ֱ

How Much Does Bias Factor Into the Paclitaxel Mortality Issue?

— New analyses presented as PAD field waits for the other shoe to drop

Last Updated December 13, 2019
MedpageToday

LAS VEGAS -- The question of whether paclitaxel-coated stents and balloons really cause late mortality among patients with peripheral artery disease (PAD) has multiple angles being explored by researchers but with few answers found.

Noticeably missing from this year's (VIVA) panel was Konstantinos Katsanos, MD, of Patras University Hospital in Rio, Greece, whose findings in a -- a 68% increase in risk of death at 2 years with paclitaxel devices in lower-limb arteries compared to controls -- first ignited the paclitaxel issue.

A common critique of the Katsanos meta-analysis was that it had relied on several randomized trials plagued by missing follow-up data.

If control-arm patients who went missing were actually misclassified deaths, the risk ratio for paclitaxel would drop to 1.37 in the case of 5% misclassification (95% CI 1.04-1.81) and even turn insignificant if misclassification reached 20% (RR 1.15, 95% CI 0.88-1.49), according to Ramon Varcoe, MBBS, MS, PhD, of the Prince of Wales Hospital in Sydney.

Furthermore, these trials were not blinded, possibly resulting in bias.

To illustrate this point, Varcoe discussed his analysis in which taking paclitaxel out of the equation by comparing non-paclitaxel device treatment and control groups still led to the same trend of improved survival out to 3 years among controls -- suggesting that paclitaxel by itself did not cause the deaths.

"Could it be that SFA [superficial femoral artery] interventions cause increased mortality amongst our patients?" he asked. "Or, more likely, could it be that this association between experimental SFA interventions and higher rates of death [is] due to the introduction of bias, more tenacious follow-up of those subjects in the experimental arms, and higher rates of medical interaction [for those patients] with more frequent [target lesion revascularizations]?"

Notably, FDA advisors had concluded in June that a mortality signal attributed to paclitaxel balloons and stents was real, even if no one could provide a biological mechanism to explain these deaths.

Since then, the recovery of more patient follow-up data incrementally dropped the risk of late mortality (HR 1.27, 95% CI 1.03-1.58; down from prior HR 1.38, 95% CI 1.06-1.80) in the VIVA/NAMSA IPD project, the most comprehensive meta-analysis to date by Krishna Rocha-Singh, MD, of Prairie Heart Institute at St. John's Hospital in Springfield, Illinois, and colleagues.

Rocha-Singh said that this drop in the hazard ratio may look good "symbolically" but is likely to be statistically similar to the earlier risk estimate. Thus, a "modest and consistent" mortality signal continues to be found in patients exposed to paclitaxel, he concluded.

These latest findings were first announced at the Transcatheter Cardiovascular Therapeutics meeting in September and are pending acceptance at a peer-reviewed journal. "We captured all the data we can capture" and this now represents the "cleanest dataset" possible, the presenter told the audience here.

The next step would be to consider complementary real-world datasets, Rocha-Singh said.

To that end, population-based data from the Centers for Medicare and Medicaid Services, the private insurer Optum, the Vascular Quality Initiative, and the German insurance company Barmer have all found no survival disadvantage from the paclitaxel stents and balloons, according to Eric Secemsky, MD, MSc, of Beth Israel Deaconess Medical Center in Boston.

"The biggest issue is ascertaining mortality," Secemsky told the audience, citing the difficulty in linking patients to deaths when there is a lag in the datasets being updated.

FDA continues to work with device manufacturers, professional societies, and clinical researchers to collect missing data from past studies and capture as much patient information as possible in ongoing and future ones, said agency medical officer Donna Buckley, MD, of MedStar Georgetown University Hospital in Washington, D.C.

In the meantime, the FDA has healthcare providers to discuss the potential risks and benefits of these devices with their patients.

  • author['full_name']

    Nicole Lou is a reporter for ֱ, where she covers cardiology news and other developments in medicine.

Disclosures

Rocha-Singh disclosed receiving honoraria and research funds from Medtronic; consulting for Alucent Biomedical, Abbott Vascular, Medtronic, ROX Medical, and Soundbite Medical; and holding stock in PQ Bypass.

Secemsky reported receiving honoraria from Medtronic and Cook Medical; consulting for Medtronic; and getting research funding from Bard Peripheral Vascular, CSI, Cook Medical, Medtronic, and AstraZeneca.

Varcoe disclosed receiving honoraria from and/or consulting for Abbott Vascular, Boston Scientific, Medtronic, Shockwave Medical, SurModics, and Intact Vascular; and receiving research support from Penumbra and the NIH.