Author + information
- Received February 6, 2017
- Revision received April 28, 2017
- Accepted May 4, 2017
- Published online August 21, 2017.
- Hans Krankenberg, MDa,∗ (, )
- Thomas Zeller, MDb,
- Maja Ingwersen, DVMa,
- Josefin Schmalstieg, MDc,
- Hans Martin Gissler, MDd,
- Sigrid Nikol, MDe,
- Iris Baumgartner, MDf,
- Nicolas Diehm, MDg,
- Estell Nickling, MDa,
- Stefan Müller-Hülsbeck, MDh,
- Rainer Schmiedel, MDi,
- Giovanni Torsello, MDj,
- Willibald Hochholzer, MDk,
- Christian Stelzner, MDl,
- Klaus Brechtel, MDm,
- Wulf Ito, MDn,
- Ralph Kickuth, MDo,
- Erwin Blessing, MDp,
- Marcus Thieme, MDq,
- Jaroslaw Nakonieczny, MDr,
- Thomas Nolte, MDs,
- Ragnar Gareis, MDt,
- Harald Boden, MDu and
- Sebastian Sixt, MDf
- aDepartment of Angiology, Asklepios Klinikum Harburg, Hamburg, Germany
- bDepartment of Angiology, University Heart Center Freiburg–Bad Krozingen, Bad Krozingen, Germany
- cDepartment of Anesthesiology, Intensive Care and Pain Medicine, BG Klinikum Unfallkrankenhaus Berlin, Berlin, Germany
- dDepartment of Radiology, Kantonsspital Aarau, Aarau, Switzerland
- eDepartment of Angiology, Asklepios Klinik St. Georg, Hamburg, Germany
- fDepartment of Angiology, Inselspital, Universitätsspital Bern, Bern, Switzerland
- gZentrum für Gefäßmedizin Mittelland, Aarau, Switzerland
- hDepartment of Radiology, Diakonissenkrankenhaus Flensburg, Flensburg, Germany
- iPraxis für Interventionelle Angiology, Kaiserslautern, Germany
- jDepartment of Vascular Surgery, St. Franziskus-Hospital Münster, Münster, Germany
- kDepartment of Cardiology and Angiology II, University Heart Center Freiburg - Bad Krozingen, Bad Krozingen, Germany
- lDepartment of Angiology, Städtisches Klinikum Dresden-Friedrichstadt, Dresden, Germany
- mJoint Practice for Radiology, Berlin, Germany
- nCardiovascular Center Oberallgäu-Kempten, Hospital Immenstadt, Immenstadt, Germany
- oDepartment of Diagnostic and Interventional Radiology, Universitätsklinikum Würzburg, Würzburg, Germany
- pDepartment of Internal Medicine, Klinikum SRH Karlsbad, Karlsbad, Germany
- qDepartment of Angiology, Cardiology and Diabetology, MEDINOS Klinik Sonneberg, Sonneberg, Germany
- rDepartment of Vascular Surgery, GPR Klinikum Rüsselsheim, Rüsselsheim am Main, Germany
- sDepartment of Vascular Surgery, Herz- und Gefäßzentrum Bad Bevensen, Bad Bevensen, Germany
- tCardiologicum Stuttgart, Stuttgart, Germany
- uDepartment of Internal Medicine, Ilm-Kreis-Klinikum, Ilmenau, Germany
- ↵∗Address for correspondence:
Dr. Hans Krankenberg, Department of Angiology, Asklepios Klinikum Harburg, Eißendorfer Pferdeweg 52, 21075 Hamburg, Germany.
Objectives Atherosclerosis of iliac arteries is widespread. As inflow vessels, they are of great clinical significance and increasingly being treated by endovascular means. Most commonly, stents are implanted.
Background So far, due to a lack of comparative data, no guideline recommendations on the preferable stent type, balloon-expandable stent (BE) or self-expanding stent (SE), have been issued.
Methods In this randomized, multicenter study, patients with moderate to severe claudication from common or external iliac artery occlusive disease were assigned 1:1 to either BE or SE. The primary endpoint was binary restenosis at 12 months as determined by duplex ultrasound. Key secondary endpoints were walking impairment, freedom from target lesion revascularization (TLR), hemodynamic success, target limb amputation, and all-cause death.
