Propensity Score Analysis Comparing Clinical Outcomes of Drug-Eluting vs Bare Nitinol Stents in Femoropopliteal Lesions

Abstract

Purpose: To present a propensity score matching analysis comparing the 1-year outcomes of de novo femoropopliteal lesions treated with drug-eluting stents (DES) or bare nitinol stents (BNS). Methods: A retrospective review was conducted of 452 limbs in 389 patients (mean age 74±8 years; 284 men) treated with DES implantation and 1808 limbs in 1441 patients (mean age 72±9 years; 1023 men) implanted with BNS for de novo femoropopliteal lesions. One-year follow-up data were available on all patients. The primary endpoint was 12-month restenosis assessed by duplex ultrasonography or follow-up angiography within ±2 months. Secondary endpoint was major adverse limb events (MALE) including major amputation, any reintervention, and restenosis. Results: The BNS group was more likely to have current smoking, chronic total occlusion, and poor below-the-knee runoff. The stratification analysis demonstrated that diabetes mellitus (DM) and reference vessel diameter (RVD) had a significant interaction on the association of DES vs BNS implantation with restenosis (interaction p<0.05). Thus, the population was stratified into 4 subgroups (1: –DM, RVD ≥5 mm, 2: +DM, RVD ≥5 mm, 3: –DM, RVD <5 mm, and 4: +DM, RVD <5 mm); the RVD threshold was empirically determined. There were no significant intergroup differences in baseline variables after matching. There was no significant difference in restenosis risk between DES and BNS in the RVD ≥5 mm subgroup regardless of the presence of DM. The DES group had a significantly higher restenosis risk in the RVD <5 mm subgroup regardless of the presence of DM. No significant difference was observed in the risk of major amputation, reintervention, or MALE in any subgroup. Conclusion: These results suggest that a first-generation DES was not superior to a conventional BNS for femoropopliteal lesions.

Keywords

adverse events diabetes drug-eluting stent femoropopliteal segment nitinol stent peripheral artery disease reference vessel diameter reintervention restenosis stenosis stent

Introduction

The introduction of a bare nitinol stent (BNS) into endovascular therapy (EVT) has decreased the incidence of restenosis in the short and midterm compared with balloon angioplasty, and several studies have shown excellent outcomes.^1–4 The TASC II (TransAtlantic Inter-Society Consensus II) guidelines recommend EVT for TASC-II A/B lesions and bypass for C/D lesions.⁵ However, a recent study showed the efficacy of a BNS for treatment of TASC-II A-C lesions.⁶ Thus, the scope of EVT has tended to expand,⁷ and the procedure has been applied to more complex lesions in actual clinical settings.

Good outcomes have also been reported for procedures using a drug-eluting stent (DES),^8,9 and treatment of complex lesions, such as TASC-II C/D lesions, is likely to be effective using a DES.¹⁰ The efficacy of the Zilver PTX DES (Cook Medical, Bloomington, IN, USA) and a bare Zilver stent have been compared in a randomized clinical trial in simple and relatively short lesions.^8,9 However, this trial did not include cases with a more complicated background and pathology, which are often found in clinical settings, and there are no comparative studies on complex lesions. Here, we compare historical data for 1-year outcomes of de novo femoropopliteal lesions treated with the Zilver PTX DES or a BNS using propensity score matching.

Methods

Study Design and Population

The protocol for this retrospective analysis was developed in accordance with the Declaration of Helsinki and was approved by the ethics committee of Kansai Rosai Hospital, Amagasaki, Japan. Every patient provided written informed consent for use of their anonymized data.

