Randomized Trial of the SMART Stent versus Balloon Angioplasty in Long Superficial Femoral Artery Lesions: The SUPER Study
- First Online:
- Cite this article as:
- Chalmers, N., Walker, P.T., Belli, A. et al. Cardiovasc Intervent Radiol (2013) 36: 353. doi:10.1007/s00270-012-0492-z
- 1.9k Views
To determine whether primary stenting reduces the rate of restenosis compared with balloon angioplasty alone in the endovascular treatment of long superficial femoral artery lesions; and to assess the effect of treatment on quality of life.
A total of 150 patients with superior femoral artery occlusion or severe stenosis of 5–22 cm length from 17 UK centers were randomized to either primary stenting with the SMART stent or balloon angioplasty (i.e., percutaneous transluminal angioplasty, PTA). Bailout stent placement was permitted in case of inadequate result from PTA. The primary end point was restenosis measured by duplex ultrasound at 1 year. Quality-of-life assessments were performed by the EuroQol (EQ)-5D questionnaire.
Mean lesion length was 123.0 mm in the stent group and 116.8 mm in the PTA group. A total of 140 (93.3 %) of 150 had total occlusions. At 12 months’ follow-up, restenosis measured by Duplex ultrasound was not significantly different between the stent and PTA groups by intention-to-treat or as-treated analyses: 47.2 versus 43.5 % (p = 0.84) and 40.8 versus 46.7 % (p = 0.68), respectively. There were fewer target lesion revascularizations in patients randomized to stenting, but this did not reach statistical significance (12.5 vs. 20.8 %, p = 0.26). There was no difference in the rate of amputation. Patients in both groups reported improved quality of life.
Primary stenting of long lesions in predominantly occluded superficial femoral arteries does not reduce the rate of binary restenosis compared with balloon angioplasty and bailout stenting. Both treatment strategies conferred a meaningful and sustained improvement to the quality of life of patients with severe superficial femoral artery disease.
KeywordsBalloon angioplastyLong lesionsRandomized trialSMART stentSuperficial femoral artery
When this study was initiated, the SIROCCO [1, 2] trials and other studies [3, 4] had demonstrated that the Cordis SMART stent was associated with high patency rates in the superior femoral artery (SFA), but at that time, there were no controlled trial data demonstrating any advantage of Nitinol stent placement in the SFA compared with balloon angioplasty (i.e., percutaneous transluminal angioplasty, PTA) alone. Our hypothesis was that stents would deliver improved patency compared with PTA in patients with long segments of disease and total occlusions. We therefore undertook a multicenter randomized, controlled trial to test this. We also sought to assess whether stent placement was more effective than angioplasty in improving quality of life.
Materials and Methods
The study device was the Cordis SMART Nitinol stent, a self-expandable stent indicated for use in a vessel with a diameter 1–2 mm smaller than the nominal stent diameter. Stents of up to 120 cm in length were supplied in 6.0 and 7.0 mm unconstrained diameters, constrained in a 6F delivery system that accommodated a 0.035 in. guide wire.
Patient Selection and Randomization
Inclusion and exclusion criteria
Age ≥30 years
Symptomatic leg ischemia by Rutherford/Fontaine classification (category 1, 2, 3, 4, or 5); duration of intermittent claudication (category 1–3) should be at least 6 months
One superficial femoral artery de novo or restenotic lesion (≥70 % stenosis or occlusions), with a lesion length of ≥5–≤22 cm
Patent popliteal artery on the index side, i.e., single vessel runoff or better with at least 1 of 3 vessels patent to the lower third of the calf before the day of the procedure. Additional intervention to further improve blood flow to the lower limb is acceptable during the index procedure, but after successful treatment of the study lesions
Vessel diameter ≥4 and ≤6 mm
Aneurysm in the superficial femoral artery or popliteal artery
Tandem lesion requiring nonoverlapping stents
Procedures which require stent in stent placement, to obtain patency, such as severe calcification resistant to stenting
Poor aortoiliac or common femoral inflow, which would be deemed inadequate to support a femoropopliteal bypass graft
Revascularization involving the same limb within 7 days before the index procedure or a planned revascularization within 7 days after the index procedure
Subject having total occlusions of the iliac artery on the same side must be excluded. However, intervention to restore adequate blood flow before the index procedure is allowed during the same procedure and before the treatment of the study lesion
Previously implanted stent or stents in the to-be-treated artery at the same site
Requiring stent placement in the popliteal artery
Presence of a femoral vascular prosthesis
Procedures and Follow-up
Antegrade or contralateral retrograde arterial puncture was followed by placement of a sheath introducer. The target lesion was traversed with a catheter and guide wire according to the operator’s usual practice in the transluminal or subintimal plane. Cases of failure to cross the lesion and reenter the true lumen at an appropriate level were considered treatment failures and were analyzed according to the intention-to-treat principle. For patients randomized to PTA, balloon selection and dilatation was performed according to the operator’s normal practice. Bailout stenting was allowed if the postangioplasty angiogram showed an unsatisfactory result such as flow-limiting dissection or persistent stenosis, according to the operator’s local practice.
