Safety of paclitaxel-coated devices in patients with peripheral artery disease: an updated systematic review and meta-analysis of randomized controlled trials

Summary

Background

Paclitaxel-coated devices have been widely used to reduce restenosis in patients with peripheral artery disease (PAD). However, their long-term safety, particularly regarding mortality and amputation risks, remains controversial.

Methods

A systematic search of PubMed, Web of Science, the Cochrane Library, and Embase was conducted to identify studies published from the inception of each database through September 2025. Randomized controlled trials (RCTs) comparing paclitaxel-coated devices with noncoated devices in patients with PAD were included with a minimum clinical follow-up duration of 24 months. Primary outcomes were all-cause mortality and limb amputation. Secondary outcomes included target lesion revascularization (TLR) and primary patency. Risk ratios (RRs) with 95% confidence intervals (CIs) were pooled using fixed- or random-effects models. Heterogeneity was assessed using the I² statistic. Risk of bias was evaluated with the Cochrane tool. This PROSPERO-registered review (CRD420251155841) was conducted between September and December 2025.

Findings

A total of 15 RCTs involving 5859 patients were included. No significant differences were observed in all-cause mortality between the paclitaxel and control groups at 1 year (RR: 1.04, 95% CI: 0.87–1.24), 2 years (RR: 1.28, 95% CI: 0.95–1.74), or 5 years (RR: 1.13, 95% CI: 0.93–1.38). Similarly, no significant differences in amputation rates were observed at 1 year (RR: 1.07, 95% CI: 0.88–1.31), 2 years (RR: 0.65, 95% CI: 0.34–1.24), or 5 years (RR: 1.03, 95% CI: 0.88–1.20). Paclitaxel-coated devices significantly reduced TLR at 1 year (RR: 0.64, 95% CI: 0.48–0.87) and 2 years (RR: 0.45, 95% CI: 0.38–0.54), but this benefit was not sustained at 5 years (RR: 0.81, 95% CI: 0.64–1.01). Primary patency at 2 years was significantly improved with paclitaxel devices (RR: 1.57, 95% CI: 1.24–1.99).

Interpretation

This updated meta-analysis indicates that paclitaxel-coated devices are not associated with increased risks of all-cause mortality or major amputation for up to 5 years. Although these devices offer mid-term benefits in reducing TLR and improving patency, these advantages decrease over the long term. These findings support the continued use of paclitaxel-coated devices in selected patients with symptomatic femoropopliteal disease, while emphasizing the need for individualized decision-making and long-term clinical follow-up.

Funding

None.

Research in context.

Evidence before this study

Paclitaxel-coated devices (drug-coated balloons and drug-eluting stents) are widely used to reduce restenosis after endovascular revascularisation in patients with peripheral artery disease. However, their long-term safety remains controversial. A 2018 meta-analysis by Katsanos and colleagues raised concerns about increased late mortality associated with these devices, prompting regulatory alerts and widespread debate. Subsequent studies, including the SWEDEPAD trials and individual patient data meta-analyses, have reported conflicting results. We conducted an updated systematic review and meta-analysis of randomized controlled trials (RCTs) to reassess the long-term safety and efficacy of paclitaxel-coated devices, with a literature search up to September 2025.

Added value of this study

To our knowledge, this is the most updated meta-analysis incorporating the largest and most contemporary RCTs, including the recently published SWEDEPAD 1 and SWEDEPAD 2 trials, which together contribute over 3500 patients with up to 5 years of follow-up. Our analysis of 15 RCTs involving 5859 patients found no significant increase in all-cause mortality or major amputation with paclitaxel-coated devices at 1, 2, or 5 years. While these devices reduced target lesion revascularisation and improved primary patency at 2 years, the TLR benefit was no longer significant at 5 years. These findings provide updated reassurance on long-term safety while highlighting the attenuation of efficacy over time.