The DES group was a subgroup of patients from the prospective multicenter ZEPHYR (Zilver PTX for the Femoral Artery and Proximal Popliteal Artery) study,¹¹ which is registered in the University Hospital Medical Information Network–Clinical Trial Registry (UMIN000008433). This study included 607 limbs with symptomatic de novo femoropopliteal lesions treated with Zilver PTX DES from July 2012 to April 2013. Of these, 155 were ineligible for this study because of death within 1 year (n=58), dropout within 1 year (n=50), and data missing (n=47). Therefore, data available on 452 limbs in 389 patients (mean age 74±8 years; 284 men) were included in this analysis.

The BNS group was taken from the REAL-FP (Retrospective Multicenter Analysis for Femoropopliteal disease) registry approved by the institutional review boards at all 13 participating cardiovascular and vascular institutions in Japan (UMIN000010986). Data from these patients were used as a historical control. The registry included 2499 limbs that had de novo femoropopliteal lesions treated with a variety of stents: S.M.A.R.T. (Cordis Corporation, Bridgewater, NJ, USA), Luminexx (Bard Peripheral Vascular, Tempe, AZ, USA), ZILVER (Cook Medical), or Misago (Terumo, Tokyo, Japan). Of these cases, 691 limbs were ineligible because of death within 1 year (n=148), dropout within 1-year (n=414), or data missing (n=129). Therefore, 1808 limbs in 1441 patients (mean age 72±9 years; 1023 men) were eligible for this analysis. Table 1 shows the baseline characteristics of the study population.

Table 1.

Baseline Characteristics of the Study Population Based on Limbs Treated Before Matching.^a

	DES (n=452)	BNS (n=1808)	p^b
Age, y	74±8	72±9	0.004
Men	330 (73)	1289 (71)	0.484
Current smoking	91 (20)	499 (28)	0.001
Hypertension	396 (88)	1593 (88)	0.747
Dyslipidemia	361 (80)	1355 (75)	0.031
Diabetes mellitus	298 (66)	1189 (66)	0.956
Regular dialysis	124 (27)	386 (21)	0.007
Cerebrovascular disease	90 (20)	336 (20)	0.896
Coronary artery disease	229 (51)	914 (51)	1.000
Chronic heart failure	76 (17)	162 (9)	<0.001
Critical limb ischemia	128 (28)	442 (24)	0.090
Chronic total occlusion	212 (47)	971 (54)	0.010
Calcification	283 (63)	992 (55)	0.001
Lesion length, cm	16±9	14±9	<0.001
RVD, mm	5.3±1.0	5.3±0.9	0.927
≤1-vessel BTK runoff	167 (37)	764 (42)	0.042

Abbreviation: BNS, bare nitinol stent; BTK, below the knee; DES, drug-eluting stent; RVD, reference vessel diameter.

Continuous data are presented as the means ± standard deviation; categorical data are given as the counts (percentage).

Intergroup difference was tested using the unpaired t test for continuous variables, the Fisher exact test for dichotomous variables, and the Mann-Whitney U test for the other discrete variables.

Procedures and Follow-up

The indication for revascularization included symptomatic disease with >50% diameter stenosis determined by angiography and mean pressure gradient >10 mm Hg determined with a 4-F diagnostic catheter. For EVT, a 6-F sheath was inserted into the femoral artery, typically using a contralateral approach. After infusion of 5000 units of heparin, 0.035- or 0.014-inch guidewires were used to cross the lesion. Stent size was ~1 mm larger than the vessel diameter assessed by angiography or intravascular ultrasound.

Dual antiplatelet therapy (DAPT) with 100 mg/d aspirin and either 200 mg/d ticlopidine or 75 mg/d clopidogrel was recommended to start at least 1 week before and continue for at least 2 months after stent implantation in the DES group and 1 month in the BNS group. Antiplatelet and anticoagulant drugs were used at the physician’s discretion based on the patient’s condition. Duplex ultrasound, computed tomography angiography (CTA), and/or angiography were conducted to evaluate the presence of restenosis at least 1, 6, and 12 months after EVT.

