Outcomes of Thoracic Endovascular Aortic Repair Using Aortic Arch Chimney Stents in High Risk Patients

Abstract

Introduction

Aortic arch disease is a challenging clinical problem, especially in high-risk patients where open repair can have morbidity and mortality rates of 30–40% and 2–20%, respectively. Aortic arch “chimney”(AAC) stents used during thoracic endovascular aortic repair(TEVAR) are a less invasive treatment strategy than open repair, but the current literature is inconclusive about the role of this technology. The focus of this analysis is to describe our experience with TEVAR and AAC stent(s).

Methods

All TEVAR procedures performed from 2002–15 were reviewed to identify those with AAC stents. Primary end-points were technical success, as well as 30-day and 1-year mortality. Secondary end-points included complications, reintervention, and endoleak. Technical success was defined as a patient surviving the index operation with deployment of the AAC stent(s) at the intended treatment zone with no evidence of type 1 or 3 endoleak on initial postoperative imaging. The Kaplan-Meier method was used to estimate survival.

Results

Twenty-seven patients(age:69±12 years[male 70%]) were identified, and all were described as prohibitive risk for open repair by the treating team. Relevant comorbidity rates were: coronary artery disease/myocardial infarction(59%), O2-dependent emphysema(30%), preoperative creatinine>1.8mg/dL(19%), and congestive heart failure(15%). Presentations included: elective-67%(n=18), symptomatic-26%(n=7), and ruptured-7%(n= 2). Eleven(41%) had prior endovascular and/or open arch/descending thoracic repair. Indications were: degenerative aneurysm(49%), chronic residual type A dissection with aneurysm(15%), type 1a endoleak after TEVAR(11%), post-surgical pseudoaneurysm(11%), penetrating ulcer(7%), and acute type B dissection(7%). 32 BC vessels were treated: innominate, n=7; left common carotid artery(LCCA), n=24; left subclavian artery(LSA), n=1. Five patients(19%) had simultaneous innominate-LCCA chimneys. BCC stents were planned in 75%(n=24) with the remainder placed for either LCCA or innominate artery encroachment(n=8). Overall technical success was 89%(1-intraoperative death, 2-persistent type 1a endoleaks in follow-up).

30-day mortality was 4%(n=1; intraoperative in a patient with a ruptured arch aneurysm) and median LOS was 6[IQR 4, 9] days. Seven(26%) patients experienced a major complication(stroke-3[all with unplanned BCC], respiratory failure-3, and death-1). Nine(33%) patients underwent aorta-related reintervention, and no chimney occlusion events occurred during follow-up(median follow-up:9[IQR 1,23] months). One and 3-year survival is estimated to be 88±6% and 69±9%, respectively.

Conclusions

TEVAR with AAC can be performed with high technical success and acceptable morbidity and mortality in high-risk patients. Unplanned AAC placement during TEVAR results in an elevated stroke risk, which may be related to the branch vessel coverage necessitating AAC placement. Acceptable mid-term survival can be anticipated, but aorta-related reintervention is not uncommon, and diligent follow-up is needed.

Introduction

Open surgical repair of proximal descending and aortic arch pathology has historically been reported to have high morbidity(30–40%) and mortality(2–20%) rates, depending on patient comorbidities, indication for repair and acuity of the presentation¹. Staged or hybrid approaches are often used to mitigate the risk of repair, including arch vessel debranching and/or extra-anatomic bypass with simultaneous or staged thoracic endovascular aortic repair(TEVAR)^2–4. Outcomes of these techniques have mixed results, with morbidity and mortality rates of 30–40% and 10–15%, respectively^{5, 6}. Although branched/fenestrated arch devices are less invasive, they are not widely available, not designed to treat the full gamut of arch pathologies, and have ill-defined durability⁷.

An alternative approach to proximal aortic disease management is use of “chimney”(AAC) stents as adjuncts to TEVAR^8–10. The concept of parallel chimney stents was first described as a “bailout” maneuver after inadvertent visceral vessel coverage during endovascular abdominal aortic repair¹¹. Since the initial description, there has been rapid proliferation and application of “chimney” techniques in the management of paravisceral aortic disease^{9, 11–13}. Not surprisingly, this technique is now being increasingly applied to more proximal aortic/arch disease treated during TEVAR^14–17. The allure of this approach is that it provides a near total endovascular solution and can be completed with readily available technologies using implantation techniques familiar to most operators performing TEVAR. Additionally, this is a versatile technique that is applicable to elective and non-elective presentations for a variety of aortic diseases.