Implications of all the available evidence

The totality of current evidence suggests that paclitaxel-coated devices are safe with respect to mortality and amputation risks up to 5 years, and offer meaningful short-to mid-term efficacy benefits. However, the waning of these benefits over time supports the need for long-term clinical surveillance and individualised treatment decisions. These findings support the continued use of paclitaxel-coated devices in appropriately selected patients with symptomatic femoropopliteal disease, while future research should focus on optimising device technology and patient selection to improve durable outcomes.

Introduction

Peripheral artery disease (PAD) is a major cause of morbidity and mortality globally, affecting more than 230 million individuals.¹ It is characterized by atherosclerotic narrowing or occlusion of the arteries, leading to symptoms ranging from intermittent claudication to chronic limb-threatening ischemia.^2,3 Revascularization procedures such as percutaneous transluminal angioplasty (PTA) and stenting are commonly performed to improve blood flow and alleviate symptoms in patients with PAD.^4,5 However, restenosis remains a considerable challenge, often necessitating repeat interventions.⁶

Paclitaxel-coated devices, including drug-coated balloons and drug-eluting stents, have been developed to reduce restenosis and improve long-term patency.^7,8 Several randomized controlled trials (RCTs) and meta-analyses have demonstrated the efficacy of these devices in reducing the need for target lesion revascularization (TLR) and improving primary patency when compared with the use of uncoated devices.9, 10, 11, 12, 13 However, concerns about the safety of these devices have emerged.

In 2018, a meta-analysis by Katsanos et al.¹⁴ raised concerns about a potential increase in late mortality with paclitaxel-coated devices. This finding sparked widespread debate and prompted regulatory agencies to issue safety alerts. Subsequent studies, including the SWEDEPAD 1 trial and an additional meta-analysis, have provided conflicting results regarding the association between paclitaxel-coated devices and mortality.^15,16

Given the conflicting evidence and the importance of balancing efficacy and safety in clinical practice, we performed an updated systematic review and meta-analysis with the goal of providing a comprehensive evaluation of the safety and efficacy of paclitaxel-coated devices in patients with PAD. The analysis includes data from multiple RCTs and extends the follow-up period to provide a more accurate assessment of long-term outcomes. The findings should help to inform clinical decision-making and guide the appropriate use of paclitaxel-coated devices in the management of PAD.

Methods

This meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines.¹⁷ This systematic review protocol was registered with the PROSPERO database of systematic reviews (CRD420251155841; www.crd.york.ac.uk/PROSPERO/).

Search strategy

We searched PubMed, Web of Science, the Cochrane Library, and Embase for studies published between the inception of each database through September 2025. The search strategy incorporated Medical Subject Headings (MeSH) and free-text terms related to “paclitaxel,” “paclitaxel-eluting balloons,” “paclitaxel-coated balloons,” “paclitaxel-eluting stents,” “paclitaxel-coated stents,” “peripheral artery disease,” “peripheral vascular disease,” “femoral artery,” “popliteal artery,” “femoropopliteal artery,” “ischemia,” “intermittent claudication,” “amputation,” and “randomized.” Boolean operators (AND/OR) were used to combine terms. The detailed search strategy is shown in Supplementary Table S1. The search was limited to human studies and RCTs published in English. The reference lists of included studies and relevant reviews were manually screened.

Inclusion and exclusion criteria

Two independent investigators (Y.L. and S.C.) performed the study selection, including title/abstract screening and full-text assessment, with any disagreements resolved through consensus or by consultation with a third reviewer (Z.J.). The eligibility of studies was assessed based on the prespecified PICOS framework.

Studies were included if they were RCTs that enrolled adult patients with symptomatic PAD (Rutherford categories 1–6) undergoing endovascular intervention for femoropopliteal artery lesions. The intervention of interest was the use of a paclitaxel-coated device (either a drug-coated balloon or a drug-eluting stent), which must have been compared against a non-paclitaxel-coated control device such as plain balloon angioplasty or a bare-metal stent. To be eligible, studies were required to report on at least one of the prespecified primary or secondary outcomes—namely, all-cause mortality, major limb amputation, TLR, or primary patency—and to have a minimum clinical follow-up duration of 24 months to ensure the assessment of mid-to long-term outcomes.