Results Six hundred sixty patients with 660 lesions were enrolled at 18 German and Swiss sites over a period of 34 months; 24.8% of the patients had diabetes and 57.4% were current smokers. The common iliac artery was affected in 58.9%. One hundred nine (16.5%) lesions were totally occluded and 25.6% heavily calcified. Twelve-month incidence of restenosis was 6.1% after SE implantation and 14.9% after BE implantation (p = 0.006). Kaplan-Meier estimate of freedom from TLR was 97.2% and 93.6%, respectively (p = 0.042). There was no between-group difference in walking impairment, hemodynamic success, amputation rate, all-cause death, or periprocedural complications.
Conclusions The treatment of iliac artery occlusive disease with SE as compared with BE resulted in a lower 12-month restenosis rate and a significantly reduced TLR rate. No safety concerns arose in both groups. (Iliac, Common and External [ICE] Artery Stent Trial; NCT01305174)
- balloon-expandable stent(s)
- common iliac artery
- external iliac artery
- peripheral artery disease
- randomized trial
- self-expanding stent(s)
Atherosclerosis is common in iliac arteries. Autopsy studies have revealed severe atherosclerosis in 15% of men and 5% of women, increasing with age (1,2). Iliac arteries, in their role as inflow vessels, affect the entire downstream blood flow and are therefore of great clinical relevance.
The Trans-Atlantic Inter-Society Consensus (TASC II) committee for the management of peripheral artery disease recommended the endovascular approach over surgery as the preferred treatment of TASC A and B aortoiliac lesions (3,4). Over time, it became apparent that iliac artery stenting compared favorably with percutaneous transluminal balloon angioplasty (5). However, no recommendation on the type of stent has been made until now.
Whereas self-expanding stents (SE) are of high elasticity but apply low radial outward force, balloon-expandable stents (BE) are rigid but support high radial outward force and allow to be placed with greater precision. With respect to iliac artery wall pulsatility, stents should be able to not only resist but also adapt to movements. This raised the question of whether both types of stents are comparably suitable, efficacious, and safe for iliac arteries.
Reekers et al. (6) showed a 12-month primary patency of 89% after BE implantation, on the other hand, SE resulted in a 12-month primary patency of 91.1% to 98% (7,8). These results suggested superiority of SE over BE. However, no data from randomized studies were available to date. Therefore, we initiated the randomized ICE (Iliac Artery Stents for Common or External Iliac Artery Occlusive Disease) trial to assess acute and midterm effectiveness and safety of primary stenting with SE compared with BE.
The study was a prospective, multicenter, block-randomized, nonblinded, investigator initiated trial. Consecutive patients with common or external iliac artery stenosis or occlusion were allocated 1:1 to either SE or BE. Enrollment was scheduled over a period of 2 years. The ICE trial was approved by the Freiburg Ethics Commission International. All patients provided written informed consent. The study complies with the Declaration of Helsinki and is registered on ClinicalTrials.gov (NCT01305174).
Patients were eligible for inclusion if they had peripheral artery disease of Rutherford stage 1 to 4 due to a single significant (≥70% diameter stenosis or occlusion by duplex ultrasound [DUS]) common or external iliac artery lesion of 10 to 200 mm in length, not extending into the aorta or the common femoral artery. Sequential lesions at a distance of <10 mm counted as a single lesion. Ipsilateral superficial or deep femoral artery had to be patent. However, parallel treatment of the ipsilateral superficial femoral artery was allowed. Key exclusion criteria were dialysis dependent end-stage renal disease and treatment with oral anticoagulants other than antiplatelet agents.
Access to the target lesion was achieved by either the retrograde ipsilateral or the contralateral crossover approach from the femoral artery with a 6-F introducer sheath. If neither of these was feasible, transbrachial access was permitted. After placement of the sheath and successful passage of a 0.018- or 0.035-inch guidewire across the target lesion, patients were randomly assigned to undergo SE or BE implantation. Pre-dilation by standard percutaneous transluminal balloon angioplasty techniques was left to physician’s discretion. If overlap of sequential stents was necessary, the amount of overlap had to be kept at a minimum. Stented length had to extend the target lesion by at least 3 mm proximally and distally.