For patients randomized to stenting, predilatation of the lesion was recommended with a balloon smaller than the diameter of the reference vessel and with a length shorter than the anticipated lesion length. The appropriate stent size was selected after the investigator’s review of the patient’s baseline angiogram with a requirement that the diameter of the stent be 1–2 mm larger than the reference vessel diameter. A maximum of two stents was permitted, and where a second stent was required, an overlap of 0.5–1 cm was specified. Stents were to be placed such that they extended at least 5 mm proximal and distal to the diseased segment. Angiograms were taken in the same projections as the preimplant dilatation to compare the preimplant and postimplant minimum lumen diameters. The procedure was considered to have been completed once all the delivery material, including catheter sheath introducer had been removed. A Doppler ultrasound (DUS) of the treated segment was completed before discharge.
Antithrombotic medication was provided according to normal clinical practice at each site. Periprocedurally, anticoagulation with at least 3000–5000 U of heparin was recommended, with additional infusion or overnight treatment also permitted. A combination of two antiplatelet drugs such as clopidogrel and aspirin, or one antiplatelet drug combined with a low-molecular-weight heparin, for at least 3–4 weeks after the procedure, was recommended. Aspirin was recommended to be continued indefinitely. Follow-up clinical evaluations were conducted before hospital discharge, at 3 months + 7 days, at 6 months + 14 days, and at 12 months + 30 days after the procedure.
Study End Points
The primary end point was the binary restenosis rate at 1 year after the index procedure as measured by DUS, the results of which were evaluated by an independent core laboratory (Harvard Clinical Research Institute). Binary restenosis by DUS was defined as >50 % restenosis on the basis of a peak systolic velocity ratio ≥2.5, or absent flow indicating total occlusion. Secondary end points prespecified in the protocol, but not subject to power calculation, included postprocedural residual stenosis measured by angiography, technical success, primary patency, ankle–brachial index (ABI), Rutherford/Fontaine grade, and quality of life. Clinical event recording included all-cause mortality, myocardial infarction, ipsilateral amputation, target vessel revascularization (TVR), and target lesion revascularization (TLR). All events were adjudicated by a blinded clinical events committee.
Technical success was defined as achievement of a residual diameter stenosis of <30 % by angiography using the assigned device only. A procedure-related serious adverse event was defined as an adverse event related to the treated limb or index procedure that required the patient to be hospitalized, that prolonged hospitalization, that necessitated intervention, or that was fatal. ABI was calculated as the highest systolic pressure in the ankle of the index limb/highest brachial pressure. Health-related quality of life was measured by the EuroQol (EQ)-5D utilities and the visual analog scale (VAS) [6, 7].
With 60 patients in each treatment arm, a clinically meaningful odds ratio of 3.03 in the primary end point could be detected with a power of 80 % at a two-sided significance level of 0.05, assuming a 1 year binary restenosis rate of 63 % in patients who received balloon angioplasty and 36 % in patients who received a stent. The intention was to recruit 150 patients (75 patients per treatment group) to allow for an estimated 20 % noncompliance to follow-up and unavailable data at 12 months.
