Medicare Analysis Offers More Reassurances in Hullabaloo Over Paclitaxel-Based PAD Devices

Among Medicare patients whose PAD is treated with paclitaxel-based balloons or stents, there is no evidence of increased mortality in comparison to non-drug-coated devices through about 20 months of follow-up, according to a new study. It is one of several recent research efforts aimed at responding to a recent meta-analysis that sent the peripheral endovascular community reeling by implicating paclitaxel in long-term mortality.

Eric A. Secemsky, MD (Beth Israel Deaconess Medical Center, Boston, MA), lead author of the new study, published online February 12, 2019 in JAMA Cardiology, said one advantage of the Medicare analysis is that the use of ICD-10 codes enables evaluation of DES and DCB patients separately, rather than as a combined drug-coated device group.

“There's still more work to be done, but this is the first real-world study to demonstrate a safety signal for both stents and balloons in both claudicants and critical limb ischemia [CLI] patients,” Secemsky said in an interview with TCTMD. “The concern here is whether people are dying or not dying, and if they're not dying more disproportionately in the drug-coated device groups then hopefully we're accomplishing our goal of reducing the need for more interventions or stenting or surgical procedures . . . and giving these patients a little bit more quality of life."

In early December 2018, the meta-analysis of 28 trials of paclitaxel DES or DCBs led by Konstantinos Katsanos, MD, PhD (Patras University Hospital, Rion, Greece), reported a 68% relative risk increase in all-cause death in the device group compared with an uncoated device group, with a number-needed-to-harm of 29. At 5 years, the relative risk increase was 93%, with a number-needed-to-harm of 14.

Since then several studies have focused on analyzing deaths in the individual trials using patient-level data. Last month an entire session at LINC 2019 was devoted to the unveiling of new data from the individual studies, none of which supported the death-signal hypothesis. A pooled patient-level meta-analysis of the trials, organized by VIVA Physicians, is being planned with the support of the five major trial sponsors, and a forum for further discussion is planned for early next month. Meanwhile, clinicians are left wondering what to tell their patients.

No Mortality Differences by Device or Disease Severity

Secemsky and colleagues examined Medicare data on 16,560 patients who underwent femoropopliteal artery revascularization during the 2016 calendar year. Half of all patients were current or former smokers, 51% had CLI, and nearly 60% had diabetes.

Compared with non-drug-coated device therapy, paclitaxel-coated balloons and stents had a lower cumulative incidence of all-cause mortality through 600 days of follow-up (32.5% vs 34.3%; log-rank P = 0.007). The findings were similar when patients were stratified by drug-coated balloon (DCB) versus plain old balloon angioplasty (POBA; log-rank P = 0.06) and by DES versus BMS (log-rank P = 0.56).

When patients were analyzed according to whether they did or did not have CLI, the CLI subgroup treated with drug-coated devices had a lower cumulative incidence of mortality through 600 days than did those treated with non-drug-coated devices (38.1% vs 40.1%; log-rank P = 0.04). However, among those without CLI there was no statistical difference in survival between drug-coated and non-drug-coated device groups (26.5% vs 29.0%; log-rank P = 0.07).

In multivariable analysis there was no association between drug-coated devices and all-cause mortality (adjusted HR 0.97; 95% CI 0.91-1.04). Importantly, mortality also did not differ by treatment with DCB alone (adjusted HR 0.94; 95% CI 0.86-1.03) or DES with or without DCB (adjusted HR 0.97; 95% CI 0.89-1.06). The results were similar in CLI patients (adjusted HR 0.93; 95% CI 0.85-1.01) and those without CLI (adjusted HR 0.94; 95% CI 0.85-1.03).

“The safety signal we see gives me more support to continue to use these devices right now,” Secemsky said. But he added that for now he will continue to communicate the concern raised by the Katsanos meta-analysis to patients, with the caveat that based on the available data the benefits appear to outweigh the risks.

Whether these findings will be enough to placate concerns in this space is unclear.

In an accompanying editorial, Jay Giri, MD, MPH (Hospital of the University of Pennsylvania, Philadelphia), says direct comparisons between the Katsanos meta-analysis and the Medicare population study are difficult to draw.

As reported by Katsanos, differences in mortality did not become evident until 24 months. The Medicare analysis had a median follow-up of 389 days, with some patients followed out to 600 days, or just under 20 months. Additionally, the Medicare analysis included a much higher proportion of CLI patients than were treated in the trials analyzed by Katsanos, which “raises the concern that a relatively small absolute increase in mortality associated with paclitaxel may be washed out by competing risks of mortality that are not associated with paclitaxel but rather are inherent to such a high-risk population,” Giri writes.

To TCTMD, Secemsky said there are two ways to look at it. “Yes, it can wash out any mortality signal from the device, but only the other hand, the drug's not killing these patients. It's good news in the sense that they're not dying earlier,” he said. “If they continue to live and are not dying disproportionately more than those not treated with the device, it seems unlikely that the device is altering their lives or prognosis.”

A Call for More ‘Life Cycle’ Research

To TCTMD, Giri said the current “groundswell of studies” aimed at countering the Katsanos conclusion, including the Medicare analysis, is probably a good thing that will “greatly improve our total understanding clinically of what these products are, what they do, and what they don't do in a way that none of the initial studies told us.”

But taken as a whole, the situation is representative of something that may have been preventable, yet may happen again and again unless industry and researchers change their approach to “the life cycle of research,” he observed.

I give [Katsanos] credit for raising the issue. Jay Giri

"Look at the furor that's happened because of this one [meta-analysis] with publicly available data,” Giri said. “From an industry standpoint and for physicians on advisory boards of various companies they need to think much more carefully about what the life cycle of research means . . . They've got to have it in mind from day one what the ongoing plan is for research. It doesn't end when you get that FDA clearance.”

While Giri acknowledged that companies take a risk in letting their data out, he said the Katsanos meta-analysis proves the risk is potentially even greater when patient-level data are withheld and long-term postmarketing safety follow-up is not clearly defined.

In the end, endovascular medicine may find that it needs to look to how life cycle research has been handled for valve therapies and left-atrial appendage closure devices, for example, where Giri said there have been documents drafted by experts in the field ahead of time that have laid out long-term plans and called for postmarket registries. Doing so, he said, produces "stakeholders" in real-world usage of the devices, a group that includes patients, clinicians, researchers, and industry all focused on ongoing safety and certainty.

“I give [Katsanos] credit for raising the issue, " Giri said. "He may end up being proven wrong by more rigorous analyses, but there's no reason this ever had to happen in the first place."