The BE stent (Visi-Pro, ev3 Endovascular, Inc., Plymouth, Minnesota) was pre-mounted on a balloon catheter and expanded and deployed by inflation of the balloon. The nominal stent diameter had to match the reference vessel diameter (RVD) of the target lesion. Post-dilation was permitted.
The nominal diameter of the SE stent (Protege, ev3 Endovascular, Inc.) had to exceed the RVD at least by 1 mm. Post-dilation was mandatory. The inflated post-dilation balloon should approximate the RVD.
Patients not on chronic antiplatelet therapy were pre-medicated with acetylsalicylic acid (100 mg/day) and clopidogrel (75 mg/day) for at least 10 days. Patients not on this regimen were administered an intravenous bolus of 500 mg acetylsalicylic acid and an oral loading dose of 600 mg clopidogrel before or immediately after the procedure. After sheath placement, a bolus of 5,000 U of heparin was administered. All patients had to be maintained on clopidogrel (75 mg/day) for at least 1 month and on acetylsalicylic acid (100 mg/day) indefinitely after the procedure.
Follow-up clinical evaluations were performed at 6 and 12 months. Evaluation included a medical examination and an interview on cardiovascular events as well as the assessment of Rutherford stage, ankle-brachial pressure index at rest (at 6-month follow-up), and relative claudication distance (RCD) by treadmill test at 2 mph on a 12% incline, not exceeding a walking distance of 1,000 m.
Color DUS examination of the target vessel was mandatory within 1 week before the procedure and at the 6- and 12-month follow-up visits. The ratio of intrastenotic and pre-stenotic peak systolic velocity (proximal peak systolic velocity ratio [PSVR]) corresponds to the degree of diameter stenosis. A PSVR ≥3.4 indicated ≥70% diameter stenosis.
The primary study endpoint was the cumulative incidence of binary restenosis at 12 months. Clinically relevant restenosis was defined as PSVR ≥3.4 assessed by DUS.
Secondary procedural effectiveness endpoints were acute technical success and Kaplan-Meier estimate of primary patency at 12 months. Acute technical success was defined by access and deployment of the stent with <30% residual diameter stenosis on the procedural angiogram without flow-limiting dissection, using the assigned stent only. In cases of doubt about the hemodynamic significance of the residual stenosis, it had to be confirmed by a translesional pressure gradient. A gradient of ≥20 mm Hg was defined as significant. Kaplan-Meier estimate of primary patency was defined by the absence of binary restenosis and target lesion revascularization (TLR). Secondary clinical and hemodynamic effectiveness endpoints were Kaplan-Meier estimate of freedom from TLR (based on clinical presentation and DUS findings), walking impairment (percentage of patients in whom Rutherford stage improved by ≥1 without the need of TLR), RCD at 6 and 12 months, and hemodynamic success (share of patients in whom the ankle-brachial index improved by ≥0.15 at 6 months without the need of TLR). Secondary safety endpoints were periprocedural complications, binary restenosis at 6 months, TLR (not including procedural bailout), target vessel revascularization, all-cause death, target limb amputation, stroke, and the composite endpoint of major adverse vascular events (MAVE) (target limb amputation, cardiovascular death, and stroke) at 6 and 12 months.
Based on the assumption of a cumulative incidence of 12-month binary restenosis of 11% after BE implantation (6) and of 2% to 5% after SE implantation (7,8) and based on an anticipated DUS dropout rate of 30%, we calculated, that the assignment of 339 patients per group would provide a power of 80% to detect significance with a 2-sided alpha level of 0.50. Analyses were performed in the intention-to-treat population, which consisted of all patients who underwent randomization, regardless of the treatment received. Patients who were lost to follow-up were not included in the denominator for calculations of binary endpoints.
