Paclitaxel-Coated Devices for PAD: Mortality Signal Shrinks in Patient-Level Meta-Analysis

A comprehensive meta-analysis using individual patient-level data has found the use of paclitaxel-coated devices for peripheral artery disease (PAD) to be associated with a smaller increased risk of mortality than was found by the initial aggregate-data meta-analysis published by Katsanos and colleagues in December 2018. Additionally, the dose-response relationship between paclitaxel dose and mortality risk found by Katsanos et al was no longer present in the patient-level analysis.

The new meta-analysis, conducted by VIVA Physicians and published online in Circulation in May, also found that inclusion of data from subjects previously lost to follow-up further reduced the observed increase in mortality.

“This new analysis represents the most complete set of available data on mortality from randomized controlled trials of paclitaxel-coated devices that are approved by the FDA,” says study co-author Sean Lyden, MD, Chair of Vascular Surgery at Cleveland Clinic. “It has helped fill some gaps in our understanding of the safety of these devices and identified gaps that may not be filled without further randomized trials.”

Backdrop to the analysis

The initial meta-analysis by Katsanos et al pooled data from 28 randomized controlled trials of paclitaxel-coated balloons and stents for treating PAD in the femoropopliteal artery. These researchers reported a significantly increased rate of all-cause death at two years in claudicants receiving paclitaxel-coated devices relative to controls, and they identified a potential dose-related signal.

These findings prompted controversy, in view of the lack of a suggested plausible mechanism of harm and the fact that a substantial share of patients — up to 30% — had been lost to follow-up, as the studies were not designed to assess long-term mortality.

In response, VIVA Physicians announced that it would work with industry to locate as much of the missing data as possible and conduct the newly published independent meta-analysis using individual patient-level data.

The new analysis at a glance

The new analysis by VIVA Physicians included all trials of FDA-approved paclitaxel-coated devices with at least 2 years of follow-up — eight studies in all (ILLUMENATE, ILLUMENATE EU, IN.PACT SFA I/II, IN.PACT SFA Japan, LEVANT I, LEVANT II, LUTONIX Japan and Zilver PTX). De-identified individual data (baseline characteristics and mortality) were provided for 2,185 patients, 1,382 of whom received paclitaxel-coated devices and 803 of whom received non-drug-coated balloons or bare metal stents.

For the primary outcome of all-cause mortality, Cox proportional hazards one-stage meta-analysis models were employed, using intention-to-treat methodology. A secondary analysis of additionally recovered missing vital status data was also performed. Additional factors assessed included the relationship between paclitaxel dose and mortality.

Over median follow-up of 4 years, 386 deaths were identified among the 2,185 subjects included. The primary analysis revealed a 38% relative increase in mortality risk with paclitaxel-coated devices compared with control devices over 5 years (hazard ratio [HR] = 1.38; 95% CI, 1.06-1.80). The absolute increase in 5-year mortality risk was 4.6% (18.3% in the paclitaxel group vs. 13.7% for controls).

These differences were substantially lower than the 93% relative risk increase and 6.6% absolute risk increase for mortality at 4 to 5 years reported in the meta-analysis by Katsanos and colleagues.

When recovered supplemental vital status data were included, reducing loss of follow-up from 24% to 9.8% of the control cohort and from 23% to 9.3% of the paclitaxel cohort, the relative increase in death in the paclitaxel group declined further, to 27% at 5 years (HR = 1.27; 95% CI, 1.03-1.58). Dr. Lyden notes that the Zilver PTX study employed a secondary randomization for failed angioplasty in which the risk declined to 19% (HR = 1.19; 95% CI, 0.89-1.60), with the confidence interval spanning 1.0 and thus losing its significance.

Notably, the analysis identified no relationship between paclitaxel dose and mortality risk, with no mortality differences according to whether patients were exposed to high, medium or low doses of the drug.

Where will practice go from here?

“This meta-analysis with individual patient data demonstrated a 38% relative increase in risk of death in patients treated with paclitaxel-coated devices for femoropopliteal PAD compared with controls, which is considerably smaller than the risk reported in the earlier aggregate meta-analysis,” observes Dr. Lyden. “The increased risk was attenuated further, to 27%, when the share of patients lost to follow-up was reduced to about 10%. In the Zilver PTX secondary randomization, the risk fell to 19% and was no longer significant. The overall observations remained consistent across multiple scenarios, including intention-to-treat and as-treated analyses, with no evidence of increased risk over time with paclitaxel-coated devices.”

He notes that the lack of evidence of a relationship between paclitaxel dose and mortality “further underscores the absence of a clear mechanism by which paclitaxel would cause harm in this setting.” He adds that the signal for mortality risk is difficult to fully understand when it continues to weaken with additional follow-up data and detail.

The VIVA authors acknowledge that their analysis still demonstrated an increased, albeit weaker, risk of late mortality with paclitaxel-coated devices. And they add that their analysis, like earlier analyses of this question, is limited by an insufficient number of events to adequately evaluate an unplanned safety endpoint. “Future endovascular device trial designs must incorporate strategies and methods to maximize patient retention and facilitate long-term assessment and reporting of vital statistics,” they write.

In the meantime, Dr. Lyden says, these findings argue for continued diligence in carefully advising patients about the risk/benefit calculation when considering paclitaxel-coated devices for treatment of femoropopliteral PAD.