However, AAC used during TEVAR is still an unproven strategy, and concerns regarding patient selection, device choice, operative technique, durability and long-term outcomes remain unresolved. Thus, we sought to review our experience with AAC techniques used during TEVAR.

Methods

The study was approved by the University of Florida Institutional Review Board(#838–2014). The need for informed consent was waived due to the retrospective nature of the analysis.

Study cohort

A prospectively collected database was queried for TEVAR procedures performed with AAC stent(s) performed between January 2002 and June 2015. AAC procedures were first attempted in our practice after 2009 due to our experience with visceral aortic chimney procedures. During this time period, 968 patients underwent TEVAR at our institution, of whom 115(12%) had Ishimaru¹⁸ Zone 0/1 deployments. Of this subset, 27(23%) were deemed unfit for direct open repair or extra-anatomic arch debranching and underwent AAC stent(s) placement. Planned or unplanned AAC procedures were included in the study, and patients who underwent simultaneous sternotomy, thoracotomy, or placement of a fenestrated/branched arch device (N=4) were excluded.

Patient demographics, comorbidities, and operative variables were extracted from the database and electronic medical record. The definitions and severity of comorbidities were described as per the Society of Vascular Surgery guidelines¹⁹. All additional concurrent adjunctive procedures were described per reporting standards¹⁹. Postoperative computed tomographic angiograms(CTA) were reviewed to verify chimney patency and determine presence of endoleak. Reintervention was defined as any unplanned return to the operating room and was dichotomized into aorta and non-aorta related indications.

Patient selection and clinical practice

All patients were considered prohibitively high risk for open surgical repair²⁰ due to the unique constellation of medical and anatomic factors that characterized each patient’s presentation. Consensus opinion was obtained regarding risk for open repair in each case among members of the vascular surgery and thoracic/cardiovascular surgery groups as previously reported²¹. For planned AAC procedures, patients and/or their families were thoroughly informed of the “off label” nature of the procedure.

Preoperatively, all patients underwent CTA with center-line, three-dimensional reconstruction(TeraRecon Inc., San Mateo, CA) for planning. The treating surgeon was responsible for device selection and implantation technique. During the study period, our practice evolved from selective to routine, pre-emptive revascularization of the left subclavian artery(LSA) in cases of anticipated long-segment aortic coverage(>200mm) in an effort to reduce spinal cord ischemia(SCI) and stroke risk^{22, 23}. Similarly, pre-emptive spinal drainage was increasingly utilized in extensive aortic coverage cases if the patient’s clinical presentation allowed.

Brachiocephalic chimney technique

All operations were performed under general anesthesia in hybrid operating rooms using fixed imaging, single plane systems(Infinix VC-I, Toshiba Medical Systems, Tokyo, Japan;Artis zeego system, Siemens Medical Solutions USA, Inc., Malvern, PA). Depending on the planned configuration of the reconstruction, the brachiocephalic target vessel(s) were exposed for retrograde delivery of branch stents, or a carotid-subclavian(CS) bypass was performed, through which a retrograde carotid stent-graft was placed. Routinely, percutaneous access of the femoral arteries was used for TEVAR delivery as previously described²⁴. Patients were systemically heparinized(80–100 Units/kg) to achieve an activated clotting time ≥ 300 seconds prior to any wire manipulation of the aortic arch. A combination of digital subtraction angiography with intravascular ultrasound(IVUS) was used to minimize risk of unintended branch vessel compromise. The main thoracic endograft was delivered and deployed, with concomitant or subsequent deployment of the chimney stent(s), depending on the type of main aortic graft and the landing zone.

For Zone 0 or 1 implants, selective utilization of right atrial balloon occlusion and/or rapid ventricular pacing was employed. In cases of planned LSA coverage, the CS bypass was routinely completed prior to graft implantation. Typically, self-expanding stent-grafts(e.g. Gore® Viabahn®;W.L. Gore, Inc., Flagstaff, Arizona) were used for left common carotid artery(LCCA) chimneys due to their flexibility; however these are often supported proximally with a balloon and/or self-expandable stent/graft(e.g. iCAST™; Atrium Maquet Getinge Group, Inc., Germany or Zilver 518®;Cook Medical, Inc., Bloomington, Indiana). Notably, due to the large diameters of the innominate artery that are frequently encountered, alternative self-expanding stent graft choices such as iliac endograft limbs(e.g. Zenith Flex® TFLE/ZSLE; Cook Medical, Inc., Bloomington, Indiana) were utilized(Figure 1 and 2)(Supplementary Table I).