Continuous variables are reported as mean ± SD and categorical variables are reported as count and percent. Differences between continuous variables were assessed with the Student t test or the Mann-Whitney U test. Fisher exact test was used for comparison of categorical variables. Kaplan-Meier analysis was performed to estimate primary patency and freedom from TLR with the Wilcoxon test for equality of functions. Results are presented as parameter estimates and their corresponding 95% confidence interval (CI). Logistic regression was used to assess predictors of binary restenosis. Established candidate variables including the treatment arm were pre-screened based on univariable analysis with a p value cutoff of 0.25 based on the Wald test from logistic regression. Subsequently, variable selection for multivariable modeling was continued by stepwise backward regression with an entry and removal threshold p value of 0.10. To assess consistency of treatment effect among patient and lesion characteristics, a post hoc subgroup analysis was performed with formal interaction testing (8 patients and 5 lesion characteristics tested). The p value threshold for interaction was adjusted for multiple analyses (p < 0.004). Results are presented as odds ratios (ORs) and their corresponding 95% CIs. A 2-sided value of p < 0.05 indicated statistical significance. Statistical analyses were performed with XLSTAT software, version 2015.6.01.24026 (Addinsoft SARL, New York, New York).
Study population and treatment
From August 2010 to June 2013, 660 patients were enrolled at 18 German and Swiss centers and randomly assigned to undergo iliac artery angioplasty with either a SE or a BE stent (Figure 1). Due to delayed recruitment, the number of included patients fell slightly below the targeted sample size.
The 2 treatment groups were well matched at baseline with respect to coexisting conditions, morbidities, and lesion characteristics. Mean lesion length was 41.1 ± 32.9 mm in the SE group and 33.7 ± 26.5 mm in the BE group (p = 0.005). The treated length was longer with SE (56.7 ± 35.1 mm) than with BE (47.6 ± 28.5 mm; p < 0.0001). Vice versa, reference vessel and stent diameter were larger in the SE group than in the BE group (RVD: 8.3 mm vs. 7.7 mm; p < 0.0001; stent diameter: 9.1 mm vs. 7.9 mm; p < 0.0001). There was no significant difference in the rate of acute technical success (Tables 1 and 2).
The cumulative incidence of binary restenosis at 12 months, assessed by DUS, was 6.1% in the SE group and 14.9% in the BE group (between-group difference 8.8 percentage points; 95% CI: −15.1 to −2.5; p = 0.006) (Table 3).
Twelve-month Kaplan-Meier estimates of primary patency were 94.5% in the SE group and 87.0% in the BE group (p = 0.026) (Figure 2A). Multivariable logistic regression identified BE treatment (OR: 2.7; 95% CI: 1.3 to 5.7) and in-stent restenosis (OR: 5.7; 95% CI: 1.5 to 21.4) as predictors of binary restenosis. Patient-related risk factors such as diabetes, severe claudication, or ischemic pain at rest as well as lesion-related factors such as lesion location, RVD, lesion length, or calcification were not associated with binary restenosis (Figure 3). Neither clinically relevant patient characteristics nor key lesion characteristics proved to interact with the treatment effect (Figures 4 and 5).
Clinical, hemodynamic, and safety outcomes
In accordance with the incidence of binary restenosis, the 12-month incidence of TLR was significantly lower after SE implantation than after BE implantation (3.0% vs. 6.9%; p = 0.041). Twelve-month Kaplan-Meier estimates of freedom from TLR were 97.2% in the SE group and 93.6% in the BE group (p = 0.042) (Figure 2B). Not including patients who required TLR, there was no significant between-group difference in walking impairment or in the RCD at 6 and 12 months.
Ankle-brachial index at 6 months was significantly higher after SE implantation than after BE implantation (0.88 ± 0.21 vs. 0.84 ± 0.22; p = 0.022). It improved in 81.1% of the SE patients and in 77.3% of the BE patients (p = 0.264) by 0.32 ± 0.20 and 0.28 ± 0.17 (p < 0.0001), respectively.
