Vessel Patency and Associated Factors of Drug‐Coated Balloon for Femoropopliteal Lesion

Abstract

Background

Although clinical trials have reported favorable outcomes after drug‐coated balloon (DCB) therapy for femoropopliteal lesions, their real‐world performance and predictors have not been well evaluated. This study aimed to elucidate 1‐year freedom from restenosis and to explore the associated factors after a DCB for femoropopliteal lesions in clinical settings.

Methods and Results

This multicenter, prospective cohort registered 3165 de novo or restenotic femoropopliteallesions (mean lesion length, 13.5±9.3 cm; chronic total occlusion, 25.9%; severe calcification, 14.6%) that underwent successful DCB (Lutonix [24.2%] and IN.PACT Admiral [75.8%]) treatment between March 2018 and December 2019. Patency was assessed at 12±2 months. The primary outcome measure was 1‐year freedom from restenosis and its associated factors. Bailout stenting was performed in 3.5% of patients. The postprocedural slow flow phenomenon was observed in 3.9% of patients. During a median follow‐up of 14.2 months, 811 patients experienced restenosis. The Kaplan–Meier estimate of freedom from restenosis was 84.5% at 12 months (79.7% at 14 months). Focal, tandem, diffuse, and occlusive restenosis accounted for 37.4%, 9.8%, 18.9%, and 33.9%, respectively. Freedom from target lesion revascularization was 91.5% at 12 months. Risk factors independently associated with 1‐year restenosis were a history of revascularization, smaller distal reference vessel diameter, severe calcification, chronic total occlusion, low‐dose DCB, and residual stenosis.

Conclusions

The 1‐year clinical outcomes after DCB use for femoropopliteal lesions in real‐world practice was favorable. The additive risk factors were associated with a lower rate of freedom from restenosis.

Nonstandard Abbreviations and Acronyms

BNS

bare‐nitinol stent

CTO

chronic total occlusion

DCB

drug‐coated balloons

DES

drug‐eluting stents

EVT

endovascular therapy

TLR

target lesion revascularization

Clinical Perspective.

What Is New?

• Our study demonstrated the true performance of drug‐coated balloons in real‐world practice, with a very low rate of bailout stenting and no use of atherectomy devices.

• In addition, it also elucidated morphologies associated with restenosis and the risk factors for restenosis after using drug‐coated balloons.

What Are the Clinical Implications?

• This large‐scale multicenter prospective clinical drug‐coated balloon study indicated that 12‐month freedom from restenosis and target lesion revascularization was 84.5% and 91.5%.

• Two thirds of restenosis had a stenotic pattern, and independent predictors of restenosis were prior revascularization, smaller vessel, presence of severe calcification, chronic total occlusion, use of low‐dose drug‐coated balloons, and residual stenosis.

In recent years, the advent of drug‐eluting devices has dramatically reduced restenosis and target lesion revascularization (TLR) of femoropopliteal lesions after endovascular therapy (EVT) compared with conventional percutaneous transluminal angioplasty and bare‐nitinol stent (BNS) implantation. ¹ , ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ Currently, there are 2 types of drug‐eluting devices: drug‐coated balloons (DCBs) ² , ³ , ⁴ and drug‐eluting stents (DESs). ⁵ , ⁶ , ⁷ Although both are used in various clinical settings, DES is a permanent metallic implant, such as BNS, and shares a future risk of stent fractures. ⁸ , ⁹ Recently, aneurysmal degeneration after DES placement has been reported, ¹⁰ which is another concern regarding primary DES implantation. In contrast, DCBs are free from these concerns; however, their restenosis rate reported in clinical trials are higher than those after DES implantation. Moreover, large‐scale clinical studies on restenosis, its associated factors, and morphology after DCB use are warranted.

To assess the restenosis risk associated with a specific device, it is important to minimize the influence of other devices. In this context, evaluating the real‐world efficacy of DCBs alone is difficult in Europe and North America, where bailout stenting is commonly performed, especially in complex lesions, ¹¹ and atherectomy devices are also available. ¹² , ¹³ However, in Japan, bailout stenting after DCBs is rare because it is not reimbursed, and no atherectomy devices have been approved yet. Therefore, Japan is an appropriate field for evaluating the true effectiveness of DCBs in real‐world clinical practice.

