The Safety of Paclitaxel-Coated Devices for Patients with Peripheral Artery Disease

Abstract

Purpose of Review

Peripheral artery disease (PAD) is a common, debilitating disease that impacts 8.5 million Americans and carries a poor prognosis. The most common manifestation of lower extremity PAD is claudication—a condition which significantly reduces quality of life and functional status. Paclitaxel-coated balloons and stents (PCBs and PESs) represented a breakthrough in the ability to treat medication-refractory patients relative to bare metal stents (BMSs) and percutaneous transluminal angioplasty (PTA) because they improve primary patency rates, reduce target lesion revascularization (TLR), and minimize late-lumen loss for femoropopliteal lesions. As a result, paclitaxel-coated devices (PCDs) were swiftly established as the standard of care for revascularization of femoropopliteal artery disease. A recent meta-analysis of summary-level data demonstrated a late mortality signal for patients treated with paclitaxel-coated devices relative to uncoated devices. This has had a major impact on the vascular community and for the treatment of patients with PAD. Herein, we provide a detailed review of the available data on the late mortality signal associated with paclitaxel.

Recent Findings

In December of 2018, Katsanos et al. J Am Heart Assoc 7: e011245, 2018) published data from randomized-controlled trials (RCTs) that demonstrated an increase in mortality at 2 and 5 years in patients treated with PCDs involving the femoropopliteal arterial segment relative to patients treated with uncoated devices. As a result of this analysis, randomized trials were stopped and the FDA sent a letter to healthcare providers recommending restriction of use of these devices to patients at the highest risk of restenosis. As additional data emerged supporting the safety of these devices, the FDA organized an advisory committee meeting to review the available data and to determine a pathway forward. The FDA concluded that there were insufficient data to make a final decision regarding the safety of PCDs. They allowed these devices to remain on the market, but with revised safety labeling and updated their letter to healthcare providers to continue to restrict use to patients at highest risk of reintervention. The FDA also called for additional long-term data, including from RCTs and real-world data. To date, an updated patient-level meta-analysis of clinical trial data, RCTs with longer-term follow-up, and large observational studies have been conducted.

Summary

While meta-analyses conducted using overlapping clinical trial data have found a persistent increase in mortality for those treated with PCDs, individual industry-sponsored RCTs and large observational studies have consistently failed to detect a corresponding mortality increase. To date, no mechanism linking paclitaxel to mortality has been observed. We are currently at an impasse for drawing definitive conclusions regarding the long-term safety of paclitaxel-coated devices. As we await enrollment in ongoing clinical trials, we must proceed with making reasonable decisions for our patients’ care from the available data, as these devices have important clinical implications for our patients. A critical lesson that can be learned from this controversy is that, for future device trials, committing to long-term follow-up is crucial.

Introduction

PAD is a common disease impacting 8.5 million Americans and 230 million people worldwide [1,2]. Its prevalence is increasing over time, with a 17.1% increase from 2010 to 2015 [2]. Patients with PAD suffer from high mortality due to elevated rates of myocardial infarction, stroke, and cardiovascular death [3]. For patients with lower extremity disease, PAD often results in clinical symptoms ranging from claudication to ischemic rest pain and critical limb ischemic (CLI). These symptoms result in reduced quality of life and limitations of functional status [4, 5]. In addition, depressive symptoms are frequently identified in patients with PAD [6]. Overall, PAD is underdiagnosed and severely undertreated relative to patients with coronary artery disease [7].

Evidence-based therapies for PAD involve medical management, lifestyle interventions, and revascularization [8]. The goals of medical therapy are to reduce cardiovascular events, claudication, and progression of local atherosclerotic disease. Medical therapies include statins, anti-hypertensives (ACE-inhibitors and angiotensin-receptor blockers in particular), antiplatelet therapy, low-dose rivaroxaban, PCSK9 inhibitors, and cilostazol. Cilostazol is the only FDA-approved drug for the specific treatment of claudication symptoms.

