Choice of revascularization technique was determined by patient anatomy and physician preference. All patients in revascularized group were performed TEVAR and concomitant LSA revascularization. F-TEVAR includes pre-fenestration and situ fenestration. The stentgrafts were individually modified to preserve the flow of LSA by surgeons. For some unrevascularized and CSB cases, proximal occlusion was performed via catheterbased embolization if intraoperative type II endoleak occurs.

Carotid-subclavian bypass technique
Under general anesthesia, CSB was performed as the first stage of the procedure using standard supraclavicular incision and an 8-mm Dacron graft as bypass conduit (Fig. 1A). Then, TEVAR was performed with one femoral artery incised or directly punctured for access. The incision is usually packed and left open until placement of the TEVAR. The angiographic examination was used to ascertain treatment outcomes (Fig. 1B).

Situ fenestration technique
Under general anesthesia, one femoral artery was exposed for the aortic stent grafts, and left brachial artery was then exposed for the fenestration of LSA. Then TEVAR was deployed with its proximal end next to the ostium of the LCCA. An 8F angle-adjustable sheath (Lifetech, Inc., Shanghai, China) was inserted retrograde from the left brachial artery until its tip reached the aortic stent graft. The tip was adjusted to be as perpendicular as possible to the greater curvature of the aortic stent graft. Once the sheath was located in the expected position, a flexible aspiration biopsy needle was used to penetrate the aortic stent graft to create the fenestrations. After puncturing, a 0.014-inch guidewire was advanced through the aperture into the ascending aorta ( Fig. 2A). The initial aperture was then expanded by a 4-mm balloon, and then a 0.035-inch stiff guidewire was exchanged for 8-mm balloon expansion. Under these circumstances, self-expanding Fluency stent was implanted with its proximal end in the aortic stent graft. The angiographic examination was used to ascertain treatment outcome (Fig. 2B).

Pre-fenestration technique
Under general anesthesia, one femoral artery was incised or directly punctured for access; using CTA and intraprocedural angiography, the distance of the disease area from the LSA and the distal end of the LCCA, and the diameter of the proximal anchorage zone of the disease area, the diameter of the LSA. The stent was released and underwent pre-fenestration based on the latter measurements (Fig. 3A). Precise deployment of the fenestrated arch stent-grafts was confirmed by postprocedural angiography, and angiographic examination was used to ascertain treatment outcome (Fig. 3B).

Outcomes
The outcomes of interest were mortality, stroke, left upper extremity ischemia, spinal cord ischemia (SCI). All were assessed clinically within 30 days after TEVAR and at midterm follow-up. The latter was defined as > 30 days and < 5 years. Stroke was defined as any new global or focal neurologic deficit lasting for > 24 h with an acute lesion on brain imaging [4]. SCI was defined as any new lower extremity deficit that was unrelated to an intracerebral event [4]. Left upper extremity ischemia was defined as weakness or numbness of the left upper extremity. All revascularized patients were treated with strict antiplatelet therapy for at least 3 months postoperatively. Most

Statistical analysis
Categorical variables were analyzed with Pearson χ2 test, continuity corrected χ2 test or Fisher's exact test. Continuous variables with normal distributions are presented as mean, standard deviation, and comparisons are made using Student t-test with significance accepted at p < 0.05. Skews distributions continuous variables are summarized as median, quartiles, and comparisons are made using non-parametric test with significance accepted at p < 0.05. Survival analysis was performed to estimated allcause mortality. Statistical analysis was carried out using SPSS 25 (SPSS, Inc., Chicago, IL, USA).
Demographics were similar between the unrevascularized and revascularized groups, except for procedure urgency. The unrevascularized group was more likely to have undergone TEVAR for urgent or emergent cases (p = 0.005). Demographics were similar between CSB and F-TEVAR groups, except there was a predominance of males in CSB group (90% vs. 76.7%, p = 0.036) ( Table 1).
Spinal cord ischemia developed in 2.9% (n = 5) of patients during their hospitalization and there is no occurrence of spinal cord ischemia at follow-up. In addition, 1 patient underwent LSA embolization 2 weeks after TEVAR due to the occurrence of type II endoleak in unrevascularized    Table 3). Comparison of 30-day and midterm overall study outcomes stratified by urgency showed no difference for stroke, left upper extremity ischemia, spinal cord ischemia, or mortality (Table 4). Kaplan-Meier estimates, stratified by procedure, were performed for survival. Survival estimates did not reach significance for unrevascularied LSA, CSB, or F-TEVAR (p = 0.176) (Fig. 4).

