Treatment of Aortoiliac Occlusive Disease with the Endologix AFX Unibody Endograft

(1)

Treatment of Aortoiliac Occlusive Disease with the Endologix AFX Unibody

Endograft

T.S. Maldonadoa,*, G.G. Westina, O. Jazaerib, M. Mewissenc, M.M.P.J. Reijnend, A.J. Dwivedie, H.E. Garrett, Jr.f, A. Dias Pererag, T. Shimshakh, V. Mantesei, C.J. Smolockj, Z.M. Arthursk

a

New York University Langone Medical Center, New York, NY, USA b

University of Colorado Denver, Denver, CO, USA c

Vascular Center at St. Luke’s Medical Center, Milwaukee, WI, USA d

Rijnstate Hospital, Arnhem, The Netherlands e

University of Louisville, Louisville, KY, USA f

University of Tennessee, Baptist Memorial Hospital Memphis, Memphis, TN, USA g

University of Tennessee, Cardiovascular Surgery Clinic, PLLC, Memphis, TN, USA h

Wheaton Franciscan Healthcare, Milwaukee, WI, USA i

Mercy Clinic Vascular Specialists, St. Louis, MO, USA j

Cleveland Clinic, Cleveland, OH, USA k

San Antonio Military Medical Center, Uniformed Services University of the Health Sciences, San Antonio, TX, USA

WHAT THIS PAPER ADDS

Treatment of aortoiliac occlusive lesions with a unibody stent-graft positioned at the aortic bifurcation has several potential advantages for the treatment of aortoiliac occlusive disease (AIOD): it can preserve the aortic bifurcation, avoid limb competition in the distal aorta, allow for future endovascular interventions, and protect against potentially fatal aortoiliac rupture in heavily calcified lesions. The current study examines the safety, efficacy, and early patency rates of the Endologix AFX unibody stent-graft for treatment of AIOD. This is the largest study examining the use of the Endologix AFX unibody stent-graft for the treatment of AIOD. The results of this study offer a significant contribution to the treatment of AIOD.

Objective/Background:Aorto-bifemoral bypass remains the gold standard for treatment of aortoiliac occlusive disease (AIOD) in patients with advanced (TASC D) lesions, but has signiﬁcant associated morbidity and mortality. Treatment with a unibody stent-graft positioned at the aortic bifurcation is a potential endovascular option for the treatment of AIOD. The current study examines the safety, efﬁcacy, and early patency rates of the Endologix AFX unibody stent-graft for treatment of AIOD.

Methods:A multicenter retrospective review was conducted of patients treated exclusively for AIOD with the AFX device. Primary, assisted primary, and secondary patency rates were noted. Clinical improvement was assessed using Rutherford classiﬁcation and ankle brachial index. Mean duration of follow-up was 22.2 11.2 months. Ninety-one patients (56 males [62%]) were studied.

Results:Sixty-seven patients (74%) presented with lifestyle-limiting intermittent claudication and the remaining 24 (26%) had critical limb ischemia. Technical success was 100%. Complications included groin infection (n¼ 4 [4%]), groin hematoma (n¼ 4 [4%]), common iliac rupture (n ¼ 4 [4%]), iliac dissection (n ¼ 4 [4%]), and thromboembolic event (n¼ 3 [3%]; one femoral, one internal iliac artery, and one internal iliac with bilateral popliteal/tibial thromboemboli). Thirty-day mortality was 1% (1/91) resulting from a case of extensive pelvic thromboembolism. At 1 year, 73% of patients experienced improvement in Rutherford stage of3 or greater compared with baseline. Nine patients (10%) required 16 secondary interventions. At all time points, primary patency rates were> 90%, assisted patency rates were > 98%, and secondary patency rates were 100%.

Conclusion:This is the largest study to examine the use of the Endologix AFX unibody stent-graft for the treatment of AIOD. Use of the AFX stent-graft appears to be a safe and effective endovascular treatment for complex AIOD.

Keywords:Aorto-iliac disease, Claudication, Stent-graft

* Corresponding author.

(2)

INTRODUCTION

In recent years, the shift in vascular surgery from open surgical intervention to endovascular therapy has been particularly pronounced in patients with atherosclerotic occlusive disease of the aorta and iliac arteries (aortoiliac occlusive disease [AIOD]), with an 8.5-fold increase (from 0.4 to 3.4 per 100,000 adults) in iliac artery angioplasty and stenting between 1996 and 2000 alone, while aorto-bifemoral bypass rates fell by only 15% (from 5.8 to 4.9 per 100,000 adults).1 However, for those lesions most difficult to treat with endovascular techniques (Trans-Atlantic Inter-Society Consensus [TASC II] type D lesions), aorto-bifemoral bypass (AFB) is the preferred revasculari-zation strategy. Traditional endovascular interventions for complex AIOD, such as “kissing” stents, pose technical challenges and increase procedural complexity and risk in patients with heavily calcified aortic bifurcations or aortic thrombus. Moreover, patency may be compromised in more complex lesions owing to radial size mismatch be-tween stents and certain stent configurations within the distal aorta.2

However, despite these technical limitations, endovas-cular therapy has the potential to reduce significantly the morbidity and mortality associated with the treatment of complex AIOD. Indeed, endovascular therapy as afirst-line treatment is becoming more commonplace, with open surgery reserved as a secondary option.3One alternative to treat disease of both the distal aorta and the iliac arteries is the use of stent-grafts. The Endologix AFX stent-graft is a unibody, low-profile stent-graft that preserves the aortic bifurcation and has a number of the advantages inherent to endovascular techniques. The current study examines the safety, efficacy, and early patency rates of the AFX unibody stent-graft for the treatment of AIOD.