Figure 1.

A. This image demonstrates a patient with a proximal descending thoracic aneurysm requiring a double chimney into the ascending aorta (Zone 0). The carotid subclavian bypass has been completed, and the Left common carotid artery (LCCA) chimney, in this case a Gore Viabahn, is placed through the bypass graft. The innominate chimney stent, in this case an iliac limb device, has been delivered through the Right axillary artery. The thoracic aortic endograft is also in place. Graft sizing is generally 30% greater than the outer to outer centerline aortic measurement at the anticipated landing zone (0–10% in cases of dissection). Overlap between the chimney and thoracic stent-grafts is generally 3cm or greater (especially if there is significant arch curvature). B. This image demonstrates the device partially deployed, with the proximal aspect still constrained while the chimney stents are deployed, which is demonstrated in image C. Image D demonstrates concomitant ballooning of all three stents with a compliant balloon in both the innominate and aortic stent-grafts and a non-compliant balloon in the L CCA stent-graft. This final ballooning often takes multiple surgeons working simultaneously to complete. With proximal deployment such as this, we often use a right atrial inflow occlusion balloon (not shown) that is delivered via right common femoral vein access to lower the blood pressure and avoid shifting of the devices during ballooning. We advocate internal reinforcement of the chimney stent-grafts with self-expanding stents. In this image, the left subclavian artery has been embolized using a vascular plug.

Figure 2.

83 year old patient status post previous sternotomy for coronary artery bypass presented with a large arch aneurysm. A left-carotid-subclavian bypass was performed and access to the left common carotid artery was obtained through the bypass graft (A). Access to the right axillary artery was obtained for delivery of a Zenith Flex® TFLE (Cook Medical, Bloomington, IN) iliac limb into the innominate artery (B). After deployment of the stent grafts, a small type Ia endoleak was noted, however this resolved after concurrent aortic/branch stent balloon angioplasty (C/D). Follow-up CTA demonstrated successful thrombosis of the aneurysm sac and widely patent stented branch vessels (red circle) (E/F). Figure F also demonstrates an Amplatzer™ vascular occlusion plug (St. Jude Medical, St. Paul, MN) in the left subclavian artery (white arrow).

Postoperative management

Spinal drain management and blood pressure goals were based upon a previously published protocol²³. Postoperative surveillance included CTA prior to discharge for patients undergoing non-elective procedures, followed by CTA at 1-month, 6-months and annually thereafter. Unless contraindicated, patients were started on dual anti-platelet therapy(aspirin 81 mg/day and clopidogrel 75mg/day) and a statin(e.g. simvastatin 20mg/day) post-procedure. Clopidogrel was discontinued after 3 months, while aspirin and the statin agent were prescribed indefinitely. Need and timing of reinterventions were left to the discretion of the operating surgeon but were all adjudicated by multiple experienced cardiothoracic and vascular surgeons in the practice. Frequently, innominate chimney interventions were performed via right axillary artery access while left carotid chimney interventions were completed via left brachial artery access.

End-points and definitions

The primary end-points of the analysis were technical success, as well as 30-day and 1-year mortality. Secondary end-points included complications, reintervention and endoleak. A significant stenosis of a chimney stent was defined as ≥ 50% decrease in luminal diameter on follow-up CTA. Technical success was defined by survival of the index operation with deployment of the AAC stent(s) at the intended treatment zone and no evidence of type Ia or III endoleak on the first follow-up CTA. Mortality events were verified by query of the Social Security Death Masterfile.

Statistics

All statistical analyses were performed using STATA software(version 9.2, College Station, Texas). Categorical factors were summarized using frequencies and percentages, and continuous variables were described using means, medians, and standard deviations. Categorical variables and continuous measures were compared between groups using Fischer’s exact test or two-sampled t-tests, when indicated. Survival analysis was completed using Kaplan-Meier methodology. A significance level of P<.05 was assumed for all tests.

Results

Preoperative patient characteristics

From 2002–15, 27 patients underwent TEVAR with AACC. The majority were male(70%;n=19) with mean age of 69±12 years(range 40–90) and 41%(n=11) had prior history of either endovascular and/or open aortic repair. Most had a history of hypertension(96%), coronary disease(59%), and dyslipidemia(56%). Additional details regarding patient demographics and comorbidities are presented in Table I. A majority(67%;n=18) were treated electively, while seven(26%) were urgent/symptomatic and two(7%) had a diagnosis of rupture(Table II). Virtually all patients were deemed medically high-risk as determined by an American Society of Anesthesiologists classification of IV(96%;n=26). The most frequent presenting diagnosis was degenerative thoracic aortic aneurysm(48%).

