CREST Trial: Carotid Stenting versus Endarterectomy for Carotid-Artery Stenosis
In patients with symptomatic or asymptomatic carotid stenosis at average surgical risk, does carotid-artery stenting (CAS) with embolic protection differ from carotid endarterectomy (CEA) on the 4-year composite of periprocedural stroke, MI, or death plus ipsilateral stroke?
Brott TG, Hobson RW 2nd, Howard G, et al. Stenting versus endarterectomy for treatment of carotid-artery stenosis. N Engl J Med. 2010;363(1):11–23. · doi:10.1056/NEJMoa0912321 · 2,502 patients
Population
Included
- Symptomatic carotid stenosis: TIA, amaurosis fugax, or minor nondisabling stroke within 180 days, with ipsilateral stenosis ≥50% by angiography, ≥70% by ultrasound, or ≥70% by CTA/MRA
- Asymptomatic carotid stenosis: ≥60% by angiography, ≥70% by ultrasound, or ≥80% by CTA/MRA
- Average surgical risk (distinct from the high-surgical-risk SAPPHIRE population)
- Eligible for either CAS or CEA in the judgment of the treating team
Symptomatic patients
- Transient ischemic attack, amaurosis fugax, or minor nondisabling stroke involving the study carotid artery within 180 days before randomization
- Stenosis of 50% or more on angiography, 70% or more on ultrasonography, or 70% or more on CT angiography or magnetic resonance angiography if the stenosis on ultrasonography was 50 to 69%
Asymptomatic patients (eligibility added in 2005)
- No qualifying ipsilateral neurologic event within 180 days
- Stenosis of 60% or more on angiography, 70% or more on ultrasonography, or 80% or more on CT angiography or magnetic resonance angiography if the stenosis on ultrasonography was 50 to 69%
General
- Clinical and anatomical suitability for management by means of either of the study revascularization techniques (carotid stenting or carotid endarterectomy)
- Average surgical risk (the high-surgical-risk SAPPHIRE population was not enrolled)
Excluded
- Previous disabling stroke
- Chronic atrial fibrillation
- Inability to comply with dual antiplatelet therapy in the CAS arm
- Anatomy precluding safe execution of either CAS or CEA
- Previous stroke that was sufficiently severe to confound the assessment of end points
- Chronic atrial fibrillation
- Paroxysmal atrial fibrillation that had occurred within the preceding 6 months or that necessitated anticoagulation therapy
- Myocardial infarction within the previous 30 days
- Unstable angina
Source: Brott TG et al., NEJM 2010;363(1):11-23; PMID 20505173; PMCID PMC2932446· Retrieved 2026-06-09
Primary Outcome
4-year composite of periprocedural stroke, MI, death + ipsilateral stroke
Negligible absolute difference
4-year primary composite
Superiority not demonstrated for either approach on the composite. No prespecified noninferiority margin; failure to reject the null is not proof of equivalence. The clinically meaningful story is the component split: CAS roughly doubles periprocedural stroke (4.1% vs 2.3%, HR 1.79); CEA roughly doubles periprocedural MI (2.3% vs 1.1%, HR 0.50 favoring CAS); cranial-nerve palsy is essentially a CEA-only harm (4.7% vs 0.3%, HR 0.07). Treatment-by-age interaction (P=0.02) crosses over near 70 years.
Study Arms
- Agent
- RX Acculink self-expanding nitinol stent with RX Accunet distal embolic-protection device (Abbott Vascular)
- Route
- Endovascular
- Frequency
- Single procedure
- Duration
- One-time, with embolic protection used whenever feasible (96.1% of patients)
- Co-interventions
- Dual antiplatelet therapy: aspirin 325 mg twice daily plus clopidogrel 75 mg twice daily for at least 48 hours before the procedure. Interventionalists were certified through a formal CREST lead-in credentialing program.
