JAMA September 16, 2026

Revascularization in Asymptomatic Carotid Artery Stenosis With Hemodynamic Impairment and Cognitive Outcomes

CREST-2 Substudy Investigators

Bottom Line

In a substudy of the CREST-2 trial, carotid revascularization did not improve cognitive dysfunction over 1 year compared to intensive medical management alone in patients with asymptomatic high-grade carotid stenosis and baseline reduced cognition.

Key Findings

1. Of the 392 participants enrolled, 291 had complete imaging and cognitive data, and 51 were identified as having reduced cognition at baseline (cognitive z score ≤-1.0 SD).

2. In the reduced cognition subgroup, mean (SD) cognitive z scores improved slightly but similarly in both groups: from -1.36 (0.51) to -1.17 (0.47) in the revascularization arm, and from -1.34 (0.83) to -1.01 (0.97) in the intensive medical management (IMM) arm.

3. There was no significant difference in cognition between the treatment groups over 1 year (adjusted difference in cognitive z score, -0.15 SD [95% CI, -0.54 to 0.24]).

4. There was no significant interaction between the treatment group and the degree of baseline hemodynamic impairment regarding cognitive outcomes (β estimate = 0.018 [95% CI, -0.025 to 0.061]; P = .40).

Study Design

Design

RCT Substudy

Single-Blind

Sample

392

Patients

Duration

1 yr

Median

Setting

Multicenter, North America

Population Adults aged 35-86 years without dementia, with an eighth-grade education or higher, and unilateral asymptomatic internal carotid artery stenosis ≥70%.

Intervention Carotid revascularization (endarterectomy or stenting) plus intensive medical management (IMM).

Comparator Intensive medical management (IMM) alone.

Outcome Difference in composite cognitive z score over 1 year based on 5 cognitive tests, testing for interaction with baseline hemodynamic impairment among those with reduced baseline cognition.

Study Limitations

• The primary analysis cohort was limited to only 51 participants with baseline reduced cognition and complete data, which may restrict the statistical power to detect subtle cognitive changes.

• A substantial proportion of enrolled participants (101 of 392) were excluded due to missing baseline imaging or incomplete cognitive examinations, which introduces the potential for selection bias.

• The follow-up duration was limited to 1 year, precluding the evaluation of longer-term cognitive trajectories or delayed benefits from revascularization.

Clinical Significance

These findings challenge the longstanding hypothesis that correcting chronic cerebral hypoperfusion through carotid revascularization can reverse or improve cognitive impairment. Consequently, prophylactic carotid endarterectomy or stenting should not be performed primarily to treat cognitive dysfunction in patients with asymptomatic internal carotid artery stenosis.

Historical Context

For decades, it has been debated whether asymptomatic high-grade carotid stenosis contributes to insidious cognitive decline via chronic cerebral hypoperfusion (often referred to as 'vascular dementia' or 'hypoperfusion dementia'). The CREST-2 trial was designed to compare modern intensive medical management to revascularization for stroke prevention. The CREST-H substudy was prospectively designed to rigorously investigate whether reversing hemodynamic impairment via revascularization could yield cognitive benefits, ultimately finding no such advantage over a 1-year period.

Guided Discussion

High-yield insights from every perspective

Med Student

Medical Student

Why might high-grade asymptomatic carotid artery stenosis theoretically lead to cognitive impairment, and what compensatory mechanisms often prevent ischemic symptoms in these patients?

Key Response

Cognitive impairment in these patients is historically hypothesized to be driven by chronic microvascular hypoperfusion or silent microemboli. However, the Circle of Willis and pial collaterals provide collateral blood flow, often preserving overall macro-perfusion and preventing focal stroke symptoms, which may explain why mechanically restoring flow in one vessel does not dramatically reverse diffuse cognitive changes.

Resident

Based on this CREST-2 substudy, if a patient with 80 percent asymptomatic carotid stenosis and mild cognitive impairment asks if surgery will help their memory, how should you counsel them?

Key Response

You should counsel the patient that revascularization (endarterectomy or stenting) combined with medical management has not been shown to improve cognitive function at 1 year compared to intensive medical management alone. The focus should remain strictly on aggressive risk factor modification to prevent stroke, not to reverse cognitive decline.