METHODS

Design and data collection

An institutional review board-approved retrospective re-view of patients treated exclusively for AIOD with the Endologix AFX unibody stent-graft was conducted at nine centers in the USA and one in the Netherlands. This device is approved for the treatment of abdominal aortic aneu-rysms (AAAs) and holds unique advantages for use in pa-tients with aortoiliac occlusive disease, owing to its low-profile delivery system. All risks and benefits to using this device, including its off-label indication, as well as all al-ternatives (endovascular and open) were discussed with the patients in detail prior to surgery. Any patient with a concomitant aortic aneurysm (defined as an aortic diameter > 3.5 cm) was excluded so as to assure that intention to treat was for AIOD and not for aneurysmal disease. Trained chart abstractors reviewed medical records, including operative notes, to record patient demographics, presenting symptoms, anatomic and procedural details (including need for adjunctive procedures), secondary interventions after the initial intervention, and complications.

Technique

Techniques for delivery and deployment of the AFX device have been well described when treating AAAs.4 The AFX system consists of bifurcated unibodies with short, inte-grated iliac limbs. All AFX endografts feature a cobalt chromium stent frame with a multilayer expanded poly-tetraﬂuoroethylene material external to the stent. The implant is delivered using a dedicated 17-Fr introducer sheath on the ipsilateral side and an auxiliary 9-Fr intro-ducer sheath on the contralateral side. The unibody design avoids the need to cannulate the contralateral gate and allows for placement of theﬂow divider of the bifurcated component of the AFX system directly on the native aortic bifurcation.

Choice of anesthesia is physician- and patient-dependent, and may depend on the decision to perform adjunctive surgical interventions such as femoral endarterectomy. In this study, 41.4% of patients were treated percutaneously. Hemostasis was achieved using the Proglide (Abbott Vascular, Abbott Park, IL, USA). Intraoperative heparin dosing varied by center. However, generally, heparin was administered so as to maintain an activated clotting time> 250 seconds, when possible.

There are some important nuances in technique that are unique when using the AFX device to treat occlusive dis-ease. Most centers represented in this study used exclu-sively femoral access, although a brachial approach can be useful for crossing total occlusions. Generally, even for TASC B and C lesions, predilation of the iliac lesions is advisable, to minimize resistance as the AFX is delivered into the aorta and to allow for easier rotation of the device as needed. For TASC D lesions, the procedure demands that recanalization of the distal aorta occurs as close as possible to the aortic bifurcation. This will ensure that the AFX device sits directly on the aortic bifurcation and that it is not wedged on aortic debris. For these reasons, a key technique is to recanalize from one common iliac artery into the opposing common iliac artery, and then snare the wire to create a femoral-femoral wire (seeFig. 2). In this experience, the use of re-entry devices was not necessary, although this may be useful to help achieve through and through femoral-femoral wire access in some cases. The aortic occlusion is then recanalized from one of the common iliac arteries. The femoral-femoral wire is then replaced with a Kumpe cath-eter; the SurePass wire attached to the contralateral limb is then passed through the Kumpe. This step avoids trying to snare the wire within the distal aortic occlusion. In addition, it is not necessary to place the 17-Fr sheath up to the aortic bifurcation. The sheath can be placed within 5 cm of the aortic bifurcation, and the AFX can be delivered through the common iliac and aortic occlusion. It is imperative that the chronic occlusions are predilated, preferably with bilateral 6e8 mm 150 mm balloons. Without predilatation, there is severe resistance as the device is placed into position, and it is very difﬁcult to rotate the device to ensure the limbs seat appropriately. Most centers represented in this study recommend that all common iliac arteries be treated with

(3)

adjunctive stents. Adjunctive stenting of the distal aorta is less often done and depends on the degree of compression or calciﬁcation visualized intraoperatively on completion angiogram and/or intravascular ultrasound. Proximal aortic extension is rarely needed.

All patients are discharged on either aspirin or clopi-dogrel, if not already on one of these medications.

Outcomes

Routine practice at all centers included clinical follow-up at a minimum of approximately 1 month, 6 months, and annually postoperatively. Outcomes (primary, assisted pri-mary, and secondary patency rates) were assessed throughout the follow-up period, using computed tomo-graphic angiography and/or aortoiliac duplex in 76% of cases. Twenty-four percent of patients did not have any imaging studies on follow-up. Restenosis was defined as > 50% diameter reduction (peak systolic velocity [PSV] ratio 2.5e4 and/or PSV of > 180 cm/second) or 75% area reduction on axial imaging. Patency was defined according to suggested Society for Vascular Surgery standards.5 Pri-mary patency was defined as uninterrupted flow in the treated aortoiliac segment without occlusion or reinter-vention. Primary-assisted patency was defined as uninter-rupted flow in the treated segment, allowing for reintervention for hemodynamically significant lesions in order to prevent occlusion. Secondary patency was defined as patency of the treated segment, allowing for reinter-vention for occlusion.

In addition to these patency outcomes, we evaluated technical success, adverse events, and clinical improve-ment. Technical success was deﬁned as insertion of the endograft withﬂow into both iliac arteries at the conclusion of the procedure. Thirty-day major adverse cardiovascular and cerebrovascular events (MACCE) included any death, myocardial infarction (MI), or stroke. Thirty-day major adverse limb events (MALE) included any amputation above the ankle and any major vascular reintervention (throm-bectomy, thrombolysis, or surgical bypass) to either lower extremity.6 MI was determined based on postoperative biomarkers and/or changes on electrocardiogram. Other adverse events assessed included groin infections, hema-tomas, vessel ruptures, dissections, and thromboembolic events.

Ankle brachial indices (ABI) were measured at all follow-up time points. Clinical improvement was assessed using the Rutherford classiﬁcation (RC) of chronic limb ischemia.5 Given the possibility of patient follow-up at outside in-stitutions, all patients, including those who did not com-plete follow-up imaging to conﬁrm patency, were contacted to determine whether or not they had subsequent in-terventions to their grafts.