Lifestyle interventions are critical in patients with PAD. Supervised exercise therapy (SET) improves functional status, claudication onset time, and quality of life [9]. SET received a class 1A recommendation from the 2016 ACC/AHA Guidelines for the Management of Lower Extremity Peripheral Artery Disease [10] and is reimbursed by Medicare. Of note, previous studies have shown that SET achieves the best results followed by home-based exercise programs in terms of change in pain-free walking time over 6 months [9, 11]. Smoking cessation is essential for patients with PAD. The prevalence of PAD in active smokers is 2- to 3-fold higher than in non-smokers. Patients who continue to smoke after receiving a diagnosis of PAD show an increased mortality at 5 years. Smoking cessation reduces rates of CLI and amputation in patients with PAD [12].

Endovascular or surgical revascularization may be necessary for limb salvage, but also may be undertaken to improve symptoms among patients with claudication that is refractory to lifestyle interventions. The 2017 ESC Guidelines on the Diagnosis and Treatment of Peripheral Arterial Diseases recommend that endovascular therapy be considered initially for femoropopliteal lesions <25 cm [13]. Studies have shown that endovascular therapy is associated with patency rates that are comparable to those of surgical revascularization, but with lower morbidity and mortality [14, 15].

Despite improvements in endovascular technology, rates of restenosis proved to be very high for PTA and BMS affecting approximately 40–60% of treated patients within 1 year [16, 17]. Inspired by success in coronary vascular beds, dedicated peripheral drug-eluting balloons and stents were developed to improve patency rates and reduce clinically driven target lesion revascularization. Paclitaxel, the first drug used for coronary stents before moving to limus-based agents, was specifically chosen for devices in the periphery due to its chemical properties. Femoropopliteal artery disease is an aggressive atherosclerotic process involving a long vessel segment subject to the mechanical stresses of leg movement and prone to heavy calcification. Paclitaxel combines high lipophilicity (facilitating rapid tissue update) and a long-term inhibitory effect on the proliferation of vascular smooth muscle cells and fibroblasts which was felt to be optimal for this disease process [18].

In randomized clinical trials, PCB/PES have been reliably found to meaningfully reduce rates of restenosis and TLR. For instance, in a meta-analysis of devices, 2-year TLR rates for POBA were 38.5%, BMS 26.9%, and DCB 17.6% (Fig. 1) [19]. Since the entry of PES into the US market in 2012 and DCBs in 2014, PCD use has increased significantly with reductions in the use of POBA and BMS [20]. PCDs have also been designated first-line devices for femoropopliteal artery revascularization society guidelines [21].

Fig. 1.

Target lesion revascularization at 24 months after treatment with an uncoated balloon (PTA), bare metal stent (BMS), paclitaxel-eluting stent (PES), or paclitaxel-coated balloon (PCB). Randomized clinical trials have shown that paclitaxel-coated devices (PCDs) (depicted in red) meaningfully reduce rates of restenosis and TLR relative to non-PCDs (depicted in blue). (Data derived from Katsanos K. et al. [19])

In December of 2018, a meta-analysis published by Katsanos et al. [22••] found a 68% increased risk of 2-year mortality with PCDs and a 93% increased risk of 5-year mortality. Despite concerns about the methodology of the analysis and lack of mechanism for how PCDs influence mortality, the publication has had a major impact on the treatment of patients with PAD. The use of PCDs has plummeted since its publication [23] and two clinical trials, BASIL-3 and SWEDEPAD, were stopped while awaiting safety analyses [24]. The FDA became involved early on, cautioning clinician to reserve use of these devices until additional safety analyses could be performed. Following an advisory committee meeting in June 2019 to review the available data and determine a pathway forward, the FDA concluded that there were insufficient data to make a definitive statement regarding the safety of PCDs. As a result, they allowed these devices to remain on the market, but with revised safety labeling and updated their letter to healthcare providers to continue to restrict use to patients at highest risk of reintervention. The FDA also made a call for additional long-term data, including RCTs and real-world data.

