November 11, 2025

SAFE-PAD Trial Final Report Finds No Increased Mortality With Drug-Coated Devices

November 11, 2025—In this retrospective cohort study designed to assess long-term safety outcomes of drug-coated devices (DCDs) versus non–drug-coated devices (NDCDs) for femoropopliteal revascularization, Kim et al found no evidence of a long-term mortality risk and observed similar rates of major amputation with a lower need for surgical revascularization. Results of this final analysis of SAFE-PAD were published in the European Heart Journal.¹

KEY FINDINGS

Use of DCDs was not associated with increased long-term mortality compared with NDCDs.
Safety outcomes were consistent across inpatient, outpatient, CLTI, and low-risk patient subgroups.
Rates of major amputation and hospitalization were similar between groups, although repeat revascularization was slightly higher with DCDs.

Investigators used Medicare fee-for-service claims data to identify 168,553 beneficiaries aged ≥ 66 years who underwent femoropopliteal revascularization with either DCDs (drug-eluting stent [DES] ± drug-coated balloon [DCB], bare-metal stent [BMS] with DCB, or DCB alone) or NDCDs (BMS ± percutaneous transluminal angioplasty [PTA] or PTA alone) between 2015 and 2018. Device exposure (DCD vs NDCD) was determined via procedural codes. The study was undertaken in collaboration with industry partners and the United States FDA to evaluate potential long-term mortality risks associated with DCDs following earlier meta-analyses suggesting possible harm.

Baseline covariates included age, sex, race, comorbidities, chronic limb-threatening ischemia (CLTI), and inpatient or outpatient treatment setting. Inverse probability treatment weighting (IPTW) was used to balance groups. Sensitivity analyses incorporated instrumental variable approaches, falsification endpoints, and hypothetical unmeasured confounder simulations.

The primary endpoint was all-cause mortality, assessed up to 9 years postprocedure. Secondary endpoints included hospitalizations, major amputations, repeat revascularizations, and cardiovascular medication use. Subgroup analyses evaluated results by clinical setting, disease severity, and patient risk profiles.

The analysis included 168,553 patients treated at 2,978 institutions (mean age, 77.0 ± 7.6 years; 44.9% female; 81.9% White; 46.7% had CLTI; 7.9% had prior amputation). During the index femoropopliteal revascularization procedure, 41.9% received DCDs and 58.1% received NDCDs. Among DCD-treated patients, 60.4% underwent procedures with DCBs and 39.6% with DES. In the NDCD group, 59.5% were treated with PTA alone and 40.5% with BMS. Before statistical weighting, demographic and clinical characteristics were similar between groups, although DCD-treated patients more often underwent adjunctive atherectomy (42.5% vs 29.8%) and were less frequently treated as inpatients (41.4% vs 52.1%). DCD procedures were also performed at higher-volume hospitals on average. After weighting adjustments, all baseline and procedural characteristics were well balanced between the DCD and NDCD groups.

At a median follow-up of 4.3 years (IQR, 1.6-6.7 years; maximum, 9 years), DCD use was not associated with increased mortality compared with NDCDs (adjusted hazard ratio [HR], 0.98; 95% CI, 0.97-0.99), meeting the prespecified 5% noninferiority margin. Sensitivity analyses confirmed the robustness of findings. Secondary outcomes showed similar hospitalization and amputation rates but slightly higher repeat revascularization rates in the DCD group (HR, 1.09; 95% CI, 1.07-1.10). Subgroup analyses—including inpatient, outpatient, younger/low-risk, and CLTI populations—demonstrated consistent safety across all groups.

Investigators noted some limitations, including potential residual confounding despite robust adjustments, lack of procedural detail (eg, lesion characteristics, paclitaxel dose), restriction to an older Medicare population, and absence of cause-specific mortality data. The median follow-up exceeded 4 years, with 32.5% of patients contributing data beyond 5.5 years.

This final report of the SAFE-PAD study found no evidence of increased long-term mortality associated with DCDs, supporting their noninferiority to NDCDs across diverse clinical subgroups. These findings aligned with subsequent randomized trial data (eg, SWEDEPAD 1 and 2) and informed the FDA’s 2023 reversal of warnings against DCD use. The authors highlighted SAFE-PAD as a model for leveraging large-scale, real-world data to guide cardiovascular device safety evaluations.

1. Kim JM, Yeh RW, Song Y, Secemsky EA. Drug-coated vs non-drug-coated devices for femoropopliteal artery interventions: long-term outcomes of the SAFE-PAD study. Eur Heart J. Published online September 25, 2025. doi: 10.1093/eurheartj/ehaf721

ENDOVASCULAR TODAY ASKS...