This study aimed to elucidate the 1‐year freedom from restenosis and to explore the associated factors after EVT with a DCB for symptomatic femoropopliteal lesions in real‐world clinical settings.

METHODS

Study Population

This study used the clinical database of POPCORN (Prospective Multicenter Registry of Drug‐Coated Balloon for Femoropopliteal Disease). POPCORN is an ongoing prospective multicenter observational study that registered patients (≥20 years) undergoing DCB treatment for femoropopliteal lesions (either de novo or restenotic) with symptomatic peripheral arterial disease (Rutherford category 2 to 5) at 81 cardiovascular centers across Japan between March 2018 and December 2019. Annual follow‐ups with a time window of ±2 months for 5 years were scheduled. A total of 3165 lesions in 2827 limbs of 2507 patients undergoing treatment with Lutonix or IN.PACT Admiral were included in the registry. The study was conducted in accordance with the Declaration of Helsinki and approved by the institutional review boards of the participating centers. Informed consent was obtained from the participants or, if not possible, their families. This study reported the 1‐year clinical outcomes from the registry. The data, analytic methods, and study materials will not be made available to other researchers for purposes of reproducing the results or replicating the procedure.

Procedures and Follow‐Up

All patients were recommended to receive dual antiplatelet therapy (aspirin 81–100 mg/day and clopidogrel 75 mg/day) the day before EVT or earlier. Either ipsilateral or contralateral femoral punctures were performed. After insertion of a 6‐ or 5‐Fr sheath, an intra‐arterial bolus of 5000 IU heparin was injected and supplemented as required to maintain an active clotting time of >200 seconds.

After the lesion was successfully crossed with a 0.035‐, 0.018‐, or 0.014‐inch guidewire, predilatation was performed with a standard angioplasty balloon that was the same size as the reference vessel or 0.5–1 mm smaller. Following successful predilatation (without flow‐limiting dissection and <50% of residual stenosis), 2 types of DCBs were used: Lutonix (Bard, New Hope, MN, USA) and IN.PACT Admiral (Medtronic, Santa Rosa, CA, USA). The DCB type was determined by each operator and the DCB size was the same as the reference vessel diameter. In principle, DCB was used to cover the entire lesion. Postdilation was performed if needed and sized appropriately according to reference vessel diameter. Intravascular ultrasound (IVUS) was performed at the discretion of the operators based on their usual practice. Atherectomy devices are not commercially available in Japan; therefore, they were not used in this study.

After the procedure, lifelong aspirin (81–100 mg/day) and prolonged (at least 1 month) clopidogrel (75 mg/day) were recommended. Restenosis was routinely monitored by duplex ultrasound at 1 month and 1 year, regardless of the presence of ischemic symptoms. The 1‐year follow‐up was scheduled at 12±2 months (ie, between 10 and 14 months) after the procedure.

Outcome Measures

The primary outcome measure was freedom from restenosis and its associated factors. The secondary outcome measures were freedom from TLR, limb salvage rate, freedom from any reintervention and major adverse limb events, and all‐cause mortality.

Definitions

Restenosis was defined as >2.4 times of the peak systolic velocity ratio on duplex ultrasound or >50% of the arterial diameter based on angiography. ¹⁴ Restenosis was classified as focal (<30 mm), tandem (multiple focal), diffuse (>30 mm), and occlusive. An undetectable signal in the treated segments on duplex ultrasound was graded as a complete occlusion. TLR was clinically driven and defined as reintervention performed for lesions with >50% diameter stenosis identified by angiography within ±5 mm of the target lesion after the documentation of recurrent clinical symptoms. ¹⁴ Therefore, all TLR cases were included as restenosis cases. Residual stenosis was defined as ≥25% stenosis after DCB, based on angiography findings. The dissection grade was determined based on angiographic dissection patterns. ¹⁵ Severe dissection was defined as grade D or greater. Limb salvage was freedom from major amputation, which was defined as surgical excision of the limb above the ankle. ¹⁶ Major adverse limb events were defined as a composite of major amputation, bypass conversion, or any reintervention during the study period. Arterial calcification was assessed using angiography before the procedure, and below‐the‐knee runoffs were angiographically assessed after femoropopliteal treatment. Arterial calcification was graded using the peripheral arterial calcium scoring system ¹⁷ and severe calcification was denoted as grade 4. The reference vessel diameter was assessed via angiography at healthy distal sites, free of atherosclerotic plaques. Chronic kidney disease was referred to as an estimated glomerular filtration rate of <60 mL/min per 1.76 m².