In this review, we examine the initial meta-analysis and limitations of the available data. We then summarize the results of PCD investigations that have been published since the meta-analysis, including longer-term data from RCTs, an individual patient-level meta-analysis, and a number of large observational studies.

Advantages of Drug-Coated Devices

Balloon expansion and stent placement cause injury to the vascular endothelium which sets off a local inflammatory response aimed at healing the vascular endothelium. This process results in proliferation of underlying smooth muscle cells and migration of cells into the neointima leading to restenosis. Studies in PTA and BMS show high rates of restenosis, affecting approximately 40–60% of treated patients within 1 year [16, 17].

The application of paclitaxel and sirolimus to coronary devices has been very successful in reducing rates of restenosis. Both agents locally inhibit the proliferation of smooth muscle cells. In, particular, paclitaxel has been found to be efficacious at preventing restenosis of the femoropopliteal artery. Paclitaxel binds to tubulin which prevents tubulin depolymerization and stops mitotic cell division [18]. This precludes vascular smooth muscle proliferation and migration, therefore reducing restenosis. Paclitaxel is highly lipophilic which facilitates its rapid uptake into tissues and its retention in tissues after initial application. Its lasting effects over time are also mediated by its irreversible binding to tubulin. Paclitaxel localizes preferentially in the deeper smooth muscle and fibroblast layers which allows it to successfully inhibit the source of neointimal hyperplasia. These properties of paclitaxel make it particularly conducive to use with drug-coated balloons, which require that a device deliver sufficient drug on a single application yet have a durable effect on restenosis over time.

The effects described above occur at low concentrations of paclitaxel. Mean total doses delivered by drug-coated devices in clinical trials range primarily from 1 to 20 mg [25]. A trial examining a spectrum of doses of paclitaxel delivered during stent placement documented systemic effects such as moderate neutropenia, sensory neuropathy, and alopecia, only at doses much higher than are delivered by drug-coated devices [26]. Paclitaxel has been extensively studied as a chemotherapeutic agent at total doses of up to 230–300 mg for individual treatments and 1200 mg over several treatments. At these higher concentrations, documented side effects include neutropenia, neuropathy, hypotension/hypertension, bradycardia, myalgia myelotoxicity, anaphylaxis, and nausea.

Multiple clinical trials have demonstrated the superiority of PCDs relative to BMS and PTA in terms of primary vessel patency, need for TLR, and late-lumen loss. The THUNDER trial, a multicenter RCT, evaluated late-lumen loss in 154 patients with femoropopliteal lesions. They were assigned to PCB, POBA with paclitaxel in the contrast medium, or POBA alone. At 6 months, they reported significantly less late-lumen loss in the PCB group relative to those treated with POBA (0.4 ±1.2mm vs. 1.7 ±1.8, p<0.001) and significantly lower rates of TLR in the PCB group relative to control at 6 and 24 months (4% vs. 37%, p<0.001 at 6 months and 15% vs. 52% at 24 months, p<0.001) [27]. At 5-year follow-up, the PCB group had significantly lower rates of TLR compared with control (21% vs. 56%, p=0.0005), but there was no difference in late-lumen loss [28]. A smaller group of patients for whom imaging was available demonstrated significantly lower rates of binary restenosis in the PCB arm relative to the POBA arm (17% vs. 54%, p=0.04). In terms of safety, serious adverse events including rates of major amputation and death did not differ between groups at 5 years.

In the multicenter IN.PACT SFA trial, 331 patients with femoropopliteal lesions were randomly assigned to receive a DCB or PTA [29]. At 5 years of follow-up, patients treated with DCB had greater freedom from TLR relative to those treated with PTA (Kaplan-Meier estimate of 74.5% vs. 65.3%, log-rank p=0.02). There was no difference in major adverse events between the groups. All-cause mortality was not significantly different between the groups and no deaths were deemed to be device or procedure related.