Data analysis

Statistical analysis was carried out using SPSS 19 (IBM, Armonk, NY, USA). Univariate analyses were performed using chi-square tests for discrete variables and Student t

tests for continuous variables. A p-value < .05 was considered signiﬁcant.

RESULTS

Study population

Ninety-one patients with AIOD presented to the 10 participating centers between 2012 and 2014 and under-went treatment with an Endologix AFX unibody device. Baseline patient characteristics are details in Table 1. Further risk stratification of patients was determined ac-cording to American Association of Anesthesiologists phys-ical status classification system, as well as by cardiac and pulmonary function (Table 2). Thirty-four patients (37%) of patients were deemed unfit for open surgery, owing to high-risk comorbidities.

According to RC at baseline, 67 patients (74%) presented with lifestyle-limiting intermittent claudication (RC1, n¼ 2 [2%]; RC2, n¼ 9 [10%]; RC3, n ¼ 56 [62%]), 16 (18%) with ischemic rest pain (RC4), and eight (9%) with tissue loss

Table 1. Baseline patient characteristics and details (n¼ 91).

Mean SD age (y) 67.9 9.7

Sex (n¼ 91) Female 35 (38) Male 56 (62) Ethnicity (n¼ 91) White 73 (80) African American 7 (8) Hispanic 3 (3) Other 8 (9) Smoking (n¼ 89) 81 (91) Hypertension (n¼ 89) 71 (80) CAD (n¼ 89) 46 (52) Prior PTCA (n¼ 89) 27 (30) CHF (n¼ 88) 4 (5) Diabetes (n¼ 91) Insulin-dependent 10 (11) Noninsulin medications 16 (18) Diet controlled 2 (2) COPD (n¼ 89) 27 (30) Hypercholesterolemia (n¼ 88) 79 (90) Stroke/TIA (n¼ 88) 14 (16)

Renal function (creatinine) (n¼ 89)

< 1.5 mg/dL 78 (88) 1.5e3.0 mg/dL 11 (12) Ambulatory status (n¼ 90) Ambulatory 84 (93) With assistance 4 (5) Wheelchair-bound 2 (2) ASA (n¼ 91) 75 (82)

P2Y12 antagonist (e.g., clopidogrel) (n¼ 91) 31 (34) ASA or P2Y12 antagonist (n¼ 91) 84 (92)

Statin use (n¼ 91) 70 (77)

Mean lowest resting ABI 0.57

Note. Data are n (%) unless otherwise indicated. CAD¼ coronary artery disease; PTCA ¼ percutaneous transluminal coronary angioplasty; CHF ¼ congestive heart failure; COPD ¼ chronic obstructive pulmonary disease; TIA ¼ transient ischemic attack; ASA¼ aspirin; ABI ¼ ankle brachial index.

(4)

(RC5, n ¼ 7 [8%]; RC6, n ¼ 1 [1%]). All patients with symptoms of intermittent claudication were initially managed conservatively with implementation of an exercise regimen (supervised when possible) and treatment with cilostazol when appropriate (with the exception of the Dutch Center which did not prescribe cilostazol routinely). Ultimately, the decision to intervene in patients with clau-dication was made only when symptoms were debilitating. There were 11 patients with only mild-to-moderate inter-mittent claudication (RC1 and RC2) who were younger in age and had persistent symptoms that were affecting life-style and underwent intervention.

The mean resting ABI in the most severely affected limb prior to intervention was 0.57. The majority of patients had TASC D lesions (n¼ 74 [81%]) and nearly one-quarter (18/ 74 [24%]) of these TASC D lesions were chronic aortoiliac occlusions (Fig. 1). No patient underwent intervention for acute obstruction of the aortoiliac segment. A representa-tive total aortic occlusion (TASC D1) is shown inFig. 2.

Interventions

Percutaneous access was used in 36 patients (41%) and general endotracheal anesthesia in the majority of cases (n¼ 90 [99%]). General anesthesia was used owing to the complexity of the procedure, including open surgical femoral artery access in 59% of patients and the need for concurrent common femoral endarterectomy. Fifty-six pa-tients (64%) required adjunctive procedures at the time of

initial intervention (Table 3). This included an open surgical procedure in 39% of patients (endarterectomy, n ¼ 30; bypass, n¼ 4) and an endovascular procedure in 59% of patients (iliac stent, n¼ 53 [40 balloon expandable, 13 self-expanding]; aortic stent, n¼ 10 [Palmaz]).

Other procedural details, including median ﬂuoroscopy and operative times, are listed inTable 4.

Outcomes

The mean duration of follow-up was 22.2 11.2 months. Technical success was achieved in 100% of patients. The 30-day MACCE rate was 2% (one death, one stroke); no MIs were observed. The one patient who died within 30 days had chronic aortic occlusive (TASC C) disease with signi ﬁ-cant thrombus in the aorta. The patient developed pelvic arterial thromboembolism and subsequent intestinal and gluteal necrosis, and died on postoperative day 28. This thromboembolic complication and ultimate mortality was a result of thrombus in the aorta at the time of intervention that was dislodged as a result of excessively aggressive wire or catheter manipulation, and represents a technical complication.

The overall procedure-related complication rate was 22% (20 patients) and included six groin infections (7%), four hematomas (4%), four vessel ruptures (4%), four dissections (4%), and three thromboembolic events (3%;Table 5). None of the complications appeared to be device-related. The majority of these complications (vessel rupture, dissections, and thromboembolic events) were treated with either endovascular therapy or supportive medical therapy (anti-biotics for infection). The four hematomas occurred in groins in which percutaneous access had been achieved, were managed expectantly, and did not require exploration/ evacuation. The four ruptures (one aortic and three com-mon iliac artery) were treated with covered stent exten-sions. The four iliac dissections were treated with bare metal stents. Of the three thromboembolic events, one patient embolized to the right hypogastric artery and was managed expectantly, with no sequelae, including buttock claudication, noted. One required open thrombectomy of a femoral occlusion. The ﬁnal aforementioned patient suf-fered massive thromboemboli to the pelvis and bilateral lower extremities. The peripheral emboli were treated with open thrombectomy of bilateral popliteal and tibial vessels. All cases of embolization had the entire disease segment covered but likely embolized during delivery of device and/ or as a result of wire or catheter manipulation.

