As ‘Paclitaxelgate’ Loses Steam, Experts Mull Lasting Impact for Future Trials, Practice

Efforts by investigators, industry, and regulatory authorities to understand a mortality signal in patients with PAD who were treated with paclitaxel-based balloons and stents has yielded lessons that will improve data acquisition and analysis for PAD devices going forward, a working group has concluded. That view echoes the voices from experts participating in dedicated panel discussions at recent meetings.

The RAPID Pathways Collaborative Paclitaxel Project Working Group, convened after a 2-day public meeting of an FDA advisory panel back in June 2019, included PAD experts from academia, industry, regulatory, and the clinical community. This month, the group, led by Aaron E. Lottes, PhD (Purdue University, West Lafayette, IN), describe lessons learned from the fallout surrounding the paclitaxel mortality signal, which they say will help guide future studies. They published their conclusions earlier this month in the American Heart Journal.

‘Paclitaxelgate,’ as it was called by some, stemmed from the publication of a meta-analysis nearly 2 years ago by Konstantinos Katsanos, MD, PhD (Patras University Hospital, Rion, Greece). The paper showed a 68% relative risk increase in all-cause death with paclitaxel-coated devices compared with an uncoated balloon at 2 years, and a relative risk increase of 93% at 5 years. What followed was a flurry of research activity, turning up a treasure trove of missing and incomplete data in randomized controlled trials (RCTs). But large, subsequent real-world Medicare analyses found no support for the mortality signal; neither did long-term data sets from Germany, Italy and Japan.

Recommendations for Moving Forward

The RAPID group reviewed 138 clinical studies, which included seven complete RCTs with 36- to 60-month data, 12 ongoing RCTs with 36- to 60-month data, two RCT-like data sources with 36- to 60-month data (BEST-CLI and VOYAGER PAD), 17 RCTs with 24-month data, 24 single-arm/observational data sources with 36- to 60-month data, 15 single-arm/observational data sources with 24-month data, and 60 other data sources with less than 24-month data.

Based on their review, Lottes and colleagues isolated four focus areas surrounding data quality where, they say, more work is needed: trial design, trial conduct, data analysis, and data capture.

“Streamlined case report forms (CRF) with consistently defined and structured data elements that can be used across individual trials are important to collect the necessary data for endpoint analyses, reduce the likelihood of missing data, and promote comparability of results, when appropriate,” the group writes. Some elements specific to PAD trials, they add, include listing all concomitant medications, information on repeat procedures, prior exposure to drug-coated therapies, consistent capture of comorbidities, detailed cause of death, consistent reporting of adverse events, and long-term follow-up plans.

The RAPID group also recommends that national and international consortia work together on recommendations for PAD trial designs, important data elements, and uniformity in definitions, which they say can “facilitate signal detection, data analysis and clinical and regulatory decision-making.”

Other recommendations include:

• Minimize missing data to the furthest extent possible
• Blind research personnel to the greatest extent feasible
• Include long-term follow-up, especially for vital status
• Pre-specify statistical methods and analyses to account for imbalances in data
• Consider and understand the impact of different analysis cohorts such as intention-to-treat, as-treated, etc, on major endpoints when deciding on final analysis population
• Account for potential confounders when analyzing non-randomized data
• Consider novel data sources and collection methods to ascertain endpoints and perspectives that are important to both patients and clinicians
• Consider how to accrue larger sample sizes in PAD studies or supplemental datasets that can potentially assess for the unexpected, such as a mortality signal
• Carefully consider the importance of real-world data and other large data sources outside of RCTs on critical endpoints

Circling back to the idea of working together, Lottes et al encourage collaboration between industry, clinical investigators, and real-world data owners to “provide important additional data that can contribute significantly to better understanding the benefit/risk profile of these devices for the proposed indication and elucidating potential signals in the future.” The group also says they are in the early phases of developing an analysis plan for emerging data “to further inform the safety signal and provide more insight into the benefit/risk profile of these devices.”

Over the last few months, new data have been emerging at virtual meetings such as a VIVA webcast and at TCT Connect 2020. Most recently, a session this month at VIVA 2020 addressed what is still needed to understand the mortality signal and the controversy surrounding it in a more granular way.

In my mind, we're kind of at a tipping point where the original meta-analysis is now the spurious finding here and all these other studies are supporting the safety. Eric Secemsky

Sean P. Lyden, MD (Cleveland Clinic, OH), said recent long-term mortality updates from VOYAGER PAD and the ILLUMENATE series of trials “continue to really reassure me that there is no mortality difference.” He added that he feels confident that as the RCTs report their long-term data, the mortality rate will continue to drop and be a shadow of what was reported in the Katsanos meta-analysis.

While all the data that trickle in provide pieces of the mortality signal puzzle, Peter Schneider, MD (University of California, San Francisco), said it’s important to remember that none of the studies are geared to give the entire answer.

“You can't stay awake at night adding data bits to every little thing that comes in,” he said. “So, what else do you have your eye on and how often do you regroup and look at this?”

FDA: Still Watching and Reviewing

Misti Malone, PhD (US Food and Drug Administration, Silver Spring, MD), also speaking during the VIVA session, said the expectation of the FDA is that much of the data from the US and outside the US that have been presented at meetings in the last few months will be published in peer-reviewed journals. As those complete data sources become available, she said the agency will assess them and update their advice when appropriate.

“And though I can't say when—I don't have my magic ball—I can say we are continually looking at [the data],” Malone said. “We do take a bit of a conservative approach here, but we also understand clinical practice, which is why our communications have allowed clinicians to continue to use these devices in patients that they believe the benefit will outweigh the risk.”

Malone added that the questions and challenges posed by the mortality signal led to broader questions regarding the best clinical approach for PAD.

“I appreciate that because of the signal we are having these discussions and [seeing] additional data coming from studies in order to better address both regulatory decision-making and clinical decision-making,” she said.

“In my mind, we're kind of at a tipping point where the original meta-analysis is now the spurious finding here and all these other studies are supporting the safety,” added Eric Secemsky, MD (Beth Israel Deaconess Medical Center, Boston, MA).

The lingering practical question, noted Bret N. Wiechmann, MD (North Florida Regional Medical Center, Gainesville), is how to explain the new data and the changing opinion on the mortality signal to patients.

“I think that the deeper we dig into the data, the more we have in our camp, so to speak, to support a lack of mechanism to explain the original meta-analysis,” he said. “So I think we continue to go about it the way we always have, in the short term, until we get the definitive FDA decision.”