Table I.

Patient characteristics and comorbidities

Characteristic	N=27, No (%)
Age, years (mean ± SD)	69±12
Male	19 (70)
BMI	27±6
Comorbidities	No (%)
Hypertension	26 (96)
Smoking (current/prior)	21 (78)
Coronary artery disease	16 (59)
Dyslipidemia	15 (56)
COPD (O2-dependent)	8 (30)
Renal insufficiency (Cr>1.8mg/dL)	5 (19)
Congestive heart failure	4 (15)
CVOD	4 (15)
PVOD	3 (11)
Arrhythmia	2 (7)
Diabetes	2 (7)
Composite total, mean ± SD	3.7±1.5

SD, standard deviation; BMI, Body Mass Index; COPD, chronic obstructive pulmonary disease; CVOD, Cerebrovascular Occlusive Disease; PVOD, Peripheral Vascular Occlusive Disease

Table II.

Clinical presentation, aortic pathology, and aortic-surgery history

Clinical Presentation	N=27, No (%)
Elective	18 (67)
Urgent	7 (26)
Rupture (Emergent)	2 (7)
ASA IV	26 (96)
Aortic Pathology	No (%)
Degenerative thoracic aneurysm	13 (48)
CTAD with aneurysm	4 (15)
Type 1a endoleak post-TEVAR	3 (11)
Post-surgical pseudoaneurysm	3 (11)
Penetrating ulcer	2 (7)
Acute dissection	2 (7)
Aortic Variables	No (%)
Maximum aortic diameter at presentation (cm ± SD)	6.1 ± 1.6
Prior aortic surgery history	11 (41)
Open thoracic	5 (19)
TEVAR	2 (7)
Open abdominal	3 (11)
EVAR	1(4)

ASA, American Society for Anesthesia Physical Status Classification; CTAD, chronic Type A aortic dissection; TEVAR, thoracic endovascular aortic repair; EVAR, endovascular infrarenal aortic repair; SD, standard deviation

Aortic arch chimney vessel data and operative characteristics

In total, 32 aortic arch chimney stents were implanted including the left subclavian artery(LSA=1), left CCA=24, innominate artery(IA=6), and one additional chimney placed into a pre-existing ascending aortic arch innominate/carotid debranching graft. A single chimney was utilized in 22 patients(83%) and 75%(n= 24) of chimney placements occurred electively. Additional details regarding Ishimaru¹⁸ zone deployment, chimney target vessel and implant strategy are highlighted in Figure 3 and Table III.

Figure 3.

This figure depicts the aortic arch deployment zones (e.g. Ishimaru¹⁸) where the various brachiocephalic chimney stents were implanted. Due to a practice bias to routinely employ a left carotid-subclavian bypass, only a single subclavian chimney stent was placed during an elective procedure to treat a saccular transverse proximal thoracic aneurysm. The vast majority of the chimney procedures occurred for Zone 0 or 1 thoracic endovascular aortic stent deployments.

Table III.

Chimney target vessel, anatomic combinations and indications

Chimney target	No.
Total number chimneys stents implanted	32
L CCA	24
IA	6
L SA	1
Debranching graft (other)	1
Chimney stent combination	N=27, No. (%)
L SA alone	1 (4)
L CCA alone	19 (70)
L CCA + IA	5 (19)
IA alone	1 (4)
Other	1 (4)
Chimney insertion scenario	N = 32, No. (%)
Planned	24 (75)
Rescue (unplanned)	8 (25)
Unplanned chimney indication	N = 8, No. (%)
L CCA coverage	6 (75)
IA coverage	1 (13)
Debranching graft coverage	1 (13)

LSA, left subclavian artery; LCCA, left common carotid artery; IA, innominate artery

Twenty patients(74%) had either pre-emptive or concomitant left CS bypass. Excluding the unplanned LSA chimney, the remaining 6 patients in the series did not undergo subclavian revascularization. These cases occurred early in the experience and before guidelines²² became available regarding LSA revascularization during TEVAR. Multiple combinations of TEVAR graft and AAC stent combinations were used(Supplementary Table I).

