Commentary Regarding Katsanos et al’s Meta-Analysis: Finding a Silver Lining in the Controversy

What is your reaction to the recent meta-analysis by Katsanos et al¹?

It is “a bit of déjà vu all over again.” I suspect many people may have the same feeling, because since Dr. Katsanos’ original 2018 meta-analysis², the whole field has been flipped on its head.

In 2006, we had a similar scare in cardiology with drug-eluting stents, and to me, this past year has felt similar in that there's a lot of smoke, but it's not clear if there's really any fire. Dr. Katsanos and his team have stirred the pot and this latest meta-analysis brings the same kind of pot-stirring. However, in contradistinction to the original meta-analysis, which had more than 20 trials evaluated in a primarily claudicant population, the authors of this study looked at the high-risk critical limb ischemia (CLI) population specifically, and as a consequence, have far fewer studies to consider. Despite that, the authors have been able, at least at the macro level, to show a difference in one year amputation-free survival (AFS) between  paclitaxel-coated device treated patients and those treated without a drug-eluting technology.

As with the original meta-analysis, I think the devil is in the details. What’s interesting is that there is no clear, statistically significant mortality signal at the one-year time point. The amputation signal is also not clearly, on its own, a positive signal, but when you sum the two using AFS as the metric, that is when we see a difference. It is different than what we are seeing with the claudication population where the all-cause mortality was the primary outcome that was statistically significantly greater in the paclitaxel-treated group. The other issue is that the reduced AFS signal is on a much faster timeline. Whereas in the 2018 meta-analysis, there was a signal identified at two years for mortality, here the AFS difference is seen at one year. Finally, there has yet to be a plausible mechanism for mortality identified in the claudicant population. With respect to amputation, there has been a great deal of discussion regarding the role of embolic debris from drug-coated balloons, but there hasn’t been compelling or clear head-to-head randomized data suggesting that drug-eluting devices cause meaningful differences in amputations either. Not only is the mortality mechanism not clear, but the amputation mechanism is likewise not clear. Without a mechanism, it is really hard to understand the association, much less causal relationship, of paclitaxel-eluting devices and mortality or amputation.

In addition, there were a number of methodological flaws in the 2018 meta-analysis which have subsequently been highlighted. In fact, many of the criticisms levied against the original meta-analysis are likely to be applicable here, since the methodologies are similar. For example, I would ask, what is the level of ascertainment in these patients? This is a very sick population. The mortality risk for CLI patients is much higher than for the claudicant population and the loss to follow-up rates are generally higher. What proportion of patients were available for follow-up at the evaluated time points? There is a great deal of “missingness” in adjudication of vital status. As such, it is possible some of the criticisms levied against the original 2018 meta-analysis would also apply here. We won't know until there is more of a deep dive into the individual studies that comprise the analysis to see the level of vital status ascertainment. Moreover, I am also dismayed that Katsanos et al persevere in their dosing analysis in this paper. In the original manuscript for the claudicant population, the dose analysis was resoundingly rejected by people like Elazer Edelman, MD, PhD, of MIT and others who thoroughly understand the relevant biophysics, physiology, and pharmacology of local vascular drug delivery. Here yet again in their sensitivity analysis, Katsanos et al are levying the charge that high dose paclitaxel-eluting devices are more dangerous than low dose devices without really analyzing the actual doses of drug used or delivered per patient. There isn't any good evidence for that in this CLI population or even in the claudicant population for a dose response, and this is not further characterized here. As such, I am surprised that the manuscript doesn’t address these issues in the conduct of the meta-analysis and it is curious that this meta-analysis passed peer review, because these concerns have certainly received attention in the last year, and the methodological flaws in the original meta-analysis have been more clearly vetted and reviewed. I would reserve judgment on the veracity of this analysis until methodological issues have been ironed out, as we have begun to do with the claudicant populations.

It’s worth adding that none of these products are specifically FDA approved for the United States market. The only device that has been studied in the U.S. was the Lutonix below-knee device and that device is not yet FDA approved. So unlike the 2018 meta-analysis, this manuscript is less relevant to the U.S. provider and patient population, in the sense that these devices are not available here in the United States.