Statistical Analysis

Data were presented as mean±SD for continuous variables and as percentage for categorical variables, unless otherwise indicated. A P value <0.05 was considered statistically significant, and 95% CIs were reported when appropriate. Time‐to‐event was estimated using the Kaplan–Meier method. The Kaplan–Meier method was originally used for right‐censored data. In contrast, data on restenosis were interval censored; restenosis, especially asymptomatic restenosis, could not be confirmed without examination. In other words, restenosis was overlooked unless objectively evaluated. When the interval‐censored data were analyzed by the Kaplan–Meier method, cases in which restenosis had developed but was left undetected were treated as those free from restenosis, resulting in an artificially high event‐free rate (or low event incidence rate). As mentioned, in the current registry, the 1‐year follow‐up was scheduled at 12±2 months, and not all participants were expected to complete the 1‐year examination at 12 months. Therefore, we presented the Kaplan–Meier estimate of freedom from restenosis not only at 12 months, but also at 14 months, when the scheduled 1‐year examination was completed.

The association of baseline characteristics with the 1‐year restenosis risk was investigated using the Cox mixed‐effect model, in which the interhospital variability was treated as random effects, whereas baseline characteristics were treated as fixed effects. The R package coxme was used for model development. The hazard ratio (HR) of a baseline characteristic without adjustment for the other baseline characteristics, named an “unadjusted” HR, was obtained from the mixed model, whose fixed effect was the covariate alone. We conveniently named this model the “univariate” model. We also developed a mixed model, in which all baseline characteristics were entered as fixed effects. This model was named the “multivariate” model, and the HRs derived from the model were called adjusted HRs. Missing data were addressed using multiple imputations by the chained equations method during model development. In this procedure, we generated 5 imputed data sets and combined the analytic results based on Rubin's rule. The predictive performance of the multivariate model was evaluated using the C‐statistic, which was calculated as the area under the time‐dependent receiver operating characteristic curve. ¹⁸ Based on the multivariate model, we tentatively developed a simple risk score that was calculated as the number of accumulated independent risk factors. The predictive performance of the risk score was also assessed using the C‐statistic. The 95% CIs and P values for C‐statistics were obtained using a 2000‐time bootstrapping method. All statistical analyses were performed using software R, version 4.1.1 (R Development Core Team, Vienna, Austria).

RESULTS

The patient (n=2507), limb (n=2827), and lesion (n=3165) characteristics of the study population are summarized in Table 1. The mean patient age was 75±9 years, and 64.9% of the study participants were men. The prevalence of patients with diabetes and chronic kidney disease on dialysis was 65.4% and 29.0%, respectively. Chronic limb‐threatening ischemia accounted for 31.2% of the cases. The mean lesion length and distal reference vessel diameter were 13.5±9.3 cm and 4.8±0.9 mm, respectively. The frequencies of chronic total occlusion (CTO) and severe calcification were 25.9% and 14.6%, respectively.

Table 1.

Baseline Characteristics

Patient characteristics	(n=2507)
Age, y	75±9
Male sex	1626 (64.9%)
Nonambulatory	323 (12.9%)
Smoking	516 (20.6%)
Diabetes	1639 (65.4%)
CKD*
None	762 (30.4%)
CKD without dialysis	1015 (40.5%)
CKD on dialysis	728 (29.1%)
Heart failure*	463 (18.5%)
Aspirin use*	1964 (78.9%)
P2Y12 inhibitor use*	2155 (86.3%)
Cilostazol use*	623 (25.2%)
Anticoagulant use*
None	2050 (82.2%)
Warfarin use	191 (7.7%)
Direct oral anticoagulant use	254 (10.2%)
Statin use*	1514 (61.0%)

Limb characteristics	(n = 2827)
Rutherford classification
Category 2	739 (26.1%)
Category 3	1207 (42.7%)
Category 4	291 (10.3%)
Category 5	590 (20.9%)
Ankle‐brachial index*	0.61±0.23
Aortoiliac lesion*	616 (22.0%)
No below‐the‐knee runoff*	357 (12.7%)

Lesion characteristics	(n=3165)
History of EVT
None (de novo)	2370 (74.9%)
1 EVT	474 (15.0%)
≥2 EVTs	321 (10.1%)
In‐stent restenosis*	442 (14.0%)
Popliteal lesion	1069 (33.8%)
Distal reference vessel diameter (mm)*	4.8±0.9
Lesion length (cm)*	13.5±9.3
Severe calcification (peripheral arterial calcium scoring system grade 4)*	463 (14.6%)
Chronic total occlusion*	819 (25.9%)
Lutonix use	765 (24.2%)
Intravascular ultrasound use*	2196 (73.4%)

CKD indicates chronic kidney disease; and EVT, endovascular therapy.