In the SE group, 2 patients died. One of them due to myocardial infarction at 12 months and the other for unknown reasons at 3 months. In the BE group 8 patients died: 2 of them from myocardial infarction at 5 months or cardiogenic shock at 6 months, and 4 of them from cerebral aneurysm, pneumonia, sepsis, or multiple myeloma within 12 months. Two patients from the BE group died for unknown reasons. However, 12-month cumulative incidence of index limb amputations (0% SE vs. 0.4% BE; p = 0.476) and stroke (0.4% vs. 1.2%; p = 0.351) did not differ significantly between the groups. The same applied to all-cause death (0.7% SE vs. 3.3% BE; p = 0.053) and the composite endpoint of MAVE (0.7% SE vs. 2.9% BE; p = 0.092) (Table 3).
Periprocedural complications occurred in 6.3% (20 of 340 patients) after SE implantation and in 3.8% (12 of 320 patients) after BE implantation (p = 0.211). Most of these complications were minor bleeding events at the access site or occurred due to perforation or dissection. In 4 SE patients, a distal embolization occurred and could be resolved during the index procedure (Table 4). Three of the patients with embolization were treated due to an occlusion and 1 of them was treated due to an in-stent restenosis. Thus, occlusions or in-stent restenosis were more frequently affected by distal embolization than restenosis (3.0% vs. 0.0%; p = 0.002).
Endovascular intervention with primary stenting is the preferred treatment for aortoiliac TASC A, B, and C lesions. However, the choice between SE and BE is left to the interventionalist’s preference.
This randomized trial demonstrated superiority of SE over BE for iliac artery occlusive disease regarding 12-month binary restenosis rate. This translated into a significant clinical benefit in terms of a lower 12-month TLR rate after SE stenting. Despite a considerably low DUS completion rate, but due to a clear difference between restenosis rates, the statistical power for the primary endpoint analysis was sufficient.
Mechanical properties of the stents might have been causally related to the primary outcome. Less radial forces from the SE possibly helped to prevent circumferential stress and a subsequently enhanced neointimal proliferation. The higher elasticity might have preserved the arterial distensibility to a greater extent than did the more rigid BE. With respect to the procedure, post-dilation was mandatory for SE, reflecting common practice to complete stent expansion. Therefore, post-dilation was part of the assessed SE treatment just as the expansion by balloon for BE deployment.
Results from previously conducted studies on the efficacy of SE or BE are inconsistent. A subgroup analysis from data of a retrospective registry with more than 2,000 patients, three-quarters of whom were treated with SE, showed no difference in the 12-month primary patency between SE and BE (9). On the other hand, results on efficacy from 5 prospective studies on SE (7,8,10–12) and 3 prospective studies on BE (6,12,13) support the assumption of a difference. The primary outcome of binary restenosis and Kaplan-Meier estimates of primary patency of the present trial confirm the superiority of SE over BE. This was also reflected in a significantly lower need for TLR after SE implantation. However, not including patients who received TLR, neither the present trial nor previous studies could demonstrate a difference in hemodynamic success or walking impairment.
Independent of the type of stent, in-stent restenosis was significantly associated with recurrent restenosis. However, this condition is challenging in general and worth investigating separately in future trials. Recently published results from 2 small-scale, single-center experiences with drug-coated balloons are promising (14,15). The significantly larger RVD of SE group lesions could initially be interpreted as an advantage. However, logistic regression and subgroup interaction test showed that RVD was unrelated to outcome.
The post hoc subgroup analysis should be considered as exploratory. However, results demonstrated consistency of the treatment effect across patient and lesion subgroups of clinical importance. Due to multiplicity, the interaction p = 0.043 for calcification could have been due to chance. However, for heavily calcified lesions, benefit from SE remained doubtful. A possible explanation would be that the higher radial outward force of BE compensated its potential disadvantages in these lesions (16).
The slightly increased rate of pre-dilations in the SE group is likely related to the somewhat higher amount of occlusions, in-stent restenosis, and heavy calcifications in the SE lesion group. Therefore, pre-dilation is directly related to the properties of the lesion and does not independently affect the outcome.
In two-thirds of all patients with restenosis, TLR was deemed clinically necessary. This resulted in a significantly higher TRL rate after BE stenting within 12 months. However, the overall cumulative incidence of TLR was in line with results from published data (4% to 6.2%) (9,11,17).
