May 27, 2020
VIVA's Paclitaxel Analysis Finds Increased Mortality Risk, But No Dose Association; Lead Author Encourages Close Read
May 27, 2020—Earlier this month, VIVA Physicians announced the publication of an analysis of mortality and paclitaxel-coated devices by Krishna Rocha-Singh, MD, et al, which is available online ahead of print in Circulation. The data were also presented by Gary Ansel, MD, during a Charing Cross virtual session on May 25. The investigators performed an individual patient-level data (IPD) analysis of the safety of paclitaxel-containing devices (PTXD), further exploring the increased mortality signal first identified by Konstantinos Katsanos, MD, et al in their December 2018 Journal of the American Heart Association publication. The VIVA analysis, which included 2,185 patients across eight studies with a median follow-up of 4 years, also identified an increased mortality risk associated with PTXD use.
However, it found a weaker mortality signal, and no drug-dose relationship was established, in contradistinction to the findings of Katsanos et al's study-level analysis.
In collaboration with the FDA and other groups, VIVA Physicians requested that manufacturers of United States FDA-approved and commercially available devices in the United States provide deidentified data from their investigational device exemption (IDE) trials for a combined independent analysis, producing what lead author Krishna J. Rocha-Singh, MD, described as “a very homogeneous patient population with detailed demographic and procedural data.” Cox proportional hazards were used for the primary analysis, and a secondary analysis of previously missing vital status data was performed. The impact of control subjects crossing over to PTXD, cause-specific mortality, and drug-dose relationships to mortality were all assessed.
Although the primary findings of the VIVA meta-analysis were originally presented in June 2019 at the panel convened by the FDA to explore the safety of paclitaxel-based PAD devices, as well as at key fall 2019 meetings, their publication in May 2020 has stirred considerable response online, the nature of which Dr. Rocha-Singh finds concerning.
In comments to Endovascular Today, he put the data into context and urged a thorough reading of the paper beyond the abstract.
“What was different between prior presentations [of the VIVA IPD analysis] and the recent publication is that we really dove into the issues of paclitaxel dose-response and found no association; this was new peer-reviewed data,” explained Dr. Rocha-Singh.
“I think the reader of this study must be diligent and look beyond the title that says, ‘Paclitaxel and Mortality,’” continued Dr. Rocha-Singh. “You have to get past reading just the abstract Background, then glancing down to the Conclusion that states we found an absolute 4.6% increase in mortality risk with paclitaxel use. Because if you only get through the abstract, you totally miss what we fully dive into in the Discussion section. We call out a variety of blind spots, including the consideration of selection bias, the drug dose analysis, the many sensitivity analyses, and the fact that we did not go into these randomized trials expecting to see a mortality concern—[so] these randomized controlled trials were not powered to assess mortality.”
In the wake of the Katsanos et al meta-analysis, the FDA conducted its own analysis of the data from United States IDE trials of paclitaxel devices, which also identified a mortality signal. These findings were informally presented at VIVA’s Vascular Leaders Forum in March 2019. However, similar to the Katsanos et al study, the FDA's analysis was conducted with study-level data and did not capture data on the patients lost follow-up—two of the biggest concerns raised in evaluating the original findings—and the relative risk ratio was higher than that observed in the analysis performed by VIVA Physicians, said Dr. Rocha-Singh.
“One of the essential takeaways from our publication is that the observed association between paclitaxel use and mortality was not as significant as what was reported by Katsanos, but I think it points to the directionality, and it reflects the veracity of data included in the analyses and consideration of the impact of bias,” he continued. “For example, the absolute risk of mortality dropped from 6.6% to 4.6% by reducing the selection bias and minimizing the vital status loss to follow-up rate in both treatment arms to less than 10%.
“If we had reduced the loss to follow-up rate further, to around 5% in both the control and paclitaxel arms, I believe the paclitaxel mortality association would have dissipated,” concluded Dr. Rocha-Singh to Endovascular Today.
For more information, visit Continuing Coverage: Paclitaxel in PAD, Endovascular Today's online compendium on the paclitaxel safety controversy.