Highlights
In the final SAFE-PAD report, paclitaxel-coated drug-coated devices used for femoropopliteal revascularization were not associated with excess long-term all-cause mortality compared with non-drug-coated devices.
The adjusted hazard ratio for death was 0.98 (95% confidence interval 0.97 to 0.99), satisfying the pre-specified 5% non-inferiority margin and remaining consistent across multiple sensitivity analyses and clinical subgroups.
Hospitalization and major amputation rates were broadly similar between treatment groups, while repeat revascularization was higher with drug-coated devices in this claims-based analysis.
These findings are clinically important because they address a major safety controversy that reshaped device use, regulatory oversight, and informed consent in peripheral arterial interventions.
Background
Femoropopliteal peripheral artery disease remains one of the most challenging vascular territories for durable endovascular treatment. Restenosis after balloon angioplasty or stent implantation is common because the superficial femoral and popliteal arteries are long, mobile, calcified, and exposed to repetitive torsion and compression. Paclitaxel-coated technologies, including drug-coated balloons and drug-eluting stents, were developed to reduce neointimal hyperplasia and improve patency.
Early randomized trials and meta-analyses suggested that these devices improved lesion-level efficacy, especially by lowering restenosis and target lesion revascularization. However, the field changed abruptly after a 2018 meta-analysis by Katsanos and colleagues raised the possibility of increased late mortality after paclitaxel-coated device exposure in the femoropopliteal segment. That signal triggered intense debate because the proposed biological mechanism was uncertain, individual trials were not designed for mortality, long-term follow-up was incomplete, and dose-response modeling was controversial. Even so, regulators responded cautiously. In 2019, the US Food and Drug Administration issued communications warning of a potential late mortality signal, which led many clinicians and hospitals to restrict use of paclitaxel-coated devices.
SAFE-PAD was designed in this context. Working with the FDA, investigators used a large real-world Medicare cohort to evaluate whether femoropopliteal drug-coated devices were associated with excess long-term death. The present publication is the final report and therefore carries particular importance for clinicians, health systems, and policy makers who have had to balance procedural efficacy against unresolved safety concerns.
Study Design
Structure and data source
SAFE-PAD was a retrospective cohort study using Medicare fee-for-service claims. The analysis included 168,553 beneficiaries aged 66 years or older who underwent femoropopliteal artery revascularization between 2015 and 2018. Device exposure was classified as drug-coated devices or non-drug-coated devices based on claims coding.
Population
The cohort reflects routine US practice in an older population with substantial comorbidity burden. This is a major strength because the mortality controversy concerned real-world patients, not only selected trial participants. The investigators also examined clinically relevant subgroups, including inpatient versus outpatient procedures, lower-risk populations, and patients with high-risk chronic limb-threatening ischaemia.
Endpoints
The primary endpoint was all-cause mortality. Secondary endpoints included all-cause hospitalization, repeat revascularization, major amputation, and cardiovascular medication use. Mortality was compared using inverse probability treatment weighting, an approach intended to balance measured baseline differences between the treatment groups.
Methods to address bias
Because device selection in observational data is not random, the investigators performed several pre-specified sensitivity analyses. These included instrumental variable methods, falsification endpoints, and simulation of hypothetical unmeasured confounding. This layered analytic strategy is important because the central scientific question is not only what the hazard ratio was, but how believable it is in the presence of treatment-selection bias and coding limitations.
Key Findings
Primary outcome: mortality
At a median follow-up of 4.3 years, with maximum follow-up extending to 9.0 years, drug-coated devices were not associated with increased all-cause mortality. The adjusted hazard ratio was 0.98, with a 95% confidence interval of 0.97 to 0.99. This result met the study’s pre-specified non-inferiority criterion of a relative 5% margin.
Clinically, the most important message is straightforward: this large, regulator-engaged, real-world analysis found no evidence that paclitaxel-coated femoropopliteal devices increase long-term mortality in older Medicare patients. The point estimate was slightly below 1.0, but the correct interpretation is not that drug-coated devices improve survival. Rather, the data argue strongly against a clinically meaningful mortality hazard from device use.
