Acknowledgment of Safety Signal for Paclitaxel Devices in PAD May Mark Sea Change in Controversy

Washington, DC—In a Town Hall here at the CRT 2019 meeting, researchers, clinicians, and statisticians all agreed that the safety signal for paclitaxel-based devices used in the treatment of femoropopliteal artery disease cannot be dismissed. That consensus comes even in the face of incomplete and possibly erroneous data, confusion as to whether paclitaxel itself is harmful, and uncertainty about how to proceed.

Another thing was crystal clear, based on a unanimous vote taken at the end of the day: patients should be informed about the suggestion of a mortality risk before being offered one of the devices.

In an interview with TCTMD after the all-day session concluded, co-moderator David Kandzari, MD (Piedmont Heart Institute, Atlanta, GA), said the discussions had been very productive and agreed that the perspective seemed to have morphed over the last few weeks and months from criticizing and attempting to disprove the findings, to bringing together experts who "came in more with a fundamental agreement that we do see some signal of safety.” What’s needed now, he continued, is a better understanding of the next steps needed to move forward.

Controversy has dogged the meta-analysis by Konstantinos Katsanos, MD, PhD (Patras University Hospital, Rion, Greece), since its publication in late 2018. Its release stunned the endovascular community by suggesting there was a signal of increased death beyond 1 year, continuing out to 5 years, in patients treated with paclitaxel-eluting stents or drug-coated balloons (DCBs) compared with uncoated balloon therapy. Since that presentation, there have been two high-profile gatherings, one at the LINC meeting in Leipzig, Germany, and a second dubbed the VIVA Physicians’ Vascular Leaders Forum (VLF), that addressed the topic.

Every Possible Perspective

The Town Hall session entailed hours of detailed discussions about theories of paclitaxel toxicity, industry presentations on updated data sets, input from US Food and Drug Administration representatives, and two separate talks by Katsanos. Despite the range of viewpoints, a patient advocate on the panel noted the shift in tone.

The advocate, Naftali Frankel, MS (Kew Gardens, NY), acknowledged the obvious strife that has surrounded the issue. He had watched Katsanos’ presentation at the recent LINC meeting and said: “It was striking to me after he finished presenting that someone from industry interrupted his statistical analysis very sharply disputing the validity of his analysis. Someone harped on the software that he used, that it was free. Then I noticed at the VLF meeting that [an FDA spokesperson] noted that she was involved in the approval and supported the approval of these devices [and] that based on their analysis, taking multiple different methods, that they consistently found that there indeed is a signal, that there is legitimacy to the analysis, and that there was not a statically glitch.

“Today at CRT,” Frankel continued, “we are moving forward from that and discussing more on a granular level as far as what exactly is the reason why we're seeing this in the meta-analysis, not whether there is foundational legitimacy to the analysis.”

For his part, Katsanos told TCTMD in an interview yesterday that he is pleased with the response from the endovascular community.

"The industry, the amount of information that they have collected and have started to share, is invaluable and is tremendous,” he observed. “Over the 3 months since the publication, everybody seems to be taking this very, very seriously. There is uncertainty about our findings and limitations, but at the end of the day I think we are going down the right direction.”

Dose-Response, Doubts, and Unaccounted Deaths

The day followed a format similar to the VLF session last week, with considerable discussion devoted to all aspects of the original meta-analysis, omissions that have since been made public regarding certain trial data, long-term patient-level data, including from Medicare sources, and the involvement of the FDA, which has said it is looking at all available data from the RCTs and registries in an effort to consider how best to weigh the benefit and risk to patients.

FDA Medical officer Donna Buckley, MD, noted that “the mortality signal is enhanced or muted depending on how you look at the data and how you analyze it.” She added that while there may be some frustration about how the FDA is currently proceeding, “final decisions regarding actions and communications thereof is made at a higher agency level and the timelines of that might not perfectly sync up with conferences and talks . . . . We'll let you know as soon as we can, and we are honestly trying to be as transparent as possible.”

Two other important takeaways from the CRT session were acknowledgment from the panel that there has been some retreat from the use of the devices in everyday practice until more information becomes available as well as uniform agreement that patients need to be made aware of the potential mortality risk before they agree to have a procedure.

“We did not have the opportunity to delve into what that conversation should be,” Kandzari noted to TCTMD. “But as I've suggested, I think it's fair for doctors to begin the discussion with their patients before the procedure to simply indicate that there is some data that would suggest higher mortality with the use of these technologies, there’s still uncertainty regarding the accuracy of that data, studies are ongoing, and until we have those studies we know that these technologies do offer a benefit of avoiding revascularization. But [also that] there still remains this outstanding concern of higher mortality that we need to resolve.”