^*<5.5% of values were missing.

Perioperative outcomes are shown in Table 2. Slow flow or no reflow after DCB was observed in 3.9% of the patients. Severe dissection after DCB was found in 4.6% (95% CI, 3.9%–5.4%), and bailout stenting was performed in 3.5% (95% CI, 2.9%–4.2%). Perioperative complications were observed in 3.6% (95% CI, 2.9%–4.3%) of patients, of which the incidence of distal emboli and acute occlusion were similar with 0.6% (95% CI, 0.3%–0.9%).

Table 2.

Perioperative Outcomes

Clinical outcome	Estimate [95% CI]
Perioperative outcomes
Residual stenosis (≥25%)	26.4% [24.8%–27.9%]
Blood flow after DCB
No change	96.0% [95.4%–96.7%]
Slow flow	2.8% [2.2%–3.4%]
No reflow	1.1% [0.8%–1.5%]
Dissection after DCB
None	33.4% [31.8%–35.1%]
Grade A	27.1% [25.5%–28.6%]
Grade B	23.6% [22.2%–25.1%]
Grade C	11.2% [10.1%–12.3%]
Grade D	3.9% [3.2%–4.5%]
Grade E	0.6% [0.4%–0.9%]
Grade F	0.1% [0.0%–0.3%]
Dissection grade C or greater	15.8% [14.6%–17.1%]
Dissection grade D or greater	4.6% [3.9%–5.4%]
Bailout stenting	3.5% [2.9%–4.2%]
Mean ankle‐brachial index after the procedure	0.89 [0.88–0.89]
Perioperative complication	3.6% [2.9%–4.3%]
Perioperative death	1.0% [0.6%–1.4%]
Target lesion revascularization (endovascular therapy)	0.7% [0.4%–1.0%]
Target lesion revascularization (bypass)	0.0% [0.0%–0.1%]
Distal embolism	0.6% [0.3%–0.9%]
Transfusion for bleeding	0.6% [0.3%–0.9%]
Acute occlusion	0.6% [0.3%–0.8%]
Vessel rupture	0.1% [0.0%–0.2%]
Blue toe syndrome	0.1% [0.0%–0.3%]
Major amputation	0.2% [0.0%–0.4%]
Myocardial infarction	0.2% [0.0%–0.4%]
Stroke	0.1% [0.0%–0.3%]
Renal impairment	0.2% [0.0%–0.3%]

DCB indicates drug‐coated balloon.

Outcome Measures

During the follow‐up period, 172 patients died and 206 patients were lost to follow‐up. One‐year follow‐up rate was 84.9% (2129/2507) on a patient basis. The median follow‐up period was 14.2 (interquartile range, 10.1–23.0) months. Restenosis was detected in 811 patients, of whom 545 were found to have restenosis in the first year post procedure. Figures 1 and 2 illustrate the lesion, limb, and patient prognoses after EVT. The Kaplan–Meier estimate of freedom from restenosis was 84.5% (95% CIs, 83.1%–85.8%) at 12 months (79.7% [95% CI, 78.1%–81.2%] at 14 months). Focal, tandem, diffuse, and occlusive restenosis accounted for 37.4% (95% CI, 29.9%–44.9%), 9.8% (95% CI, 8.4%–11.2%), 18.9% (95% CI, 17.0%–20.9%), and 33.9% (95% CI, 26.0%–41.8%) of restenosis cases, respectively (Figure 1A). The 12‐month rate of freedom from TLR was 91.5% (95% CI, 90.5%–92.5%) (Figure 1B).

Figure 1. Freedom from restenosis and TLR.

A, Kaplan–Meier estimates of 1‐year freedom from restenosis and morphology of restenosis. B, Kaplan–Meier estimates of 1‐year freedom from TLR. Dotted lines represent 95% CIs. TLR indicates target lesion revascularization.