Endpoints and Definitions

The primary endpoint was 12-month restenosis assessed by duplex ultrasound or angiography within ±2 months. Restenosis was defined as recurrence of ≥50% diameter stenosis determined by angiography or a peak systolic velocity ratio >2.4 by duplex.¹² An undetectable signal in the stented segments on duplex ultrasound was graded as complete occlusion. Any reintervention or major amputation within 1 year was automatically included in the restenosis endpoint.

The secondary endpoint was major adverse limb events (MALE), which included major amputation or any reintervention (surgical or endovascular). Major amputation was surgical excision of the limb above the ankle. Any amputation at or distal to the Lisfranc level was not considered a limb salvage failure. Amputation data were obtained through outpatient clinic follow-up contact.

Reference vessel diameter (RVD) was estimated by quantitative vessel angiography or visual estimate in each institution. Coronary artery disease was defined as stable angina with evidence of coronary stenosis documented by coronary catheterization or functional or perfusion studies, history of percutaneous coronary intervention or coronary artery bypass graft surgery, or previous myocardial infarction. Cerebrovascular disease was defined as a hospital or neurologist report with diagnosis of a transient ischemic attack or ischemic stroke. Heart failure (HF) was judged to be present based on a previous diagnosis of HF, a history of hospitalization for HF, or current treatment for HF. Diabetes was defined as hemoglobin A1c ≥6.5%, fasting plasma glucose ≥126 mg/dL, random plasma glucose ≥200 mg/dL, past history of diagnosis, or current treatment with oral hypoglycemic agents or insulin.

Statistical Analysis

Data are presented as the mean ± standard deviation for continuous variables and the number (percentage) for discrete variables, if not otherwise noted. Logistic regression analysis was employed to calculate the odds ratios (OR) and 95% confidence intervals (CI) of DES vs BNS implantation for 1-year restenosis.

To examine whether any background characteristics would influence the prognostic impact of DES and BNS implantation, a stratification analysis was performed including the following background characteristics: sex, age, smoking, hypertension, dyslipidemia, diabetes mellitus (DM), regular dialysis, cerebrovascular disease, coronary artery disease, chronic heart failure, critical limb ischemia, chronic total occlusion, arterial calcification, lesion length, RVD, and the below-the-knee runoff. The p values for interaction between these variables and the treatment strategy were calculated using logistic regression analysis.

The study population was stratified according to the variables with a significant interaction effect (p<0.05), and the study endpoints were compared between DES and BNS in each stratified subgroup. To minimize intergroup differences in baseline characteristics, the comparison was performed after propensity score matching. The propensity score was developed using the logistic regression model in which the variables listed above were entered. The one-to-one pair matching was performed with the nearest neighbor method without replacement. According to Austin’s recommendation,¹³ the groups were matched on the logit of the propensity score within the caliper of 0.2 standard deviations of the logit of the propensity score. Intergroup differences after matching were tested using paired analysis. Bootstrap resampling (100,000 times) was performed to estimate the 95% CI of 1-year incidence of clinical outcomes and to test the intergroup difference in the incidence. A p<0.05 indicated a statistically significant difference. Statistical analyses were performed using IBM SPSS Statistics (version 22; IBM Corporation, Somers, NY, USA), except for the propensity score matching and the bootstrap resampling, which were performed using R (version 3.1.0; R Development Core Team, Vienna, Austria; http://www.r-project.org ).

Results

Compared with the BNS patients, the DES group was older and had higher proportions of dyslipidemia, regular dialysis, chronic heart failure, calcification, and longer lesions, whereas the BNS group was more likely to have current smoking, chronic total occlusion, and poor below-the-knee runoff (Table 1).