Conversely, studies were excluded for the following reasons: nonrandomized design; a focus on coronary artery interventions; duplicate publications or secondary analyses of already included trials without novel outcome data; unavailability of full text or inability to extract relevant outcome data despite contact with authors; the use of a paclitaxel-coated device or another active antiproliferative agent in the control group; and publications such as abstracts, conference proceedings, or review articles that did not present original trial data.

Data extraction and endpoints

Following the initial screening and eligibility assessment, a standardized data extraction process was implemented to ensure the systematic and accurate collection of relevant information from the included studies. Data extraction was independently conducted by Y.L. and L.Q. using a predesigned form, and discrepancies were resolved by discussion with J.Z.

For all studies, information was collected regarding first author, year of publication, trial name, registration number, study design, randomization methodology, blinding procedures, follow-up duration, sample size, and baseline patient demographics.

The endpoints for this meta-analysis were predefined. The primary safety endpoint was all-cause mortality, defined as death from any cause occurring at any time after the index procedure. The secondary safety endpoint was limb amputation, defined as an amputation performed above the ankle on the index limb. Efficacy was assessed through two primary measures: TLR, which encompassed any repeat percutaneous intervention or surgical bypass of the original target lesion; and primary patency, defined as the absence of clinically driven restenosis within the treated segment without the need for repeat revascularization. Data for these endpoints were extracted at all reported time intervals to facilitate a comprehensive analysis of both short- and long-term outcomes.

Risk of bias

The risk of bias assessment was independently performed by Y.L. and T.X. using the Cochrane Collaboration's tool, with disagreements resolved by consulting Z.J.¹⁸ This tool evaluates seven domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other potential sources of bias. Each domain was judged as “low risk,” “high risk,” or “unclear risk” of bias.

Statistical analysis

Dichotomous outcomes were pooled using risk ratios (RRs) with 95% confidence intervals (CIs). Continuous outcomes were analyzed by calculating the pooled mean difference (MD) and 95% CIs. Heterogeneity across studies was assessed using the Cochran’s Q test and quantified by the I² statistic and tau-squared (τ²). A fixed-effects model was applied if I² < 50% and the chi-squared test P value was >0.10; otherwise, a random-effects model was used.¹⁹ When significant heterogeneity was observed, we conducted sensitivity analyses to test the stability of the pooled results. Publication bias was assessed visually with funnel plots and statistically with Egger’s test. A P value < 0.05 with Egger’s test was considered suggestive of significant publication bias. All statistical analyses were performed using Review Manager (RevMan) software, version 5.4 (The Cochrane Collaboration, 2020) and Stata, version 15.1 (StataCorp LLC). All statistical tests were two-sided, and a P value < 0.05 was considered statistically significant.

Role of the funding source

This research received no external funding. The authors had full access to all of the data in the study and had final responsibility for the decision to submit for publication.

Results

Characteristics of included RCTs

A total of 15 RCTs were included in the systematic review and meta-analysis.^10,12,15,20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 The PRISMA flowchart illustrating the study selection process is depicted in Fig. 1. The main characteristics of the included studies, which collectively enrolled a total of 5859 patients, are summarized in Table 1. The maximum follow-up duration varied across studies: 8 RCTs reported outcomes up to 2 years; 2 RCTs, up to 3 years; and 5 RCTs, up to 5 years.

Fig. 1.

Flowchart of study identification and selection.

Table 1.

Studies included in the meta-analysis and characteristics of participants.

