‘It works in practice,but does it work in theory?’ An important addition to the body of real-world evidence supporting carotid endarterectomy

Carotid artery stenosis remains a significant underlying cause of stroke. Accordingly, the optimal management and treatment paradigms remain the focus of multispecialty and multidisciplinary care. Currently, there are three primary revascularization procedures to reduce long-term stroke risk when utilized in an evidenced-based fashion: carotid endarterectomy (CEA), transfemoral carotid stenting (CAS), and transcarotid artery revascularization (TCAR). ¹ As advances in medical therapy and surgical technology evolve over time, the optimal application and utilization of different revascularization procedures remain a focus of debate concerning the nuances of best treatments for patients with either asymptomatic or symptomatic carotid disease, respectively.^2–4 Moreover, these controversies highlight the benefits and/or limitations of randomized trials and observational studies as to how they can inform practice.

In this issue of Vascular Medicine, Willot and colleagues report on 10-year nationwide practice patterns and outcomes after carotid revascularization with CEA and CAS in France. ⁵ The investigators used data from the French National Health Data System, which integrates with payor records, private and public health records, and the national death register. As a result, the study included nearly 150,000 patients who underwent CEA or CAS with 9 years of follow-up, which compares impressively with the number of procedures included in many prior observational studies on this topic. Specifically, the authors documented that patients with asymptomatic carotid stenosis comprised 85% of this group; those who underwent CEA incurred a 1.5% 30-day stroke or death risk, whereas patients who underwent CAS experienced a comparative 3.0% 30-day stroke or death risk. In contrast, symptomatic patients comprised 15% of the study, where patients who underwent CEA exhibited an 11.4% 30-day stroke or death risk, compared to a 14.2% 30-day stroke or death risk for those who underwent CAS. Notably, at 5 years, patients who underwent CAS demonstrated greater cumulative stroke and mortality risks than patients who underwent CEA in both the asymptomatic and symptomatic groups. In addition, patients in this study achieved excellent medical compliance, with 98% of all patients meeting their postoperative definition.

This real-world, well-conducted study is particularly timely in light of the recent Medical Management and Revascularization for Asymptomatic Carotid Stenosis (CREST-2) trial, and highlights the utility of high-quality observational analyses and their important role in supplementing randomized trials to best inform practice. ⁶ Specifically, CREST-2 demonstrated that medical management plus CAS had a stroke-risk reduction benefit over medical management alone, whereas medical management plus CEA did not demonstrate benefit over medical management alone. It is worth emphasizing that the recent CREST-2 results represent a notable departure from a growing body of literature demonstrating CEA confers equivalent or superior results compared to CAS among asymptomatic patients.^7,8

The apocryphally coined phrase ‘it works in practice, but does it work in theory?’ is an aptly chosen expression that highlights a potential limitation of randomized trials. Specifically, the results of Willot et al., in addition to many other prior observational studies, document that CEA appears to offer excellent real-world outcomes despite comparatively underperforming in CREST-2, where more rigid inclusion and exclusion criteria may diverge from everyday practice. The ultimate underlying reasons for these observational outcome disparities remain somewhat unclear and are a focus of continued debate.^9,10

It is also worth highlighting the excellent postoperative medical therapy compliance reported by Willot et al., with 98% of patients meeting their definition of postoperative medical therapy compliance, as well as documenting the high levels of preoperative adherence. Typically, clinical trials, such as CREST-2, have been criticized for sporadically unrealistic medical therapy adherence rates, thought achievable only through rigorous study governance and oversight. However, this study demonstrates that excellent medical therapy rates can also be achieved in real-world practice.

Moreover, observational studies capture outcomes from proceduralists who may not have had the requisite procedure volume that is often required to serve as a participating trial interventionalist. This element of trial design raises appropriate questions about the generalizability of outcomes documented in randomized trials and may serve to effectively calibrate true outcome benchmarks. Therefore, although the results reported in the CREST-2 trial are laudable, the findings presented herein by Willot et al. offer an important look into real-world clinical practice, which may in part account for some of the differential findings.

In conclusion, although randomized trials historically represent the highest level of evidence to inform practice, it is important that they be interpreted through the lens of generalizability. When thoughtfully executed, real-world studies provide an important supplement to bridge the limitations of randomized trials, despite being subject to various potential biases and confounding. Accordingly, this study by Willot and colleagues adds to a growing body of literature documenting superior stroke outcomes for CEA when compared to CAS, as well as documenting that excellent rates of medical therapeutic compliance are achievable in a real-world setting.