The most common TEVAR device utilized was a Zenith® TX2®(Cook Medical, Bloomington, IN) (70%;n=19), and the most common BCC devices were a self-expanding covered stent graft(Gore® Viabahn®, W.L. Gore, Inc., Flagstaff, AZ) and/or balloon-expandable covered stent graft(Atrium iCAST™, Maquet Getinge Group, Inc., Germany)(n= 26 of 32;81%). Internal reinforcement with a balloon or self-expandable stent(Zilver 518®, Cook Medical, Inc., Bloomington, IN) was employed in 31%(n=10) of AACC. Specific information on intraoperative adjunct, procedural details and chimney stent types are listed in Table IV.

Table IV.

Perioperative adjuncts and procedure related variables

Adjunct Type	N, No (%).
Total femoral vessels accessed	44
Open access	12 (27)
Percutaneous access	32 (73)
Total subclavian vessels covered	28
Subclavian vessels revascularized	23 (82)
Prior C-SC bypass	6
Concomitant C-SC bypass	15
Prior debranching graft	1
LSA Chimney	1
Access vessel adjunct
Common carotid artery exposure	5
Right subclavian artery exposure	1
Axillary artery exposure	6
Brachial artery exposure	5
Brachial artery percutaneous access	6
Open aortic access	1
Intraprocedural adjunct
Subclavian artery embolization	12
Endovascular conduit	1
Open iliac conduit	2
Right atrial inflow occlusion balloon	4
Simultaneous EVAR	4
Femoral endarterectomy	1
Femoral-brachial wire	1
Procedure variable	mean ± SD or No (%)
Cerebrospinal fluid drain	15 (56)
Procedure time, min	240±115
Fluoroscopy time, min	42±28
Contrast material exposure, mL	160±100
Estimated blood loss, mL[median, range]	300[150, 3000]
Completion angiogram Ia endoleak	2 (7)
Chimney stent details	N = 32
Self-expanding stent graft alone	7
Balloon expandable stent graft alone	9
Self-expanding stent alone	4
Combination/reinforced (internally stented)	10

EVAR, endovascular aortic aneurysm repair; C-SC, carotid-subclavian; L SA, left subclavian artery

Postoperative outcomes and complications

Mean length of stay was 9±9 days(median 5.5;range 1–36). Of surviving patients, the majority(65%;n=17) were discharged to either home or short-term rehabilitation while 4 patients(15%) were transferred to a long-term acute care facility. The 30-day and overall in-hospital mortality rate was 4%(n=1). A description of the major postoperative neurologic, cardiopulmonary and renal complications is listed in Table V. Seven patients(26%;n=4, elective;n=2, urgent-symptomatic;n=1, emergent-ruptured) had a major post-operative complication with 4 patients suffering multiple complications. No episodes of spinal cord ischemia occurred. However, three patients(11%) suffered a postoperative stroke, two of which had complete clinical resolution, with one having a persistent severe neurologic deficit at discharge(major stroke rate:4%). Notably, all 3 patients experiencing a stroke were in cases of unplanned AAC deployment. Details of these three cases are further outlined in Supplementary Table II.

Table V.

Perioperative outcomes of aortic arch chimney procedures

Characteristic	N=27, No. (%)
30-day mortality	1 (4)
Any complication	16 (59)
Any major complication	7 (26)
Stroke	3 (11)
Resolved (MRS 0)	2
Persistent deficit	1
Respiratory failure	3(11)
Spinal cord ischemia	0
Acute kidney injury	0
Cardiac	0
Intraop endoleaks
Type I/III	3 (11)
Type II	1 (4)
False Lumen flow	1 (4)
Indeterminate	0 (0)
Disposition of survivors	N=26, No. (%)
Home	17 (65)
Inpatient facility	9 (35)
Length of stay by urgency, days	N=26, [median, IQR]
Elective	6 [4,9]
Urgent and Emergent	7 [6,10]

MRS, Modified Rankin Scale

Follow-up, endoleak and reintervention

The median clinical follow-up time was 9(range 1–23) months. At least one postoperative CTA was available for 24 of 26 surviving patients. No stent migration, component separations, fractures, retrograde dissections and/or aneurysm ruptures occurred during follow-up. Nine(33%) patients underwent aorta-related reintervention and no AACC occlusion events were identified during the study interval. Three patients underwent open conversion during follow-up [type 1a endoleak, N =1; false lumen aneurysm expansion/persistent perfusion, N=2]. Estimated freedom from any aorta and/or non-aorta related reintervention at 12 months was 48±13%(Figure 4). Additional detailed descriptions of perioperative events, as well as timing and nature of reinterventions are highlighted in Table VI.

