No Paclitaxel Death Signal in VA Data Out to 3 Years

Long-term data from the Veterans Health Administration (VHA) are in line with the conclusions of several recent trials and registries that dispute the existence of an increased mortality signal with paclitaxel-based devices for treatment of peripheral artery disease.

Among more than 10,000 patients included in the VHA series, survival at 3 years was similar regardless of whether they received paclitaxel-based device or a nonpaclitaxel balloon or stent.

“I think right now we can say to patients that yes there was a signal, but the original data were flawed, data were missing, [and] every study that's come along since has basically shown no difference,” lead investigator Jorge Antonio Gutierrez, MD (Duke University Medical Center, Durham, NC), told TCTMD.

In addition to data reconciliation efforts on behalf of multiple RCTs, real-world Medicare analyses have found no support for a paclitaxel mortality signal; neither have long-term data sets from Germany, Italy, and Japan. These have all been undertaken to better understand the meta-analysis published in 2018 by Konstantinos Katsanos, MD, PhD (Patras University Hospital, Rion, Greece), and colleagues suggesting that paclitaxel-based balloons and stents were associated with increased all-cause death compared with uncoated devices at 2 years and beyond.

This latest VHA data were published online February 6, 2021, ahead of print in the Journal of the American Heart Association.

No Difference in the CLI Subgroup

Gutierrez and colleagues looked at the outcomes of 10,505 PAD patients treated with revascularization procedures within the VHA from October 2015 through June 2019. Men comprised 98% of the study population. Procedures involved a paclitaxel-based device in 21.6% and a nonpaclitaxel device in 78.4%. In the paclitaxel group, 39.6% were treated with drug-coated balloon (DCB) alone, 53% with DES alone, and 7.4% with both DCB and DES. Atherectomy was used in approximately one-quarter of patients treated with paclitaxel and 20% of those treated with a nonpaclitaxel device.

For the group as a whole, investigators found similar rates of survival among patients who received a paclitaxel- versus nonpaclitaxel device out to 3 years (70.7% vs 71.8%; log-rank P = 0.17). Sensitivity analyses further confirmed no significant difference in all-cause mortality between the groups at either the 2-year (HR 0.98; 95% CI 0.86-1.12) or 3-year follow-up (HR 1.11; 95% CI 0.91-1.35).

While the 35% of patients who had critical limb ischemia (CLI) showed poorer survival overall than non-CLI patients, survival rates in the CLI group did not differ between those treated with or without paclitaxel. After adjustment, the hazard for all-cause mortality for patients treated with a paclitaxel-based device versus a nonpaclitaxel device was 1.06 (95% CI 0.95-1.18).

Information on cause of death was available for 771 patients. Of 169 deaths over a median of 20 months, 12.7% occurred in the paclitaxel group and 12.3% in the nonpaclitaxel group. In the group as a whole, CV disease was the most common cause at 38.7%, followed by complications of diabetes at 13.4%, malignancy at 11.2%, and infection at 8.7%. For each cause of death, there were no significant differences between patients treated with or without paclitaxel.

I don't know necessarily if the FDA is ever going to reconvene on this, but at least the scientific community can say we looked and we tried, but we really don't see a difference. Jorge Antonio Gutierrez

“It was pretty much what we expected to see in a population like this,” Gutierrez said. “It’s reasonable, and it makes sense.”

The VHA analysis, he added, contributes to the paclitaxel literature by confirming the lack of mortality signal in a population that is relatively stable in terms of their treatment and outcomes. “Vets don’t really leave the system where they are being treated, so you don’t have the loss to follow-up and it’s relatively easy to track [the entirety of their care],” Gutierrez noted. “Then you add to that the ability to link to real-world causes of death through the [National Death Index].”

The study cannot however account for operator selection bias or unmeasured confounders, among other things.

According to Gutierrez, the recent results from the 2,289-patient SWEDEPAD study showing no mortality signal for paclitaxel, taken with all the other efforts that have refuted the Katsanos meta-analysis, tell a consistent story.

“I think we have a good body of evidence now that is all going in the same direction,” he said. “I don't know necessarily if the FDA is ever going to reconvene on this, but at least the scientific community can say we looked and we tried, but we really don't see a difference.”