New Data Add Fuel to Paclitaxel-Device Firestorm in PAD

Patrice Wendling

January 25, 2019

Last month's controversial meta-analysis that suggested a higher late mortality risk with the use of paclitaxel-based balloons and stents for femoropopliteal artery disease has prompted a federal investigation and now a flurry of new research from the major device makers.

A new patient-level meta-analysis shows no correlation between mortality and paclitaxel exposure in patients treated with the IN.PACT Admiral (Medtronic) paclitaxel drug-coated balloon (DCB) in two single-group and two randomized controlled trials (RCTs).

There was no significant difference in mortality when paclitaxel dosage was stratified into low (mean, 5,019.0 μg), medium (mean, 10,007.5 μg), and high (mean, 19,978.2 μg) tertiles (P = .70), with a suggestion of better survival in the highest tertile group, according to study author Peter Schneider, MD, Hawaii Permanente Medical Group, Honolulu.

The report was published online January 25 in the Journal of the American College of Cardiology (JACC) and was part of a phalanx of hurriedly prepared studies and even a lecture on statistical methodology highlighted in a special session at this week's Leipzig Interventional Course (LINC) 2019.

"All the major companies have presented more data, and we still have to check that data in depth, but my first impression is that the 2-year and the 5-year result that we presented in our analysis doesn't change in terms of significance," lead author of the controversial meta-analysis, Konstantinos Katsanos, MD, PhD, Patras University Hospital, Rion, Greece, told | Medscape Cardiology.

After analyzing 28 RCTs, Katsanos and colleagues reported that paclitaxel-based devices were associated with a 68% relative risk increase in all-cause death at 2 years and a 93% relative risk increase at 5 years compared with noncoated devices.

Within days, two European peripheral artery disease (PAD) trials suspended enrollment and the US Food and Drug Administration (FDA) subsequently announced that it is evaluating long-term follow-up data from studies that supported the approval of paclitaxel-based balloons and stents in the United States, along with other available data.

Although critics have questioned how DCBs that release paclitaxel relatively quickly could affect long-term mortality, Katsanos explained that the half-life of paclitaxel in the tissue is between 1 week and 2 months, and increases over time. Paclitaxel content also varies widely between devices, from 0.4 mg with the ELUVIA device to 8.4 mg with IN.PACT.

Devices delivering at least 3.5 mg of paclitaxel are associated with a lower 2-year risk for target lesion revascularization (TLR) than those delivering less than 3.5 mg (relative risk [RR], 0.35; 95% CI, 0.28 - 0.45 vs RR, 0.61; 95% CI, 0.46 - 0.81), but this comes with a higher 2-year risk for death (RR, 2.31; 95% CI, 1.23 - 4.33 vs RR, 1.38; 95% CI, 0.87 - 2.19), according to unpublished research Katsanos presented at LINC.

He suggested that the latent effects observed with paclitaxel-based devices might be related to the fact that paclitaxel interferes with mitosis, causes cell death, and increases aneuploid daughter cells.

Katsanos buttressed this argument with a second unpublished pooled analysis of 1-year FDA-submitted data from the Zilver PTX, Lutonix, Stellarex, and IN.PACT devices that shows consistently higher risks for cardiovascular (RR, 1.73), gastrointestinal (RR, 1.69), infectious (RR, 1.75), and respiratory (RR, 2.23) adverse events, compared with control subjects.

"I don't think we will go back to the Stone Age and not use paclitaxel at all," Katsanos said. "But I don't think we can really say much until the FDA or some other organization does a proper in-depth analysis of all the data."

The new meta-analysis in JACC was performed independently by the Baim Institute for Clinical Research and includes a larger, deeper dataset with longer follow-up than the Katsanos summary-level meta-analysis, Schneider said at LINC.

All-cause mortality was similar between the 1837 patients treated with the IN.PACT DCB and the 143 treated with percutaneous transluminal angioplasty (PTA) through 5 years (9.3% vs 11.2%; = .399). In addition, the mean paclitaxel dose did not differ between patients treated with DCB who died and those who survived (11.8 vs 11.4mg; P = .529).

Increasing age, renal insufficiency, previous target limb amputation, and Rutherford category (4 - 6 vs 1 - 3) were all independent predictors of mortality in multivariable analysis, but paclitaxel dose was not. Patients treated with paclitaxel DCB who died were older with more comorbidities, noted Schneider.

Follow-up visit compliance, a surrogate for repeat contact with the healthcare system, however, was better in patients treated with PTA. It was also lower in DCB-treated patients who died than in those who survived, suggesting an alternative hypothesis for the lower mortality risk observed in control subjects in the Katsanos meta-analysis, he said at LINC.

Katsanos pointed out that follow-up through 5 years was available only from the IN.PACT SFA trial (3 years for IN.PACT Japan and IN.PACT Global; 1 year for IN.PACT China) and that detailed comparisons were made only for paclitaxel doses.

"So basically, in terms of whether paclitaxel increases death or not, we have to stick to the original data that we presented in our paper and are again presented in table 4 of this paper; and it is actually the very same thing," he said.

The IN.PACT SFA results show dramatically higher mortality rates in the DCB group than the PTA group at 2 years (8.1% vs 0.9%), 3 years (10.7% vs 1.9%), 4 years (13.0% vs 6.8%), and 5 years (15.8% vs 9.6%).