Study	Enrollment timeframe	Study design	Region	Allocation in study arms	Maximum follow-up period	Inclusion criteria	Paclitaxel-coated device
AcoArt I20	2013–2014	Multicenter single-blind (1:1)	China	DCB (n = 100) vs. PTA (n = 100)	5 years	•Rutherford class 2–5 •Femoropopliteal artery lesions ≤40 cm	Orchid by Acotec Scientific
BATTLE21	2014–2016	Multicenter open-label (1:1)	France	DES (n = 85) vs. BMS (n = 84)	2 years	•Rutherford class 2–5 •Target lesion length between 2 and 14 cm •De novo atherosclerotic lesions (stenosis or occlusion) in the SFA and/or proximal popliteal artery	ZILVER-paclitaxel (PTX) Stent by Cook Medical
BIOPAC22	2014–2019	Multicenter single-blind (1:1)	Poland	DCB (n = 33) vs. PTA (n = 33)	3 years	•Rutherford class 2B-5 •De novo or nonstent restenotic lesions •Lesion length >4 cm and <15 cm	Microcrystalline paclitaxel-coated balloon, PAK; Balton
CONSEQUENT23	2013–2015	Multicenter single-blind (1:1)	Germany	DCB (n = 78) vs. PTA (n = 75)	2 years	•Rutherford class 2–4 •Lesion length 4–27 cm •De novo, restenotic, or occluded lesions (excluding in-stent restenosis)	SeQuent Please by B. Braun Late
EffPac24	2015–2016	Multicenter single-blind (1:1)	Germany	DCB (n = 85) vs. PTA (n = 86)	2 years	•Rutherford class 2–5 •Single stenosis or occlusion ≤150 mm in length	Luminor by iVascular
ILLUMENATE12	2012–2015	Multicenter single-blind (3:1)	Germany and Austria	DCB (n = 222) vs. PTA (n = 72)	2 years	•Claudication or rest pain due to ≥70% stenosis of the superficial femoral and/or popliteal artery	Stellarex by Spectranetics
IN.PACT SFA25	2010–2013	Multicenter single-blind (2:1)	United States, European Union	DCB (n = 220) vs. PTA (n = 111)	3 years	•Rutherford class 2–4 •Stenosis 70%–99% or total occlusion (≤10 cm) •Lesion length 4–18 cm in superficial femoral or proximal popliteal artery	IN.PACT Admiral by Medtronic
ISAR-PEBIS26	2010–2013	Two-center open-label (1:1) for ISR	Germany	DCB (n = 36) vs. PTA (n = 34)	2 years	•Rutherford class 2–5 •ISR > 70% or occlusion of the SFA	IN.PACT Admiral by Medtronic
ISAR-STATH29	2009–2013	Two-center open-label (1:1:1)	Germany	DCB + BMS (n = 48) vs. PTA + BMS (n = 52)	2 years	•Rutherford class 2–6 •Symptomatic de novo stenosis >70% or occlusion of the SFA	IN.PACT Admiral by Medtronic
LEVANT I30	2009	Multicenter single-blind (1:1)	Germany, Belgium	DCB (n = 49) vs. PTA (n = 52)	2 years	•Rutherford class 2–5 •Symptomatic de novo or nonstent restenotic femoropopliteal lesions •Lesion length ≥4 cm and ≤15 cm	Lutonix Drug-Coated Balloon
MDT-211310	2013–2015	Multicenter single-blind (2:1)	Japan	DCB (n = 68) vs. PTA (n = 32)	2 years	•Rutherford class 2–4 •Stenosis severity 70%–99%	IN.PACT Admiral by Medtronic
PADI31	2007–2013	Multicenter open-label (1:1)	The Netherlands	DES (n = 73) vs. PTA-BMS (n = 64)	5 years	•Rutherford class ≥4 •Infrapopliteal lesions (below the knee)	TAXUS Liberté paclitaxel-eluting stent
SWEDEPAD 115	2014–2023	Multicenter participant-masked (1:1)	Sweden	Paclitaxel-coated devices (n = 1206) vs. uncoated devices (n = 1194)	5 years	•Rutherford class 4–6 •Eligible for infrainguinal endovascular revascularization	CE-marked paclitaxel-coated balloons and stents
SWEDEPAD 228	2014–2023	Multicenter participant-masked (1:1)	Sweden	Paclitaxel-coated devices (n = 565) vs. uncoated devices (n = 571)	5 years	•Rutherford class 1–3 •Eligible for infrainguinal endovascular revascularization	CE-marked paclitaxel-coated balloons and stents
Zilver PTX27	2005–2008	Multicenter open-label (1:1)	United States, Japan, Germany	DES (n = 241) vs. PTA (n = 238)	5 years	•Rutherford class 2–5 •≥50% diameter stenosis •Reference vessel diameter 4–9 mm •Lesion length ≤14 cm	ZILVER-PTX Stent by Cook Medical