Discussion
Our study demonstrates there was a higher rate of 30-day SCI in patients with LSA coverage during TEVAR, and CSB and F-TEVAR were equivalent in terms of mortality, stroke, left upper extremity ischemia, and spinal cord ischemia at 30-day and mid-term.
There was a high rate of LSA unrevascularization in the urgently treated patients in our cohort. We performed selective revascularization in emergent cases and we were more likely to perform LSA revascularization in elective cases. For patients with acute thoracic emergency, TEVAR is required urgently and coverage of the LSA is necessary. LSA revascularization should be individualized and addressed on the basis of patient's anatomy and urgency [5]. Revascularization of the left subclavian artery during TEVAR has led to various controversies among vascular groups and organizations initiating guideline recommendations [3,6]. Nevertheless, there are no high-quality data to tip that revascularization is beneficial. Several studies are suggestive that revascularization is beneficial [1,7,8]. Other's counterargue that there are is no statistical differences [9][10][11][12]. Our study favors the former for the reason that there was a higher incidence of 30-day SCI in unrevascularized group. However, LSA coverage does not increase the incidence of mortality, stroke and left upper extremity ischemia in our study.
In those patients who underwent LSA origin coverage, the incidence of SCI was significantly higher in those patients who did not have surgical revascularization of the LSA (10.7% vs. 1.4%, p = 0.032). This result is consistent with previous studies [13,14]. A report from the European Collaborators in Stent-Graft Techniques for Abdominal Aortic Aneurysm Repair (EUROSTAR) registry of 606 patients demonstrated an SCI rate of 2.5% and found coverage of the LSA without revascularization was an independent risk factor for SCI (OR, 3.9; P = 0.027) [13]. The risk factors for SCI have been reported including long segment coverage of thoracic aorta, LSA coverage, prior abdominal aortic repair, and perioperative hypotension, but length of aortic coverage is the only independent predictive factor of SCI after TEVAR [15]. This study did not describe precise length of aortic coverage, but the number of stents used is a reasonable surrogate for this and there is no difference in the number of stents between two groups.
Data from the Medtronic Thoracic Endovascular Registry (MOTHER) database suggested that coverage of the LSA without previous revascularization significantly increased the risk of stroke, most specifically in the posterior territory [7]. However, a large retrospective cohort that included patients with various aortic diseases failed to find any benefit in terms of postoperative stroke for prior selective LSA revascularization when covered during TEVAR [10]. Our study seems to support the latter. In the MOTHER database, patients with an aortic aneurysm and presumable more extensive atherosclerotic disease had a higher stroke rate compared with patients who had chronic dissection. It might be due to differences in patient selection and type of aortic disease inducing this difference outcomes. A recent meta-analysis demonstrated that revascularization of covered LSA in TEVAR is associated with significantly reduced risks Fig. 4 Kaplan-Meier estimates for survival, stratified by procedure of cerebrovascular accident, SCI, and left upper limb ischemia but not with increased risk of 30-day mortality [16]. Left subclavian revascularization theoretically avoids a low flow state of vertebral arteries which may induce stroke especially in those people whose Willis circle is intact, but it is difficult to accurately assess the integrality of Willis circle for all patients, especially in an emergency.
In our study cohort, comparisons between the unrevascularized and revascularized groups in left upper extremity ischemia were not statistically significant. Klocker et al. [17] investigated left arm ischemia and functional status and quality of life after LSA coverage during TEVAR using the Disabilities of the Arm, Shoulder, and Hand questionnaire and 12-Item Short Form Health Survey. Their results showed that TEVAR with LSA coverage is associated with low risk of left arm ischemia and no impact on left arm function and quality of life after the investigation. However, unrevascularization group had 2 patients need subsequent carotid-subclavian bypass for severe left arm ischemia which needed our attention.
Revascularization technique, CSB, and F-TEAVR did not reveal any significant difference in complication incidence in our sample. Thus, our findings would suggest that LSA revascularization during zone 2 TEVAR may be safely and effectively performed with either approach. Despite the increased arch manipulation and procedural complexity required for totally endovascular therapy of the LSA during zone 2 F-TEVAR, this may not translate into worse neurological outcomes. Besides, graft manipulation and the clamping of left common carotid could also increase rates of stroke in the CSB group [18]. Although unable to comment on the exact role of specific anatomic features in determining the prognosis of individual outcomes, the findings would suggest that, provided physician's judgment of patients' fitness for one technique (which is a complex decision based on the need to balance lowest possible perioperative morbidity in top of ensuring the longest durable outcomes possible), 30-day and midterm outcomes are equally satisfied with both CSB and F-TEVAR.
In our cohort, 30-day and midterm complications appeared to occur more commonly in CSB, but this difference was not statistically significant. It also might be argued that the additional open procedure might cause extra complications (particular local bleeding, nerve palsy, graft infection and lymphatic leak) [19,20], while a totally endovascular approach to LSA revascularization would theoretically not be subject to these complications and would maintain the minimally invasive nature of TEVAR [21]. In our study, all patients underwent LSA fenestration during TEVAR were scheduled for elective surgery. CSB still serves as gold stand for late efficacy data of emerging endovascular solutions due to excellent patency rates and low rate of proximal landing zone reinterventions [19,20]. Surgical subclavian bypass can be performed in urgent or even emergent cases and allow simpler exclusion of the aortic pathology by a more straightforward stent graft procedure [22]. Both situ fenestration and homemade fenestrated stent-graft need anatomic considerations. For situ fenestration, A 90° puncture or dilatation angle can avoid irregular fenestrations and fabric tears [23], which means more strict aortic arch morphology is required. For homemade fenestrated stent-graft, Graft rotation and misalignment of the vessel ostium interface can still occur [24]. Precise deployment of these fenestrated arch stent-grafts to correctly orient the fenestrations toward the branches for which they are intended also put higher demands on physicians. In addition, the durability of the stent-graft (Metal fatigue and material deterioration) after fenestration was a significant problem. Long-term monitoring of patients is required to avoid major complications resulting. In our series, no stent fractures were detected by follow-up radiologic examinations.

Limitations
It was a retrospective study and the selection of the procedure lends itself to selection bias. The number of patients with fenestrated grafts increased with clinical experience. The unrevascularization groups accounting for only 16.3% of the entire study which may have influenced outcomes. Besides, the influence of aortic arch morphology on choice of surgical methods for the management of LSA needed further evaluation.

Conclusions
Based on present study, LSA revascularization in zone 2 TEVAR is necessary and important which is associated a low 30-day rate of spinal cord ischemia. When LSA revascularization is required during TEVAR, CSB and F-TEVAR are all safe and effective methods, and F-TEVAR appears to offer equivalent clinical outcomes as a less time-consuming and minimally invasive alternative.