A meta-analysis published in the Journal of the American Heart Association (JAHA) late last year by Konstantinos Katsanos (Patras, Greece) and colleagues, suggesting an increased risk of death at two and five years following the use of paclitaxel-coated balloons and stents in the femoropopliteal artery, has generated months of debate amongst the international interventionalist community. The conversation is still ongoing, with every key vascular conference of 2019 to date hosting discussion on the future of paclitaxel-coated and paclitaxel-eluting devices. Since then, prominent physicians have pointed out flaws in Katsanos et al’s methodology. Patient-level data published after the JAHA meta-analysis by industry have revealed no increased mortality associated with the use of paclitaxel-releasing devices. Medtronic and Cook Medical have also announced corrections to their published data. Meanwhile, the continued investigation by the US Food and Drug Administration (FDA) indicates that the story is not yet over.
Speaking from the floor of the Vascular Leaders Forum [see our live coverage here] (VLF; 1–2 March, Washington, DC, USA), hosted by Vascular Interventional Advances (VIVA), FDA investigator Donna Buckley, who has “reviewed most of these devices and recommended them to be approved”, explained the regulatory body’s perspective: “Our first response to the meta-analysis was, I think, similar to everybody else’s: we were a little bit surprised. Nonetheless, it was compelling information that we took very seriously.”
The ongoing FDA evaluation is focusing primarily on US-based randomised controlled trials with data the regulatory body can validate. The FDA’s preliminary evaluation confirmed the mortality signal reported by Katsanos et al. Speaking at VLF, Buckley said the FDA’s own analysis of five US pre-market approval (PMA) trials has “converged to where we feel like it [the meta-analysis findings] is not a statistical glitch—but we still have all this incongruous information as well with all the statistical analysis that we have gone through.”
Also at VLF, Peter Schneider (Honolulu, USA) called the process that the FDA is undertaking to review the data “judicious, straightforward and spot on”.
Indeed, the FDA issued a letter to healthcare providers in January stating that it was evaluating the “recent information regarding the potential for increased long-term mortality” following paclitaxel-coated balloon or paclitaxel-eluting stent treatments in the femoropopliteal artery for patients with peripheral arterial disease (PAD). The agency said that whilst the data review was ongoing, the FDA recommends “continued surveillance” for patients treated with paclitaxel. The regulatory body stated in the letter that it believes the “benefits continue to outweigh the risks” for approved devices within their indications.
At the CRT summit (2–5 March, Washington, DC, USA), which directly followed the release of the FDA’s preliminary evaluation at VLF, polling of panellists in the drug-coated balloon (DCB) safety townhall indicated that the majority of voters believed there to be a “mortality signal” in the femoropopliteal arteris (detailed polling results below).
Katsanos et al’s JAHA meta-analysis
In the original JAHA paper, published 6 December 2018, Katsanos et al conducted a systematic review and meta-analysis of 28 randomised controlled trials investigating paclitaxel-coated balloon angioplasty or paclitaxel-coated metal stents in the femoral and/or popliteal arteries. In all, 4,663 patients (89% intermittent claudication) were analysed. The primary safety measure was all-cause patient death, analysed at different time points.
Previous randomised controlled trials have evidenced that paclitaxel-releasing balloons and stents significantly reduce the rates of vessel restenosis and target lesion revascularisation after lower extremity interventions.
At one year, all-cause patient death (28 randomised controlled trials with 4,432 patients) was similar between 17 paclitaxel-coated devices and control arms (2.3% vs. 2.3% crude risk of death). All-cause death at two years (12 randomised controlled trials with 2,316 patients) was significantly increased in case of paclitaxel vs. control (7.2% [101 deaths in 1,397 patients] vs. 3.8% [35 deaths in 919 patients] crude risk of death, number-needed-to-harm [NNH]: 29 patients [95% CI: 19–59]). Long-term risk of all-cause death up to five years (three randomised controlled trials with 863 patients) increased further in case of paclitaxel (14.7% [78 deaths out in 529 patients] vs. 8.1% [27 deaths in 334 patients] crude risk of death NNH: 14 patients [95% CI: 9–32]).
Katsanos and colleagues wrote that meta-regression showed a significant relationship between exposure to paclitaxel (dose-time product) and absolute risk of death (0.4±0.1% excess risk of death per paclitaxel mg-year; p<0.001). “Trial sequential analysis excluded false-positive ﬁndings with 99% certainty (2-sided alpha, 1.0%),” they said.
These data led the authors to conclude that there is increased risk of death following application of paclitaxel-coated balloons and stents in the femoropopliteal artery. “Further investigations are urgently warranted,” they wrote.
This finding prompted the cessation of patient enrolment of three randomised controlled trials within days of its release: SWEDEPAD 1, SWEDEPAD 2, and BASIL 3. At the date of publication of this newspaper, the status of the SWEDEPAD trials is still uncertain, with principal investigator Mårten Falkenberg (Gothenburg, Sweden) saying at the recent VLF: “We are at the moment struggling to make up our mind on how to proceed.”
Industry respond that their data do not support the potential association of increased mortality with paclitaxel use
Several large industry players presented their own patient-level analysis at the Leipzig Interventional Course (LINC; 22–25 January, Leipzig, Germany), unanimously concluding that they could find no association between paclitaxel dose and mortality.
- An independent, third party, pooled analysis of all the IN.PACT Admiral DCB (Medtronic) clinical programmes, published in the Journal of the American College of Cardiology and including 1,980 patients, demonstrated that at five years, there is no statistically significant difference in all-cause mortality between the DCB and the control arm (9.3% vs. 11.2% respectively, p=0.399).