Continuous variables were presented as mean ± standard deviation (SD) or median and 25/75 percentiles. Comparisons of continuous variables used the Student’s t test or the Wilcoxon rank sum (Mann–Whitney) test. For categorical variables, the absolute (n) and relative frequency (%) were tabulated, and binary variables were compared by Fisher’s exact test.
Primary analyses were conducted on an intention-to-treat basis according to randomization. As-treated results are also reported. Additional quality-of-life analyses, prespecified in the economic evaluation plan, were conducted to identify and validate the EQ-5D utilities and VAS scores associated with preprocedural Rutherford/Fontaine classifications and differences between patients who did and did not incur binary restenosis or require TLR/TVR at follow-up. These data are shown irrespective of which treatment an individual patient was randomized to, because the purpose of this analysis was to determine the utilities associated with each of these end points. Categorization by presence or absence of restenosis excluded patients who had a TLR/TVR in order to assess whether DUS restenosis alone, in the absence of symptom recurrence sufficient to warrant TLR/TVR, had any impact on quality of life.
Recruitment and Patient Population
Baseline patient demographics and clinical characteristics
Stent (n = 74)
PTA (n = 76)
65.9 ± 9.0
69.8 ± 8.5
Coronary heart disease
Critical limb ischemiaa
0.61 ± 0.24
0.57 ± 0.15
103.4 ± 116.5
88.1 ± 55.5
Baseline lesion characteristics
Stent (n = 74)
PTA (n = 76)
123.0 ± 54.3
116.8 ± 52.2
Length of occlusion (mm)
83.9 ± 46.3
62.8 ± 37.1
Reference vessel diameter (mm)
5.4 ± 0.7
5.3 ± 0.6
Procedural and Clinical Outcomes
Procedural and clinical outcomes
Stent (n = 74)
PTA (n = 76)
Stenosis after procedure (%)
10.0 ± 16.4
21.2 ± 19.6
Clinical outcomes to 12 mo
Restenosis by DUS only (primary end point)
As-treated restenosis by DUS
As-treated primary patency
Clinically driven TLR
0.83 ± 0.27
0.81 ± 0.29
There were five deaths in the stent arm and two in the PTA arm of the study, which were unrelated to the procedures or to peripheral arterial disease. There were four ipsilateral amputations in total, two in each group, which all occurred in patients who initially presented with critical limb ischemia. Two patients randomized to stenting underwent amputation at the foot, with one requiring a further amputation below the knee. One patient randomized to PTA required amputation at the foot and one above the knee. There were fewer clinically driven TLRs in the patients randomized to stenting, but this did not reach statistical significance (12.5 vs. 20.8 %, p = 0.26). ABI increased from baseline in both groups, but there was no significant difference between the groups at 12 months’ follow-up. As-treated outcomes for stent vs. PTA at 12 months were restenosis by DUS 40.8 versus 46.7 % (p = 0.68), primary patency 51.8 versus 39.3 % (p = 0.29), and TLR 11.9 versus 22.2 % (p = 0.12).
Quality of Life
EQ-5D utility scores at baseline grouped by Rutherford/Fontaine category
0.55 ± 0.25 (16)
66.4 ± 16.2 (15)
0.43 ± 0.30 (56)
65.7 ± 17.8 (56)
0.36 ± 0.31 (42)
59.6 ± 17.1 (42)
0.27 ± 0.28 (8)
41.6 ± 22.1 (7)
0.38 ± 0.38 (17)
47.6 ± 24.6 (17)
EQ-5D quality-of-life scores at baseline and at follow-up
EQ-5D utility score, median (IQR)
0.52 (0.09–0.69) (69)
0.59 (0.09–0.69) (70)
0.69* (0.52–0.85) (59)
0.73* (0.69–1.00) (62)
0.69 (0.52–1.00) (57)
0.69 (0.52–0.80) (58)
0.69 (0.52–0.81) (58)
0.71 (0.62–0.78) (54)
EQ-5D visual analog score, mean ± SD
59.4 ± 21.0 (68)
61.5 ± 18.6 (69)
63.3 ± 25.0 (58)
70.3 ± 15.2 (62)
66.4 ± 19.4 (57)
65.4 ± 18.5 (61)
66.1 ± 20.9 (60)
62.2 ± 21.6 (56)
Since starting the SUPER study, four comparable studies involving Nitinol stents [5, 8–10] and a Cochrane review  have been published. The Cochrane review concluded that routine stenting of the SFA after PTA could not be recommended, although the authors did note that there was a small, short-term benefit of stenting over PTA when measured by DUS or angiographic patency. Only two of the studies included in this meta-analysis used self-expanding Nitinol stents [5, 8], which are now the stent technology of choice for the SFA as a result of the superior resistance of Nitinol stents to compression, torsion, flexion, contraction, and extension when compared with stainless steel stents .