Fellow

How does the presence of 'hemodynamic impairment' theoretically alter the risk profile in asymptomatic carotid stenosis, and why might revascularization fail to rescue cognition even in this high-risk subset?

Key Response

Hemodynamic impairment (e.g., impaired cerebrovascular reserve) indicates exhausted collateral capacity. While this increases stroke risk, revascularization might fail to improve cognition because the cognitive deficits are likely driven by irreversible microstructural damage, chronic white matter changes, or concomitant neurodegenerative pathology (like early Alzheimer disease) rather than purely reversible hypoperfusion.

Attending

How do these findings shift the paradigm of 'prophylactic' revascularization for atypical symptoms, and how can we use this data to curb surgical overutilization in older adults with cognitive decline?

Key Response

Proceduralists sometimes justify revascularization in asymptomatic patients with cognitive complaints under the guise of improving cerebral blood flow. This study directly refutes the utility of 'flow restoration' for cognitive benefit at 1 year, providing a strong evidence-based mandate to rely on intensive medical management and avoid operative risks in this specific clinical scenario.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD

In evaluating cognitive outcomes following revascularization, what are the methodological limitations of a 1-year follow-up period, and how might practice effects or ceiling effects confound the trajectory of cognitive testing?

Key Response

A 1-year horizon may be too brief to detect meaningful differences in the trajectory of chronic cognitive decline. Furthermore, repeated cognitive testing is heavily influenced by practice effects, which can mask true decline or falsely suggest improvement. A rigorous design must account for baseline cognitive trajectories, utilize sensitive, ecologically valid neuropsychological batteries, and extend follow-up over several years.

Journal Editor

As a reviewer, what threats to validity arise from using a substudy of patients with baseline cognitive impairment from a larger trial like CREST-2, particularly regarding selection bias and statistical power?

Key Response

Substudies often suffer from reduced statistical power, significantly increasing the risk of Type II errors (failing to detect a true difference). Additionally, patients who consent to intensive neuropsychological testing and meet the criteria for baseline reduced cognition may not represent the general population of asymptomatic carotid stenosis, limiting external validity. Reviewers must rigorously scrutinize the post-hoc power calculations for this specific cognitive endpoint.

Guideline Committee

Current AHA and ESVS guidelines emphasize intensive medical therapy for asymptomatic carotid stenosis but allow revascularization in highly selected patients. Should guidelines explicitly recommend against revascularization if the sole indication is attempting to reverse cognitive impairment?

Key Response

Current guidelines recommend intensive medical therapy (Class I) and consider CEA/CAS in highly selected asymptomatic patients (Class IIa/IIb) primarily for stroke prevention. The guidelines currently lack strong directives on cognitive indications. This high-quality evidence supports adding a specific recommendation (potentially Class III: No Benefit) against performing CEA/CAS with the primary goal of reversing cognitive dysfunction or treating hemodynamic impairment-associated cognitive decline.

Clinical Landscape

Noteworthy Related Trials

1995

ACAS Trial

n = 1,662 · JAMA

Tested

Carotid endarterectomy (CEA) plus aspirin

Population

Asymptomatic patients with >=60% carotid stenosis

Comparator

Aspirin alone (Medical therapy)

Endpoint

Ipsilateral stroke or perioperative stroke or death

Key result: CEA significantly reduced the 5-year risk of ipsilateral stroke from 11.0% to 5.1% compared to medical management alone.

2004

ACST-1 Trial

n = 3,120 · Lancet

Tested

Immediate carotid endarterectomy (CEA)

Population

Asymptomatic patients with severe carotid stenosis

Comparator

Indefinitely deferred CEA (Medical management)

Endpoint

Any stroke or perioperative death

Key result: Immediate CEA halved the 5-year stroke risk compared to deferred CEA, confirming robust long-term benefits.

2010

CREST Trial

n = 2,502 · NEJM

Tested

Carotid artery stenting (CAS)

Population

Symptomatic and asymptomatic patients with significant carotid stenosis

Comparator

Carotid endarterectomy (CEA)

Endpoint

Composite of stroke, myocardial infarction, or death

Key result: Rates of the primary composite endpoint were similar between CAS and CEA, though CAS had a slightly higher stroke risk and CEA had a higher MI risk.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis