June 20, 2019
FDA Advisory Meeting Day 2: Panel Weighs the Future of Paclitaxel and Clinical Trials in PAD
June 20, 2019—After a full first day of presentations, questions, deliberations, and recommendations, US FDA advisors reconvened to address additional queries posed by the agency as to the safety and efficacy of paclitaxel therapies in peripheral artery disease (PAD), as well as future steps needed to ensure and communicate their status. (See Day 1 coverage here)
The agenda for the advisory committee's second day was ambitious, with seven new questions posed to the panel. Some carried wide-ranging implications with the potential to shape the future of paclitaxel therapy use, and possibly the entire PAD clinical trial landscape.
However, although each question was met with considerable feedback, many of the collective responses focused on the inconclusiveness of data to support the panel making strong recommendations. Others saw new areas of consensus on paths forward emerge—a welcome late-session trend.
DAY 2 SUMMARY
FDA Presents Animal Safety Data and Potential Future Study Mechanisms
The FDA opened Day 2 with a summary of their reevaluation of preclinical animal testing data from each approved device. Presenter Eleni Whatley, PhD, noted that some of the studies were more than a decade old, and noted changes in designs and feedback since that time, as well as challenges in assessing long-term outcomes such as the late mortality signal in this setting. Their analysis showed that paclitaxel was resident in local and downstream tissues for ≥ 60 days, and as long as 270 days. Although no relationship was gleaned between concentration and local tissue effects, Dr. Whatley noted the potential that the drug could affect various tissue systems during its residence, with chronic effects unknown.
Her colleague, Karen Manhart, VMD, DACVP, continued the discussion on animal findings, concluding that no systemic pathologic changes appeared to be device- or drug-related. There were low rates of morbidity and mortality, and no evidence of device-related bone marrow suppression or hepatic or renal toxicity, nor reports of malignancy or unusual findings. Medial smooth muscle cell loss was noted to be common. There were drug-related changes in downstream skeletal muscle and coronary band arterioles noted in drug-coated balloon (DCB)–treated limbs at low levels, sometimes containing crystalline material; downstream findings were rare or absent for drug-eluting stents (DES).
Overall, the reviewed data also did not suggest a potential mechanism for late mortality in human subjects, though it was again acknowledged that the time points were shorter compared to when the signal was observed.
Donna Buckley, MD, MS, then discussed dose analyses, benefit-risk profiles, and next steps such as the logistics of conducting a postapproval study and the potential for device labeling changes.
No relationship between mortality and paclitaxel dosage was observed in the agency's review of RCT data, which also confirmed the effectiveness of paclitaxel-coated devices compared to uncoated devices in terms of reduced target lesion revascularization.
Dr. Buckley also addressed a frequent topic of Day 1 concern—the lack of availability of complete long-term data from the IDE trials. Owing to both trial size and loss to follow-up, long-term data are collectively available for fewer than 1,000 patients, which limits the ability to evaluate the mortality signal observed in the Katsanos meta-analysis and the FDA's own review. There are also the questions of additional/control patient paclitaxel exposure across trials and cause of death adjudication.
With a trend of increased mortality at 3 to 5 years but no causal link established and considerable limitations in the analyzability of the data, the agency is considering additional study mechanisms and asked the panel to weigh in on what might be required. To this end, Dr. Buckley showed potential models for a postapproval randomized trial to explore the mortality—with population figures in excess of 40,000—and also discussed the pros and cons of real-world evidence collections.
The industry's collaborative response to addressing the mortality signal was often referred to as a positive byproduct of the current paclitaxel landscape. The manufacturers suggested continued collaboration across societies, investigators, patients, regulators, and industry to improve both investigation and patient care in PAD. Throughout the later panel deliberations, FDA Director of Cardiovascular Devices Bram Zuckerman, MD, frequently called back to the essentiality of these efforts.
Representing the manufacturers in a Day 2 summary presentation, Laura Mauri, MD, MSc, of Medtronic, vascular surgeon Daniel Clair, MD, and interventional cardiologist Eric Secemsky, MD, provided the collective take on potential next steps. The most significant request they made was that the FDA "update its March 15 Letter to Healthcare Providers to convey the totality of evidence in support of the benefit-risk profile of these devices for their indicated use."
The effect of the letter has been significant in the US and globally, they noted, with some institutions shelving the devices in practices and clinical trial settings and many physicians concerned with potential lawsuits. Whether the FDA modifies its communication after reviewing the proceedings and recommendations of the advisory committee will be one of the most significant outcomes of the session and the data collection efforts that preceded it.
Regarding future PAD studies, the manufacturers focused on the need for complete follow-up and consent for vital status ascertainment as well as recording all prior and additional revascularizations, including treatment type, throughout follow-up, and adherence to optimal medical therapy. They further recommended implementation of a structured minimum core data set and definitions, as well as new initiatives in patient and physician guidance, including a consortium of stakeholders creating clinical guidelines.
Open Public Session: Robust Large-Scale Data Sets, Oncology, and the Potential Effects of Bias
More than a dozen presentations from interested parties were then given. These included the emergence of a handful of intriguing data sets and collection initiatives for paclitaxel in PAD, a perspective from an oncologist with experience studying paclitaxel in cancer settings, and questions as to whether factors outside of paclitaxel may contribute to a statistical signal of increased mortality in PAD trials. One newly presented French study pitting a drug-eluting stent against a bare-metal stent found higher mortality in the BMS arm. Re-running the Katsanos meta-analysis at 2 years with these data, Yann Gouëffic, MD, PhD, found that the meta-analysis data set no longer supported an increased risk of death at that juncture.
Although compelling and potentially impactful on future study and understanding, summary of these presentations is beyond the scope of this coverage.