In this study, the primary end point of 12 month binary restenosis was not significantly reduced by a strategy of primary stenting compared with balloon angioplasty in long (approximately 12 cm) SFA lesions in predominantly totally occluded vessels. This is similar to the findings of the FAST  study but in contrast to the ABSOLUTE  and ASTRON  studies. The SUPER study recruited patients with lesions 20 % longer than those of any other randomized trial published so far, with 2½ times as many total occlusions and a higher proportion of patients with critical limb ischemia. All of these factors have been reported as predisposing toward poorer outcomes with endovascular treatment [13–17] and may have reduced the effect size that was expected.
There are several possible explanations for the absence of a greater difference in outcome between stenting and PTA. It is likely that in most cases recanalization was in the subintimal plane. The role of uncovered stents in the subintimal plane is debatable [18–20]. It may be that there is less advantage to stent placement in the subintimal plane than in the true lumen, particularly if the added benefit of the stent is to tack down flow limiting intimal flaps. These generally do not occur in subintimal recanalization except at the entry tear. There were few cases of crossover from PTA to stent placement in this trial. This reflects the observation that the subintimal track generally looks satisfactory on the immediate postdilatation angiogram. There were four instances of crossover from stent to PTA. The usual explanation was low reentry into the true lumen, beyond the distal end of the occlusion, such that the operator considered stent placement to be contraindicated. There may have been instances of suboptimal stent technique, including failure to stent the entire diseased segment from normal to normal vessel, the result of there being no truly disease-free arterial segment in the SFA in some cases, with the associated risk of disease progression in the unstented portion. In addition, despite randomization, patients assigned to stenting had a preprocedural occlusion length that was 21 mm longer than those assigned to PTA, which may have disadvantaged the stenting arm of the study. There were also a higher proportion of total occlusions and a larger number of failures to traverse the occlusion in the stent group. The as-treated analysis removes the cases of failure to cross the lesion. This leads to a small, nonsignificant numerical benefit from stenting.
Primary stenting did not result in improvements in ABI at 12 months compared with PTA alone, and EQ-5D utility scores were virtually identical at each point of measurement, indicating that neither treatment strategy conferred any measurable health benefit over the other. Quality-of-life measures are highly relevant to current health policy as the collection of patient-reported outcomes is becoming an increasingly important measure of the effectiveness and quality of treatment. A previous study has shown that quality of life measured by the SF-36 questionnaire improves significantly after SFA intervention and that restenosis adversely affects quality of life . The EQ-5D is the preferred generic measure of health status by both the National Health Service (NHS) in England  and by the National Institute for Health and Clinical Excellence (NICE) .
There was good agreement between the established Rutherford/Fontaine clinical categorization of peripheral arterial disease and preprocedural EQ-5D utility scores, which lends support to the validity of using the EQ-5D to assess outcomes in these patients. A similar relation between EQ-5D and preprocedural Rutherford classification has been reported by de Vries et al. , although the mean EQ-5D scores in this Dutch population were higher than those found in SUPER study patients. The BASIL study , which similarly recruited UK patients, reported a mean EQ-5D utility score of 0.26 ± 0.32 for patients with ischemic rest pain or tissue loss before angioplasty, which is comparable with the scores recorded for SUPER study patients categorized as Rutherford 4/Fontaine III (0.27 ± 0.28) and Rutherford 5/Fontaine IV (0.38 ± 0.38). These data highlight the serious impact of symptomatic peripheral arterial disease on patients when compared with other cardiovascular conditions that warrant intervention. For example, Longworth et al.  reported mean EQ-5D utilities of 0.36 (95 % confidence interval 0.25–0.48) in patients with Canadian Cardiovascular Society grade IV angina pectoris, whereas Serruys et al.  reported preprocedural utilities of 0.69 ± 0.20 in patients with multivessel coronary artery disease in the ARTS study. If comparability of utilities between disease states is assumed, even patients with mild intermittent claudication in the SUPER study scored themselves as having a worse quality of life than patients with symptomatic multivessel coronary artery disease.