Secondary outcomes
All-cause hospitalization and major amputation rates were similar between drug-coated and non-drug-coated device groups. These findings support the broader safety profile of drug-coated technology in routine practice.
One finding deserves careful attention: repeat revascularization was increased with drug-coated devices in this analysis. At face value, this is somewhat counterintuitive because randomized trials generally showed superior patency and lower repeat target lesion revascularization with paclitaxel-coated devices. Several explanations are possible. Claims-based definitions may capture repeat procedures at the limb or patient level rather than lesion-specific target lesion revascularization. Residual confounding is also plausible, since operators may preferentially choose drug-coated devices for more complex lesions or restenotic disease. In addition, real-world practice patterns, staged procedures, or coding variation could influence this endpoint more than mortality.
Subgroup findings
The mortality findings were consistent across subgroup analyses, including inpatient and outpatient settings, younger or lower-risk groups, and high-risk patients with chronic limb-threatening ischaemia. This consistency matters because concerns about harm would be expected to surface most clearly in vulnerable populations if the risk were genuine and treatment-related.
Robustness analyses
The study’s sensitivity analyses supported the main conclusion. Instrumental variable analysis is particularly helpful when unmeasured confounding is suspected, though it relies on strong assumptions and the validity of the chosen instrument. Falsification endpoints provide a pragmatic check: if the treatment appears to affect outcomes that should not plausibly be related to it, residual bias may be driving the result. The simulation analyses addressing hypothetical unmeasured confounders further strengthen confidence that a hidden mortality signal large enough to overturn the primary conclusion is unlikely.
Clinical Interpretation
SAFE-PAD is best viewed as a decisive real-world safety study rather than an efficacy trial. It does not replace lesion-level randomized evidence on patency, but it does address the question that most altered practice: are paclitaxel-coated femoropopliteal devices associated with late death? Based on this final report, the answer is no.
These findings are highly relevant to contemporary procedural decision-making. For many operators, the mortality controversy created a paradox. Devices with clear anti-restenotic benefits were being withheld because of a safety concern that remained biologically unexplained and methodologically debated. The SAFE-PAD data, especially in conjunction with other patient-level meta-analyses and registry-based studies published since the original signal, support a return to evidence-based use of drug-coated technology where restenosis prevention matters.
That said, the study should not be overinterpreted. A hazard ratio slightly below 1.0 does not prove a survival benefit. Nor does this analysis determine which paclitaxel device, dose, lesion subset, or procedural strategy is best. The practical takeaway is narrower and more important: clinicians should not avoid femoropopliteal drug-coated devices out of concern for excess late mortality based on currently available evidence.
Strengths
The most obvious strength is scale. More than 168,000 Medicare beneficiaries provide substantial statistical power and broad clinical representation. The long follow-up, extending up to 9 years, is particularly valuable because the original safety concern centered on delayed mortality rather than peri-procedural harm.
The study also benefits from its regulatory alignment. SAFE-PAD was pre-specified with the FDA, which increases confidence that the analysis was designed to answer the safety question prospectively rather than retrospectively after inspecting the data.
Finally, the analytic approach was appropriately rigorous for an observational study. Inverse probability weighting, instrumental variable methods, falsification endpoints, and simulations of unmeasured confounding collectively provide a coherent framework for testing whether the headline result is stable under different assumptions.
Limitations
As with all claims-based retrospective studies, residual confounding cannot be eliminated. Device choice is influenced by anatomy, lesion length, calcification, chronic total occlusion status, prior interventions, operator preference, and center expertise, many of which are incompletely captured in administrative data.
The endpoint of all-cause mortality is robust, but claims data provide limited mechanistic insight. If there were concern about cause-specific death, such as malignancy or cardiovascular death, this dataset would be less ideal than a trial with adjudicated outcomes. The study also primarily reflects US fee-for-service beneficiaries aged 66 years or older; extrapolation to younger populations, Medicare Advantage enrollees, or non-US systems should be thoughtful rather than automatic.