DCB, drug-coated balloon; PTA, Percutaneous Transluminal Angioplasty; DES, drug-eluting stent; BMS, bare-metal stent; ISR, In-Stent Restenosis; SFA; Superficial Femoral Artery.

Baseline patient demographics and lesion characteristics were largely comparable across the included studies. Most of the study population consisted of patients with intermittent claudication (Rutherford category 1–3; n = 3175), whereas a smaller proportion presented with chronic limb-threatening ischemia (Rutherford category 4–6; n = 2711). Overall, 61.0% (3575/5859) of the participants were male, and the mean age ranged from 65.5 to 77.0 years. A pooled analysis of baseline characteristics demonstrated no significant difference in the proportion of male patients between the paclitaxel-coated device group and the control group (61.4% [1889/3079] vs. 60.6% [1686/2780]; P = 0.87). The mean age also did not differ significantly between groups (MD: −0.20, 95% CI: −0.96 to 0.57, P = 0.62). Heterogeneity for these comparisons was low (sex: I² = 9%, P = 0.35; age: I² = 0%, P = 0.56). Most studies involved lesions of moderate length. With few exceptions, the protocols followed in the studies recommended a short course (1–3 months) of dual antiplatelet therapy after the intervention.

The design features of the included RCTs are detailed in Table 1. Briefly, the paclitaxel-based devices were predominantly used for the treatment of femoropopliteal artery disease. The devices included a range of paclitaxel-coated balloons (e.g., IN.PACT Admiral, Stellarex, Luminor, Orchid) and stents (ZILVER PTX, TAXUS Liberté). All trials compared paclitaxel-coated devices against noncoated control devices (plain balloon angioplasty or bare-metal stent) and were designed to report on intention-to-treat outcomes for efficacy and safety endpoints. Data were available for the endpoints of primary patency, TLR, all-cause mortality, and amputation at various follow-up time points.

Risk of bias assessment

The assessments for each study are summarized in Supplementary Fig. S1. Overall, the risk of bias arising from the randomization process and allocation concealment was low for most of the included trials, indicating adequate sequence generation and concealment mechanisms.

Blinding of participants and personnel (performance bias) was a significant source of potential bias. Most studies were classified as having a high risk of bias in this domain due to the inherent challenge of masking the intervention. However, the risk of bias for the blinding of outcome assessment (detection bias) was generally low for the endpoints of all-cause mortality and major amputation.

Regarding incomplete outcome data, the risk of bias was low for most studies, as the studies reported low rates of loss to follow-up and conducted intention-to-treat analyses. The risk of selective reporting bias was also low across the studies, as data for all prespecified outcomes for this meta-analysis were reported. No other significant sources of bias were identified.

All-cause mortality

At 1 year, data from 9 trials involving 4897 patients were available.^15,21, 22, 23^{,25,27,28,30,31} Mortality rates were comparable between the two groups (8.5% [214/2507] in the paclitaxel-coated device group vs. 8.5% [202/2390] in the control group). The pooled estimate indicated no increased risk associated with paclitaxel exposure (RR: 1.04, 95% CI: 0.87–1.24, P = 0.70). Heterogeneity among the studies was low (I² = 0%, P = 0.68).