Periprocedural stroke was higher with stenting (4.1% vs 2.3% for endarterectomy, P=0.01), while periprocedural myocardial infarction was lower (1.1% vs 2.3%, P=0.03). This stroke-versus-MI trade-off is the clinically actionable signal from CREST.
- Agent
- Standard carotid endarterectomy by credentialed vascular surgeons
- Route
- Surgical
- Frequency
- Single procedure
- Duration
- One-time; general anesthesia in 90.0%, patch repair in 62.4%, shunt use in 56.7%
- Co-interventions
- Aspirin 325 mg daily for at least 48 hours before the procedure, continued for one year or longer.
Periprocedural myocardial infarction was higher with endarterectomy (2.3% vs 1.1% for stenting, P=0.03), while periprocedural stroke was lower (2.3% vs 4.1%, P=0.01). The 4-year primary composite did not differ between arms (7.2% CAS vs 6.8% CEA, HR 1.11, 95% CI 0.81-1.51, P=0.51); superiority of neither approach was demonstrated.
Safety
Periprocedural stroke (30 day)
4.1%
2.3%
Any periprocedural stroke: 4.1% (CAS) vs 2.3% (CEA), HR 1.79 (95% CI 1.14–2.82), P=0.01. CAS roughly doubles periprocedural stroke risk versus CEA. Minor ipsilateral stroke drove much of the difference (2.9% vs 1.4%, HR 2.16, P=0.009); major ipsilateral stroke was directionally consistent but not significant (0.9% vs 0.3%, HR 2.67, P=0.09).
Periprocedural death (30 day)
0.7%
0.3%
Death within the periprocedural window: 0.7% (CAS) vs 0.3% (CEA), HR 2.25 (95% CI 0.69–7.30), P=0.18. Numerically higher with CAS but not statistically significant; absolute event counts are small.
Periprocedural MI (30 day)
1.1%
2.3%
Periprocedural myocardial infarction: 1.1% (CAS) vs 2.3% (CEA), HR 0.50 (95% CI 0.26–0.94), P=0.03. CEA approximately doubles periprocedural MI risk versus CAS. Note: this row uses the majorBleeding slot to surface a non-bleeding harm that is the mirror image of the stroke signal. Read it as "MI rate" not "major bleeding".
Trial Design
Type
- Multicenter randomized open-label superiority trial
- 117 sites (108 US, 9 Canada)
- 1:1 allocation CAS vs CEA
- Symptomatic (~52.7%) and asymptomatic (~47.3%) carotid stenosis combined
- Formal operator credentialing / lead-in phase for both arms
- Composite primary endpoint; time-to-event analysis
Timeline
Enrolled December 2000 – July 2008; median follow-up 2.5 years (planned to 4 years); primary results published NEJM July 2010
N
2,502
Enrollment
2,502 patients (1,262 CAS / 1,240 CEA, ITT after exclusion of 1 site for data fabrication) at 117 sites (108 US, 9 Canada). Symptomatic 52.7% / asymptomatic 47.3%. Formal operator credentialing / lead-in phase for both arms. Median follow-up 2.5 years (planned 4 years). Enrolled December 2000 to July 2008. Published NEJM 2010.
ClinicalTrials.gov
NCT00004732Bedside Pearl
CREST compared carotid stenting with endarterectomy in average-surgical-risk patients. The 4-year primary composite was 7.2% (CAS) vs 6.8% (CEA), HR 1.11 (95% CI 0.81–1.51, P=0.51). No significant difference; superiority not demonstrated. The actionable signal is the component split: CAS roughly doubles periprocedural stroke (4.1% vs 2.3%, HR 1.79), CEA roughly doubles periprocedural MI (2.3% vs 1.1%, HR 0.50 favoring CAS), and cranial nerve palsy is essentially a CEA-only harm (4.7% vs 0.3%, HR 0.07). Treatment-by-age interaction (P=0.02) crosses over near 70 years; CAS tends to do better in younger patients, CEA in older.