The panel, co-chaired by Jeffrey J. Popma, MD (Beth Israel Deaconess Medical Center, Boston, MA), wrestled with the seemingly illogical conclusion that a small dose of paclitaxel on a device in the peripheral artery could contribute to mortality, while paclitaxel in much greater doses has been found to be safe for pregnant women, fetuses, and cancer patients.

“Is there a chance that we’re missing the villain by focusing on paclitaxel?” Popma mused. “Have we ruled out other components that are not paclitaxel?”

Presenter Aloke V. Finn, MD (CV Path, Gaithersburg, MD), who spoke about the pharmokinetics of paclitaxel, said the chance that the excipient is responsible was extremely unlikely. Finn also noted that there has been no clear signal of one type of mortality—due to neutropenia, for example—that is increased in patients treated with paclitaxel-based devices.

Much talk was devoted to the idea of a dose-dependent effect of paclitaxel, and it was clear that there was no uniform consensus. “We need to understand the exposure to the procedure, not necessarily to paclitaxel,” Popma noted.

Another major point of interest was the lack of complete follow-up and loss to follow-up in the RCTs.

"I don't think we can even start to implicate [specific] factors until we resolve the issue of not knowing in some studies what has happened to 20% or more of patients who were lost to follow up, not knowing how patients who underwent subsequent procedures were treated, and not knowing what the cause of death is for many of these patients who are listed as dead at follow up," Kandzari told TCTMD.

Another issue that was been brought to the forefront by presenter William A. Gray, MD (Lankenau Heart Institute, Wynnewood, PA), is that the Katsanos meta-analysis assumes that the percutaneous transluminal angioplasty (PTA) group was paclitaxel-naive for the entirety of follow-up.

In both the United States and the rest of the world, paclitaxel-based devices were commercially available and were likely used if control patients developed restenosis, he contended. However, those data were not recorded. According to Gray, the average excess of TLR in the PTA groups across all the trials was 10-15% at 1 year. That would translate to between 190 and 285 patients who should not have been exposed to paclitaxel potentially having exposure that should have increased their mortality risk, if the theory about paclitaxel being the culprit is correct.

For Katsanos, the issue is an important one to bring to light. To TCTMD, he added that he also has concerns about the poolability of the existing trial data, something that is about to happen soon as VIVA Physicians facilitates an independent patient-level meta-analysis with the cooperation of all trial sponsors. Panelist Gary Ansel, MD (OhioHealth/Riverside Methodist Hospital, Columbus, OH), said part of that cooperation includes providing intention-to treat and as-treated data. According to statistician Sharon-Lise Normand, PhD (Harvard School of Public Health, Boston), the as-treated group is the most important to look at from a safety perspective, but not everyone agreed.

"The intention-to-treat principal seems to me to be very, very valuable because otherwise you may be contaminating or perverting the data sets," Katsanos told TCMTD. However, the controversy surrounding his meta-analysis has created an opportunity to lobby for release of all anonymized patient data sets for additional research, he noted.

"It cannot be only VIVA, it cannot be only a nominated, select group of physicians. It has to be to the public domain. It's going to be an amazing precedent and perhaps the most valuable thing to gain from this whole discussion if the endovascular community starts to share this data for . . . any kind of independent analysis in the future, and I think the industry might be headed that way.

Asked if he thought physicians should be using the lowest-dose device available for now until more data come in, Katsanos said, “I think we need to use paclitaxel very judiciously and very sparingly. Perhaps only in young people because age has been one of the predictors of mortality as well.” Overall, the most important point is to avoid high cumulative doses of paclitaxel, he advised.

At the end of the Town Hall, the panel voted on four questions. The responses were anonymous, although some voters volunteered their response and why they voted the way they did. Again, it was clear there was no uniform consensus except for whether to tell patients about the controversy.

1)      Question: Given the data from the meta-analysis and other data that was presented today, is there a safety signal for DCB and DES in the SFA?

Answer: 10 voted yes, four voted “I don’t know,” zero voted no

2)      Question: Should the current labeling for paclitaxel DES/DCB be changed?

Answer: three voted it should be restricted, three voted it should be changed, eight voted it should stay unchanged

3)      Question: Should patients be informed about the controversy prior to implantation with a paclitaxel device?

Answer: 14 voted yes

4)      Question: Is the patient-level data of all paclitaxel systems poolable for definitive analysis?

Answer: nine voted yes, one voted no, four voted “I don’t know”