A Vascular Annual Meeting (VAM) sponsored session on paclitaxel safety drew a mix of skeptical and cautious responses to the latest meta-analysis from interventional radiologist Konstantinos Katsanos, MD, that reported a heightened risk of major amputation after use of paclitaxel-coated balloons (PCBs) in femoropopliteal and infrapopliteal arteries.
A question posed from the floor of the Boston Scientific and Medtronic co-sponsored panel discussion raised the recent in-press publication in the European Journal of Vascular and Endovascular Surgery (EJVES) from Katsanos and colleagues following a series of presentations that covered harm signals raised by the assistant professor from Patras University Hospital, Greece, in 2018.
“This is a really interesting discussion, and you’ve convinced me that there’s no mortality—great. What about the most recent Katsanos meta-analysis, in press, EJVES, 66% increase in amputation, dose dependent and potential mechanism of embolization? Now what?” an audience member asked as the breakfast session Aug. 19 in San Diego entitled “Paclitaxel safety: A view from multiple perspectives” wound down.
Panel members Peter Schneider, MD, professor of vascular and endovascular surgery at the University of California San Francisco, Daniel Clair, MD, surgery department chair at the University of South Carolina School of Medicine in Columbia, and Darren Schneider, MD, chief in the division of vascular surgery and endovascular therapy in the University of Pennsylvania Health System in Philadelphia, tackled the question along with audience participant Joseph Mills, MD, professor and chief in the division of vascular surgery and endovascular therapy at Baylor College of Medicine in Houston.
Peter Schneider opened the round of responses, stating that there was “good and bad” to be taken from the publication. If “real,” he said, the findings are “important,” continuing: “One good thing is at least there’s a mechanism that you can postulate [with amputation]. With mortality, I couldn’t imagine how [the use of PCBs] causes death. The interesting thing about the mortality issue that we haven’t discussed is that paclitaxel is one of the most common chemo-therapeutic agents for breast cancer. Breast cancer patients get not 10x, not 100x— but a 1,000x higher dose than what we can give in DCB [drug-coated balloon] or DES [drug-eluting stent] treatment. And, in fact, the patients who get that live longer, and a lot of them are treated for early-stage breast cancer … and there is no amputation risk in that group … To make a long story short, I would feel more comfortable if that data had been published by virtually anyone else in the entire world.”
As Vascular Specialist reported when news of the latest Katsanos et al meta-analysis emerged, the harm signals suggested for paclitaxel devices from previous meta-analyses have not been substantiated by a wealth of real-world datasets that have shown no long-term safety concerns, and instead suggested increased benefits with the use of PCDs. Importantly, an interim mortality analysis of the SWEDEPAD registry randomized trial did not confirm a heightened mortality risk in cases of paclitaxel treatment, as noted in the EJVES paper.
As reported in EJVES, the authors analyzed 21 RCTs with 3,760 lower limbs treated nearly equally for intermittent claudication (52%) and chronic-limb threatening ischemia (CLTI; 48%). The median follow-up period was two years. The authors reported 87 major amputations in 2,216 limbs in the paclitaxel arm (4% crude risk) compared with 41 major amputations in the 1,544 limbs in the control arm (2.7% crude risk). The risk of major amputation was significantly higher for PCBs with a hazard ratio (HR) of 1.66 (95% CI 1.14–2.42; p=0.008, one stage stratified Cox model). The observed amputation risk played out equally across femoropopliteal (p=0.055) and infrapopliteal (p=0.055) arteries. In this systematic review and meta-analysis, the authors suggested the summary effect demonstrated a 66% higher relative risk of major amputation in the limbs treated with PCBs. Katsanos and colleagues also reported that there was good evidence of a significant non-linear dose relationship with accelerated risk per cumulative paclitaxel dose. The results were also stable across sensitivity analyses (clarified as pertaining to different models and subgroups, based on anatomy and clinical indication and excluding unpublished trials).
Schneider further stated that he did not believe journal review processes are equipped to manage such papers. “We have to go over it with a fine tooth comb,” he said. “I guarantee you none of the statistical reviewers did that yet.”
Clair pointed to the fact the amputation risk-focused meta-analysis—like the previous one trained on mortality—does not use patient-level data from the trials analyzed. “There’s a real flaw in that methodology because you don’t know what you’re comparing,” he said. “It’s saying, ‘We’re just going to take all of these trials and put them together and see what we come up with.’ But there are real differences in how the patients are managed in those trials, how they present, which patients qualify for the different trials, and, in addition, like Peter [Schneider] said, if you look at the published results of those trials, they differ from the information that [Katsanos] has in his meta-analysis. So, I am really struggling.” Clair added that more information would have to be amassed in order for specialists to continue to advance therapy.
Speaking from the audience, Mills pointed out that EJVES—where he is the section editor for peripheral arteries on the journal’s Editorial Board—boasts an extensive team fully focused on epidemiology, biostatistics and meta-analyses. He said Katsanos’ paper clearly states how he and colleagues decided on what they would analyze, adding, “If you look at levels of evidence, a meta-analysis of randomized controlled trials is the second highest thing there is—the only thing higher is if you can actually get the patient-level data, which is hard to get.”
Turning to the results of the study, Mills said care should be taken to sort through the findings “because there are signals here; we can’t understand them all, but the amputation one is worrisome.” He referred to a reported amputation risk ratio of 1.7 for patients treated with PCDs in femoropopliteal and below-knee arteries—with the risk higher for those with CLTI—commenting that “when you start looking at the actual data, the number of amputations was relatively low. Which gets to my point: We need studies that are re-designed to look at what Peter [Schneider] was getting at, and that is patients in clinically significant endpoints.”
As specialists interested in patency, Mills continued, vascular providers tend to think that if a treated lesion is more patent, then the treatment “has got to be better for the patient. But the actual endpoint for a claudicant would be how far they can walk, what their quality of life is; and, for a CLTI patient, it’s going to be: Do they heal, do they have recurrent wounds? Those questions are a little bit different than the actual lesion-specific thing. I think if we can do anything from this, it would be to get industry to start looking at different endpoints.”
Darren Schneider, meanwhile, agreed that the amputation signal was “a concern,” but cited flaws—including the study’s small numbers and non-patient level data. “But it does concern me that, at least in my first-order-thinking brain, that there is a mechanism that you could put together with amputation, and it may very well be very device-specific as well,” he said.