February 21, 2020
No Mortality Difference Observed in JEVT Meta-Analysis of CLTI Patients Treated With Paclitaxel
February 21, 2020—Investigators seeking to explore the risks of paclitaxel-coated devices (PCDs) in patients with chronic limb-threatening ischemia (CLTI) have published a new meta-analysis finding no difference in short- to midterm mortality versus outcomes in patients treated with uncoated devices. The study comes on the heels of a similar meta-analysis published in the Journal of Interventional Radiology (JVIR) by Katsanos et al in January 2020, although the studies differ in their inclusions, methods, and results.
Published online ahead of print in the Journal of Endovascular Therapy (JEVT), a meta-analysis by Dinh et al searched databases on November 5, 2019, to identify trials randomizing patients to PCD or uncoated devices with least 6 months of follow-up data available for analysis. Of note, the meta-analysis also includes trials that were not strictly limited to CLTI, but the inclusion criteria required that trial populations were either at least half CLTI patients, or had extractable data on a CLTI subgroup if < 50% of the total population had CLTI. This resulted in an analysis of 1,450 patients randomized across 11 trials (866 treated with PCD, 584 with uncoated devices), 1,367 of which were determined to have CLTI.
Eight of the trials evaluated drug-coated balloons (DCBs) while three involved drug-eluting stents (DESs), and follow-up periods ranged from 6 months to 5 years, with incremental analyses conducted where available at 6, 12, 24, 36, and 60 months.
In six studies reporting mortality data on a total of 1,036 patients at 6 months, the mortality rates were identical between the two groups at 6.5% (relative risk [RR], 1.06; 95% confidence interval [CI] 0.65–1.72; P = .82). At 12 months, 10 studies reported data on 992 total subjects, at which point 10.9% of paclitaxel-treated patients had died compared with 10.7% of control patients (RR, 1.07; 95% CI, 0.74–1.52; P = .73). Three studies reported mortality data from at a total of 212 subjects at 24 months, with rates of 22.9% in the PCD arm and 34% in the control arm (RR, 0.65; 95% CI, 0.31–1.32, P = .23). The data collected from two studies (176 patients) at 36 months included mortality rates of 35% and 45.6% for the PCD and control arms, respectively (RR, 0.76; 95% CI, 0.45–1.28, P = .30). And, at 60 months, data from 537 patients collected in three trials included mortality rates of 37.5% in subjects treated with PCDs and 46% for those randomized to uncoated therapies (RR, 0.91; 95% CI, 0.75–1.10, P = .32). Of note, the 5-year analysis included data provided by the authors of a forthcoming publication of the IN.PACT DEEP trial.
The investigators report that there was no statistical evidence of heterogeneity among studies at each of these time points.
Sensitivity and subgroup analyses reflected no correlation between the lowest paclitaxel doses and those with the lowest death rates, and studies involving the highest doses did not correlate with the studies that had the highest mortality rates. Similarly, the studies with the shortest and longest lesion lengths did not correlate with those with the lowest and highest mortality rates, respectively.
DATA IN CONTEXT
Although questions and criticisms of the designs and methodology used in the 2018 meta-analysis by Katsanos et al on femoropopliteal PCD interventions published in the Journal of the American Heart Association (JAHA) were raised and subsequently discussed by the authors, explorations of its alarming safety signal and subsequent scrutiny of all past and present paclitaxel studies dominated the field of peripheral vascular intervention in 2019, a trend that continues in early 2020. In contrast, criticisms of the methodology of the more recent JVIR meta-analysis focusing on below-the-knee interventions, including its design divergences from the group’s previous study in JAHA, have thus far seemed to cloud the response to its findings. However, this is also likely due in part to the relatively limited use of PCDs below the knee compared with the femoropopliteal anatomy. For extensive coverage of the JVIR meta-analysis with commentary from the authors and other leading experts, click here.
Important distinctions exist between all three meta-analyses, each of which must be taken into consideration in evaluating and contrasting their findings. Overall, the study conducted by Dinh et al in the JEVT meta-analysis appears to share some common methodologies with the JAHA paclitaxel meta-analysis by Katsanos et al (aside from the key difference of the JEVT study’s inclusion of CLTI patients versus the JAHA paper’s largely claudicant group).
As was the case with the JAHA meta-analysis, the JEVT study includes DCBs and DESs, whereas the JVIR publication included only DCBs. However, the JEVT publication is distinct in that it allowed for procedures done either above or below the knee, as opposed to below-the-knee only, which was the case in the JVIR study. Only studies that were published or accepted for publication (currently in press) were included in the JEVT meta-analysis; the 5-year data from the large-scale IN.PACT DEEP trial were included, whereas the trial’s 1-year results were included in the JVIR publication.
Perhaps the most significant distinction is the focus on the single primary endpoint of all-cause mortality. Although the original JAHA meta-analysis focused primarily on mortality, the JVIR meta-analysis employed a combined endpoint of amputation-free survival.
Of note, one of the authors of the JEVT publication, Ramon Varcoe, MBBS, MS, FRACS, commented to Endovascular Today that their meta-analysis was conducted without knowledge of the second meta-analysis by Katsanos et al, which was published 2 months after the JEVT paper’s literature search was conducted.
“What we’ve tried to do here is take a scientifically sound approach to investigating safety concerns using paclitaxel in patients with CLTI,” said Prof. Varcoe. “We decided to precisely replicate the methodology of the 2018 meta-analysis. The thinking being that if there truly was a higher incidence of death, we should be able to demonstrate it without selecting new endpoints, including unpublished work or adjusting inclusion criteria. However, we found no mortality difference, which led us to conclude that these devices with proven benefits should continue to be used in patients with CLTI.”