Figure 4.

This Kaplan-Meier curve with 95% confidence intervals demonstrates the estimated freedom from any aorta and/or non-aorta related reintervention. The 12-month estimated freedom from reintervention is 48±13% (standard error of the mean exceeds 10% at 11 months). Due to the complexity of the pathology that was selected for these procedures, patients underwent various types of reintervention including remote aortic site operations for synchronous or metachronous disease, as well as remediation of the index aortic repair.

Table VI.

Description of Aortic Arch Chimney Patient Complications, Endoleak and Reintervention

Patient	Intraoperative Complication	Intraoperative Endoleak	Postoperative Complication	Endoleak in follow-up	Aortic and/or Chimney Reintervention
1	None	Type 1a	Stroke, respiratory failure	Type 1a @ 15.2mo	None (stable TAA @ 19.2mo)
2*	None	None	None	Persistent FL flow	TEVAR extension, chimney re-stent for stenosis @ 3 mo
3*	R SC injury	None	Respiratory failure, DVT, C.difficile	None	None
4	None	None	None	None	Open AAA repair @ 1.9mo
6	None	None	None	Persistent FL Flow	Open conversion @ 29.2mo
8	None	None	None	Type 2 @ 6mo	Aortic root replacement @ 8.4mo
10	Femoral thromboembolism requiring thromboembolectomy	Type 2	Neck hematoma aspiration	Persistent FL flow	TEVAR extension @ 13.8mo
12	None	None	Delirium, protracted LOS	None	None
13	None	None	None	Type 1a @ 12mo	Open conversion @ 14mo
15*	None	None	Stroke, GI bleed	None	None
16*	None	None	None	None	None
17*	Cardiac arrest/death	None	Death	N/A	N/A
18	None	None	None	None	EVAR for AAA @ 8.6mo
20	None	None	Stroke	None	None
21	None	Type 1a	None	None	None
22	None	None	None	Persistent FL flow	Open conversion @ 21.2mo
23	None	None	None	Type 2 @ 1 mo.	Left carotid chimney extension @ 2.1mo-Coronary/LIMA steal
24	None	None	None	Persistent FL flow	Open AAA repair planned
25	None	None	Respiratory failure	None	None
27	None	None	None	Type 2 @ 1 mo.	Left carotid chimney re-stent for deformation @ 2mo

^*unplanned chimney procedures; TAA, thoracic aortic aneurysm; FL, false lumen; TEVAR, thoracic endovascular aortic repair; R SC, right subclavian artery; DVT, deep venous thrombosis; AAA, abdominal aortic aneurysm; LIMA, left internal mammary artery; N/A, not applicable

Notably, the rate of any type of intraoperative endoleak was 11%(N=3; type 1a, N =2). The 2 patients experiencing intraoperative type 1a endoleak did not undergo subsequent remediation in follow-up as their aortic diameters decreased and the endoleaks resolved. There were 4 additional patients who developed an endoleak during follow-up. Only one of these patients was noted to have a type 1a endoleak at 12 months postoperatively, however this did lead to open conversion with arch reconstruction.

AAC related reinterventions occurred in 3 patients. One case was due to coronary-subclavian steal from LCCA chimney compression requiring chimney extension 2 months postoperatively. The second AAC related reintervention occurred 1 month postoperatively was also for significant LCCA chimney compression that was remediated using a self-expanding stent. The final patient underwent proximal TEVAR extension requiring extension of a carotid chimney 3 months after the index procedure. The AAC primary patency is estimated to be 89±6% at 12 months(Figure 5).

Figure 5.

The primary patency of brachiocephalic chimney stents is estimated to be 89±6% at 1-year. Three patients underwent chimney-related reintervention during follow-up, however no occlusion events occurred. Reintervention occurred for either stent compression (>50% stenosis) on follow-up computed tomographic imaging or proximal extension of the chimney to facilitate more proximal TEVAR. All displayed intervals have < 10% standard error of the mean. The image also depicts the 95% confidence interval.

Fourteen patients have ≥6 months of clinical follow-up time with appropriate imaging follow-up. In these patients, diameter stabilization was detected in 11(79%) while significant reduction(≥ 5mm) occurred in 7(50%). Overall, mean survival time was 62(SE±6; 95% C.I. for mean: 50–74) months. One and 3-year survival is estimated to be 88±6% and 69±9%, respectively(Figure 6). There were 5 late deaths, however all were due to non-aorta related pathology(pneumonia resulting in respiratory failure-2; myocardial infarction-3, none related to supra-aortic trunk vessel patency/left internal mammary graft compromise).

