The Neurosurgery Signal — Signal, Not Noise

Rapid review note

Journal Club is a rapid, AI-assisted appraisal layer. It highlights study design, effect estimates, and practice relevance, but it is still a briefing, not a replacement for the paper.

For education only. Not medical advice.

Paper snapshot

Rapid study overview

Open paper

DOI

10.1056/NEJMoa0912321

PMID

N/A

PICO

Population

Patients with symptomatic (TIA, amaurosis fugax, or minor nondisabling stroke within 180 days) or asymptomatic extracranial carotid stenosis meeting imaging criteria

Intervention

Carotid-artery stenting with embolic protection device

Comparator

Carotid endarterectomy

Outcomes

Composite of stroke, myocardial infarction, or death during periprocedural period or ipsilateral stroke within 4 years

Design

Type

Randomized controlled trial

Randomized

Yes

Multicenter

Yes

Blinded

Blinded end-point adjudication

Follow-up

Median 2.5 years, primary endpoint at 4 years

Primary endpoint

Composite of stroke, myocardial infarction, or death during periprocedural period or ipsilateral stroke within 4 years

Secondary endpoints

Stroke or death at 4 years
Cranial-nerve palsy during periprocedural period
Health status at 1 year (SF-36)

Practice impact

What this means

CREST trial: 2502 patients randomized to carotid stenting vs endarterectomy. No difference in 4-year composite of stroke/MI/death (7.2% vs 6.8%). Stenting had higher periprocedural stroke (4.1% vs 2.3%), endarterectomy had higher MI (1.1% vs 2.3%). Age matters: stenting better for <70, endarterectomy for >70. Both require highly skilled operators.

Bottom line

Carotid stenting and endarterectomy have similar composite outcomes but differ in periprocedural risks: higher stroke with stenting, higher MI with endarterectomy.

Strength of evidence

high

Recommendation

consider change

Why it matters

Composite primary endpoint showed no significant difference (HR 1.11, p=0.51)
Age interaction suggests stenting may be better for patients <70, endarterectomy for >70
Stroke had greater impact on quality of life than MI based on SF-36 analysis

What would change my mind

A validated full-text appraisal with explicit effect estimates and page-linked citations.

Critical appraisal

How strong is the paper?

Methods critique

Risk of bias

Randomization, credentialing of operators, and blinded end-point adjudication make internal validity strong, although crossover, differential procedural expertise, and exclusion of an underperforming site still matter when applying the results.

Confounding

Baseline characteristics well-balanced except dyslipidemia (82.9% vs 85.8%, p=0.048); intention-to-treat analysis used

Missing data

Multiple imputation used to assess bias from differential withdrawal; 9.2% lost to follow-up overall

Multiplicity

Two interim analyses with O'Brien-Fleming boundaries; prespecified subgroup analyses for symptomatic status, sex, and age

Notes

Operator certification required >12 procedures/year with complication rates <3-5%
Single stent system used (Acculink with Accunet)
Data from one center excluded due to fabrication (20 patients)

Stats check

NNT

N/A

Effect sizes

Primary endpoint HR 1.11 (95% CI 0.81-1.51, p=0.51)
Stroke or death at 4 years HR 1.50 (95% CI 1.05-2.15, p=0.03)
Periprocedural stroke HR 1.79 (95% CI 1.14-2.82, p=0.01)

Absolute effects

Primary endpoint: 7.2% stenting vs 6.8% endarterectomy (absolute difference 0.4%)
Periprocedural stroke: 4.1% vs 2.3% (absolute difference 1.8%)
Periprocedural MI: 1.1% vs 2.3% (absolute difference -1.1%)

Concerns

Age interaction found (p=0.02) with crossover at ~70 years
Lower than expected event rate for asymptomatic patients
No adjustment for multiple secondary endpoints

External validity

Who it applies to

Patients similar to the study population once inclusion criteria are verified.

Who it does not

Patients outside the verified eligibility criteria.

Generalizability notes

External validity depends on operative workflow, center expertise, and patient selection.

Evidence trace

Source trace and metadata

Citations (5)

claim_id

methods_critique.risk_of_bias

locator

p. 2 Methods

quote

certification was achieved by 477 surgeons... documenting that they performed more than 12 procedures per year

claim_id

methods_critique.risk_of_bias

locator

p. 1 Results

quote

hazard ratio with stenting, 1.11; 95% confidence interval, 0.81 to 1.51; P = 0.51

claim_id

methods_critique.risk_of_bias

locator

p. 8 Results

quote

interaction between age and treatment efficacy was detected (P = 0.02)... carotid-artery stenting tended to show greater efficacy at younger ages

claim_id

stats_check.concerns

locator

p. 1

claim_id

practice_impact.bottom_line

locator

p. 1

Metadata

Generated at

2026-03-09T22:32:33.777Z

Version

pdf-archive-ingest-v1