The stratification analysis (Figure 1) demonstrated that DM and RVD had significant interaction effects on the association of DES and BNS with restenosis risk (interaction p<0.05 for both groups). The study population was thus stratified into the following 4 subgroups for further analysis: (1) those without DM and with a RVD ≥5 mm, (2) those with DM and a RVD ≥5 mm, (3) those without DM and with a RVD <5 mm, and (4) those with DM and a RVD <5 mm. The RVD threshold was empirically determined. After stratification, propensity score matching was performed in each subgroup. The matching extracted 107 pairs in the –DM and RVD ≥5 mm subgroup, 200 pairs in the +DM and RVD ≥5 mm subgroup, 34 pairs in the –DM and RVD <5 mm subgroup, and 87 pairs in the +DM and RVD <5 mm subgroup (Figure 2). There was no significant intergroup difference in baseline characteristics after matching in any subgroup (Tables 2 and 3).

Figure 1.

Odds ratios of drug-eluting stents (DES) vs bare nitinol stents (BNS) for restenosis before matching.

Figure 2.

Distribution of propensity scores in matched and unmatched populations. BNS, bare nitinol stent; DES, drug-eluting stent; DM, diabetes mellitus; RVD, reference vessel diameter.

Table 2.

Baseline Characteristics of the Matched Study Population.^a,b

	−DM, RVD ≥5 mm (n=107 each)		+DM, RVD ≥5 mm (n=200 each)		−DM, RVD <5 mm (n=34 each)		+DM, RVD <5 mm (n=87 each)
	DES	BNS	DES	BNS	DES	BNS	DES	BNS
Age, y	75±8	75±8	73±8	73±8	77±10	76±8	72±9	72±10
Men	87 (81)	83 (78)	167 (84)	171 (86)	19 (56)	20 (59)	45 (52)	47 (54)
Current smoking	20 (19)	19 (18)	44 (22)	47 (23)	7 (21)	6 (18)	18 (21)	19 (22)
Hypertension	93 (87)	88 (82)	178 (89)	176 (88)	26 (76)	28 (82)	79 (91)	78 (90)
Dyslipidemia	77 (72)	75 (70)	173 (86)	172 (86)	22 (65)	25 (74)	68 (78)	73 (84)
Diabetes mellitus	0 (0)	0 (0)	200 (100)	200 (100)	0 (0)	0 (0)	87 (100)	87 (100)
Regular dialysis	21 (20)	14 (13)	65 (32)	62 (31)	6 (18)	5 (15)	26 (30)	27 (31)
Cerebrovascular disease	16 (15)	19 (18)	43 (22)	49 (24)	3 (9)	4 (129	24 (28)	26 (30)
Coronary artery disease	45 (42)	47 (44)	112 (56)	110 (55)	11 (32)	17 (50)	48 (55)	41 (47)
Chronic heart failure	14 (13)	16 (15)	30 (15)	28 (14)	3 (9)	3 (9)	16 (18)	13 (15)
Critical limb ischemia	22 (21)	22 (21)	57 (28)	51 (25)	12 (35)	6 (26)	26 (30)	31 (36)
Chronic total occlusion	57 (53)	62 (59)	82 (41)	88 (44)	16 (47)	18 (53)	46 (53)	39 (45)
Calcification	57 (53)	55 (51)	144 (72)	138 (69)	16 (47)	12 (35)	52 (60)	48 (55)
Lesion length, cm	16±9	18±9	14±9	15±9	16±9	16±6	16±9	16±8
RVD, mm	5.8±0.7	5.9±0.8	5.8±0.7	5.7±0.7	4.3±0.4	4.4±0.2	4.2±0.5	4.2±0.4
≤1-vessel BTK runoff	33 (31)	30 (28)	78 (39)	81 (41)	10 (29)	12 (35)	33 (38)	36 (41)

Abbreviation: BNS, bare nitinol stent; BTK, below the knee; DES, drug-eluting stent; DM, diabetes mellitus; RVD, reference vessel diameter.

Continuous data are presented as the means ± standard deviation; categorical data are given as the counts (percentage).

There were no intergroup differences, which were tested using the paired t test for continuous variables and the McNemar test for dichotomous variables.

Table 3.