The secondary result showing more repeat revascularization with drug-coated devices should therefore be interpreted cautiously. It may reflect confounding by indication, coding artifacts, or the difference between patient-level claims outcomes and lesion-specific efficacy outcomes from randomized trials.
How SAFE-PAD Fits With the Broader Evidence
The paclitaxel controversy has moved through three phases: initial enthusiasm based on improved patency, abrupt concern after a mortality signal in trial-level meta-analysis, and subsequent reassessment using patient-level trial data, registries, and administrative datasets. Over time, the weight of evidence has shifted away from the hypothesis of a true causal mortality effect.
SAFE-PAD contributes powerfully to that reassessment because it combines scale, long follow-up, and a pre-specified regulatory framework. Its findings also align with the broader trend in the literature: once follow-up matured and larger real-world data sources became available, the mortality signal weakened or disappeared. In that sense, SAFE-PAD is not an isolated reassuring result; it is a culminating study that helped close a major safety debate.
Practice Implications
For clinicians performing femoropopliteal interventions, the study supports routine consideration of drug-coated balloons and drug-eluting stents when clinically indicated, especially in lesions where restenosis risk is high and durability matters. Shared decision-making should still include discussion of expected benefits, alternatives, and device-specific considerations, but the counseling emphasis can now shift from unresolved mortality concern back toward individualized effectiveness, limb outcomes, lesion complexity, and cost.
For hospitals and payers, the findings support normalization of access to paclitaxel-coated technology after years of caution. For regulators, SAFE-PAD is a reminder that safety signals from indirect meta-analytic comparisons, while important, require confirmation in appropriately designed datasets before they are allowed to reshape practice for prolonged periods.
Conclusion
The final SAFE-PAD report provides strong real-world evidence that drug-coated devices used for femoropopliteal revascularization are not associated with increased long-term mortality. In an older Medicare population followed for a median of 4.3 years and up to 9 years, the adjusted hazard ratio for death was 0.98, satisfying the study’s pre-specified non-inferiority margin. Secondary outcomes showed similar hospitalization and amputation rates, although repeat revascularization was higher in the claims-based analysis and warrants cautious interpretation.
Overall, SAFE-PAD substantially strengthens the case that paclitaxel-coated femoropopliteal devices are safe from a long-term mortality standpoint. Its greatest contribution may be practical: it helps restore confidence in using anti-restenotic technology where the clinical need remains significant and the burden of recurrent femoropopliteal disease is high.
Funding and ClinicalTrials.gov
The abstract provided does not specify funding details. SAFE-PAD was conducted as a retrospective Medicare claims analysis designed in collaboration with the US Food and Drug Administration. A ClinicalTrials.gov registration number is not applicable based on the information available for this observational claims-based study.
References
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. European Heart Journal. 2026;47(17):2093-2103. PMID: 40996087.
Katsanos K, Spiliopoulos S, Kitrou P, Krokidis M, Karnabatidis D. Risk of death following application of paclitaxel-coated balloons and stents in the femoropopliteal artery of the leg: a systematic review and meta-analysis of randomized controlled trials. Journal of the American Heart Association. 2018;7(24):e011245.
Rosenfield K, et al. Risk of mortality with paclitaxel-coated devices in the femoropopliteal artery: patient-level meta-analysis of randomized trials. Circulation. 2020;141(23):1859-1869.
Secemsky EA, Kundi H, Weinberg I, et al. Association of survival with femoropopliteal artery revascularization with drug-coated devices. JAMA Cardiology. 2019;4(4):332-340.
Gerhard-Herman MD, Gornik HL, Barrett C, et al. 2016 AHA/ACC guideline on the management of patients with lower extremity peripheral artery disease. Circulation. 2017;135(12):e726-e779.