At 2 years, the analysis included a cohort of 2119 patients from 12 trials.^10,12,20,21,23, 24, 25, 26, 27^,29, 30, 31 The risk of all-cause death remained similar between the intervention and control arms (RR: 1.28, 95% CI: 0.95–1.74, P = 0.11). The corresponding mortality rates were 8.1% and 6.5% for the paclitaxel and control groups, respectively, with no evidence of significant heterogeneity (I² = 34%, P = 0.12).

For long-term follow-up at 5 years, results from 5 trials continued to show no significant difference in mortality between the groups.^{15,20,27,28,31} The cumulative incidence was 36.2% (753/2075) in the paclitaxel group vs. 33.0% (685/2077) in the control group. The pooled RR was 1.13 (95% CI, 0.93–1.38, P = 0.21). Given the considerable heterogeneity observed (I² = 62%, P = 0.03, τ² = 0.03), a random-effects model was used for this analysis.

These findings are summarized in the forest plots shown in Fig. 2.

Fig. 2.

Forest plots of the meta-analysis for all-cause mortality at (a) 12 months, (b) 24 months, and (c) 60 months. M-H, Mantel-Haenszel; Fixed, fixed-effects model; Random, random-effects model; CI, confidence interval.

Amputation

At 1 year, the amputation rate was 7.1% (177/2507) in the paclitaxel group compared with 6.8% (163/2390) in the control group, based on data from 9 trials.^15,21, 22, 23^{,25,27,28,30,31} The meta-analysis demonstrated no significant benefit associated with paclitaxel-coated devices (RR: 1.07, 95% CI: 0.88–1.31, P = 0.49). Heterogeneity was low (I² = 14%, P = 0.32).

The 2-year outcomes, incorporating data from 13 trials (n = 2171), were consistent with the 1-year findings.^10,12,20, 21, 22, 23, 24, 25, 26, 27^,29, 30, 31 The RR for amputation remained nonsignificant (RR: 0.65, 95% CI: 0.34–1.24, P = 0.19), with event rates of 1.1% and 1.9% in the paclitaxel and control groups, respectively. Heterogeneity was low (I² = 0%, P = 0.78).

After 5 years of follow-up in 4 trials, the cumulative amputation incidence was 13.3% (254/1907) in the paclitaxel-coated device group and 12.9% (245/1901) in the uncoated device group.^15,20,28,31 The long-term pooled estimate confirmed the absence of a significant difference between the interventions (RR: 1.03, 95% CI: 0.88–1.20, P = 0.75), with no substantial heterogeneity detected (I² = 22%, P = 0.28).

The forest plots for these analyses are presented in Fig. 3.

Fig. 3.

Forest plots of the meta-analysis for amputation at (a) 12 months, (b) 24 months, and (c) 60 months. M-H, Mantel-Haenszel; Fixed, fixed-effects model; CI, confidence interval.

TLR

At 1 year, data from 8 trials demonstrated an advantage for paclitaxel-coated devices.^{15,21,23,25,27,28,30,31} The TLR rate was significantly lower in the paclitaxel group (12.8% [313/2446]) than in the control group (17.9% [417/2331]). The pooled analysis showed a significant 36% reduction in the risk of TLR associated with paclitaxel-coated devices (RR: 0.64, 95% CI: 0.48–0.87, P = 0.004). Of note, considerable heterogeneity was present at this time point (I² = 65%, P = 0.005, τ² = 0.10), necessitating the use of a random-effects model.

At the 2-year mark, data from 12 trials (n = 1945) continued to show a lower risk of TLR in the paclitaxel group (13.5%) than in the control group (31.9%).^10,12,20,21,23, 24, 25, 26, 27^,29, 30, 31 The pooled risk ratio remained significant (RR: 0.45, 95% CI: 0.38–0.54, P < 0.001). Heterogeneity was low (I² = 23%, P = 0.21).

For 5-year assessment, data from 5 trials (n = 4111) were available.^{15,20,27,28,31} The cumulative TLR rates were 25.1% for the paclitaxel group and 28.0% for the control group. The meta-analysis for this time point, which employed a random-effects model due to substantial heterogeneity (I² = 68%, P = 0.01, τ² = 0.04), yielded a nonsignificant risk ratio (RR: 0.81, 95% CI: 0.64–1.01, P = 0.06).