Standardized Differences in Baseline Characteristics for the Subgroups in the Matched Study Population.

	−DM, RVD ≥5 mm		+DM, RVD ≥5 mm		−DM, RVD <5 mm		+DM, RVD <5 mm
	SD	p	SD	p	SD	p	SD	p
Age, y	0.001	0.993	−0.070	0.450	0.013	0.961	−0.040	0.799
Men	0.093	0.571	−0.055	0.651	−0.059	1.000	−0.046	0.855
Current smoking	0.024	1.000	−0.036	0.807	0.075	1.000	−0.028	1.000
Hypertension	0.130	0.441	0.031	0.877	−0.146	0.754	0.039	1.000
Dyslipidemia	0.041	0.885	0.015	1.000	−0.192	0.629	−0.147	0.441
Diabetes mellitus	0.000	1.000	0.000	1.000	0.000	1.000	0.000	1.000
Regular dialysis	0.178	0.265	0.032	0.820	0.080	1.000	−0.025	1.000
Cerebrovascular disease	−0.076	0.690	−0.071	0.561	−0.097	1.000	−0.051	0.868
Coronary artery disease	−0.038	0.883	0.020	0.923	−0.364	0.146	0.161	0.310
Chronic heart failure	−0.054	0.845	0.028	0.880	0.000	1.000	0.093	0.629
Critical limb ischemia	0.000	1.000	0.068	0.571	0.192	0.549	−0.123	0.522
Chronic total occlusion	−0.113	0.470	−0.061	0.590	−0.118	0.804	0.161	0.391
Calcification	0.037	0.888	0.066	0.496	0.241	0.503	0.093	0.635
Lesion length, cm	−0.221	0.114	−0.100	0.328	0.086	0.700	0.016	0.909
RVD, mm	−0.042	0.761	0.063	0.524	−0.221	0.408	−0.118	0.456
≤1-vessel BTK runoff	0.062	0.770	−0.031	0.832	−0.126	0.804	−0.071	0.760

Abbreviation: BTK, below the knee; DES, drug-eluting stent; DM, diabetes mellitus; RVD, reference vessel diameter; SD, standardized difference.

The 1-year incidences of the study outcomes in the matched population are shown in Figure 3. There was no significant difference in restenosis risk between DES and BNS in the subgroups with RVD ≥5 mm regardless of the presence of DM (Figure 3A and B). On the other hand, the DES group had a significantly higher restenosis risk in the subgroups with RVD <5 mm regardless of the presence of DM (Figure 3C and D). No significant difference was observed in the risk of major amputation, reintervention, or MALE in any subgroup.

Figure 3.

One-year clinical outcomes after femoropopliteal Intervention in the matched population. DES, drug-eluting stent; DM, diabetes mellitus; MALE, major adverse limb events; RVD, reference vessel diameter.

Discussion

In this study, historical data for 1-year outcomes of femoropopliteal lesions treated with a DES (Zilver PTX) or a BNS were compared using propensity score matching to reduce potential bias. The use of DES showed no clear advantage; instead, the restenosis rate in a subgroup with small vessels (RVD <5 mm) was greater in the DES group, which was an unexpected result.

A previous study⁸ found excellent outcomes using a Zilver PTX, but the mean lesion length was short (66.4±38.9 mm), the vessel diameter was well preserved with at least 1 patent runoff vessel (inclusion criteria), and patients with renal failure, dialysis, or cardiac failure were excluded (exclusion criteria). Thus, the study included many patients in whom patency would be expected to be relatively good, and this is substantially different from lesions found in actual clinical settings. The mean lesion length in the DES group in the current study was long (16 cm), and the rates of critical limb ischemia, dialysis, and ≤1 runoff vessel were 28%, 27%, and 37%, respectively. Therefore, the current study, using exclusion criteria similar to the randomized trial, included significantly more complex lesions that place higher expectations on the efficacy of a DES.

