Speaking to investors at the JP Morgan healthcare conference in San Francisco, executives from Medtronic and Boston Scientific said their data do not show the safety risks cited in Katsanos (Patras, Greece) et al’s published analysis.
According to the Star Tribune, the companies stated that they are not backing away from drug-eluting devices used in blood vessels in the legs, despite the recent JAHA (Journal of the American Heart Association) meta-analysis linking the devices to a significantly increased risk of death at two and five years following the use of paclitaxel-coated balloons (DCBs) and stents in the femoropopliteal artery.
Katsanos and colleagues concluded their study by saying that “further investigations are urgently warranted”, but also claimed that their findings were backed up by strong statistical evidence; the paper synthesised the results of 28 randomised controlled trials of stents and DCBs (including Medtronic’s In.Pact Admiral and Boston Scientific’s Ranger DCB) used above the knee in the legs. Boston Scientific’s Eluvia drug-eluting stent also contains paclitaxel, though was not used in the analysis.
However, looking “deeply” at their own internal data, Medtronic and Bostin Scientific officials claimed they could find no support for an association between paclitaxel and higher death rates.
As reported in the Star Tribune, the Medtronic cardiac and vascular group president Mike Coyle told investors: “We have generated a tonne of clinical evidence: 1,800 patients, we have data out to five years, randomised controlled study for the USA, randomised controlled clinical study for Japan, global registry data. And all of these data have been analysed. We have not seen this safety signal in our data”. These data are likely to be published in coming months.
Boston Scientific global chief medical officer Ian Meredith said meta-analyses like the paclitaxel-device study in JAHA sometimes reach findings that are not replicated in subsequent studies. The Star Tribune also reports that Meredith noted that the broad, population-level data offered no explanation for the potential mechanism behind the deaths.
A statistical observation which generates important questions, not an answer
“I think at this stage, good clinical studies are appropriate,” Meredith said. “There does not seem to be a plausible mechanism to understand how a dose that is imperceptible in plasma or tissue at 30 days could actually really affect mortality two years and beyond.”
In response to the industry push-back, Katsanos commented in an email to the Star Tribune: “I do understand the scepticism of the industry, but at the end of the day, the statistical signal is too strong to be ignored by reasonable doubt”.
Speaking to Interventional News, Peter Schneider (Kaiser Permanente, Honolulu, USA) echoes Meredith’s concern, stating: “The recent meta-analysis associating paclitaxel with late mortality yields a statistical observation. In my opinion, there is a mismatch between the available data and the conclusion. The study generates several important questions but does not propose a mechanism and does not provide a definitive answer.
“The following factors should be carefully considered as we analyse this study. All reported mortality data in superficial femoral-popliteal DCB and drug-eluting stent (DES) studies are within the range of reported results for mortality in other vascular device trials and expected results from population-based studies of peripheral arterial disease patients. The meta-analysis assumes that paclitaxel is present for the length of patient follow-up, and there is much evidence that contradicts this assumption. Some of the data from studies included in the meta-analysis are inferred from the published papers, rather than actual event numbers. It combines DCB and DES data to assess mortality, but this is probably not scientifically valid based upon the very different preparation and mechanism of delivery between DCB and DES.
“Only three studies have extended to five years and would be potentially reliable sources for information on long-term mortality. Although the paclitaxel treatment groups (DCB and DES) are garnering a lot of attention, we must also look at the angioplasty (PTA) control groups. The PTA groups are small at five years and none of the studies were powered to evaluate mortality, especially long-term mortality; 109 patients for DCB (IN.PACT and THUNDER) and 121 patients for DES (ZILVER PTX). By chance, the PTA group in IN.PACT had among the lowest mortality ever reported in a vascular study. The PTA group in THUNDER had 46% lost to follow-up. It is not clear whether PTA groups were followed with the same tenacity as the paclitaxel groups, especially after a patency end point was reached, and this happened significantly more often in the PTA groups than in the paclitaxel treatment groups. Another possibility that needs investigation is whether PTA patients, who underwent many more target lesion revascularisations, also received better medical management in association with more frequent encounters? When deaths were adjudicated by clinical events committees, as in the IN.PACT and ZILVER PTX trials, there was no clustering around any particular cause or timeline.
“In summary, we have a mathematical observation based upon heterogeneous data that prompts many questions but does not answer any.”
Meredith noted that oncologists have more than two decades of experience with paclitaxel as a chemotherapy agent, at much higher doses than used on a stent or balloon, and have found no similar increased risk of death. Katsanos and colleagues wrote in their JAHA article that whereas paclitaxel for chemotherapy is solvent-based and has a half-life of about six hours, the paclitaxel on medical devices has a half-life of weeks to months.