Aside from TLR, no safety concerns arose in both study groups. Periprocedural complications, target limb amputations, and all-cause deaths were in the same range as expected from previous studies (2.6% to 9.0%, 0.0% to 2.0%, and 1.8% to 10.0%, respectively) (9,13,18). Occlusions and in-stent restenosis were linked to an increased rate of distal embolization. No such association could be detected for the treatment, despite an imbalance in favor of BE. The incidence of MAVE did not differ significantly between groups. Overall in this study endovascular stent-supported treatment of iliac arteries proved to be safe.
DUS completion for primary outcome analysis was lower than anticipated. Nevertheless, this study was statistically powered to detect a difference in the primary endpoint of binary restenosis at 12 months. However, it was of limited power to detect differences in rare but serious safety events, predictive values of risk factors, and interaction of subgroups. Moreover, none of the parties were blinded and no independent core laboratory was involved. Thus, interpretation of angiographic findings and DUS was left to the investigators. Overall, the primary outcome was at low risk of performance and detection bias.
The treatment of iliac artery occlusive disease with SE led to a decreased incidence of restenosis at 12 months and thus provided superior primary patency than the treatment with BE. No safety concerns arose in both groups. From this, a treatment recommendation in favor of SE rather than BE in iliac artery occlusive disease can be derived.
WHAT IS KNOWN? Endovascular stent-supported treatment of iliac arteries is effective and safe. It compared favorably with percutaneous balloon angioplasty alone.
WHAT IS NEW? This is the first randomized comparison of SE stents with BE stents for iliac artery occlusive disease. SE were found to be superior to BE regarding the 12-month binary restenosis rate and, in consequence, resulted in a lower TLR rate. The incidence of target limb amputation, cardiovascular death, or stroke did not differ between the groups. Therefore, a treatment recommendation in favor of SE rather than BE can be derived.
WHAT IS NEXT? Advanced devices such as drug-coated balloons, drug-coated stents, covered stents, or stents with advanced designs might be examined and compared for suitability, safety, and efficacy in iliac artery occlusive disease.
The authors gratefully acknowledge Dr. Thilo Tübler for major contribution in conception, design, and supervision of this study.
The work was funded from a grant by ev3 Europe SAS (Paris, France). Dr. Gissler has received modest honoraria from Covidien and Abbott. Dr. Baumgartner has received research and educational grants from Abbott Vascular, Cook, Optimed, Terumo, Promedics, Amgen, Boston Scientific, Bayer, AstraZeneca, and Sanofi. Dr. Diehm has received a significant research grant from Biotronik; a modest research grant from Medtronic; significant expert witness fees from Biotronik; and modest expert witness fees from Medtronic and Genae; and has owned significant ownership interest in MediCut/BBraun. Dr. Hochholzer has received consulting and lecture fees from AstraZeneca, Boehringer Ingelheim, Daiichi Sankyo, and The Medicines Company. Dr. Kickuth has received modest honoraria from Abbott Vascular. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Krankenberg and Zeller contributed equally to this work.
- Abbreviations and Acronyms
- balloon-expandable stent(s)
- confidence interval
- duplex ultrasound
- major adverse vascular event(s)
- odds ratio
- proximal peak systolic velocity ratio
- relative claudication distance
- reference vessel diameter
- self-expanding stent(s)
- Trans-Atlantic Inter-Society Consensus
- target lesion revascularization
- Received February 6, 2017.
- Revision received April 28, 2017.
- Accepted May 4, 2017.
- 2017 American College of Cardiology Foundation
- Sternby N.H.
- Mitchell J.,
- Schwartz C.
- Jaff M.R.,
- White C.J.,
- Hiatt W.R.,
- et al.
- Burket M.W.,
- Brodmann M.,
- Metzger C.,
- Tan K.,
- Jaff M.R.
- Troisi N.,
- Ercolini L.,
- Peretti E.,
- et al.
- Ichihashi S.,
- Higashiura W.,
- Itoh H.,
- Sakaguchi S.,
- Kichikawa K.