- Ranger (Boston Scientific) SFA randomised trial three-year data from 105 patients showed no significant difference in all-cause mortality between DCB and control (13.8% vs. 10.7%, respectively), according to the presentation given at LINC with CEC-adjudicated cause of death.
- Results from a Cook Medical press release also demonstrated the safety of their drug-eluting stent, the Zilver PTX. Patient-level data found no increased mortality at five years with Zilver PTX compared to non-coated stents and balloons (18.7% vs. 17.6% respectively, p=0.53).
- A pooled analysis of patient-level data from Philips also demonstrates the “strong safety profile” of the company’s Stellarex DCB in above-the-knee studies, according to a company statement at LINC. The independent, third party pooled analysis evidenced low mortality rates through three years after the treatment with no device-related deaths.
- Five-year data from 1,189 patients in the LEVANT 2 trial within the Lutonix DCB (BD) programme showed no statistically significant difference in all-cause mortality between the DCB and the control group (14.2% vs. 10.6% respectively, p=0.22), as described at LINC.
However, Medtronic and Cook Medical have since issued corrections to their published data regarding the safety of their devices, the IN.PACT Admiral DCB and the Zilver PTX drug-eluting stent, respectively. Medtronic announced they have revised IN.PACT post-market study data due to a “programming error,” and Cook Medical noted a mistake made during data publication: that two mortality figures were “inadvertently reversed”.
According to Medtronic, mortality data were inadvertently omitted from the summary tables included in the statistical analysis of their data presented at LINC. These deaths were, however, previously included and reported in Medtronic’s database, captured in the appropriate study exit forms and adjudicated by an independent clinical events committee, a press release issued by the company said. Immediately upon learning of this error, Medtronic notified the FDA and the study authors. The company added in the release that while a component of the recent patient-level meta-analysis will need to be updated, it has found the revised analysis still supports earlier conclusions.
Cook Medical announced that in a paper published in Circulation in 2016 assessing the clinical effectiveness of their Zilver PTX stent, the mortality figures for the Zilver PTX arm and the percutaneous transluminal angioplasty (PTA) arm were inadvertently swapped. Responding to this error, Circulation state: “When high-resolution files were requested during production, an incorrect version of Figure 1 was mistakenly provided. Subsequently, the published version of the flow chart in Figure 1 contained incorrect numbers. The authors now provide the corrected version of Figure 1.
“The following sentence from the ‘Safety’ section of the paper is incorrect: ‘The five-year all-cause mortality rate was 13.6% (10.2% for the primary DES group and 16.9% for the PTA group, p=0.03), and no deaths were adjudicated as procedure or device related.’ […] The sentence should read: ‘The five-year all-cause mortality rate was 13.6% (16.9% for the primary DES group and 10.2% for the PTA group, p=0.03), and no deaths were adjudicated as procedure or device related.’”
Polling results and informed consent
Polling of 14 panellists at CRT revealed that the majority of voters (10) believed there is “a mortality signal” in the meta-analysis. While none voted against, four responded that they “did not know” if there was a signal or not. In an additional poll, the panel voted eight to six not to change or restrict device labelling. Three voted to restrict and three voted to change labelling. The CRT panel voted unanimously to discuss the Katsanos et al meta-analysis findings with patients during conversations of informed consent. The final poll at CRT showed that the majority of the panel believed the industry patient-level data are poolable for definitive analysis. Nine voted for, one against, and four said they did not know.
Polling from the 31st International Symposium on Endovascular Therapy (ISET; 27–30 January, Hollywood, USA) and from the more recent VLF indicate that the majority of physicians will not change their use of paclitaxel devices following the Katsanos et al meta-analysis. At VLF, 60% (21/35) of the polled audience said they would not be more conservative in recommending drug or device combinations for claudication following the discussion around the JAHA meta-analysis, and 86.5% (32 out of 37 responders) of those polled said they would not change their clinical practice as a result of the meta-analysis findings.
In addition, following a morning-long session at ISET, 79% of the audience responded to an anonymous poll to say that they do not believe the Katsanos et al findings are sufficient to change practice. Sixty-nine per cent of respondents said they would continue to use paclitaxel devices as they always have, with the remaining 31% saying they would use paclitaxel devices less.
However, several physicians voiced concerns, at ISET and since, regarding litigation due to the FDA’s continued investigation into the use of paclitaxel in the femoropopliteal artery. Speaking in general on informed consent regarding the use of paclitaxel-releasing devices at VLF, Paul Rudolf (Washington, DC, USA) said he did not want to trivialise the findings by Katsanos and colleagues, and commented: “Even if the data in the meta-analysis are true, at five years only one out of 14 patients would be affected by the mortality signal, which means that 13 out of 14 patients you would be treating with a drug-eluting stent or balloon would still do just fine and not be affected by that.”
Rudolf also said that 40–45% of malpractice lawsuits hinge on an inadequate discussion of the complications of the procedure prior to treatment. He emphasised that in the case of treatment with paclitaxel, informed consent is the most important factor to consider when thinking about the legal implications. As doctors treating patients with paclitaxel devices in the lower leg already have “complicated informed consents” due to the risk of amputation and recurring disease, Rudolf explained that, from his perspective, “I am not sure how different the informed consent will be. In my view, this is just something that is added on to what is already there.” Rudolf clarified that his comments were made in a general discussion of informed consent during the summit, and were not official legal advice given to practising physicians.
Rudolf suggested physicians stay up to date on device manufacturer information, FDA recommendations, and practice guidelines.