Figure 2. Limb salvage, freedom from any reintervention and MALE, and all‐cause mortality.

A, Kaplan–Meier estimates of 1‐year limb salvage. B, Kaplan–Meier estimates of 1‐year freedom from any reintervention. C, Kaplan–Meier estimates of 1‐year freedom from MALE. D, Kaplan–Meier estimates of 1‐year all‐cause mortality. Dotted lines represent 95% CIs. MALE indicates major adverse limb event (major amputation, bypass conversion, and reintervention).

The 12‐month Kaplan–Meier estimates of limb salvage (Figure 2A), freedom from any reintervention (Figure 2B), and freedom from major adverse limb events (Figure 2C) was 98.6% (95% CI, 98.2%–99.1%), 91.1% (95% CI, 90.0%–92.2%), and 90.0% (95% CI, 88.9%–91.2%), respectively. The 12‐month cumulative incidence rate of all‐cause mortality was 9.5% (95% CI, 7.4%–9.7%) (Figure 2D).

Table 3 demonstrates the association between baseline characteristics and the 1‐year risk of restenosis. Baseline characteristics that were independently associated with 1‐year restenosis were history of revascularization (1.32 [95% CI, 1.01–1.73] for 1 EVT and 1.70 [95% CI, 1.23–2.34] for more EVTs versus de novo lesions; P=0.044 and 0.001), distal reference vessel diameter (0.87 [95% CI, 0.78–0.97] per 1‐mm increase; P=0.012), severe calcification (1.29 [95% CI, 1.03–1.63]; P=0.027), CTO (1.28 [95% CI, 1.04–1.58]; P=0.021), Lutonix use (1.97 [95% CI, 1.61–2.41] versus IN.PACT Admiral use; P<0.001), and residual stenosis (1.51 [95% CI, 1.24–1.83]; P<0.001). The C‐statistic of the multivariate model was (0.71 [95% CI, 0.68–0.74]; P<0.001) (Figure 3A). The C‐statistic of the simple risk score (ie, the number of accumulated independent risk scores) was 0.66 (95% CI, 0.63–0.69); it was smaller than that of the multivariate model (the difference, −0.05 [95% CI, −0.07 to −0.03]; P<0.001), but was still significantly larger than 0.5 (P < 0.001) (Figure 3A). Accumulation of these risk factors was associated with a lower rate of freedom from restenosis (Figure 3B).

Table 3.