Separate analysis of the 34 patients (37%) who were deemed unfit for open surgery revealed a complication rate of 26.5% (Table 5). However, the data indicate that there is no statistically significant increase in complication rates amongst those unfit for open repair when compared with thosefit for open repair (9/34 [26.5%] vs. 11/57 [19.3%]; p¼ .243).

RC and ABI were recorded at scheduled follow-up time points (Table 6). At 6 months, 73% of patients experienced improvement in RC of three or more categories (p¼ .048).

Table 2. Risk stratiﬁcation of patients according to pulmonary and cardiac status (n¼ 91). n (%) ASA class I 5 (6) II 15 (16) III 52 (57) IV 19 (21) Cardiac status

0: asymptomatic, normal ECG 46 (51)

1: asymptomatic, remote MI by history (> 6 mo), occult MI by ECG

34 (37) 2: stable angina, controlled ectopy or

asymptomatic arrhythmia, drug compensated CHF

6 (7)

3: unstable angina, symptomatic or poorly controlled ectopy/arrhythmia, poorly compensated CHF, MI< 6 mo

5 (6)

Pulmonary status

0: asymptomatic, normal 60 (66)

1: asymptomatic/mild dyspnea on exertion, PFTs 65e80% of predicted

23 (25)

2: between 1 and 3 7 (8)

3: vital capacity< 1.85 L, FEV1< 1.2 L (< 35% predicted), supplemental O2used medically, or pulmonary hypertension

1 (1)

Note. ASA ¼ American Society of Anesthesiologists; ECG ¼ electrocardiography; MI ¼ myocardial infarction; CHF¼ congestive heart failure; PFT ¼ pulmonary function test; FEV1¼ forced expiratory volume in 1 s.

(5)

Only one patient had worsening classiﬁcation during follow-up (RC5 to RC6). (Fig. 3A). Twenty-four patients (26%) had critical limb ischemia, 16 (18%) with ischemic rest pain (RC4) and eight (9%) with tissue loss (RC5, n¼ 7 [8%]; RC6, n¼ 1 [1%]). All but one of the eight patients with tissue loss (RC5 and RC6) went on to heal their wounds. This diabetic patient had a preoperative ABI of 0.25 that improved to 0.75 after the procedure. Despite this improvement inﬂow, toe ulcerations progressed and patient underwent tibial angioplasty at 5 months via a contralateral approach; the patient was noted to have extensive tarsal occlusive disease and eventually underwent below-knee amputation at 6 months. This late complication was not considered to be related to the index procedure but rather a progression of distal disease.

Mean ABI improved from 0.5 0.2 preoperatively to 0.9 0.2 postprocedure (p < .01), obtained on ﬁrst follow-up visit at 1e6 months (Table 6). In 42% of patients, the ABI resting increased by 80e100% compared with baseline (Fig. 3B). Nine patients (10%) required subsequent angio-plasty and/or stenting (Table 7). All but one of these were iliac limb interventions. One patient developed progression

of occlusive disease proximal to the aortic stent and required placement of aortic cuff extension, with excellent technical results and clinical improvement. There were two iliac limb occlusions or thromboses noted at 6 months and 1 year postoperatively, and the rate of 30-day MALE was 0%. Patency rates are depicted in Fig. 4. Among patients who returned for follow-up at 3 years, 100% had patent grafts. Freedom from secondary intervention was 88.9% at 1 year and 87.5% at 2 years (Fig. 5).

DISCUSSION

This multicenter study evaluating the unique application of the Endologix bifurcated stent graft in a series of patients with complex AOID demonstrates favorable results. This de-vice, designed and approved for use in the treatment of aneurysmal aortoiliac disease, has properties that may be more desirable than traditional endovascular therapy for the treatment of AIOD. The present experience evaluating this device for this nonaneurysmal application represents the largest series reported thus far. The use of the Endologix bifurcated stent-graft for complex AIOD was associated with

Figure 1. Trans-Atlantic Inter-Society Consensus classiﬁcation for all patients with aortoiliac occlusive disease treated with an AFX aorto-unibody device.

(6)

a high rate of technical success, favorable patency rates to 2 years postprocedure, and low rates of MACCE and MALE. Notably, the single mortality encountered resulted from thromboembolism in a patient with TASC C occlusive disease and was likely a result of excessive wireecatheter manipu-lation in the setting of underappreciated aortic thrombus. While atheroembolism is a recognized complication inherent to endovascular procedures, one should especially avoid excessive instrumentation in cases with increased thrombus burden. An open approach to such higher risk cases may be warranted so as to protect against distal embolization to the lower extremities by clamping the femoral and profunda femoral arteries for critical portions of the case.

