Better Data Are Needed to Assess Paclitaxel-Eluting Devices

Hello, and welcome to the July 2019 issue of Vascular Disease Management. I have chosen to comment on the results of the June 19-20, 2019 FDA panel that was convened to address the late mortality safety signal associated with the utilization of paclitaxel-eluting devices in the treatment of femoral-popliteal arterial obstructive disease. 

This panel of medical experts was chosen to further explore the increased late mortality signal associated with these devices as reported by Katsanos and colleagues in a meta-analysis of 28 published studies based on intention to treat rather than as treated outcomes. Interventionists have awaited further suggestions from the FDA regarding the use of paclitaxel devices.

Katsanos and colleagues reported that mortality was the same in the paclitaxel device group and the control group (2.3%) at 1 year, but increased at 3 years (7.2% vs 3.8%, respectively) and 5 years (14.7% vs 8.1%, respectively) in the paclitaxel device group. No causality was demonstrated, and the high rate of crossover (up to 40% in the Zilver PTX study) between the groups was not addressed. Katsanos concluded that there was a “dose-time product” and absolute cause of death. Critics of the publication argued that pooling the data makes the assumption that these devices are all similar, that “as treated” rather than “intention to treat” would more accurately depict true mortality risk, that there was no causal relationship, and that it is difficult to understand how a very low dose of paclitaxel with a relatively short half-life can plausibly explain late-term mortality, as this was not demonstrated with administration of much higher repeated doses used in chemotherapy of malignancies.

The FDA performed its own preliminary analysis of mortality risk associated with paclitaxel devices of 3 trials with at least 5 years of follow-up with 975 patients. The trials analyzed were:

1. IN.PACT Admiral DCB

2. Zilver PTX DES

3. Lutonix DCB

The March 15, 2019 letter to physicians from the FDA said that the FDA noted a 50% higher risk of mortality with paclitaxel peripheral arterial disease (PAD) products. The agency suggested that alternative treatment options should be used until more data are available. The FDA acknowledged that pooling the studies may lead to “greater uncertainty in results,” and acknowledged that a specific cause and mechanism of the higher mortality is unknown. The FDA recommended that providers continue to discuss the risks and benefits of these devices. The agency further noted that “for some individual patients at particularly high risk for restenosis, the benefits of using a paclitaxel-coated device product may outweigh the risks.” The letter stated that that dose does not seem to be the key discriminant, as was suggested by Katsanos.

The June 19-20 FDA panel heard presentations from scientific experts citing mortality rates in various paclitaxel-eluting devices, including coronary paclitaxel devices as well as those used in the femoral-popliteal segment. Aaron Lottes and Michael Dake presented “real-world as treated” data on Zilver PTX drug-eluting stents, suggesting no mortality signal. The fact that there was a 40% crossover rate in the Zilver PTX studies that had been cited by Katsanos that was not addressed in his article was discussed. JD Meier and Ken Ouriel spoke on behalf of Lutonix, again citing that review of their database did not establish increased mortality and that the deaths were not via causes that have been reported as secondary to paclitaxel. William Gray, speaking on behalf of Philips, stated that there was no mortality signal at 1 or 3 years with the Stellarex DCB. There are no 5-year data with that device. Ian Meredith and Robert Lookstein spoke behalf of the Eluvia stent and cited no available data yet to suggest a mortality signal. Renowned cardiovascular pathologist Renu Virmani cited her pre-clinical experience and autopsy experience with coronary paclitaxel stents, stating that she believes clinicians should have unrestricted access to paclitaxel devices as she does not believe there is a clear signal of increased mortality with paclitaxel devices. 

The FDA panel concluded that there is a “mortality signal,” but acknowledged that “better data collection was necessary going forward as some data were missing.” The panel determined that “there were not enough data to confirm a class effect and no consensus was reached on cause of death.” No further guidance to physicians was recommended. 

I have found that meta-analyses are often misleading. Failure to account for treatment crossover, inability to establish mortality causality, pooling of data with devices that are substantially different, and a reported mortality in the control arm that is significantly lower than most PAD historical controls groups are all “red flags” to me, suggesting data are suspect. 

Mortality should be evaluated by each individual device, as each device is different. Device risks must be evaluated on an “as treated” not by “intention to treat” analysis. If there is a real mortality signal, the interventional community must know this, and it can only be evaluated by determining outcomes of all who have received devices versus those who have not.  More information is needed. 

Based on the presently available data, I am not convinced that there is a true mortality signal, but I am convinced that we must obtain “clean data” based on “as treated” analysis assessing each device separately. I was hoping that the panel would result in clear guidelines, but physicians have been left in limbo.

Additionally, I am surprised that an article of such potential clinical importance as the Katsanos article was approved for publication in less than 3 weeks from the time of submission and that there was no accompanying editorial to address potential shortcomings of the data.