Association of Baseline Characteristics With 1‐Year Restenosis Risk

	Unadjusted hazard ratio	Adjusted hazard ratio
Age (per 10 y)	0.88 [0.81–0.97] (P=0.008)	0.91 [0.82–1.01] (P=0.071)
Male sex	1.05 [0.87–1.25] (P=0.62)	1.18 [0.97–1.42] (P=0.093)
Nonambulatory	1.46 [1.14–1.86] (P=0.003)	1.15 [0.88–1.51] (P=0.30)
Smoking	0.97 [0.78–1.20] (P=0.76)	0.94 [0.75–1.18] (P=0.57)
Diabetes	1.18 [0.98–1.42] (P=0.073)	1.07 [0.88–1.30] (P=0.48)
CKD
CKD without dialysis	0.88 [0.71–1.10] (P=0.27)	0.92 [0.74–1.16] (P=0.49)
CKD on dialysis	1.48 [1.20–1.82] (P<0.001)	1.19 [0.93–1.51] (P=0.16)
Heart failure	0.98 [0.79–1.23] (P=0.89)	0.83 [0.65–1.04] (P=0.11)
Aspirin use	0.93 [0.75–1.15] (P=0.48)	0.96 [0.75–1.23] (P=0.73)
P2Y12 inhibitor use	0.97 [0.74–1.26] (P=0.80)	0.99 [0.74–1.33] (P=0.96)
Cilostazol use	0.90 [0.73–1.12] (P=0.35)	0.87 [0.68–1.11] (P=0.26)
Anticoagulant use
Warfarin use	1.36 [1.02–1.80] (P=0.034)	1.21 [0.89–1.63] (P=0.22)
Direct oral anticoagulant use	0.86 [0.63–1.17] (P=0.33)	0.93 [0.66–1.31] (P=0.68)
Statin use	0.93 [0.78–1.10] (P=0.39)	0.99 [0.82–1.19] (P=0.89)
Rutherford classification	1.27 [1.18–1.38] (P<0.001)	1.09 [0.99–1.20] (P=0.073)
Aortoiliac lesion	1.12 [0.91–1.37] (P=0.27)	1.17 [0.95–1.44] (P=0.14)
No below‐the‐knee runoff	1.63 [1.30–2.04] (P<0.001)	1.26 [0.99–1.60] (P=0.065)
History of revascularization
1 EVT	1.28 [1.02–1.60] (P=0.035)	1.32 [1.01–1.73] (P=0.044)
≥2 EVTs	1.71 [1.33–2.20] (P<0.001)	1.70 [1.23–2.34] (P=0.001)
In‐stent restenosis	1.18 [0.94–1.48] (P=0.15)	0.91 [0.66–1.24] (P=0.54)
Popliteal lesion	1.53 [1.29–1.82] (P<0.001)	1.20 [0.99–1.45] (P=0.063)
Reference vessel diameter (per 1 mm)	0.76 [0.69–0.83] (P<0.001)	0.87 [0.78–0.97] (P=0.012)
Lesion length (per 10 cm)	1.22 [1.12–1.34] (P<0.001)	1.09 [0.98–1.22] (P=0.098)
Severe calcification	1.63 [1.31–2.02] (P<0.001)	1.29 [1.03–1.63] (P=0.027)
Chronic total occlusion	1.42 [1.18–1.71] (P<0.001)	1.28 [1.04–1.58] (P=0.021)
Lutonix use	2.06 [1.69–2.51] (P<0.001)	1.97 [1.61–2.41] (P<0.001)
Intravascular ultrasound use	0.72 [0.57–0.91] (P=0.007)	0.80 [0.64–1.01] (P=0.056)
Residual stenosis	1.69 [1.40–2.05] (P<0.001)	1.51 [1.24–1.83] (P<0.001)
Blood flow after drug‐coated balloon (vs no change)
Slow flow	0.72 [0.38–1.36] (P=0.31)	0.63 [0.33–1.19] (P=0.16)
No reflow	1.42 [0.66–3.05] (P=0.37)	1.55 [0.72–3.35] (P=0.26)
Dissection	1.09 [1.02–1.18] (P=0.016)	1.04 [0.96–1.13] (P=0.29)

Data are hazard ratios [95% CIs] (P values), derived from the Cox mixed‐effect model in which the intersubject variability was treated as the random effects. Severe calcification was referred to as peripheral arterial calcium scoring system classification grade 4, whereas severe dissection indicated grade C or more severe. CKD indicates chronic kidney disease; EVT, endovascular therapy; and N/I, not included.

Figure 3. Prediction of 1‐year restenosis risk.

The simple risk score was calculated as the number of accumulated independent risk factors, that is, history of EVT, distal reference vessel diameter <5 mm, severe classification (PACSS classification grade 4), chronic total occlusion, Lutonix use, and residual stenosis (see Table 3). Kaplan–Meier estimates of 1‐year primary patency by the simple risk score. Dotted lines represent 95% CIs. EVT indicates endovascular therapy; and PACSS, peripheral arterial calcium scoring system.

Rutherford classification and ankle‐brachial index significantly improved at 1‐year follow‐up (Table 4).

Table 4.

Rutherford Classification and Ankle‐Brachial Index at 1 Year Versus Baseline

	Baseline	1 year	P value
Rutherford classification			<0.001
Category 0	0.0% [0.0%–0.0%]	46.8% [44.3%–49.2%]
Category 1	0.0% [0.0%–0.0%]	26.5% [24.5%–28.5%]
Category 2	25.7% [24.2%–27.2%]	9.6% [8.4%–10.8%]
Category 3	42.4% [40.7%–44.2%]	5.3% [4.0%–6.5%]
Category 4	10.7% [9.6%–11.8%]	3.1% [2.0%–4.2%]
Category 5	21.2% [19.7%–22.6%]	7.8% [6.4%–9.2%]
Category 6	0.0% [0.0%–0.0%]	1.0% [0.2%–1.7%]
Mean ankle‐brachial index	0.61 [0.60–0.61]	0.85 [0.83–0.87]	<0.001

Data are estimates [95% CIs], derived from the multiple imputation.