A promising rate of graft patency was found, with> 88% of patients maintaining primary patency throughout the ﬁrst 2 years of follow-up. Although the overall complication rate was high (22%), the majority of these procedure-related complications were treated with endovascular therapy at the time of the procedure, and local complica-tions may decrease with increased experience and

improvements in device design. Moreover, incidence of groin-related complications in the present study is compa-rable with that found in open surgical repair (AFB) and may reflect the fact that a significant number of patients un-derwent open surgical access and thus were exposed to the same risks found in open surgery. In a recent study by Pejkic et al. of 120 consecutive patients undergoing elective AFB for occlusive disease, early postoperative complications (30 days) affected 35 (15%) groin wounds in 29 (25%) patients.7 Lymph fistulae/lymphoceles were observed in 15 (6%), infection in 11 (5%), and noninfectious wound dehiscence in nine (4%) of groin incisions.7 The authors of the current paper believe groin complications following AFX repair of AIOD may decrease significantly as more experience is gained with the technique and limit the use of open surgical access to those patients requiring adjunctive surgical pro-cedures (i.e., femoral endaraterectomy or profundoplasty). Indeed, although general anesthesia and open surgical ac-cess were used commonly (99% and 59%, respectively) in this initial experience, more recent practice has been

Figure 2. Trans-Atlantic Inter-Society Consensus D lesion (D1) shows total aortic occlusion (A) treated successfully with an aorto-unibody stent-graft (D). Adjunctive stenting is often required of iliac limbs. A favored technique is to recanalize from one common iliac artery into the opposing common iliac artery, and then snare the wire to create a femoral-femoral wire (B). The aortic occlusion is then recanalized from one of the common iliac arteries. The femoral-femoral wire is then replaced with a Kumpe catheter; the SurePass wire attached to the contralateral limb is then passed through the Kumpe. Kissing balloons are used to fully expand the limbs and AFX main body so as to recreate the aortic bifurcation (C).

(7)

characterized by complete percutaneous access, use of conscious sedation, and local anesthetic when possible.

The use of the Endologix AFX stent-graft for complex AIOD provides an attractive alternative for patients who

might otherwise require aortobifemoral bypass. Meta-analyses of aorto-bifemoral bypass have reported local and systemic morbidity rates of 8% and 12%, respectively, and mortality rates of 3e5%.8Mortality rates are higher in those over the age of 65 years and with comorbidities, including insulin-dependent diabetes, chronic kidney dis-ease, and severe chronic obstructive pulmonary disease.9As such, the AFX stent-graft may play a unique role in those patients who are at prohibitive risk for open surgery. Nearly 40% of the patients in the present study were deemed unﬁt for surgery. However, this higher-risk cohort did not suffer an increased complication rate when compared with those patients determined to be good candidates for open sur-gery. Despite this higher-risk patient population, in this study 30-day MACCE was only 2% and 30-day MALE was 0%, though with local morbidity in 20% of patients.

Several alternative endovascular interventions have been proposed to treat AIOD. The dominant treatment in this regard is iliac angioplasty and stenting, with“kissing” stents used for patients with disease involving the aortoiliac bifurcation. Kissing balloons were first described in 1985, and kissing stents in 1991.10e12 Estimates of long-term patency after kissing stent placement vary widely based on anatomy, indications, and patient risk factors, with some studies showing primary patency as high as 82% at 5 years and 68% at 10 years, and others with patency even after reintervention (primary-assisted patency) as low as 65% at 2 years.2,13e21 Compared with iliac stenting that does not involve the iliac bifurcation, stenting of the aortic bifurca-tion is more geometrically challenging to achieve without narrowing the flow lumen; the limb competition of two “crossed” kissing stents in a diseased distal aorta can lead to significant flow compromise.15,22 Aortic bifurcations with heavy calcification, aortic thrombus, and size or geometric mismatch between the native vessels and the final stent conformation present challenges to successful treatment with kissing stents.2,23,24

There is some evidence, including the randomized COB-EST trial, that covered stents may be less likely to restenose or occlude than bare metal stents, particularly in TASC C and D lesions.25,26 This may be because the graft fabric represents a direct barrier to tissue ingrowth from neo-intimal hyperplasia. This, along with concern about limb competition in a narrow distal aorta, led to the develop-ment of the covered endovascular reconstruction of aortic bifurcation (CERAB) technique, which uses three covered stents to cover the distal aorta and bilateral common iliac arteries.27 The CERAB technique has certain clear advan-tages: it is straightforward, allows for use of relatively low profile delivery systems (7-Fr and 9-Fr sheaths), and it is likely less costly than a stent-graft. Moreover, this technique has been shown to have excellent primary patency rates (87.3% at 1 year and 82.3% at 2 years).28 Optimal con fig-uration depends on precise placement of the limbs. In vitro data suggest that this technique can reduce radial mismatch compared with nitinol stenting or covered kissing stents in the aorta.29 However, radial mismatch is increased if the limbs are placed higher than the nonflared aortic stent. In

Table 3. Adjunctive procedures performed at time of treatment, using the AFX device.

Patients, n (%) Planned vs. unplanned n¼ 87 56 (64) Endovascular procedures 51 (59) 25 planned

Aortic stent 10 (11) Palmaz

(unplanned) Arteries treated Iliac stent 53 (61) EIA¼ 25 CIA¼ 56

Surgical procedures 34 (39) 23 planned CFA endarterectomy 30 (37) 25 planned

Bypassa 4 (4) 4 planned

Note. CFA¼ common femoral artery; EIA ¼ external iliac artery; CIA¼ common iliac artery.

a

Four adjunctive bypasses were performed (Patient 1: right and left external iliac limb to profunda femoris bypasses with 10-mm Dacron graft. Sartoriusﬂaps to bilateral groins, prophylactic, re-operative groins three times with prior thrombosed prosthetics in the groins. Patient 2: left femoral-popliteal; patient presented with rest pain and known superﬁcial femoral artery occlusion. Patient 3: endarterectomy and planned femoral-tibial bypass with vein graft; presented with rest pain and dry gangrene. Patient 4: planned femoral-popliteal bypass).

Table 4. Procedural details.