The efficacy of a Zilver PTX in complex lesions has been reported,¹⁰ but another study found insufficient efficacy in a clinical setting.¹⁴ The mean follow-up periods in these studies were 5.3 months¹⁰ and 13.2±5.3 months,¹⁴ and the mean lesion length was 188±96 mm. The different findings suggest that sufficient follow-up is needed to evaluate the outcome of a DES in complex lesions. The occurrence of restenosis peaks at 1 year in patients undergoing BNS placement,¹⁵ and thus the follow-up time after DES placement should be at least 1 year. For this reason, only patients who were followed for >1 year were included in the current study.

This study was unable to demonstrate the efficacy of a DES compared with a conventional BNS, even though a DES should have much better late loss. There may be several reasons why a DES was less effective than expected. First, a DES may not be effective for calcified lesions; the study included many patients on dialysis, and many lesions were severely calcified. For such lesions, the drug may separate from the stent during delivery, and it is unclear whether a DES can be effective when the drug is in a calcified area.

Second, there were many patients (47%) with chronic total occlusion. The occlusion length was also relatively long. The randomized trial included many patients with stenotic lesions through which a guidewire was usually passed in the true lumen, while several patients with chronic total occlusion would have had stent placement in the subintimal space. The neointimal response after placement of a self-expanding stent in a femoropopliteal lesion differs from that in the coronary artery,¹⁶ which might be attributable to poor expansion due to a chronic outward force or to calcification. Initial gain may be difficult when a stent is placed in the subintimal space in patients with complex lesions, particularly in those with calcified lesions.

Since 80% of the BNS were S.M.A.R.T. stents, the radial force of the stent may have also influenced the outcome. A previous study⁸ showed that outcomes with a DES were significantly better than those with a bare Zilver stent, but the 1-year primary patency of the bare Zilver stent was ~70%, which was similar to rates reported in other studies.^1–4 The outcome may vary depending on the type or design of the BNS, and comparison studies of various types of BNS are required. Furthermore, the DES length was limited, and many overlaps were needed for long lesions, which might also cause poor outcomes. Finally, since the study included many patents with critical limb ischemia, there were many with poor runoff, which is known to have a negative effect on stent patency.^17,18 These factors may have resulted in failure to show the superiority of a DES compared to a BNS. Other factors preventing demonstration of the superiority of a DES over a conventional BNS may also be present in actual clinical settings.

Limitations

First, the study was a retrospective analysis of patients from a large-scale multicenter trial and a retrospective registry. The most appropriate cases were selected from patients in the BNS group using propensity score matching, and only cases that completed 1-year follow-up were compared between the 2 groups. Furthermore, although the subgroups were decided before analysis based on historical data, that data might not be sufficient; there are likely more determinants that may affect the outcome. The cutoff value for the RVD was empirically selected. Further investigation with many suitable factors is required to verify these findings.

Second, DAPT in the DES and BNS groups was recommended for at least 2 months and 1 month, respectively. Additionally, cilostazol administration was significantly higher in the BNS group (34% vs 52%, p<0.001) without interaction, but drug compliance was not evaluated in detail. Therefore, further research is required to analyze the influence of antiplatelet therapy in the chronic phase.

Third, the study included Japanese patients only, and it is unclear if the same results would be obtained in other races. The appropriateness of the sample size is also uncertain. Thus, a prospective randomized trial is needed to confirm the results of the current study. Within these limitations, the results of the study indicate that a DES has limited efficacy for treating complex lesions.

Conclusion

Our results suggested that a first-generation DES (Zilver PTX) was not superior to a conventional BNS for femoropopliteal lesions in a clinical setting based on a propensity matching analysis.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Dick

Wallner

Sabeti

Balloon angioplasty versus stenting with nitinol stents in intermediate length superficial femoral artery lesions. Catheter Cardiovasc Interv. 2009;74:1090–1095.

