Content area
Of further significant concern, in the SWEDEPAD 2 trial at 5 years, all-cause mortality was significantly higher in the paclitaxel-coated group than in the uncoated group (HR 1·47 [95% CI 1·09 to 1·98]), echoing our original 2018 meta-analysis. 3 The authors employed a nationwide registry-based randomisation in Sweden and stratified between critical limb-threatening ischaemia and claudication with hard patient-centred endpoints, namely quality of life, major amputations, mortality, and limb re-interventions. 1,2 We applaud the registry-RCT design that leveraged the efficiency and data linkage of the existing Swedish registry infrastructure to produce the two largest randomised studies available to date in the peripheral arterial disease arena. In the BASIL-3 trial, neither paclitaxel-coated balloons (HR 0·84 [95% CI 0·61 to 1·16]), nor paclitaxel-coated stents (HR 0·83 [95% CI 0·60 to 1·15]) conferred any significant clinical benefit in time to major amputation or all-cause death. 4 The SWEDEPAD 2 trial also did not translate the well known anti-restenotic efficacy of paclitaxel into meaningful gains of quality of life. 2 We contend that quality of life is shaped by multiple factors, including, but not limited to, instrument (questionnaire) insensitivities, lifestyle adaptations, socioeconomic environment, psychological wellbeing, systemic progression of atherosclerosis, and other influences often unrelated to the treated limb. Alternative patient-reported functional outcomes, such as pain-free walking distance, have been proposed as more appropriate primary endpoints for claudication trials. 5 The SWEDEPAD 2 trial has once again showed a markedly elevated risk of all-cause mortality—approximately 1·5-fold—in paclitaxel-treated patients at 5 years. 2 Even though this mortality signal appears attenuated in very long-term follow-up, this likely reflects the well recognised statistical properties of hazard functions, in which early differences are progressively diluted over time. 6 Its absence in the SWEDEPAD 1 trial is plausibly explained by the natural history of critical limb-threatening ischaemia, in which multiple systemic comorbidities act as competing risks and might overshadow or mute any potential paclitaxel-related harm.
Paclitaxel in the form of paclitaxel-coated balloons and stents has spearheaded innovation and medical research in the quest for a safe and effective anti-restenotic agent to inhibit neointimal hyperplasia and improve clinical outcomes following endovascular treatment for peripheral arterial disease. In The Lancet, Mårten Falkenberg and colleagues 1 and Joakim Nordanstig and colleagues 2 report two randomised controlled trials (RCTs) separately investigating the infrainguinal use of paclitaxel devices in the lower limbs (namely femoropopliteal application) with the aims of reducing major amputations in case of critical limb-threatening ischaemia (SWEDEPAD 1), 1 and improving quality of life in case of lifestyle-limiting intermittent claudication (SWEDEPAD 2). 2
In the SWEDEPAD 1 trial, 2400 patients with Rutherford 4–6 stage critical limb-threatening ischaemia were randomly assigned to treatment with paclitaxel-coated devices (n=1206) or treatment with uncoated devices (n=1194). The median age of this cohort was 77 years (IQR 71–83), 1317 (55·9%) of 2355 patients were male and 1038 (44·1%) were female, and 1761 (74·9%) of 2351 had wounds or tissue loss. After a median follow-up of 2·67 years, there was no significant difference in time to ipsilateral above ankle amputation comparing paclitaxel-coated with uncoated devices (hazard ratio [HR] 1·05 [95% CI 0·87 to 1·27]). 1
In the SWEDEPAD 2 trial, 1155 patients with intermittent claudication were randomly assigned to paclitaxel-coated devices (n=577) or uncoated devices (n=578). The median age was 73 years (IQR 68–78), 612 (53·9%) of 1136 were male and 524 (46·1%) were female, and 677 (59·6%) of 1135 had Rutherford stage 1–3 claudication. At 12 months, there were no significant differences in VascuQoL-6 scores between the paclitaxel-coated and uncoated groups (mean difference –0·02 [95% CI –0·66 to 0·62]). 2 Both randomised trials did not show any material clinical benefit with respect to their predefined primary endpoints. Of further significant concern, in the SWEDEPAD 2 trial at 5 years, all-cause mortality was significantly higher in the paclitaxel-coated group than in the uncoated group (HR 1·47 [95% CI 1·09 to 1·98]), echoing our original 2018 meta-analysis. 3
The authors employed a nationwide registry-based randomisation in Sweden and stratified between critical limb-threatening ischaemia and claudication with hard patient-centred endpoints, namely quality of life, major amputations, mortality, and limb re-interventions. 1,2 We applaud the registry-RCT design that leveraged the efficiency and data linkage of the existing Swedish registry infrastructure to produce the two largest randomised studies available to date in the peripheral arterial disease arena. Compared with the high regulatory rigour and internal validity of traditional industry-sponsored pivotal RCTs that are focused on device-centred metrics of efficacy, registry-RCTs improve external validity, evidence transfer, and real-world generalisability of clinical effectiveness and cost effectiveness. The endpoints in the SWEDEPAD trials were appropriately chosen, powered, measured, and adjudicated with respect to the clinical intricacies and targeted patient symptoms. 1,2