Figure 6.

Survival after thoracic endovascular aortic repair with brachiocephalic chimney stents is estimated to be 88±6% and 69±9%, at one and 3-years respectively, in this high risk population. All displayed intervals have < 10% standard error of the mean. The image also depicts the 95% confidence interval.

Discussion

This study represents the largest published experience of AAC stents with TEVAR for Zone 0/1 deployments, and further supports the role of these procedures for the treatment of high-risk patients with complex proximal thoracic and transverse aortic arch disease, and as a salvage maneuver for unintentional branch coverage. These data demonstrate that good short term results can be achieved relative to previously reported outcomes after direct arch reconstruction and/or debranching. Notably, despite having a high medical and anatomic risk cohort, excellent technical success with relatively low morbidity and mortality was observed, especially in patients with planned AAC procedures. No AAC occlusion events occurred, but three reinterventions were required for significant proximal stenosis related to stent compression. The overall reintervention rate was significant and also reflected the need for future treatment of synchronous or metachronous aortic disease.

The historical gold-standard for repair of proximal descending and transverse aortic arch disease is open surgical repair, but outcomes of mixed hemiarch and total arch repair with elephant trunk or frozen elephant trunk have variable elective morbidity(15–40%) and mortality rates(6–22%)^25–27. Due to mixed pathologies, presentations, and the complexity of these procedures, it is difficult to characterize which factors drive outcomes in these series^28–30. With improvements in operative technique and perioperative care, contemporary results show that 30-day mortality for patients undergoing open thoracic aortic repair ranges between 2–20%^{28, 29}. In the absence of aortic dissection, centers of excellence frequently report elective mortality rates under 5% for good risk patients. However, there are subsets of patients with proximal thoracic/transverse aortic arch disease with significant comorbidities who may be deemed poor candidates for conventional open repair³¹. The morbidity of open thoracic aortic repair, especially in poor risk candidates is sobering, with 30-day morbidity and mortality rates of 30–50% and 10–20%, respectively²⁸.

This series represents a high-risk population as evidenced by preoperative ASA scores, comorbidities, presentation and the history of previous aortic repair in 41%. Importantly, all were felt to be prohibitively high-risk for open repair by a group of experienced vascular and cardiothoracic surgeons at a tertiary care medical center with a practice that collectively treats approximately 600 aortic patients per year. Notably, 9 of our patients in this series presented non-electively, and despite that presentation, 8 of these 9(89%) not only survived repair, but were discharged to home or a short-term rehabilitation facility. Several reports highlight that TEVAR with AAC is associated with lower mortality risk^{17, 32} than that reported for open and hybrid arch reconstruction, and some authors feel that this technique should be extended to lower risk patients³¹. However, despite these promising results, we feel that the durability of TEVAR with AAC is still unproven, and we have continued to offer open arch repair as a first line therapy to patients considered to be appropriate surgical candidates.

To address high risk patients, surgeons have developed a variety of hybrid arch strategies^{27, 33}, combining debranching of the supra-aortic trunk vessels to extend the proximal landing zone with immediate or delayed deployment of a distal endograft. Although this method decreases the hypothermic circulatory arrest time, stroke and mortality rates are high at 5–10% and 6–17%, respectively^{27, 31, 34}. Additionally, our practice bias has evolved to minimize use of carotid-carotid bypass due to the known risk of dysphagia postoperatively. Moreover, a meta-analysis by Benedetto and colleagues demonstrated that hybrid arch reconstruction had greater risk of early mortality(OR 1.5, 95%CI 0.6–3.7) and stroke risk(OR 1.9;95%CI .1–17.1) compared to open repair³⁰.

An alternative novel technology that may reduce risk of complex proximal aortic reconstruction is fenestrated-branched TEVAR.^{35, 36} However, at this time, these devices are not widely available, not universally suitable to all anatomies, and require specialized operator skills. These limitations further support a role for AAC as a reasonable alternative option for more widespread use in patients deemed unfit for open repair³². Other authors have advocated laser in situ aortic graft fenestration for left subclavian revascularization, which may provide an additional treatment choice in a subset of patients, especially once this method can be extended to more proximal vessels³⁷. However, the impact of laser fenestration on device integrity and durability is unknown.

