In Carotid RCTs, Outcomes Improved Over Time for CEA but not CAS: Meta-analysis

Data drawn from four randomized controlled trials over the better part of a decade show a decrease in procedural stroke/death risk for symptomatic patients who undergo carotid endarterectomy (CEA), researchers report in a new meta-analysis. Carotid artery stenting (CAS) outcomes, on the other hand, didn’t budge.

“Treatment of symptomatic carotid stenosis within the examined trials became safer over time,” Mandy D. Müller, MD (University Hospital Basel, Switzerland), and colleagues write in their paper published online recently in Circulation: Cardiovascular Interventions. Why this positive shift was restricted to CEA, however, is unclear, they observe. “Mechanisms underlying the findings remain to be determined.”

The meta-analysis, by the Carotid Stenosis Trialists’ Collaboration, included patient-level data from EVA-3S (2000-2005), SPACE (2001-2006), CREST (2000-2008), and ICSS (2001-2008) for a total of 4,597 individuals with systematic carotid artery disease.

Commenting on the findings for TCTMD, William Gray, MD (Lankenau Heart Institute, Wynnewood, PA), expressed skepticism about the lack of improvement for CAS. In particular, he questioned the decision to combine data from CREST, which he described as “well done,” with numbers from three “problematic” studies.

“The CREST improvement is swamped by the poor outcomes in the other trials,” he said.

Gray was less wary of the evidence for CEA. “I think the CEA differences are probably very real. There’s lots of data that show CEA outcomes over the last 40 years have improved dramatically,” he said, adding that by the time of the trials, operators were quite experienced at doing this procedure that had been around for decades.

Procedural Stroke/Death

In the new meta-analysis, baseline characteristics were well-balanced between the two approximately equal groups of patients receiving CAS or CEA. Patients enrolled in the latter half of the study period were significantly more likely to have a history of hypertension, coronary heart disease, smoking, and severe ipsilateral stenosis. They also were much more likely to be on lipid-lowering therapy (74.1% vs 52.4%; P < 0.001).

Procedural stroke or death within 30 days (primary endpoint) decreased over time for carotid revascularization as a whole. Individually, the stroke/death risk decreased for CEA (unadjusted OR 0.82 per year; 95% CI 0.73-0.92) but not for CAS (unadjusted OR 0.96; 95% CI 0.88-1.04). Adjustment for numerous baseline characteristics thought to affect stroke/death risk did not substantially change the results. The disparity between CEA and CAS was seen over the entire enrollment period but became more pronounced over time.

Shifts in the Carotid Field

Müller and colleagues say their findings were unexpected.

“Widespread CAS only started a few years before the start of trials contributing to this meta-analysis. We, therefore hypothesized that technical development and increasing experience would lead to a stronger decline in procedural risk with CAS compared with CEA. Surprisingly, we found the opposite to be true,” they write.

Growing use of neurophysiological monitoring and intraoperative assessment may be what shifted the needle for CEA, they suggest. “For CAS, however, with growing experience, interventionists might have accepted patients in the trials with more difficult anatomy, which may have counteracted any learning-curve effect.”

For his part, Gray said that pharmacology, technique, and patient selection may all have improved CEA outcomes between 2000 and 2008. “And it may just be that we’re picking up the line of improvement that we’ve been seeing over the last 40 years anyway,” he commented.

Müller et al readily acknowledge various limitations to their meta-analysis. Most importantly, they say, “randomized trials provide data from highly selected centers, operators, and patients. . . . Thus the external validity of the absolute risks observed on our study is limited.” The arrival of newer stent designs, reverse-flow embolic protection, and direct transcervical access also means that procedural outcomes may well be better now than they were in 2000-2008, the researchers point out.

CREST, EVA-3S, SPACE, and ICSS

Gray emphasized to TCTMD that variations across the four trials masked the potential to detect patterns in CAS outcomes.

Each trial captured a different section of the learning curve, Gray said. CREST was slow to enroll in its first 4 years, and only gained steam in its latter half, so US experience levels had more time to grow, he explained. In Europe, CAS devices were approved earlier before operators had gained the same level of experience. Additionally, “CREST mandated embolic protection in all of its patients, so in a sense it was modern carotid stenting,” Gray pointed out. “None of the other trials [did] except for EVA-3S,” which began doing so only after the first 80 CAS patients showed an unexpectedly high stroke rate.

To make his case that CAS outcomes did not in fact stagnate, Gray pointed to his analysis of CREST data, published 7 years back in Circulation, showing that as time progressed the rate of stroke/death with CAS was reduced by half—from 4.4% in 2000-2004 to 1.8% in 2008. Another analysis, this time based on all prospective CAS trials with US Food and Drug Administration oversight between 2000 and 2011, tracked similar shifts.

Still-pending results from the CREST-2 trial will help clarify the position of medical therapy compared with revascularization, at least in asymptomatic patients. But its unique two-arm design—CEA versus medical therapy or CAS versus medical therapy—precludes direct comparisons between the two interventions. Instead, that information may come from the CREST-2 Registry, designed to provide a window into real-world outcomes in a diverse patient cohort.