Signal of Late Deaths With Paclitaxel-Coated DCBs and Stents for PAD Warrants Urgent Review, Authors Say

 

(UPDATED) In a disturbing new study, use of paclitaxel-coated balloons and stents in peripheral artery disease appeared to heighten the risk of all-cause death in patients with femoropopliteal artery disease, calling into question what has been the mainstay therapy in this hard-to-treat group.

The researchers, led by Konstantinos Katsanos, MD, PhD (Patras University Hospital, Rion, Greece), note that DES in the coronary arteries—most notably the early-generation devices, some of which included paclitaxel—“have been long incriminated for late stent thrombosis with an associated risk of death.” Hints of late mortality, however, with DES or drug-coated balloons (DCBs) in the lower limbs from several recent trials have been “dismissed by expert review panels as statistical artifacts or anomalies, and both devices are currently under clinical use with extended on label indications,” they say.

In their pooled meta-analysis of 28 of the most rigorous randomized controlled DES and DCB femoropopliteal trials of recent years, Katsanos and colleagues conclude that while the paclitaxel and control arms showed equivalent all-cause mortality at 1 year, a dramatic change occurred beyond that point. At 2 years, for example, there was a 68% relative risk increase in all-cause death with paclitaxel-coated devices compared with an uncoated balloon, with a number-needed-to-harm of 29. At 5 years, the relative risk increase was a whopping 93%, with a number-needed-to-harm of just 14.

Commenting on the study for TCTMD, Sahil A. Parikh, MD (NewYork-Presbyterian/Columbia University Irving Medical Center, New York, NY), said some may be quick to compare this meta-analysis with the study by Dr Edoardo Camezind, MD (University Hospital Geneva, Switzerland), that kicked off the 2006 “firestorm.” At that time, a meta-analysis presented at the European Society of Cardiology Congress called into question the safety of first-generation coronary-artery DES. According to Parikh, however, confirmation of the Katsanos meta-analysis is needed before any kind of panicked reaction ensues.

“I’m not sure I fully understand the data and I think we need better, more detailed information than what was provided as part of this meta-analysis,” he said. If the association with increased all-cause mortality is real, however, “this could be a reckoning for endovascular therapy.”

The study was published December 6, 2018, online ahead of print, in the Journal of the American Heart Association.

This could be a reckoning for endovascular therapy. Sahil A. Parikh

Of the 28 RCTs included in the 4,663-patient meta-analysis, four involved paclitaxel DES while 24 were testing paclitaxel DCBs.

At 1 year, the crude risk of all-cause death was 2.3% in the paclitaxel-device groups (combined DES and DCB) and 2.3% in the control groups (RR 1.08; 95% CI 0.72-1.61). At 2 years, the risk was 7.2% and 3.8%, respectively (RR 1.68; 95% CI 1.15-2.47), and at 5 years the risk was 14.7% and 8.1%, respectively (RR 1.93; 95% CI 1.27-2.93).

Meta-regression analysis indicated a significant association between paclitaxel dose-time product and absolute risk of death, with 0.4 ± 0.1% excess risk of death for every paclitaxel milligram-year (P < 0.001). There was no statistical heterogeneity or major diversity between the included studies.

“Interestingly, the risk of death beyond 1 year also seemed to vary among different paclitaxel dosages, being significantly higher in the 3.5 µg/mm2 devices compared with the lower-dose devices,” Katsanos et al write.

Claims of Toxicity and Suboptimal Data

As for what could be behind an association between paclitaxel used in the lower limbs and increased risk of all-cause death, the study authors say that remains unknown.

They do hypothesize, however, that it could be related to late paclitaxel toxicity resulting from retention of the drug in the vessel wall of the iliofemoral arteries. When used as a cancer therapy, intravenous paclitaxel has a half-life of about 6 hours, Katsanos and colleagues note, “while paclitaxel crystals loaded on DCB or DES for the peripheral arteries have a half-life of weeks to months, depending on the exact chemical properties of the applied paclitaxel formulation.” They cite “bench and animal studies” showing that microparticle formation from crystalline paclitaxel can embolize in the downstream systemic circulation, possibly leading to amputations.

To TCTMD, Parikh said the authors explanations amount to “speculation that fails to prove causality.”

Taken as a whole, he said, the study does raise important issues about the quality of trials of endovascular devices. Katsanos and colleagues point out that the meta-analysis was underpowered and unable to discern differences between the individual paclitaxel-coated devices, and that most of the trials did not have follow-up beyond 2 years. Importantly, most of the studies also did not report the actual cause of death, making it impossible to differentiate something such as an accidental death from mortality following limb amputation.

“To date we've been unfortunately dealing with less-than-optimal data because of the conduct of trials that were not as rigorous as what we expect in the coronary field,” Parikh stated. As head-to-head trials move forward between various paclitaxel DCBs, in particular, he said the meta-analysis may represent “a wake-up call to industry that you have to have very high-quality data because that’s what the standard of care is in the research arena these days. We need better and more clearly constructed randomized controlled trials.”

Bias Introduced?

Mehdi Shishehbor, DO, PhD (University Hospitals Cleveland Medical Center, OH), who also commented on the study for TCTMD, called the study results “surprising” and the meta-analysis itself rigorous and well done, adding “I don’t think this finding is by chance.”

He disagrees with the conclusion of Katsanos and colleagues, however, that there is a mechanistic explanation for the deaths that directly implicates paclitaxel.

“I think that this is most likely a statistical association rather than a biological association. One possible explanation may be that the people getting angioplasty are getting more frequent evaluation [because] they have more restenosis,” he observed. “There is a lot of follow up in the first year of these studies. After the first year it decreases, and if the patients are doing well they are not coming back.” The ones who are returning for care are potentially getting more rigorous treatment that may be preventing or controlling serious health issues that could lead to death, he added.

Again, this raises issues about the follow-up done in the individual trials, but Shishehbor argued that even if the study authors had access to patient-level mortality data it would be wrong half of the time because the only true way to know cause of death is by autopsy. For that reason, he does not feel the lack of mortality data should be considered a limitation of the study.

Shishehbor said meta-analyses like this one and the first-generation DES one that caused such controversy a decade ago are often the only way to tease out mortality differences because the individual trials are underpowered for the endpoint and any signals that were detected would be unlikely to have alarmed anyone until they were put together in pooled data.

As for what should happen next, he urged more evaluation into the care of the patients in the individual trials beyond the first year to see if the more intensive follow up of the control groups may indeed explain the mortality difference, and also a deeper look into what happened to patients who withdrew from follow up.

“I don’t think we need to panic,” Shishehbor concluded. “Unless somebody shows me something different I will continue treatment of my patients the way I have been doing.”