DISCUSSION

Our multicenter, prospective, observational study documented the 1‐year clinical outcomes of DCB treatment for femoropopliteal lesions. The study included a large number of patients (3165 lesions in 2507 patients) and used real ‐world registry data. It included patients with both short femoropopliteal lesions and CTOs, which have traditionally not been included in prior studies. The proportion of bailout stenting in this study was extremely low (3.5%), reflecting the fact that bailout stenting was not reimbursed in Japan. In addition, atherectomy devices have not yet been approved. Therefore, we believe that the results of this study illustrate the true efficacy of DCBs in femoropopliteal lesions. The 1‐year rates of freedom from restenosis and TLR were clinically favorable. It was also of note that two thirds of restenosis cases were nonocclusive.

The restenosis morphology after BNS implantation was reported to be focal (≤5 cm) in 29%, diffuse (>5 cm) in 38%, and occlusive in 33%, with approximately two thirds showing the stenotic pattern. ¹⁹ Patency after EVT for stenotic restenosis was better than that for occlusive restenosis, ¹⁹ suggesting that the stenotic pattern is the preferred form of restenosis. In this study, approximately two thirds of the restenosis cases presented with a stenotic pattern. Reintervention would be easier and the patency rate would be higher for stenotic restenosis than for occlusive restenosis. The treatment strategy (DCB, DES, etc.) for stenotic and occlusive restenosis after DCB remains unclear and further studies are needed.

It has been reported that DCB does not work well for circumferential calcification. ²⁰ In this study, severe calcification was a predictor of restenosis. Residual stenosis (>25%) was also a significant predictor of patency loss. DCB alone may be insufficient to treat patients with highly calcified lesions that were difficult to be dilated by predilatation.

In this study, factors such as the Rutherford category class, CTO, and ankle‐brachial index were not independent predictors of 1‐year freedom from restenosis. The results were considered favorable compared with those of the previous BNS era (the 12‐month patency efficacy goal after femoropopliteal BNS was 66%). ²¹ Favorable 1‐year clinical outcomes after the latest DES have been reported, despite the more complex lesions (1‐year freedom from restenosis, 87.1%). ²² Although this study was not a direct comparison to DES, it may be considered a treatment option for more complex lesions.

This study showed that patency rate was reduced in patients with risk factors of multiple restenosis (Figure 3). However, the 1‐year patency rate reached approximately 80%, even in those with 2 restenosis risks (85.8% at 12 months and 80.3% at 14 months), which was considered clinically acceptable. Considering the performance goal of the 1‐year primary patency after femoropopliteal bare‐nitinol stenting, ²¹ the efficacy of DCBs may not be sufficient in patients with 3 (65.8% at 14 months) or more risk factors (58.6% at 14 months). Such risk stratification systems play an important role in deciding the treatment strategies for individual cases. It remains unclear whether DES would be effective in cases in whom a lower rate of freedom from restenosis after DCB use was expected. Further studies are required to address this issue.

Although there are some reports of pseudoaneurysms after DCB treatment, ²³ , ²⁴ this study was not designed to evaluate aneurysmal changes after DCB. Further studies are required to clarify these details.

One reason for the favorable outcomes in this study would be the common use of IVUS (73.4%). The study had a high rate of IVUS usage for sizing (73%). In the present multivariate analysis, the use of IVUS was marginally but not significantly associated with a reduced risk of restenosis (Figure S1). It has been reported that there is a discrepancy in vessel diameter by approximately 1 mm between IVUS and angiography; this discrepancy is marked in small vessels. ²⁵ The use of IVUS may provide adequate dilatation. ²⁶ In addition to CTO and circumferential calcification, postprocedural IVUS‐measured minimal lumen area has been reported to be a significant predictor of restenosis after DCB. This report recommended a minimal lumen area ≥12.7 mm^{2. 27} The use of IVUS might help in the selection of DCBs with diameters larger than the angiographically measured vessel diameter, which may have contributed to the reduction of residual stenosis. In the future, it is necessary to investigate the IVUS parameters associated with restenosis risk. In terms of medication, anticoagulant and statin were not significantly associated with a reduced risk of restenosis (Figure S2 and S3).