Since the initial meta-analysis was published in 2018, our clinical practices have been turned upside down, and we clinicians have worried about possible risks for our patients, because we take that very seriously. And yet, we are also worried about having to do multiple repeat procedures. That's basically the tradeoff: repeat procedures on patients who have meaningful clinical symptoms versus a risk of mortality which is of an uncertain magnitude. We have been forced to talk more carefully to our patients, get a better understanding of their own personal wishes and interests, and gauge how we treat those patients on an individualized basis. However, there has been some acceptance of critical limb ischemia patients as a separate group. CLI patients are so high risk that if an endovascular procedure is necessary, a surgical procedure may not be able to be done as safely. Consequently, the risk of a repeat procedure may also be somewhat greater. We were using drug-coated balloons in the majority of our patients before the 2018 meta-analysis was published, but there has been a precipitous drop-off in utilization of these technologies since then. Some of the use of DCB/drug-eluting stents is now slowly coming back. It has been widely held, however, that CLI patients’ risk of mortality is so high that it is unlikely these patients would manifest a meaningful mortality hazard from paclitaxel as opposed to the high associated mortality of their own disease. In my own practice, while I was  cautious in younger patients, particularly those with claudication, I have  been less concerned about the risks in those patients with critical limb ischemia. In all cases, we disclose the possible risks fully, but in CLI patients, I did not receive a lot of pushback from patients about using paclitaxel-eluting devices. So this meta-analysis will certainly make our lives more complicated, but I am not sure that it has a lot of new information. Because the concept of risk and benefit can be so complicated, this information is difficult to distill into easily understandable pieces to share with our patients so that they can understand the competing risks. It makes our job as clinicians to optimally treat our patients even more difficult. I suspect that the consequence of this paper will be that many patients will be treated sub-optimally, and that is a shame, because that means we are not providing the best care that medical science can provide.

Any final thoughts?

It is important for the readers to know that the community of clinicians taking care of these patients has taken this issue extremely seriously. In fact, the manuscript was published right after an FDA Registry Assessment Of Peripheral Interventional Devices (RAPID) working group meeting on January 14^th, 2020. RAPID is a public-private partnership including industry experts, society representatives, academicians, and other interested parties, along with the FDA. RAPID has worked together for several years on broad issues relating to peripheral arterial disease-related devices. In December 2018, with the publication of Dr. Katsanos’ meta-analysis² and the subsequent firestorm after the FDA panel meeting in June 2019, the group crystalized around this controversy and is now focusing on paclitaxel-eluting devices. I have been representing the American College of Cardiology in the Multi-Society Coalition reviewing this matter independently and the RAPID working group, and I am spending several hours each week thinking about paclitaxel-eluting products and their risk and benefit as part of various research and policy groups. It is refreshing that there are a number of societies, including now all the surgical groups as well as the nonsurgical groups, working together to help understand and better treat our patients. If there is any silver lining to this controversy, it is that for the first time, all of the partisan groups have come to work together in partnership with the FDA and industry to better understand this signal and hopefully prospectively, all of the future devices being developed will have even better quality science to support their use. The controversy has also spurred on innovation in the field. People have started to take seriously the growth of other technologies that could potentially supplant paclitaxel-eluting devices with other drug-eluting technologies. That is definitely good for the field, because we had been a little stagnant in terms of innovation. Finally, I think patients must have individual conversations with their doctors, rather than be misled by any hyperbole fostered by poor data. It remains important, as always, for patients to talk to their doctors and make individualized decisions.

Disclosures: Dr Parikh receives research support from Shockwave Medical, TriReme Medical, Surmodics, Silk Road Medical, Boston Scientific, and the National Institutes of Health. He receives consulting fees/honoraria from Terumo, Asahi, Heartflow, and Merril Life Sciences. He is on the advisory board of Abbott, Medtronic, Boston Scientific, CSI, and Philips. He is a board member of the American Board of Vascular Medicine.

Dr Sahil A. Parikh can be contacted at sap2196@cumc.columbia.edu

References

1. Katsanos K, Spiliopoulos S, Kitrou P, Krokidis M, Paraskevopoulos I, Karnabatidis D. Risk of death and amputation with use of paclitaxel-coated balloons in the infrapopliteal arteries for treatment of critical limb ischemia: a systematic review and meta-analysis of randomized controlled trials. J Vasc Interv Radiol. https://doi.org/10.1016/j.jvir.2019.11.015.
2. Katsanos K, Spiliopoulos S, Kitrou P, Krokidis M, Karnabatidis D. Risk of death following application of paclitaxel-coated balloons and stents in the femoropopliteal artery of the leg: a systematic review and meta-analysis of randomized controlled trials. J Am Heart Assoc. 2018 Dec 18; 7(24): e011245.