Although this difference was no longer statistically significant at years 4 and 5, it begs the question of whether it would have been if there were more patients, said Melina Kibbe, MD, vascular surgeon and chair of surgery, University of North Carolina (UNC) at Chapel Hill.

"We need a real-world prospective randomized trial that's powered sufficiently to look at efficacy end points, as well as mortality, and that is also free of interest," she told | Medscape Cardiology.

She noted that the device industry is a multimillion-dollar business and that the trials done to date have been powered to look at efficacy outcomes, such as patency and TLR rates, but not mortality.

"Most of these trials are in patients with just claudication, and that's another big issue," Kibbe said. "They're highly controlled trials that exclude a lot of the patients who have all the comorbidities that we normally see in our practices," including those with critical limb ischemia.

"Could these drug-coated balloons improve outcomes in a very defined population? Perhaps. I don't think we know yet from the data that are out there," she said.

In the meantime, Kibbe said the meta-analysis has given clinicians at UNC pause over the use of paclitaxel-coated balloons, "so we are not aggressively using them."

In a recent statement, the Society for Cardiovascular Angiography and Interventions (SCAI) took aim at the shortcomings of the Katsanos meta-analysis and highlighted some of the new patient-level data from LINC 2019.

"At present, SCAI concurs with FDA that the benefits of paclitaxel devices continue to outweigh any potential risks," SCAI President David A. Cox, MD, said in the statement. "However, we strongly encourage our members to discuss the findings of the meta-analysis with their patients and to report any safety concerns to FDA."

At LINC 2019, Dierk Scheinert, MD, University of Leipzig, Germany, said there has been no signal of excess mortality with the Lutonix DCB in four randomized trials, including the Levant 2 study. It has the longest follow-up — at 60 months — in 1029 DCB and 145 PTA patients and showed similar rates of all-cause death between the two groups (14.1% vs 10.6%; P = .22).

Newly performed analyses stratifying death into tertiles by paclitaxel dose (<1.88 mg, 1.99 - 3.14 mg, 3.14 - 5.03 mg, >5.03 mg) showed no effect of dose on 5-year all-cause death using either a binary (13.1% vs 12.7% vs 14.5% vs 17.4%; P = .182) or Kaplan–Meier (86.5% vs 85.1% vs 84.0% vs 79.3%; log-rank P = .254) analysis.

"Generally speaking, there are no excess death rates observed in any of the programs so far," said Scheinert, who observed that cardiovascular and respiratory events were the most frequent cause of death for all patients.

A deeper dive into patient-level data from the Zilver PTX study revealed that 31 of the 479 randomized patients who required reintervention within the first year crossed over to the Zilver PTX stent and, thus, weren't accounted for in the Katsanos meta-analysis, Michael D. Dake, MD, University of Arizona, Tucson, said at LINC.

There was no difference in 5-year overall all-cause mortality between the Zilver PTX stent and control treatment with a bare-metal stent or optimal medical therapy (18.7% vs 17.6%; P = .53), nor when death was examined by tertiles of paclitaxel doses ranging from about 0.3 mg to 3 mg (11.5% to 13.2%; P = .72).

Notably, the amount of paclitaxel on a Zilver PTX stent is about 10% to 20% of the amount on a DCB for the same device size and dose density, Dake said.

Investigators for the Stellarex DCB, Eluvia DES, and Ranger DCB reported similar findings at LINC 2019.

Katsanos said he welcomes the new data but noted that each device has particular properties and each company is presenting the data in slightly different ways.

"In some cases, there have been some questionable statistical tactics used, and that doesn't really help," he said. "For the time being, this new information is very welcome but I'm not sure it clarifies the picture. We are still in muddy waters."

In a separate presentation at LINC, Schneider highlighted an invitation-only vascular leaders forum hosted by VIVA Physicians March 1 and 2 that will further examine paclitaxel-drug-eluting technologies in PAD.

"What we're hoping that will come out of this is sort of an interim plan until a full set of patient-level data is available," he said.

VIVA is collaborating with all five American companies (Medtronic, Cook, Boston Scientific, Bard, and Philips) with commercially available paclitaxel-based balloons and stents, who have a data-sharing agreement in which patient-level data will be transferred to a third-party statistical consultant selected by a five-member committee. The FDA has agreed to comment on the statistical methodologies proposed by the steering committee, Schneider said.

"I see this as an opportunity for the vascular community to come together and problem solve, to demonstrate best care for our patients, to do no harm, and to dig deeper to understand better the advantages and disadvantages of this concept of pharmacologic therapies for our patients with lower extremity ischemic disease," he said.

The analysis was funded by Medtronic. Schneider is an advisory board member for Medtronic, Abbott, and Boston Scientific, and a consultant for Surmodics, Silk Road Medical, Medtronic, Cardinal, CSI, and Profusa. He is also chief medical officer for Intact Vascular and Cagent.

Kibbe reports no relevant financial relationships.

Scheinert reports serving on the advisory board or as a consultant for Abbott, Biotronik, Boston Scientific, Cook Medical, Cordis, CR Bard, Gardia Medical/Allium, Metronic, TriReme Medical, Trivascular, and Upstream Peripheral Technologies.

J Am Coll Cardiol. Published online January 25, 2019. Article

Follow Patrice Wendling on Twitter: @pwendl. For more from | Medscape Cardiology, follow us on Twitter and Facebook.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.