PES also have a demonstrated long-term advantage over PTA and BMS. In the Zilver PTX trial, 474 patients were randomly assigned to PES or PTA with secondary randomization to BMS or DES, if needed [30]. Clinical benefit (symptoms of ischemia), patency, and freedom from TLR were significantly higher at 5 years of follow-up for PES relative to PTA (freedom from reintervention is 83.1% for PES and 67.6% for PTA, p<0.01). Similar results were observed for the group that underwent secondary randomization to PES or BMS with significantly higher rates of clinical benefit and patency. PES had higher rates of freedom from TLR relative to BMS that did not reach statistical significance (84.9% vs. 71.6%, p=0.06).

Based on the advantages and the safety-profile described above, it is unsurprising that the use of PCDs has increased significantly, and rates of PTA and BMS decreased over time [31]. In addition, consensus recommendations from professional societies supported the definitive use of PCBs/PESs [21]. These devices were considered breakthrough therapies and embraced by the vascular community as a major advance in the treatment of patients with PAD.

Mortality Signal in Drug-Coated Devices

In December of 2018, Katsanos et al. [22••] published unexpected results that dramatically changed the vascular communities’ use of PCDs. Specifically, they conducted a summary-level meta-analysis of 28 RCTs designed to compare limb-related outcomes between drug-coated devices and nondrug-coated devices (24 PCB and 4 PES) [22••]. At 1 year, the authors found no difference in mortality among 28 trials which included 4432 patients. At 2 years, the authors found a 68% increase in risk of all-cause mortality for patients treated with PCDs among 12 trials and 2316 patients. At 5 years, the authors found a 93% increase in risk of all-cause mortality among 3 trials and 863.

The Katsanos meta-analysis has been criticized for methodological flaws that may have produced biased results. First, the studies included in the meta-analysis were not able to analyze that data as time-to-event or censor patients who were lost to follow-up. This biased the analysis as these studies were originally designed to assess short-term safety, such as 30-day and 1-year mortality, and patients were not often followed long-term. It was also unclear if loss to follow-up was random or non-random. Second, the trials included heterogeneous patient populations with different burdens of comorbidities, lesion characteristics, and types of devices (stent versus balloon). Third, the authors reported a dose-response relationship between paclitaxel and risk of death, but the methods used to make these calculations were flawed as they used lesion length as a proxy for paclitaxel dose. Specific device brands, however, use different doses of paclitaxel, which are also impacted by the excipients used. Finally, for the 5-year data, not all patients had reached 5 years and only 3 trials with <1000 patients were available to analyze.

In response to the meta-analysis, the FDA held a Medical Device Advisory Panel in June of 2019 to review available data regarding the safety of PCDs. They conducted an internal meta-analysis of clinical studies of FDA-approved devices which included randomized-controlled trials with at least 200 patients and at least 2 years of follow-up. Some of these trials had been analyzed in the Katsanos meta-analysis including ZILVER PTX RCT, IN.PACT SFA 1 and 2, LEVANT 2, and ILLUMENATE. They concluded that a late mortality signal was present; they emphasized, however, that this finding was qualified by significant missing data, small sample sizes, no dose-response relationship between paclitaxel and mortality, and no mechanism to explain an increase in mortality caused by paclitaxel [32].

Observational data were also presented at this meeting, all of which observed no mortality signal related to paclitaxel. Dr. Eric Secemsky presented an expanded analysis of two large observational studies he published from the Medicare data [33, 34]. This analysis included over 150,000 patients from the inpatient and outpatient setting who underwent femoropopliteal revascularization. Dr. Secemsky found no increase in mortality for patients treated with PCDs over a median of 799 days (HR 0.94, 95% CI: 0.930.96). Dr. Robert Yeh presented an analysis of the Optum claims database. In over 20,000 patients who underwent femoropopliteal revascularization, there was no increase in mortality over a median follow-up of 763 days (HR 1.09, 95% CI: 0.98–1.22). Dr. Daniel Bertges presented an analysis performed in the Vascular Quality Initiative Peripheral Vascular Intervention Registry. He analyzed 8000 patients over a median follow-up of 12.4 months and found no relationship between PCDs and mortality (HR 0.82, 95% CI: 0.68–0.98).

