Carotid Stenting Makes a Strong Showing in the CREST-2 Registry

The results of the ongoing CREST-2 study in patients with asymptomatic, high-grade carotid artery stenosis are still several years away. But new data out of a companion registry suggest that whatever the ultimate outcomes are, they will need to be interpreted in the context of some the best carotid stenting outcomes yet seen.

In the CREST-2 Registry, which is enrolling patients with either asymptomatic or symptomatic stenosis who are undergoing stenting and either don’t quality for or don’t want to participate in the CREST-2 trial, the rate of stroke or death in the 30-day periprocedural period among patients included through the end of 2018 was 2.0% overall—1.4% in the asymptomatic population and 2.8% in symptomatic patients.

The registry was designed collaboratively with stakeholders from the federal government, medical societies, and industry to provide operators with a way to maintain competence in carotid stenting, ensure high-quality technique and outcomes, and receive credentialing to participate in the CREST-2 trial, as well as to be reimbursed for doing the procedure. The Centers for Medicare & Medicaid Services (CMS) had previously instituted restrictive criteria for reimbursement, which were blamed—in conjunction with concerns about periprocedural stroke risks—for a drop in the number of stenting procedures being performed in the United States.

What this new paper—published in the December 24, 2019, issue of the Journal of the American College of Cardiology—is trying to emphasize is not necessarily the very low periprocedural stroke/death rates, Brajesh Lal, MD (University of Maryland School of Medicine, Baltimore),  told TCTMD, but rather the question of “how do we go about developing an environment that can facilitate rollout of new technology?” Lal is lead author and principal investigator of the CREST-2 Registry.

The approach used in the registry “seems to be ensuring the best possible results that one can accomplish with transfemoral stenting,” he observed. “And that’s important because the way the [CREST-2] trial is structured, it is designed to compare the best possible intensive medical therapy against the best possible stenting, and the best possible intensive medical therapy against the best possible surgery.”

The bottom line, Lal said, is that “in many ways the results that are published are a ratification of that particular infrastructure and approach, and hopefully [the paper] either in its entirety or at least elements of it can be used to replicate similar environments, safe environments, where new technology can be gradually rolled out.”

Carotid Stenting in Contemporary Practice

Several trials—CREST and ACT I among them—have shown that long-term outcomes are similar with carotid stenting and carotid endarterectomy in patients with asymptomatic high-grade stenosis, with more stroke after stenting and more MI after surgery in the periprocedural period. Both since those trials were conducted, there have been advancements in medical therapy, stenting, and endarterectomy, and it’s unclear whether the results apply to current practice. The CREST-2 study, which is actually two parallel trials pitting stenting and endarterectomy each against intensive medical therapy alone, aims to find out whether either type of revascularization adds a benefit on top of medications.

The CREST-2 Registry, which incorporates data from the Society for Vascular Surgery’s Vascular Quality Initiative and the American College of Cardiology’s National Cardiovascular Data Registry and additional information directly submitted to the registry, provides a means to ensure that operators performing stenting in the CREST-2 trial are doing so in a high-quality way.

It is reassuring that when the trial results come out we will be dealing with probably the best possible transfemoral stenting outcomes that can be achieved. Brajesh Lal

This new report looks at the periprocedural outcomes for stenting in the registry through the end of 2018. The analysis included data on 2,141 patients (mean age 68 years; 65% men) who received carotid stents for primary atherosclerosis from 187 operators at 98 US sites. Most patients (55%) had asymptomatic disease. Average percent stenosis in the carotid artery was 84.2%.

The rate of periprocedural stroke/death seen in the registry is lower than what has been seen in previous randomized trials and is comparable to what has been observed with the new revascularization technique of transcarotid artery revascularization, or TCAR, Lal et al note.

“The results of the present study suggest that if these [prior stenting] trials were to be repeated with current carotid artery stenting experience and practice, the results observed with [stenting] may be even more favorable,” the authors write.

Questions About Ascertainment of Stroke

Whether the “staggeringly low” periprocedural stroke/death rates can be believed, however, is another issue, according to Jay Giri, MD (Hospital of the University of Pennsylvania, Philadelphia), who wrote an accompanying editorial with Jeffrey Olin, DO (Icahn School of Medicine at Mount Sinai, New York, NY).

Giri said that as a comparison, the 30-day stroke rate in ACT I—which he acknowledged is an older trial that enrolled patients between 2005 and 2013—was 2.8% in asymptomatic patients, double that seen in the CREST-2 Registry. That raises questions about how strokes were ascertained in the registry and whether some might have been missed, he noted to TCTMD.

He said the investigators were asked about the discrepancy and provided reasonable explanations involving the time period when ACT I was conducted, changes in periprocedural medical management, operator experience, and case selection over time. But it’s unclear whether those factors fully explain the much lower rates seen in the CREST-2 Registry, Giri said.

“When you sum it all up, the thing that I think you can say with confidence is that one of the major goals of the investigators setting up the CREST-2 Registry was to enable operators to maintain competence and then, number two, to report data like this to verify their competence so that the results of the trials could be trusted as being internally valid,” Giri said. “I would say that the results of this study demonstrate that they actually did accomplish both of those goals.”

If the periprocedural stroke/death rate observed in this analysis is replicated in the CREST-2 trial, “you’re going to see that carotid stenting is going to have a really good chance at beating medical management in this asymptomatic population,” he predicted. He added, however, “I find that very hard to believe that that’ll be the case in an adjudicated randomized trial of carotid stenting with multiple levels of control and stroke neurologists examining every patient. I would be shocked to see that.”

Lal et al acknowledge that the periprocedural outcomes observed in the registry may not be reproduced in the trial, although they note that in the subset of registry patients who would have been eligible for randomization but chose not to participate, outcomes were even better.

Still, Lal agreed that ascertainment of strokes is more rigorous in a randomized trial. “That’s why we’re doing a trial,” he said. “If the results were 100% reliable, everybody would stop at the level of a registry. It’s less expensive, so much easier to do, and so much more convenient. And we would stop at 1 year of follow-up. So there is no doubt, this registry cannot replace a randomized trial. That’s why we are doing a randomized trial.”

And, he added, considering that the results observed in the CREST-2 Registry are better than in other registries, “it is reassuring that when the trial results come out we will be dealing with probably the best possible transfemoral stenting outcomes that can be achieved.”