Both randomized groups showed significant improvements in quality of life at 3 months compared with the preprocedure scores, and this was maintained to 12 months’ follow-up, indicating that endovascular treatment yielded lasting health benefits. Some patients presented with very poor preprocedural health-related quality of life, as indicated by the low 25th percentile values in Table 6. This lower limit of the interquartile range was notably higher at all points of follow-up, suggesting that revascularization was beneficial in these severely symptomatic patients.
Restenosis determined by DUS is an end point in most of the comparable trials of SFA intervention [1, 2, 5, 8, 9, 28]. However, when patients were categorized according to the presence or absence of binary restenosis by DUS, EQ-5D utilities (excluding those who required repeat revascularization) were almost identical at each point of follow-up, suggesting that binary restenosis in the absence of symptoms sufficient to warrant repeat revascularization is not a good indicator of health status. Stratification of all randomized patients according to whether they underwent revascularization revealed a statistically significant reduction in quality of life at 6 months, to below the preprocedural value, indicating the negative impact of symptom recurrence on some patients with recurrent SFA disease. Reintervention provided effective relief, as evidenced by the return to 3 month values at 12 months’ follow-up. These data also suggest that when comparing different interventional strategies or different devices, binary restenosis—and composite end points heavily dependent on it—may be of limited clinical relevance because its presence or absence appears to have little meaningful effect on the average health status of patients. This is not surprising, as peak systolic velocity ratio measured by DUS is in itself a surrogate for >50 % angiographic stenosis , and there may be no difference in the symptomatic status of patients with either 51 or 49 % diameter stenosis, even though one would be categorized with binary restenosis and one without. However, a reduction in the need for repeat revascularization does appear to be associated with a health benefit, and primary stenting was associated with a trend toward reduction in TLR in both the intention-to-treat and as-treated analyses. This mirrors the findings of other comparable trials. A larger-scale trial or meta-analysis, powered on the basis of TLR or quality of life, would be of value.
The lack of significant differences between the two arms of the SUPER study does not mean that the use of stents in long SFA lesions should be abandoned. PTA with bailout stenting remains an acceptable approach, but the trend toward reduced TLR, which has also been observed in the comparable trials [8, 30, 31], adds further weight to the European Society of Cardiology recommendations  that primary stenting should be considered because the reduction in TLR appears to be a meaningful health outcome for patients.
Recruitment of patients into the SUPER study took 3 years, and it is conceivable that practice patterns may have changed over that time. Slow recruitment suggests that participation was restricted to a small, and therefore possibly unrepresentative, sample of patients with peripheral vascular disease being treated by these centers. The comparison between treatment groups is also confounded by differences in patient age and the length of occlusions, despite randomization.
The SUPER study has shown that 12 month binary restenosis is not reduced by a strategy of primary stenting compared with balloon angioplasty with bailout stenting in long SFA lesions in predominantly totally occluded vessels. Both treatment groups demonstrated a significant and durable improvement in quality of life, reinforcing the value to patients of endovascular treatment in long and complex SFA lesions. The reduction in the need for TLR associated with stenting in the SUPER study is similar to that seen in other studies, despite a more complex patient population. The EQ-5D appears to be a valid and sensitive tool to detect changes in quality of life associated with intervention and recurrent disease.
The SUPER study was sponsored and funded by Cordis Clinical Research.
Conflict of interest
SF and MvR are paid employees of Cordis, Johnson & Johnson. The other authors declare that they have no conflict of interest.