Following the advisory panel, the FDA felt that no conclusions could be drawn about the safety of PCDs. They recommended that PCDs be used for patients deemed to be at high-risk of restenosis or repeat femoropopliteal interventions if the benefits outweigh risk, that physicians initiate risk-benefit discussions with PAD patients incorporating the new information regarding a mortality signal in paclitaxel, and that device labeling be updated to reflect this mortality risk. They charged the vascular community with more rigorous long-term safety monitoring in patients who underwent RCTs and asked for the Medicare claims, Optum, and VQI analyses to be extended over time.

Current State of Investigation into the Mortality Signal

We now have multiple additional sources of data investigating the late mortality signal associated with paclitaxel exposure, including longer-term follow-up from industry-sponsored RCTs, an individual patient-level meta-analysis with additional follow-up data, and large observational studies.

A critical individual patient-level meta-analysis was published in 2020 sponsored by the VIVA society. This analysis was long-awaited, as it examined not only individual patient data from 8 paclitaxel studies (vs summary-level data analyzed in the Katsanos meta-analysis) but also included data with reduced loss to follow-up, which was approximately 10% from greater than 30% [35•]. .The author found that upon reducing the degrees of loss to follow-up, the risk estimate went from as high as 1.93 (95% CI 1.27–2.93) down to 1.27 (95% CI 1.03–1.58). Mortality risk was increased for all major causes of death and notably, in one of the most definitive paclitaxel dose-mortality assessment, no relationship was found.

The results from individual industry-sponsored clinical trials have also been reassuring. Schneider et al. [36•] published a long-term follow-up of combined data from the IN.PACT SFA (n=331) and IN.PACT Japan trials (n=100) comparing PCB with PTA. There was no significant difference between cumulative incidence of mortality between PCB and PTA at 5 years. Using a multivariable Cox regression, they evaluated predictors of mortality and found that paclitaxel dose was not an independent predictor of mortality, but did predict a significantly lower risk of TLR (HR 0.79; p<0.001). They analyzed causes of death in both treatment arms and found no difference in major causes of death [36•].

Results from the ILLUMENATE randomized trials involving 589 patients also demonstrated no difference in outcomes in follow-up through 1080 days. No difference in the cause of death between those treated with and without a PCB also found no clear etiology of death. The 4-year results from the ILLUMENATE Pivotal trial presented at the VIVA LBCT in June 2020 provided more updated data again showing no significant difference in all-cause or cardiovascular mortality between devices [37].

An updated analysis from the LEVANT clinical trials combined data from the LEVANT 1 & 2 trials, LEVANT 2 CAR trial, and LEVANT Japan trial. It included 1093 patients who received PCB and 250 patients who received PTA. There was not a significant difference between mortality rates between PCB and PTA. No dose-response relationship was found between paclitaxel and mortality and there was no clustering of a specific cause of death between PCB and PTA cohorts [38].

Dake et al. [39] 2019 published the 5-year results of the Zilver PTX trial in which 336 patients were randomized to DES and 143 were randomized to PTA and received a BMS, if needed. This analysis differed from prior analyses as the investigators examined survival of the as-treated population. This varied from the prior intention-to-treat analyses by examining anyone exposed to a PES, including those who crossed over to PES after randomization. Among these as-treated patients, the authors observed no significant difference in mortality between groups. Furthermore, they analyzed the causes of death in each arm and found no significant differences between the causes of death in each group.