The results of this analysis compare favorably with a meta-analysis of 124 TEVAR patients and 136 AAC grafts by Moulakakis and associates¹⁰, who demonstrated a perioperative mortality and stroke rate of 5% and 4%, respectively. Our series had a single perioperative mortality event(4%) in a patient with a ruptured arch aneurysm, with a stroke rate of 11%. While definitive conclusions about the stroke risk are difficult to make given the heterogeneity of the patient presentations, it is notable that all of the strokes occurred in unplanned AAC cases. Aorto-iliac tortuosity and arch angulation were felt to be significant contributing factors in each of these cases that led to inadvertent supra-aortic trunk vessel coverage. Presumably, the strokes may have been related to temporary coverage of vessel ostia, along with the additional arch manipulation that is required to perform an unplanned AAC.

Like others³⁸, despite the high technical success and acceptable short-term outcomes, we remain concerned about the durability of AAC due to endoleak and/or arch vessel patency. Concerns about ‘gutter leaks’ were initially raised by Suguira et al.³⁹ when they reported midterm outcomes after 11 AAC procedures and 2(18%) patients developed type 1a endoleak. Furthermore, the pooled endoleak rate estimates from multiple series were 19% with 11% due to type 1a endoleak^{10, 32}. Indeed, 3 patients(11%) in our experience were noted to have type 1a endoleak within the first postoperative year, with one resulting in open conversion(Table VI). Importantly, the other 2 patients had resolution of their endoleak within the initial postoperative year and continued to have stable aortic diameters during follow-up.

The reintervention rate in this study(33%) exceeds that of other AAC series where estimates of 10–25% have been reported¹⁰. We feel that several important differences likely explain the higher reintervention rate in our series. Specifically, 84% of cases in the reported literature had a single chimney with more than 50% of those in the LSA, which is much different than our series. With the exception of a single procedure, all cases in this experience utilized either zone 0 or 1 with a LCCA stent and L-CS bypass, and 19%(n=5) of cases employed a ‘double-barrel’¹⁶ chimney technique. Additionally, our series includes 6 different aortic pathologies with over 30% of them being non-elective. Further, many(n=4) of the aorta-related reinterventions occurred outside of the initial intended treatment zone.

The technical conduct of AAC implantation during TEVAR has several important considerations regarding device choice and aortic landing zones. We concur with Gehringhoff et al.¹⁷ who report a preference for covered stents given the potential for erosion of the main aortic graft with bare metal stents, and the decreased likelihood for endoleak. In this series, we used a variety of thoracic/branch stent graft devices, therefore it was not possible for us to perform intra-graft comparisons or provide specific recommendations regarding device combinations. We generally plan to achieve at least 2–3cm of seal in normal diameter aorta in most cases, especially in cases of degenerative aneurysm. This is consistent with our philosophy for endovascular treatment of visceral aortic disease^{21, 40}. In general, we oversize chimney stents by 10–20% and oversize the thoracic endograft by approximately 30%, except in acute dissection where oversizing is typically minimized to 0–10%. Finally, our threshold for internal AAC stent reinforcement is low since the device-to-device and device to vessel interaction may lead to extrinsic compression as was seen in 2 of our patients who underwent reintervention.

Limitations

The current study has several inherent limitations related to its retrospective single center design with heterogeneous patient groups. There is certainly an inherent selection bias based upon multiple physiologic, anatomic and comorbid factors. Although this is one of the largest AAC series reported to date, the sample size is small. The chimney techniques evolved with experience and with changes in device availability, so it is difficult to identify a homogeneous strategy that affords more predictable outcomes. Importantly, our overall follow-up was short and the reintervention rates certainly raise concerns about long-term durability. This issue is particularly important in a patient population that has advanced comorbidity with complex multilevel aortic disease. Despite our centers extensive experience with complex aortic disease, long-term imaging follow-up has been challenging to obtain in this complex cohort of patients referred from a large geographic region. This limitation has been recently addressed by improved coordination of image data sharing from referring providers through cloud based services

Conclusions

In conclusion, the AAC technique used during TEVAR can be performed safely with a high rate of technical success, with acceptable perioperative morbidity and mortality rates, even in high-risk patients. While the technique is highly applicable to elective and non-elective presentations, neurologic morbidity remains a concern especially when AAC placement is unplanned. Importantly, reintervention is not uncommon; thus, close post-operative follow-up is crucial and anticipated patient compliance should be part of the clinical decision-making before offering these procedures. Larger series with longer follow-up are needed to further define optimal patient and device selection, and to understand the implications of reintervention.