Critique of New Meta-Analysis of Paclitaxel Highlights Issues in Study Construction

Hollywood, FL (January 23, 2020) -- The study design of a recently published meta-analysis by Katasanos and colleagues is flawed, according to a critique presented by William Gray, MD, at the International Symposium on Endovascular Therapy (ISET).

The meta-analysis by Katsanos and colleagues¹, published on January 15 in the Journal of Vascular and Interventional Radiology, found a significant increase in all-cause death and major amputation that was associated with paclitaxel-coated balloons (PCBs) for the treatment of critical limb ischemia (CLI) below the knee (13.7% crude risk of death or limb loss with paclitaxel vs. 9.4% risk with uncoated balloon angioplasty, hazard ratio, 1.52; P=.008). PCBs were also associated with a significant reduction in target lesion revascularization.

In his critique, Dr. Gray pointed out problems in the design of the meta-analysis. A close look at the study’s flowchart reveals that there are inadequate numbers to construct a study-level meta-analysis without significant risk of Type I error (false positive). “By the time you got to this five-year statistical significance, [the Katsanos et al meta-analysis] was well below a thousand patients,” Dr. Gray said, noting that the same issue was seen in the Katsanos paper from last year.²

Also problematic is the inclusion of studies with non-standard follow-up. “[Katsanos et al] included studies with non-standard follow-up mixed between 6 months and 1 year.” Continuing, Dr. Gray explained that of the 8 studies included in the meta-analysis, 3 were conference proceedings that were not published in journals. “These are data from presentations at conventions and sessions like [ISET], not peer review, and we don't know how they were controlled or anything like that. So this is problematic,” he said.

Additional issues with the Katsanos study are mathematical in nature, noted Dr. Gray. He examined errors in the Katsanos analysis of the IN.PACT Deep study, and he also noted transcription mistakes in the number of deaths in IN.PACT Deep.

Dose analysis is also highly flawed, according to Dr. Gray. “No lesion length was taken into consideration, no number of balloons used, and no adjustments for selection bias or cross-trial differences,” he said.

He continued, “What do I mean by selection bias? As soon as you tell me I've split the patients studied between high dose and low dose, you impose a selection bias because there may be a difference in why those patients got low dose or high dose.”

“Moreover, the whole dose analysis is inconsistent with his prior methodology,” Dr. Gray said, pointing out that Katsanos used inconsistent formulas.

In summary, Dr. Gray emphasized that the Katsanos study should not affect treatment of patients with CLI. “This ‘analysis’ is very poorly constructed and conducted, and therefore should have no meaningful impact on this high-risk, in-need CLI population, especially given the marked improvement documented in the same manuscript in patency,” Dr. Gray concluded.

References

1. Katsanos K, Spiliopoulos S, Kitrou P, Krokidis M, Paraskevopoulos I, Karnabatidis D. Risk of death and amputation with use of paclitaxel-coated balloons in the infrapopliteal arteries for treatment of critical limb ischemia: a systematic review and meta-analysis of randomized controlled trials. J Vasc Interv Radiol. 2020;31:202–212.

2. Katsanos K, Spiliopoulos S, Kitrou P, Krokidis M, Karnabatidis D. Risk of death following application of paclitaxel-coated balloons and stents in the femoropopliteal artery of the leg: A systematic review and meta-analysis of randomized controlled trials. J Am Heart Assoc. 2018; 7(24):e011245.