Many observational studies have now been performed. They consistently show no increase in mortality in patients treated with PCDs relative to uncoated devices. First, updated analyses from the Optum Medicare Advantage Cohort were presented by Dr. Eric Secemsky at the VIVA LBCT Webinar in June of 2020. This analysis of 16,796 patients with median follow-up time 2.66 (IQR 2.02–3.52) and longest follow-up time of 4.75 years demonstrated no difference in mortality between the patients treated with PCDs and those treated with uncoated devices [37].

Weissler et al. [40] performed an analysis of outcomes in patients treated with PCB relative to PTA in the Medicare population over a 2-year follow-up period. There were 91,763 patients who underwent femoropopliteal peripheral vascular intervention with PCB or PTA. Patients treated with PCD had a lower unadjusted cumulative incidence of all-cause mortality compared with those who underwent PTA (22.3% vs. 28%, p<0.001). The risk of all-cause mortality remained lower among the patients treated with PCB following adjustment using inverse probability weighting (aHR 0.92, 95% CI 0.89–0.96, p<0.001).

Saratzis et al. [41] performed a retrospective cohort study in a database of 3 centers comparing patients with claudication and chronic limb-threatening ischemia (CLTI) who received PCDs compared with those treated with uncoated devices using a multivariable Cox regression. They found no association between the use of paclitaxel and mortality after adjusting for patient risk factors over a mean follow-up time of 24 months.

Bohme et al. [20] performed a retrospective propensity-matched single-center analysis of patients with a median of 51 months of follow-up who underwent PCB or PTA to femoropopliteal lesions. They found a mortality rate of 27.5% after POBA and 16.9% after PCB (p<0.001). There was no correlation between PCB length (as a semiquantitative measure of drug dose) and mortality.

Behrendt et al. [42] performed a large, retrospective analysis in the German BARMER health insurance claims database. There were 21,546 propensity score-matched patients who underwent revascularization between 2010 and 2018. The study was stratified by CLI versus IC, balloons versus stents, and paclitaxel versus no paclitaxel. In both the CLTI and CI cohorts, patients treated with PCDs had a lower mortality rate at 5 years relative to patient treated with non-coated devices (31.8% vs 35.8% for CLTI and 9.4% vs 10.5% for non-coated). In the Cox-proportional hazards analysis among patients with CLTI, they observed a survival benefit for patients who were treated with a PCB alone, treated with PES alone, and treated with PCB and PES in combination. In the group with IC, they reported a survival benefit for patients who received PCB and for the combined PCB and PES group, but no difference for the group who received PES alone. Using the same BARMER insurance database, Freisinger et al. [43] analyzed 64,771 patients over a median time period of 7.6 years. They found that neither PES nor PCB was associated with increased mortality relative to non-coated devices over the study period (HR at 5 years for PES vs non-coated devices: 1.01, 95% CI 0.83–12.3, p=0.91; HR at 5 years for PCB vs non-coated devices: 0.97, 95% CI 0.89–1.06, p=0.492).

Bertges et al. [44] analyzed mortality after treatment with PCDs and non-coated devices in the Society for Vascular Surgery Vascular Quality Initiative registry. After adjustment using propensity matching, they report that mortality was significantly lower in patients treated with PCDs (8.5%) compared with non-coated devices (11.5%) (HR 0.82, 95% CI 0.68–0.98, p=0.03) at 1 year. Mortality was lower for PCDs (1.6%) relative to non-coated devices for patients with intermittent claudication (4.4%) (HR 0.59, 95% CI 0.39–0.89, p=0.01), and there was no significant difference in mortality for patients with chronic limb-threatening ischemia treated with PCDs relative to non-coated devices (15.5%) (HR 0.85, 95% CI 0.72–1.0, p=0.05) at 1 year.