The forest plots for TLR at each time point are shown in Fig. 4.

Fig. 4.

Forest plots of the meta-analysis for target lesion revascularization (TLR) at (a) 12 months, (b) 24 months, and (c) 60 months. M-H, Mantel-Haenszel; Random, random-effects model; Fixed, fixed-effects model; CI, confidence interval.

Primary patency

Data from 10 RCTs were analyzed.^10,12,20,21,23, 24, 25^,27,30,31 The primary patency rate at 2 years was 70.1% (693/989) in the paclitaxel-coated device group compared to 40.9% (286/700) in the control group. The meta-analysis, performed using a random-effects model due to substantial statistical heterogeneity (I² = 82%, P < 0.001, τ² = 0.11), demonstrated a significant improvement in primary patency associated with paclitaxel-coated devices (RR: 1.57, 95% CI: 1.24–1.99, P < 0.001). The forest plot for 2-year primary patency is shown in Fig. 5.

Fig. 5.

Forest plot of the meta-analysis for primary patency at 24 months. M-H, Mantel-Haenszel; Random, random-effects model; CI, confidence interval.

Publication bias

Publication bias was assessed visually using funnel plots and statistically using Egger’s linear regression test. Due to the multiplicity of outcomes, a funnel plot for all-cause mortality at 1 year is presented as an illustrative example (Supplementary Fig. S2). The plot exhibited approximate symmetry, suggesting no substantial publication bias for this outcome.

Egger’s test was performed for all endpoints at each time point (Supplementary Table S2). None of the tests reached statistical significance (all P > 0.05), indicating no strong evidence of publication bias across the analyzed outcomes.

Sensitivity analysis

To assess the robustness of the meta-analysis results for outcomes with substantial heterogeneity—specifically, all-cause mortality at 5 years, TLR at 1 and 5 years, and primary patency at 2 years—a sensitivity analysis was performed by systematically excluding each study one at a time and recalculating the pooled risk ratio (Supplementary Fig. S3). Detailed results of the leave-one-out sensitivity analysis, including the pooled effect sizes and 95% CIs after exclusion of each individual study, are presented in Supplementary Table S3. The results demonstrated that no single study significantly influenced the overall effect estimates.

Discussion

This updated systematic review and meta-analysis, which included 15 RCTs, provides a comprehensive evaluation of the safety and efficacy of paclitaxel-coated devices for the treatment of femoropopliteal artery disease. The main findings indicate that, compared with uncoated devices, paclitaxel-coated devices did not significantly increase the risk of all-cause mortality or amputation at 1, 2, or 5 years of follow-up. Efficacy analysis confirmed short-to mid-term benefits, with reduced TLR and higher primary patency at 2 years. However, the initially observed reduction in TLR risk was no longer significant at the 5-year follow-up.

The absence of a significant increase in all-cause mortality across all time points is reassuring and aligns with recent large-scale randomized trials and meta-analyses.^16,32 However, two other meta-analyses suggested that there was an increased risk of death in patients treated with paclitaxel-coated devices compared with control patients.^14,33 In the SWEDEPAD 2 trial, the 5-year mortality rate was significantly higher in the paclitaxel-coated device group than in the uncoated device group (4.57 vs. 3.28 per 100 person-years; HR 1.47, 95% CI 1.09–1.98; P = 0.010).²⁸ This finding has raised concerns about the long-term safety of paclitaxel-coated devices. However, over the entire follow-up period (median 7.1 years), the mortality difference was no longer significant (HR 1.18, 95% CI 0.94–1.48; P = 0.16).²⁸ This may reflect competing mortality risks from various causes rather than a direct device effect.