The SWEDEPAD 1 trial further authenticated the findings of the BASIL-3 trial 4 that documented the dismal natural history of critical limb-threatening ischaemia patients with broadly high mortality rates in the order of 10% per annum, 1,4 but also did not show any limb salvage benefit in case of the paclitaxel group. In the BASIL-3 trial, neither paclitaxel-coated balloons (HR 0·84 [95% CI 0·61 to 1·16]), nor paclitaxel-coated stents (HR 0·83 [95% CI 0·60 to 1·15]) conferred any significant clinical benefit in time to major amputation or all-cause death. 4 The SWEDEPAD 2 trial also did not translate the well known anti-restenotic efficacy of paclitaxel into meaningful gains of quality of life. 2 We contend that quality of life is shaped by multiple factors, including, but not limited to, instrument (questionnaire) insensitivities, lifestyle adaptations, socioeconomic environment, psychological wellbeing, systemic progression of atherosclerosis, and other influences often unrelated to the treated limb. Alternative patient-reported functional outcomes, such as pain-free walking distance, have been proposed as more appropriate primary endpoints for claudication trials. 5
The SWEDEPAD 2 trial has once again showed a markedly elevated risk of all-cause mortality—approximately 1·5-fold—in paclitaxel-treated patients at 5 years. 2 Even though this mortality signal appears attenuated in very long-term follow-up, this likely reflects the well recognised statistical properties of hazard functions, in which early differences are progressively diluted over time. 6 Its absence in the SWEDEPAD 1 trial is plausibly explained by the natural history of critical limb-threatening ischaemia, in which multiple systemic comorbidities act as competing risks and might overshadow or mute any potential paclitaxel-related harm. Over the last years, the literature has been rife with conflicting reports and a heated debate about the risk of death associated with the use of paclitaxel-coated devices in the lower limbs. We originally published a study-level meta-analysis that showed a heightened risk of death following application of paclitaxel devices in the femoropopliteal artery of the lower limbs, 3 which was soon corroborated by a patient-level meta-analysis of some major industry-sponsored randomised controlled trials (HR 1·38 [95% CI 1·06 to 1·80] at 5 years). 7 Further amassed evidence about the possibility of increased mortality hazards with paclitaxel devices in the lower limbs have been contradictory between observational and randomised studies. Crucially, the latest iteration of patient-level meta-analysis of some of those RCTs with more complete vital status ascertainment at 5 years has refuted the original mortality signal (HR 1·14 [95% CI 0·93 to 1·40] at 5 years), 8 and scientific advocacy and engagement eventually led to the reversal of statements of concern by regulatory bodies on paclitaxel use in the lower limbs in several health-care systems. Now, SWEDEPAD 2 results threaten to upend that narrative once again.
What does the future hold for paclitaxel in peripheral arterial disease considering previous and present SWEDEPAD evidence taken together? For sure, the implications of the SWEDEPAD trials might be profound and call for a re-evaluation of current clinical practices. As the authors state, clinicians should carefully weigh the potential risks and benefits of drug-coated devices and broader revascularisation strategies in the context of individual background risk factors and prioritise patient-centred outcomes when testing new anti-proliferative agents for prevention of restenosis. The SWEDEPAD findings not only challenge the presumed benefits of paclitaxel-coated devices but also re-establish the mortality signal in low-risk claudicant patients—a scenario in which potential harm clearly outweighs any conceivable benefit, rendering their routine use difficult to defend. It is unknown whether this signal represents a class effect incriminating all paclitaxel-coated devices. Notably, the case-mix of devices used in SWEDEPAD was quite like in our original meta-analysis, 3 leaving it unclear whether the harm signal potentially applies to newer iterations of paclitaxel devices as well. Alternative antiproliferative agents such as sirolimus and its analogues are already on the horizon and are likely to see accelerated adoption in clinical practice.
Yet, at this new turn of paclitaxel evidence, there can be no justification for the continued production of substandard scientific research. Many industry-sponsored reports remain undermined by methodological flaws, including the systematic conflation of randomised data with registries and other observational sources, treatment-effect dilution through selective study inclusion, and the ongoing sequestration of patient-level data away from the broader scientific community. Moving forward, clinicians must seize the moment and rise to the challenges posed by this latest turn in scientific evidence.
KK reports research grants from CORDIS, personal fees from BD, Cordis, Abbott Vascular, and Sirtex, outside the submitted work. SS reports personal fees from Cordis and Balt, outside the submitted work.
© 2025 Elsevier Ltd