Hess et al. [45] recently presented a prespecified subgroup analysis of patients in the VOYAGER PAD trial as late-breaking science at TCT Connect in October of 2020. They analyzed the 4379 patients who underwent endovascular revascularization during the study, 31% of whom received a PCD. They adjusted for baseline comorbidities using inverse probability of treatment waiting. Following adjustment, they found no difference in the risk of all-cause mortality among patients treated with PCDs compared with those who were treated with non-coated devices (HR 0.95, 95% CI 0.83–1.09, p=0.49).

There is also an ongoing large-scale safety analysis being conducted by Dr. Eric Secemsky called the Safety Assessment of Femoropopliteal Endovascular treatment with Paclitaxel-coated Devices (SAFE-PAD). This longitudinal study, designed with feedback from the FDA, aims to investigate long-term survival in patients treated with drug-coated devices relative to those treated with non-drug-coated devices in the Medicare population until the median follow-up surpasses 5 years. This study also includes a series of sensitivity analyses to evaluate for the influence of unmeasured confounding, and prespecified subgroup analyses to look at key subgroups (low-risk population, inpatient vs. outpatient, CLI vs. no CLI, and stent vs angioplasty).

Discussion

PAD is a morbid disease and durable, effective treatments are sorely needed. PCDs have established superior patency relative to PTA/BMS and were considered the standard therapy for endovascular revascularization up until the safety concern associated with PCDs. Since publication of the Katsanos meta-analysis, a plethora of studies have been published examining the long-term safety of PCDs (Fig. 2). Notably, only three studies have persistently shown a mortality signal associated with PCDs (Katzanos analysis, the FDR panel’s analysis, and VIVA physician’s analysis), and these are all meta-analyses primarily involving the same randomized trial data. It is worthwhile to point out that the theoretical risk associated with PCDs has attenuated substantially with the inclusion of lost patients. This suggests that there may have been non-random loss to follow-up, as indicated by Schneider et al. [25] and there is still approximately 10% missing data. Furthermore, the magnitude of this relationship has not been as strong when examining centers outside the USA, adding further suspicion that there is truly a mortality signal or a spurious finding.

Fig. 2.

Data examining the long-term safety of PCDs. Three studies have persistently shown a mortality signal associated with PCDs and they are meta-analyses primarily involving the same randomized trial data [22••, 32, 35•]. Multiple clinical trials and large observational studies have not replicated this signal of harm

Multiple clinical trials and large observational studies have also not replicated this signal of harm associated with PCDs (Fig. 2). There is no plausible mechanism linking paclitaxel to a cause of death and no established relationship between paclitaxel dose and mortality. Many studies have analyzed causes of death, as summarized in this review, and have not found significant differences in the etiology of death between those who received PCDs compared with non-coated devices.

We are currently at an impasse on further demonstrating the long-term safety of PCDs. The clinical trials that could help settle the matter are enrolling patients slowly or not at all as they await more definitive direction from governing bodies. A dedicated de novo RCT to evaluate long-term mortality with PCDs was deemed impractical at the FDA Advisory Committee [32]. At this juncture, we are forced to make reasonable conclusions and treatment decisions from the available data, as these devices have important clinical implications for our patients.

In a field where technology evolves quickly, it is critical to have a balance between expedited delivery of treatments to patients in need while simultaneously establishing safety. In the case of paclitaxel-coated devices, early trials demonstrated efficacy and safety in the short term. The devices were approved without long-term follow-up data in order to facilitate treatments for a patient population in need. If there is a positive take-home message from this controversy, it is that, for future trials, maximizing medical therapy and committing to long-term follow-up will be a necessity.

Abbreviations

BMS

Bare metal stent

CLI

Critical limb ischemia

CLTI

Critical limb-threatening ischemia

PCB

Paclitaxel-coated balloon

PCD

Paclitaxel-coated device

PES

Paclitaxel-eluting stent

PTA

Percutaneous transluminal angioplasty

PAD

Peripheral arterial disease

POBA

Plain old balloon angioplasty

RCTs

Randomized-controlled trials

SET

Supervised exercise therapy

TLR

Target lesion revascularization