Similarly, we found in this meta-analysis that the risk of major limb amputation did not differ significantly between the two groups, regardless of follow-up duration. However, another meta-analysis found that paclitaxel-coated devices may be linked to a higher risk of amputation.³⁴ In contrast, both SWEDEPAD 1 and SWEDEPAD 2 demonstrated no significant difference in amputation rates between paclitaxel-coated and uncoated devices.^15,28 Several factors might explain these inconsistencies. For one, variations in study populations, such as the proportion of patients with chronic limb-threatening ischemia vs. those with intermittent claudication, can influence outcomes. Additionally, differences in how and when amputation events are defined and measured across studies, along with variations in the devices used (e.g., the type of paclitaxel-coated balloon or stent and the drug dose), may contribute to the mixed results. For example, one study demonstrated that high-dose paclitaxel devices show a clearer signal of harm, whereas the risk is less evident with lower doses.³⁵

In terms of efficacy, our analysis confirms the well-established short-to mid-term benefits of paclitaxel-coated devices. The significant reduction in TLR and improvement in primary patency at 1 and 2 years highlight the potent antiproliferative effect of paclitaxel in inhibiting neointimal hyperplasia and restenosis. However, the attenuation of this benefit by 5 years suggests that the biological effect of paclitaxel may be time-limited, or that late restenosis mechanisms such as negative remodeling or disease progression may eventually offset earlier gains. This underscores the importance of long-term surveillance and the potential need for adjunctive or alternative strategies to sustain patency beyond the midterm. Of note, the findings of this study regarding the 5-year TLR results differ from those of a recent systematic review, which demonstrated that paclitaxel-coated devices still significantly reduced the risk of TLR at 5 years (RR: 0.53, 95% CI: 0.45–0.62).¹³ In contrast, our analysis incorporated data from the newly published SWEDEPAD trial, which had a large sample size and long follow-up duration, potentially influencing the assessment of long-term efficacy. Future rigorously designed randomized trials with longer follow-up periods are still needed to clarify the true long-term performance of paclitaxel-coated devices.

Several limitations of this analysis should be acknowledged. First, due to the unavailability of individual patient data, we were unable to perform time-dependent risk modelling to fully assess the non-proportional hazards concern raised in previous studies. Second, the included trials varied in terms of device type, paclitaxel dose, and lesion characteristics, which may contribute to clinical and methodological heterogeneity, as reflected in the high I² values for some outcomes. Third, the inability to perform patient-level data analysis limits deeper exploration of effect modifiers, such as diabetes, renal impairment, or specific anatomic factors. Fourth, despite the extensive search, publication bias cannot be entirely ruled out, although statistical tests did not indicate its presence. Fifth, our search was limited to studies published in English, which may introduce language bias despite English being the predominant language for scientific publication in this field. Finally, the generalizability of our findings may be limited to the femoropopliteal segment, as few trials included infrapopliteal disease.

In conclusion, this updated meta-analysis provides robust evidence that paclitaxel-coated devices are safe with respect to mortality and amputation risks up to 5 years. Their efficacy in reducing repeat revascularization and maintaining patency is significant in the short to mid-term, although this benefit wanes over time. These findings support the continued use of paclitaxel-coated devices in selected patients with symptomatic femoropopliteal disease, while emphasizing the need for individualized decision-making and long-term clinical follow-up. Future research should focus on optimizing device technology, drug delivery kinetics, and patient selection to maximize durable outcomes.

Contributors

Y.L., S.C., L.Q., J.Z., T.X., and Z.J. conceived and designed the study. Y.L., S.C., and L.Q. performed the literature search and data extraction. J.Z. and T.X. conducted the statistical analyses. Y.L. and Z.J. accessed and verified the underlying data. Y.L. drafted the manuscript. Z.J. and T.X. critically revised the manuscript. All authors read and approved the final version of the manuscript.

Data sharing statement

No individual participant data were collected or generated in this systematic review and meta-analysis, as the study is based exclusively on published aggregate data from previously completed randomized controlled trials. Therefore, no data are available for sharing.

Declaration of interests

All authors have completed the ICMJE uniform disclosure form. The authors declare no competing interests.