Carotid Endarterectomy Little Help in Asymptomatic Patients: VA Study

Medications for stroke prevention have become more potent over the past 20 years, and it could be that carotid endarterectomy (CEA) no longer has an advantage over optimal medical therapy for patients with asymptomatic carotid stenosis, observational data suggest.

After adjustment, the 5-year risk of fatal or nonfatal stroke was lower in patients who received CEA versus initial medical therapy (5.6% vs 7.8%), with a significant risk difference of -2.3% (95% CI -4.0% to -0.3%), according to researchers led by Salomeh Keyhani, MD (San Francisco VA Medical Center, CA).

Not only is that difference less than half of the difference in the risk of any stroke or perioperative death between the CEA and medical therapy arms of the ACST trial (5.4%), but also the advantage for CEA statistically disappeared after accounting for the competing risk of nonstroke death, they report in a study published online June 1, 2020, ahead of print in JAMA Neurology.

“This study suggests that medical therapy might be an equally attractive option, and since there is no complication rate with medical therapy up front, it might be important to consider these findings when we formulate guidelines,” Keyhani told TCTMD.

Commenting for TCTMD, Bernhard Reimers, MD (Humanitas Research Hospital, Rozzano, Italy), said these results should not be taken to mean that CEA doesn’t have a place in the treatment of asymptomatic carotid stenosis in the contemporary era. He acknowledged that medical therapy has become more effective over time. “But in some cases, carotid revascularization is necessary to prevent such a dramatic event as a stroke for our patients,” he added. “So it’s not one thing that goes in contrast to the others. We have to carefully follow up the patient and choose the right moment to intervene with revascularization.”

Looking for a New Answer to an Old Question

Keyhani pointed out that the trials establishing the superiority of CEA over medical therapy for asymptomatic patients with significant carotid stenosis are “very old.” In fact, ACST is the most recent, and it started enrolling patients in 1993. Since then, stroke rates overall have declined and medical therapy has improved, factors that come into play when considering the risk-benefit tradeoff of a surgical procedure that comes with upfront complications.

“Because there’s this risk-benefit ratio, then it’s very important to understand—after the advent of statins and improved blood pressure and cholesterol control—is it still worth it?” Keyhani said, adding that there has also always been a concern about whether the RCT results would translate into community settings.

What our study demonstrates is that in the community setting it’s an open question whether surgery is beneficial. Salomeh Keyhani

For this study, the investigators combined data from the US Department of Veterans Affairs and Medicare to mimic a randomized trial similar to ACST. The analysis included 5,221 US military veterans 65 or older (mean age 73.6 years; 98.8% men) who received carotid imaging between 2005 and 2009 and had stenosis of at least 70%; 51.9% underwent CEA and the rest received initial medical therapy.

The rate of stroke or death within 30 days in the CEA group was 2.5%.

After 5 years, and after adjustment using propensity-score matching, the rate of fatal and nonfatal stroke was lower in the CEA group. The annualized risk difference was 0.46%, within a number needed to treat (NNT) of 43 over 5 years. That compares with an NNT in ACST of about 18, the researchers report.

Findings were similar in an analysis conducted in a sample of patients who met RCT inclusion criteria, with no significant difference in the 5-year risk of stroke between the CEA and medical therapy groups when accounting for the competing risk of nonstroke death.

“What our study demonstrates is that in the community setting it’s an open question whether surgery is beneficial,” Keyhani said, adding that the same question can be asked in the setting of symptomatic carotid stenosis as well.

It’s Not Either-or

Reimers said that he can understand how some physicians and researchers could be frustrated by the lack of more-recent data comparing CEA and medical therapy in asymptomatic carotid stenosis, noting that ongoing trials either don’t include a medical therapy arm—like ACST-2, which is comparing CEA and carotid artery stenting—or are enrolling slowly, a situation that will only be exacerbated by the COVID-19 pandemic. CREST-2, which is two parallel trials comparing both carotid stenting and CEA to optimal medical therapy in asymptomatic patients, started enrolling in 2014.

While awaiting fresh RCT data, this analysis provides some insights into the impact of CEA in contemporary practice, showing that surgery does, in fact, still provide a benefit, Reimers said. He added that the elimination of the difference in an analysis incorporating the competing risk of nonstroke death was not convincing.

“You do endarterectomy to prevent a stroke, and stroke is the first cause of disability in the Western world,” he said. “Maybe it’s less effective [in the current era] but that’s what we want to do with endarterectomy. We cannot avoid death.”

Discussing the choice between CEA and medications in asymptomatic patients, he pointed out that “most of our patients get initial medical therapy. You hardly ever have a finding of asymptomatic 60%, 70%, 80% carotid stenosis. The first thing you do is you start medical therapy and then if the stenosis gets worse, you may opt for surgery.”

In this way, CEA and medical therapy “go hand in hand,” Reimers said. “I don’t think that here we have a nail in the coffin of endarterectomy.”

CREST-2 will likely provide some clearer answers, but Keyhani said even those findings will have limited applicability to practice in community settings. That’s because CREST-2 is using stringent criteria for patient and surgeon selection and employs risk factor management that is above and beyond what can be done in routine practice. The trial, she said, would be more informative if it could have been designed as a pragmatic study in an all-comers population.