Schillinger

Sabeti

Loewe

Balloon angioplasty versus implantation of nitinol stents in the superficial femoral artery. N Engl J Med. 2006;354:1879–1888.

Zeller

Tiefenbacher

Steinkamp

Nitinol stent implantation in TASC A and B superficial femoral artery lesions: the Femoral Artery Conformexx Trial (FACT). J Endovasc Ther. 2008;15:390–398.

Laird

Katzen

Scheinert

Nitinol stent implantation versus balloon angioplasty for lesions in the superficial femoral artery and proximal popliteal artery: twelve-month results from the RESILIENT randomized trial. Circ Cardiovasc Interv. 2010;3:267–276.

Norgren

Hiatt

Dormandy

Inter-Society Consensus for the Management of Peripheral Arterial Disease (TASC II). J Vasc Surg. 2007;45(Suppl S):S5-67.

Iida

Takahara

Soga

Shared and differential factors influencing restenosis following endovascular therapy between TASC (Trans-Atlantic Inter-Society Consensus) II class A to C and D lesions in the femoropopliteal artery. JACC Cardiovasc Interv. 2014;7:792–798.

Tendera

Aboyans

Bartelink

ESC Committee for Practice Guidelines. ESC guidelines on the diagnosis and treatment of peripheral artery disease: document covering atherosclerotic disease of extracranial carotid and vertebral, mesenteric, renal, upper and lower extremity arteries: the Task Force on the Diagnosis and Treatment of Peripheral Artery Diseases of European Society of Cardiology (ESC). Eur Heart J. 2011;32:2851–2906.

Dake

Ansel

Jaff

Paclitaxel-eluting stents show superiority to balloon angioplasty and bare metal stents in femoropopliteal disease: twelve-month Zilver PTX randomized study results. Circ Cardiovasc Interv. 2011;4:495–504.

Dake

Ansel

Jaff

Sustained safety and effectiveness of paclitaxel-eluting stents for femoropopliteal lesions: 2-year follow-up from the Zilver PTX randomized and single-arm clinical studies. J Am Coll Cardiol. 2013;61:2417–2427.

10.

Leopardi

Houbballah

Becquemin

. Effectiveness of Zilver PTX eluting stent in TASC C/D lesions and restenosis. J Cardiovasc Surg (Torino). 2014;55:229–234.

11.

Iida

Takahara

Soga

One-year results of the ZEPHYR (ZilvEr PTX for tHe Femoral ArterY and Proximal Popliteal ArteRy) Registry: predictors of restenosis. JACC Cardiovasc Interv. 2015;8:1105–1112.

12.

Ranke

Creutzig

Alexander

. Duplex scanning of the peripheral arteries: correlation of the peak velocity ratio with angiographic diameter reduction. Ultrasound Med Biol. 1992;18:433–440.

13.

Austin

. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat. 2011;10:150–161.

14.

Fujihara

Utsunomiya

Higashimori

Outcomes of Zilver PTX stent implantation for the treatment of complex femoropopliteal artery disease [published online October 29, 2014]. Heart Vessels. doi:10.1007/s00380-014-0596-2.

15.

Iida

Uematsu

Soga

Timing of the restenosis following nitinol stenting in the superficial femoral artery and the factors associated with early and late restenoses. Catheter Cardiovasc Interv. 2011;78:611–617.

16.

Yahagi

Otsuka

Sakakura

Pathophysiology of superficial femoral artery in-stent restenosis. J Cardiovasc Surg (Torino). 2014;55:307–323.

17.

Davies

Saad

Peden

Percutaneous superficial femoral artery interventions for claudication—does runoff matter?

Ann Vasc Surg. 2008;22:790–798.

18.

Soga

Iida

Hirano

Utility of new classification based on clinical and lesional factors after self-expandable nitinol stenting in the superficial femoral artery. J Vasc Surg. 2011;54:1058–1066.