What distinguishes CREST from other carotid trials?


By Jon Matsumura

Several features distinguish the CREST trial from most prior published trials of this new therapeutic option for patients with severe carotid stenosis. First, it is the largest randomised trial comparing carotid stenting with routine embolic protection to carotid endarterectomy, and completed its targeted enrolment. Equally important and contrary to most multicentre randomised trials which have focused on symptomatic patients – such as CAVATAS, Wallstent trial, EVA-3S, SPACE, and ICSS – CREST recruited almost half asymptomatic subjects. Only SAPPHIRE included a majority of asymptomatic patients, but focused on patients who were high risk for endarterectomy. Further, leadership of the CREST trial included multiple disciplines, and surgical and interventional management committees used stringent criteria for investigator selection. Specifically, the interventional committee considered a stroke rate outcome threshold, as well as a prior volume threshold, and required a lead-in phase prior to allowing investigators to perform carotid stenting in the randomised phase.


CREST proscribed routine embolic protection, whereas most other trials have started with optional use of embolic protection (later required by DSMB in EVA-3S). There are also differences in endpoint definitions of stroke and myocardial infarction (CREST required enzyme elevation plus electrocardiogram changes or symptoms), and the primary composite endpoints vary in trials with some including late death and late contralateral stroke (CREST included neither).


Given these fundamental differences in trial design and heterogeneity of patient cohorts, it will be difficult to analyse combined outcomes of multiple trials. However, careful study of the individual trials suggests some common themes. In my opinion, an outcome threshold is critical for optimal results with intervention, and has been common in endarterectomy trials such as ACAS and ACST. Volume thresholds or proctoring are insufficient. Similar risk-adjusted quality thresholds must be employed in the clinical practice of carotid intervention, whether by endarterectomy or carotid stenting in order to maintain these same excellent outcomes. In my interpretation of these trials, I believe that older and recently symptomatic patients have worse outcomes with distally placed filter devices. Conversely, younger and asymptomatic patients have better outcomes with carotid artery stenting.


CREST has demonstrated that carotid artery stenting and endartectomy are both safe and effective for treatment of patients with carotid artery stenosis, and further research should follow up on hypotheses generated from subgroup analysis, so physicians may optimise the care of patients by individualising treatment options which now include carotid stenting, as well as carotid endarterectomy and medical therapy alone.


Jon Matsumura, vascular surgeon, is professor of the Department of Surgery, and chairman, Division of Vascular Surgery, University of Wisconsin, Madison, USA