Low‐dose DCB (Lutonix) use in comparison with high‐dose DCB (IN.PACT Admiral) use was associated with an increased risk of restenosis (Table 3). Indeed, the 12‐month Kaplan–Meier estimate of primary patency was72.7% (69.3% to 76.2%) for Lutonix versus 88.0% (86.7% to 89.4%) for IN.PACT Admiral (P<0.001), and the corresponding estimate of freedom from TLR was 83.8% (81.0% to 86.7%) for Lutonix versus 93.8% (92.8% to 94.8%) for IN.PACT Admiral (P<0.001) (data not shown). The results were similar to the 1‐year results of the respective DCB clinical trials conducted in Japan. ²⁸ , ²⁹ , ³⁰ It has been reported that high‐dose DCBs are associated with a lower risk of restenosis and TLR. ³¹ , ³² The results of the current study would support those findings. Because our study design did not directly compare the 2 types of DCBs, a randomized clinical trial is needed to confirm the results.

The number of patients with severe calcification (peripheral arterial calcium scoring system grade IV) was only 14.6% as compared with 46% in IN. PACT SFA trial. ³³ It was considered that the main reason for this was that there is no reimbursement for bailout stent in Japan. Therefore, successful standard percutaneous transluminal angioplasty (<50% of residual stenosis without severe dissection) was needed to avoid stenting. If suboptimal result occurred after percutaneous transluminal angioplasty, the treatment strategy would be changed to stent‐first. It is also known that the Japanese population has less incidence of medial calcinosis; the current population was associated with less calcification and good outcomes despite not using atherectomy. As known severe calcification was an independent predictor of restenosis it should be kept in mind that these results cannot be generalized to other populations with a higher degree of vascular calcium.

There was substantial loss to follow‐up and 1‐year follow‐up was 84.9% of patients. All‐cause mortality at 1 year in the current study was higher (9.5%) compared with other clinical trials (2.5%–3%). ³³ , ³⁴ However, this study was a real‐world clinical study, with older mean age (75±9 years), more diabetes (65.4%), and more chronic kidney disease (69.6%). This may have affected the mortality rate.

Limitations

This study has several limitations that may have affected the clinical outcomes. First, although we enrolled almost all consecutive patients who underwent DCB, we did not enroll patients with unsuccessful predilatation. In addition, bailout stents were not reimbursed and complex lesions that required stenting may have been minimally enrolled. This likely introduced selection bias, as this is an underestimation of what would be seen in the US or European practice. Unfortunately, we could not eliminate the selection bias completely. Furthermore, in order to clarify the true effectiveness of DCBs, a contemporary cohort of patients with primary percutaneous transluminal angioplasty in whom the same standards for inclusion into the registry were needed. Second, core laboratories were not used in this study. Although this may compromise the reliability of restenosis and other clinical assessments, participating sites had sufficient experience in clinical trials, and we believe their experience will minimize variability. Finally, only 2 types of DCBs (Lutonix and IN. PACT Admiral) were used in the treatment. It is not clear whether similar results can be obtained with other types of DCBs. Additionally, in this study, freedom from restenosis and its associated factors was selected as the primary end point in contrast to the prior trails. In the IN.PACT SFA (Randomized Trial of IN.PACT Admiral Drug Coated Balloon vs Standard PTA for the Treatment of SFA and Proximal Popliteal Arterial Disease) trial, ³³ the primary outcome was clinically driven TLR, and in the LEVANT 2 (A Prospective, Multicenter Registry of the Lutonix Drug Coated Balloon for Treatment of Femoropopliteal Arteries) trial ³⁴ was primary patency of the target lesion (which included binary restenosis and freedom from TLR). These differences in primary end points among studies are also an important limitation for comparative evaluation.

CONCLUSIONS

Our data demonstrated 1‐year freedom from restenosis after DCB therapy for femoropopliteal lesions in a real‐world setting. Independent predictors of 1‐year restenosis were a history of revascularization, smaller distal reference vessel diameter, severe calcification (peripheral arterial calcium scoring system grade 4), Lutonix use (versus IN.PACT Admiral use), and residual stenosis. The presence of these risk factors was associated with a lower rate of freedom from restenosis.

Sources of Funding

The POPCORN study was supported by the Research Association for Lower Limb Artery Revascularization (LIBERAL).

Disclosures

Drs Soga, Iida, Kawasaki, and Kozuki received honoraria from BARD Japan and Medtronic Japan. Drs Tobita, Dr Fujihara, and Ando received honoraria from Medtronic Japan. Dr Tanaka received honoraria from BARD Japan. The remaining authors have no disclosures to report.

Supporting information

Data S1

Figures S1–S3