Study investigators Joseph Kim, MD, and Eric A. Secemsky, MD, with Beth Israel Deaconess Medical Center and Harvard Medical School in Boston, Massachusetts, expanded on the impact of these results in the treatment of patients with peripheral artery disease and key lessons for the future of device innovation regulation and safety evaluation.

What motivated the design of SAFE-PAD as a large, real-world evidence study, and how did collaboration with the FDA shape the study’s methodology and transparency?

SAFE-PAD was conceived in direct response to the 2018 meta-analysis suggesting a possible mortality signal with paclitaxel-coated devices. Randomized trials lacked the size and follow-up duration to address this question due to attrition and narrow inclusion criteria. A large, real-world study was therefore essential to evaluate long-term safety across a representative United States population.

The FDA’s collaboration fundamentally shaped the study’s rigor and transparency. SAFE-PAD was prespecified, preregistered, and designed jointly with FDA input, including analytic plans, noninferiority margins, and periodic data reviews. The use of validated Medicare device codes and statistical innovation, including use of IPTW, instrumental variable analyses, and falsification endpoints, ensured robustness comparable to a trial framework. This partnership set a precedent for transparent, real-world regulatory science and was later cited by the Centers for Medicare & Medicaid as a template for future device evaluations.

There was a modest increase in repeat revascularization among patients treated with DCDs. What factors do you believe most likely explain this finding?

The modestly higher rate of repeat endovascular revascularization among DCD-treated patients likely reflects unmeasured procedural or anatomic complexity rather than device failure. Patients selected for DCDs may have had longer, more calcified, or more complex lesions—factors not captured in claims data. Furthermore, reintervention was captured for any vessel/lesion to either leg, not specific to the index lesion treated. Operator practice patterns, follow-up intensity, or treatment of more aggressive disease could also contribute. Importantly, despite this difference, surgical revascularization and major amputation were not increased, reinforcing that DCD use did not worsen limb-level outcomes.

What lessons does this collaboration with FDA on SAFE-PAD provide for future regulatory evaluations?

SAFE-PAD demonstrates that prespecified, transparent real-world evidence studies can complement or even extend beyond randomized data to guide regulatory action. Continuous engagement with the FDA—through protocol alignment, regular and periodic reporting, and open data access—ensures scientific credibility and regulatory relevance.

Future device evaluations can emulate three key features of SAFE-PAD: (1) long-term follow-up via linkage to claims or mortality registries; (2) causal inference framework with multiple sensitivity analyses including using IPTW, instrumental variables, and falsification endpoints to evaluate bias; and (3) public registration and transparency to enhance trust in real-world evidence. This collaborative approach directly informed the FDA’s 2023 reversal of its caution against paclitaxel-coated devices.

How do you view the findings of SAFE-PAD, the recent SWEDEPAD findings, and the overall body of evidence regarding paclitaxel safety and efficacy? What is your advice on how these data sets can be interpreted in their respective contexts and as a whole? In your view, how can the SAFE-PAD analytic framework serve as a model for evaluating other cardiovascular devices where randomized trial data are limited?

Across multiple data sets including but not exclusive of SAFE-PAD, the cumulative evidence is reassuring on paclitaxel device safety. SAFE-PAD, the largest and longest study to date, found no excess mortality up to 9 years, consistent with the SWEDEPAD interim and final analyses and an updated patient-level meta-analysis. SWEDEPAD-1 (CLTI) showed no mortality difference, whereas SWEDEPAD-2 (claudication) observed a transient divergence that resolved by study end.

Interpretation must consider context: SWEDEPAD was a randomized trial within a single national system and early use of drug-coated technology, whereas SAFE-PAD reflects broad U.S. practice patterns and older, higher-risk patients. Together, these results confirm safety across diverse populations. While efficacy findings have varied, the totality of evidence supports paclitaxel-coated devices as safe and effective tools for reducing restenosis when used appropriately.

In our view, SAFE-PAD provides a scalable blueprint for postmarket safety assessment using real-world data. Its framework—anchored in preregistration, prespecified endpoints, robust confounding control (IPTW, instrumental variable methods), falsification testing, and confounder simulation—demonstrates how observational evidence can meet regulatory standards.

For emerging devices with limited trial data, this approach enables timely, high-confidence evaluation of safety and performance at scale. By combining methodologic rigor with data transparency, the SAFE-PAD model shows how large administrative data sets can complement clinical trials and guide evidence-based regulatory decisions across the cardiovascular field.

Find more Literature Highlights.

Browse additional literature summaries, key findings, and investigator commentaries across the vascular interventional field.