One talk in particular prompted later panel questions and extensive discussions, as vascular surgeon Ramon Varcoe raised the possibility that biases and design nuances could be affecting the statistical signals toward mortality in the study arms PAD trials, including those without paclitaxel. Endovascular Today previously covered Dr. Varcoe's initial work after his presentation at VIVA Physicians' Vascular Leaders Forum, which also featured several of the other perspectives shared at the FDA meeting. To read more on Dr. Varcoe's findings and the perspective of oncologist Erica Mayer, MD, on the toxicity of paclitaxel in cancer and PAD settings, please click here. Video of Dr. Varcoe's previous VLF presentation can be viewed above.
To gain a better understanding on the quality-of-life benefits associated with paclitaxel devices, the panel asked for further data to support these claims. Representatives from the industry trials provided data seemed to satisfy the panel as to their benefits, which seemed to provide context for the remaining deliberations of the day.
THE PANEL WEIGHS IN
Dose-Mortality Relationship Status
The first question the panel was asked to address on the day (sixth total) dealt with the much-debated relationship of paclitaxel dose to mortality risk. Due to a variety of confounding factors such as the wide ranges in device doses and a variety of patient factors, the panel quickly declared that the data were inconclusive and firm recommendations on the relationship could not be made.
Back to the Lab?
Next, the FDA asked the panel to consider whether preclinical data provide mechanistic insights into late mortality signal, as well as what if any new animal studies might be helpful. Although the study limitations raised earlier by the FDA were reiterated, the overall response was that the currently available animal data did not provide mechanistic insights. As to whether additional studies might provide greater insight, consensus was elusive, with some panelists seeking new models better able to answer the known question.
One key dissenting panel voice was that of Col. Todd Rasmussen, MD, who voted that further animal models would not be useful in addressing the current question of paclitaxel safety. "It's the easiest thing in the world to say 'we're going to do more animal studies,'" noted Col. Rasmussen, acknowledging their overall importance but positing that with thousands of patients already studied and treated, trying to find a long-term mortality signal in the animal lab would be a step back in this continued evaluation.
Perhaps the most stimulating discussion of the day came in the panel's consideration of the benefit-risk profile of paclitaxel therapies given the totality of data, as well as considerations for the continued marketing of these devices. There was general consensus as to the benefits of the devices, both in terms of reduced target lesion revascularization and quality of life improvement. Joaquin Cigarroa, MD, said he had no doubt as to there is a "very consistent and clinically profound benefit to our patients." However, he also felt there was a clear and consistent signal of mortality in the trials that have follow-up to 5 years, though their underpowering confounded the finding.
Col. Rasmussen emphasized that the benefit was shown across all kinds of studies, an inability to establish a mortality cause, and near-unanimous public testimony during the sessions. On the other side, there are the original, well conducted meta-analysis, and the FDA's replication of the finding in a somewhat modified data set. However, the concept of ascertainment bias raised by Dr. Varcoe had him calling the mortality findings into question.
There was overall agreement with the continued availability of the devices, with much conversation centered on preserving and facilitating informed patient choice in therapies. However, this will also require further clarification of the exact presence and nature of the risk side of the equation.
Postmarket Studies and Surveillance
The panel was asked to address whether postmarket studies or surveillances are needed, and if so, to weigh the pros and cons of potential design elements. There was consensus that current trials should be encouraged to collect as much and as rigorous follow-up data as possible, particularly as to vital status. The importance of collaboration among all interested parties was again emphasized.
However, the panel did not narrow down an exact means by which to collect conclusive prospective data. One key area of divergence was the utility of registry data, with some panelists intrigued by recent collection methods and findings such as those from the CMS database and others entrenched as to their limitations to addressing the current question.
Asked next to discuss whether any potential modifications should be made to existing paclitaxel device labels to convey safety information, the panel did not provide as refined a summary response as they had in previous questions (though this was possibly due to fatigue). Most members believed that the totality of data should be presented. The challenges of doing so without general consensus as to the actual risk were discussed without definitive resolution, but most agreed that the meta-analysis findings should be provided in some context, along with the benefits. The need for patients to be informed was emphasized by panel chairperson Richard A. Lange, MD, MBA.
Changes to Future Study Design
The penultimate question to be addressed had multiple parts aimed gaining recommendations for better future trial design and follow-up requirements. A variety of viewpoints were shared, including that efficacy can likely be established and the devices able to be marketed after 1 year, but safety in particular should be rigorously followed out to 5 years.
Kevin Kip, PhD, FAHA, stated that any new study must place a premium on exposure to paclitaxel prior to and after the index procedure, echoing an earlier suggestion from the manufacturers.
Dr. Zuckerman from the FDA affirmed that going forward, mortality would be given a higher priority in trial design, and that the overall quality of clinical trials must improve substantially. The FDA would be interested in increased uniformity of reporting, he added, noting that it was difficult to meta-analyze these data. Importantly, Dr. Zuckerman reminded the panel and audience of the need for practical, achievable solutions, and that collaboration among all stakeholders was needed.
Finally, the panel addressed the relatively limited data available in non superficial femoral/proximal popliteal artery settings. The FDA requested they discuss whether risk-benefit considerations were likely to be distinct. To this end, it was encouraged that these ongoing trials strive to gather all patient data through the full follow-up period to fully address this question.
The panel was thanked for its participation, as were the presenters and physicians and industry representatives who worked to provide as much data as possible for consideration. The industry will now await the FDA's response to the panel recommendations, and new data will continue to emerge from the ongoing collections described during the proceedings.
Commenting to Endovascular Today as the session concluded, Gary M. Ansel, MD, took a positive view: "Many good things could come out of this question: A drive for better data collection, cooperation between those that provide vascular care, focusing more on the disease process than a blockage, and industry setting a precedent of working together for problem solving."