Technical success 91 (100)

Median days in hospital (Q1eQ3) 3.0 (2.0e4.0) Median (Q1eQ3) procedure time (min)

(n¼ 84) 167.0 (100.0e255.6) < 90 10 (12) 90e150 28 (33) 150e210 15 (18) > 210 31 (37)

Median (Q1eQ3) ﬂuoroscopy time (min) (n¼ 62) 21.8 (5.6e36.6) < 10 7 (11) 10e20 16 (26) 20e30 13 (21) > 30 26 (42)

Median (Q1eQ3) contrast volume (mL) (n¼ 67)

60.0 (30.0e93.0)

< 100 48 (72)

100e200 15 (22)

200e300 4 (6)

Median (Q1eQ3) blood loss (mL) 200.0 (87.5e300.0) Access type (n¼ 87)

Percutaneous 36 (41)

Anesthesia (n¼ 91)

GETA 90 (99)

Note. Data are n (%) unless otherwise indicated. GETA¼ general endotracheal anesthesia.

(8)

addition, while the CERAB technique does recreate the aortic bifurcation, future crossover interventions may be more difﬁcult. This is similar to most bifurcated aortic stent-grafts where the angle of engagement at theﬂow divider makes up-and-over access more challenging. Finally, dead space outside the proximal aortic stent can result in cases of aortic diameters > 20 mm. The aortic stent is then left protruding mal-opposed in a dilated aorta. Hence, this technique may not be indicated in the setting of even mild proximal aortic dilation.

Treatment of occlusive disease at the aortic bifurcation using a unibody stent-graft offers several potential advan-tages when compared with these endovascular alternatives. The unibody design treats the occlusive disease with a covered stent while preserving the aortic bifurcation. By using fewer stents in the narrow aorta, this technique has the added advantage of minimizing potential ﬂow distur-bances introduced by multiple covered stents. Use of a unibody stent-graft is also protective in cases of potential rupture, such as in heavily calciﬁed lesions. The use of a bifurcated stent-graft that encompasses the distal aorta and

common iliac arteries may also reduce the risk of athe-roembolic events compared with conventional endovas-cular therapy. Using this covered graft and positioning it proximal and distal to the limits of disease effectively traps atherosclerotic material that could potentially embolize during device deployment, reducing atheroembolic risk. Finally, unlike kissing stents (covered or uncovered), which can often protrude into the native aorta, disrupting ﬂow and functionally raising the aortic bifurcation, a unibody device sits on and preserves the native bifurcation. This makes future“up-and-over” interventions to the lower ex-tremities less technically challenging.

As with any new treatment alternative, this treatment strategy has limitations. Compared with the placement of kissing stents, or even the CERAB technique, the AFX device requires a larger proﬁle sheath (17-Fr ipsilateral sheath, 9-Fr contralateral). Placing a stent-graft also has a higher po-tential for coverage of collateral vessels, although in the present study this did not result in any adverse event. The procedure is more time consuming and requires a higher level of endovascular technical skill. We do not have

Table 5. Periprocedural complications following AFX for treatment of aortoiliac occlusive disease in all patients and a subset of high-risk patients deemed unﬁt for open surgery.

Complication n¼ 20/91 (22) Unﬁt for open surgery

n¼ 9/34 (26) Groin infection 6 (7) 2 (6) Respiratory failure 4 (4) 2 (6) Groin hematoma 4 (4) 3 (9) Rupture 4 (4) 3 (9) Hemodynamic instability 0 (0) 0 (0) Dissection 4 (4) 3 (9)

Thromboembolic eventa _Intestinal ₁a ₀

Spinal cord 3 (3) 0 1 (3) 0 Buttock 2a 1 Lower extremity 1 0 Femoral thrombosis 2 (2) 1 (3) Stroke 1 (1) 0 (0) Contrast nephropathy 0 (0) 0 (0) Pneumonia 0 (0) 0 (0) 30-d mortalitya 1 (1)a 0 (0)

Note. Data are n (%). a

Single mortality occurred as a result of a thromboemolic event resulting in diffuse pelvic ischemia and intestinal infarction.

Table 6. Ankle brachial index (ABI) and Rutherford classiﬁcation at follow-up.

Baseline Discharge 1 mo 6 mo 1 y 2 y 3 y ABI n 70 25 51 36 55 39 16 Mean SD 0.5 0.2 0.9 0.2 0.9 0.2 0.9 0.2 0.9 0.2 0.9 0.2 0.9 0.2 Rutherford classiﬁcation 0: Asymptomatic e 30/45 (67) 59/75 (79) 43/55 (78) 47/59 (80) 34/40 (85) 10/13 (77) 1: Mild claudication 2/91 (3) 11/45 (24) 14/75 (19) 9/55 (16) 7/59 (12) 5/40 (13) 3/13 (23) 2: Moderate claudication 9/91 (10) 1/45 (2) 2/75 (3) 2/55 (4) 3/59 (5) e e 3: Severe claudication 56/91 (62) 2/45 (4) e e 1/59 (2) e e

4: Ischemic rest pain 16/91 (18) e e e 1/59 (2) e e

5: Minor tissue loss 7/91 (8) 1/45 (2) e 1/55 (2) e e e

6: Major tissue loss 1/91 (1) e e e e e e

Patients with follow-up (n) 91 91 87 68 71 57 16

(9)

information regarding the radial force in the AFX stent relative to standard balloon or self-expandable stents approved for treatment of occlusive disease. However, in its current state, the cobalt chromium component of the graft frequently appears to lack sufﬁcient radial force, which

likely contributed to the high rate of adjunctive stenting required (59%). This represents an important limitation of the current AFX design, which should be improved upon in future generations if it is to be considered a standalone treatment for AIOD. Aflow model would be of great value to help better understand the behavior of the AFX stent in lesions of variable stiffness and thus help determine the necessary radial force required to prevent stent collapse in this setting. Finally, while cost-effectiveness cannot be evaluated without robust efficacy data and will fluctuate based upon local differences in costs, use of the AFX aorto-unibody stent-graft is likely to be more expensive than kissing stent placement.

There are also several limitations to this study itself. As a retrospective review without any direct comparison to other treatment strategies, this study should be regarded princi-pally as hypothesis generating. Furthermore, despite the fact that the vast majority of patients had extensive AIOD (TASC D), a selection bias is likely to present in this retrospective analysis. One of the key strengths of open AFB is its excellent long-term patency; while this study showed promising re-sults at a mean follow-up of 22.2 11.2months, longer-term patency data and randomized comparisons to open bypass would be useful.

In addition, while excellent patency rates were demon-strated at the 1- and 2-year follow-ups, rivaling most endovascular techniques and open surgery, the restenosis rates are less precise. Owing to the heterogeneity of prac-tice patterns across centers, the type of imaging obtained, as well as the surveillance schedule, was inconsistent, limiting the ability to comment accurately on rates of restenosis. Twenty-four percent of patients had no imaging on follow-up and assessment of need for reintervention and/or graft patency was based on clinical evaluation and/ or vascular laboratory study.

Figure 3. (A) Seventy-three percent of patients experienced improvement in Rutherford stage of3 or greater compared to baseline at 6-month follow-up. (B) Forty-ﬁve percent of patients showed an increase of ankle brachial index (ABI) of 80e100% compared with baseline.

Table 7. Secondary interventions (nine patients, 16 secondary interventions).

Total patients (n¼ 9) Target vessel Intervention 1 mo 6 mo 1 yr 2 yr 3 yr 4 yr

1 Aorta Extension cuffa 1

2 CFA Left PTA 1

Stent

Right PTA 1 1

Stent

5 EIA Left PTA 2

Stent 2

Right PTA 2 0

Stent 1

4 CIA Left PTA

Stent 1 3

Right PTA

Stent 0 1

Graft patency Nob 1 1

Yes 87 67 70 57 16 2

Note. CFA¼ common femoral artery; PTA ¼ percutaneous transluminal angioplasty; EIA ¼ external iliac artery; CIA ¼ common iliac artery. a

Patient required proximal aortic cuff to treat progression of proximal disease presenting with worsening claudication. b

One patient required open thrombectomy and stenting of occluded right limb at 6 mo, another required percutaneous stenting of occluded left CIA at 1 y.

(10)

Another limitation of the present study is the use of ABI as a metric for improvement inﬂow. While 42% of patients showed a> 80% improvement in ABI, 20% showed only < 40% improvement. This can be explained in a number of ways. Perhaps those patients who had minimal improve-ment in ABI had more signiﬁcant distal disease in addition to AIOD. Unfortunately, the study did not collect any

information regarding presence of distal occlusive disease. Alternatively, this subset of patients who had minimal improvement in ABI may have had more calciﬁed tibial vessels, making ABI less reliable. These patients may have had improvements in pulse volume recordings, which were, unfortunately, not captured for analysis.

Finally, in addition to the valuable end points of mortality, complications, and patency, any treatment for symptomatic peripheral arterial occlusive disease would beneﬁt from assessment of patient-reported outcomes such as quality of life; future studies of this technique should include such measures.

CONCLUSIONS

This is the largest study to examine the use of the AFX unibody stent-graft for the treatment of AIOD. It was found that the procedure had a high rate of technical success and low mortality rate, even in relatively ill patients with TASC C and D lesions. Patency rates throughout follow-up compare well with both open aortobifemoral bypass and other endovascular treatment options such as kissing stents and the CERAB technique, as quoted in the literature. Use of the AFX stent-graft appears to be a safe and effective endo-vascular treatment for complex aortoiliac occlusive pathol-ogy that might otherwise require aorto-bifemoral or extra-anatomic bypass.

Figure 4. KaplaneMeier survival analysis showing patency curves of patients treated with an AFX device for aortoiliac occlusive disease.

Figure 5. KaplaneMeier survival analysis showing freedom from secondary intervention curves for patients treated with an AFX device (88.9% at 1 year and 87.5% at 2 years). Ninety-ﬁve percent conﬁdence intervals are shown in the shaded area.

(11)

CONFLICT OF INTEREST

M.M.P.J.R. and Z.M.A. have served as consultants and advisory board members for Endologix. A.J.D. has served as a consultant for Endologix. H.E.G.Jr and T.S. have received honoraria from Endologix for teaching.

FUNDING

None.

REFERENCES

1 Upchurch GR, Dimick JB, Wainess RM, Eliason JL, Henke PK,

Cowan JA, et al. Diffusion of new technology in health care: the

case of aorto-iliac occlusive disease. Surgery 2004;136(4):812e8.

2 Sharafuddin MJ, Hoballah JJ, Kresowik TF, Sharp WJ,

Golzarian J, Sun S, et al. Long-term outcome following stent reconstruction of the aortic bifurcation and the role of

geo-metric determinants. Ann Vasc Surg 2008;22(3):346e57.

3 Klein AJ, Feldman DN, Aronow HD, Gray BH, Gupta K,

Gigliotti OS, et al. SCAI expert consensus statement for aorto-iliac arterial intervention appropriate use. Catheter Cardiovasc

Interv 2014;84(4):520e8.

4 Carpenter JP, Endologix Investigators. Multicenter trial of the

PowerLink bifurcated system for endovascular aortic aneurysm

repair. J Vasc Surg 2002;36(6):1129e37.

5 Rutherford RB, Baker JD, Ernst C, Johnston KW, Porter JM,

Ahn S, et al. Recommended standards for reports dealing with lower extremity ischemia: revised version. J Vasc Surg

1997;26(3):517e38.

6 Conte MS, Geraghty PJ, Bradbury AW, Hevelone ND, Lipsitz SR,

Moneta GL, et al. Suggested objective performance goals and clinical trial design for evaluating catheter-based treatment of

critical limb ischemia. J Vasc Surg 2009;50(6):1462e73. e1e3.

7 Pejkic S, Dragas M, Ilic N, Koncar I, Opacic D, Maksimovic Z,

et al. Incidence and relevance of groin incisional complications

after aortobifemoral bypass grafting. Ann Vasc Surg

2014;28(8):1832e9.

8 De Vries S, Hunink M. Results of aortic bifurcation grafts for

aortoiliac occlusive disease: a meta-analysis. J Vasc Surg

1997;26(4):558e69.

9 Hertzer NR, Bena JF, Karafa MT. A personal experience with

direct reconstruction and extra-anatomic bypass for

aortoilio-femoral occlusive disease. J Vasc Surg 2007;45(3):527e35.

10 Tegtmeyer CJ, Kellum CD, Kron IL, Mentzer RM. Percutaneous

transluminal angioplasty in the region of the aortic bifurcation. The two-balloon technique with results and long-term

follow-up study. Radiology 1985;157(3):661e5.

11 Küffer G, Spengel F, Steckmeier B. Percutaneous reconstruction

of the aortic bifurcation with Palmaz stents: case report.

Car-diovasc Intervent Radiol 1991;14(3):170e2.

12 Palmaz JC, Encarnacion CE, Garcia OJ, Schatz RA, Rivera FJ,

Laborde JC, et al. Aortic bifurcation stenosis: treatment with

intravascular stents. J Vasc Interv Radiol 1991;2(3):319e23.

13 Haulon S, Mounier-Véhier C, Gaxotte V, Koussa M, Lions C,

Haouari BA, et al. Percutaneous reconstruction of the aortoiliac

bifurcation with the “kissing stents” technique: long-term

follow-up in 106 patients. J Endovasc Ther 2002;9(3):363e8.

14 Greiner A, Dessl A, Klein-Weigel P, Neuhauser B, Perkmann R,

Waldenberger P, et al. Kissing stents for treatment of complex

aortoiliac disease. Eur J Vasc Endovasc Surg 2003;26(2):161e5.

15 Yilmaz S, Sindel T, Golbasi I, Turkay C, Mete A, Lüleci E.

Aor-toiliac kissing stents: long-term results and analysis of risk

factors affecting patency. J Endovasc Ther 2006;13(3):291e301.

16 Houston JG, Bhat R, Ross R, Stonebridge PA. Long-term results

after placement of aortic bifurcation self-expanding stents: 10 year mortality, stent restenosis, and distal disease progression.

Cardiovasc Intervent Radiol 2007;30(1):42e7.

17 Björses K, Ivancev K, Riva L, Manjer J, Uher P, Resch T. Kissing

stents in the aortic bifurcationea valid reconstruction for

aorto-iliac occlusive disease. Eur J Vasc Endovasc Surg 2008;36:424e31.

18 Abello N, Kretz B, Picquet J, Magnan P-E, Hassen-Khodja R,

Chevalier J, et al. Long-term results of stenting of the aortic

bifurcation. Ann Vasc Surg 2012;26(4):521e6.

19 Hinnen JW, Konickx MA, Meerwaldt R, Kolkert JLP, van der

Palen J, Huisman AB, et al. Long term results of kissing stents in

the aortic bifurcation. Acta Chir Belg 2015;115(3):191e7.

20 Suh Y, Ko Y-G, Shin D-H, Kim J-S, Kim B-K, Choi D, et al.

Out-comes of the single-stent versus kissing-stents technique in asymmetric complex aortoiliac bifurcation lesions. J Vasc Surg

2015;62(1):68e74.

21 Pulli R, Dorigo W, Fargion A, Angiletta D, Azas L, Pratesi G, et al.

Early and midterm results of kissing stent technique in the management of aortoiliac obstructive disease. Ann Vasc Surg

2015;29(3):543e50.

22 Greiner A, Mühlthaler H, Neuhauser B, Waldenberger P,

Dessl A, Schocke MFH, et al. Does stent overlap inﬂuence the patency rate of aortoiliac kissing stents? J Endovasc Ther

2005;12(6):696e703.

23 Sharafuddin MJ, Hoballah JJ, Kresowik TF, Sharp WJ. Kissing

stent reconstruction of the aortoiliac bifurcation. Perspect Vasc

Surg Endovasc Ther 2008;20(1):50e60.

24 De Donato G, Bosiers M, Setacci F, Deloose K, Galzerano G,

Verbist J, et al. 24-month data from the BRAVISSIMO: a large-scale prospective registry on iliac stenting for TASC A & B and

TASC C & D lesions. Ann Vasc Surg 2015;29(4):738e50.

25 Mwipatayi BP, Thomas S, Wong J, Temple SEL, Vijayan V,

Jackson M, et al. A comparison of covered vs bare expandable stents for the treatment of aortoiliac occlusive disease. J Vasc

Surg 2011;54:1561e70.

26 Sabri SS, Choudhri A, Orgera G, Arslan B, Turba UC, Harthun NL,

et al. Outcomes of covered kissing stent placement compared with bare metal stent placement in the treatment of athero-sclerotic occlusive disease at the aortic bifurcation. J Vasc

Interv Radiol 2010;21(7):995e1003.

27 Goverde PC, Grimme FA, Verbruggen PJ, Reijnen MM.

Covered endovascular reconstruction of aortic bifurcation (CERAB) technique: a new approach in treating extensive aortoiliac occlusive disease. J Cardiovasc Surg (Torino)

2013;54(3):383e7.

28 Grimme FA, Goverde PC, Verbruggen PJ, Zeebregts CJ,

Reijnen MM. Editor’s Choicedﬁrst results of the covered endovascular reconstruction of the aortic bifurcation (CERAB) technique for aortoiliac occlusive disease. Eur J Vasc Endovasc

Surg 2015;50(5):638e47.

29 Groot Jebbink E, Grimme FAB, Goverde PCJM, van Oostayen JA,

Slump CH, Reijnen MMPJ. Geometrical consequences of kissing stents and the Covered Endovascular Reconstruction of the Aortic Bifurcation conﬁguration in an in vitro model for endovascular reconstruction of aortic bifurcation. J Vasc Surg