carotid stenosis
Symptomatic (ipsilateral TIA, non-disabling stroke or amaurosis in the last 6 months) with 50–99% stenosis → urgent assessment for endarterectomy (CEA), ideally within 14 days. Benefit is largest in men, older patients and with early surgery, and falls fast with delay. Asymptomatic → intensive medical therapy for everyone. Revascularisation is only for highly selected patients. Every patient gets antiplatelet, high-intensity statin, BP control and smoking cessation.
symptomatic: the benefit is front-loaded
Recurrent stroke risk is highest in the first days to weeks after the event, so the value of surgery depends on getting there early.
- symptomatic = event in the territory of that carotid: hemispheric TIA or stroke, transient monocular vision loss. Dizziness, syncope and non-focal symptoms don’t count
- duplex ultrasound to screen; CTA or contrast MRA to confirm the degree of stenosis (NASCET method) and look for tandem lesions before deciding on surgery
- refer at once to stroke/vascular surgery — don’t wait for outpatient imaging
| Stenosis | Men | Women |
|---|---|---|
| 70–99% | CEA | CEA |
| 50–69% | CEA (clinically important benefit only within 2 weeks) | consider only if highest recurrence risk |
| <50% | medical therapy (CEA harms at <30%) | medical therapy |
| near-occlusion (collapsed distal ICA) | medical therapy — no benefit at 5 years | medical therapy |
| complete occlusion | no revascularisation | no revascularisation |
- timing: as soon as the patient is stable, ideally ≤14 days. Rothwell, Lancet. 2004 pooled NASCET + ECST: NNT to prevent one ipsilateral stroke at 5 years ≈5 when surgery was within 2 weeks, vs ≈125 beyond 12 weeks. Same analysis: NNT 9 in men vs 36 in women; 5 at age ≥75 vs 18 at <65
- surgery within 48 h after a completed stroke (as opposed to a TIA) may carry more perioperative risk. Waiting a day or two is reasonable; waiting weeks isn’t
- benefit depends on a surgeon/centre with a perioperative stroke/death rate below ~6% (the rate in the original trials)
- moderate–severe disabling stroke → excluded from the trials → benefit uncertain, individualise
- minor stroke or high-risk TIA not undergoing revascularisation (considering bleeding risk) → DAPT (ASA + clopidogrel) for at least 21 days, then single antiplatelet (ESC I) → TIA and minor stroke
- perioperative CEA → continue ASA (and lifelong post-op); if started on DAPT pre-op, stop clopidogrel after surgery (ESC)
CEA or stenting
- CEA is the default. Stenting carries more periprocedural stroke, especially at age >70 and when done in the first week after the event
- stenting → patients unfit for surgery for technical, anatomical or medical reasons (e.g. radiation-induced stenosis, restenosis after CEA), or where CEA can’t be done quickly and stenting can
- extracranial–intracranial bypass for carotid or MCA occlusion → no benefit, not recommended
- tandem lesion treated with thrombectomy but not stented at the time → reassess for CEA or stenting once recovered
asymptomatic: medical therapy first
Contemporary medical therapy has lowered stroke risk well below the rates in the original asymptomatic CEA trials (ACAS, ACST). That shrinks the room for any procedure to help.
- everyone: antiplatelet (CSBP includes in medical management; ESC IIa for >50% stenosis if bleeding risk is low), high-intensity statin → dyslipidaemia, BP to target → hypertension, diabetes control, smoking cessation
- no routine carotid screening in peripheral arterial disease (CCS PAD) — management is the same whichever bed is diseased
- CEA for 60–99% only in highly selected patients (CSBP): life expectancy >5 years, centre with complication rate <3%. Benefit is least clear in women
- features that may raise stroke risk on medical therapy, though it’s unproven that they should change management: progression on serial imaging, silent ipsilateral infarcts, microemboli on transcranial Doppler, high-risk plaque (intraplaque haemorrhage)
CREST-2 (2026), ≥70% asymptomatic stenosis, intensive medical management in all arms. Primary outcome: stroke/death within 44 days, or ipsilateral stroke to 4 years.
- stenting + medical vs medical alone: 2.8% vs 6.0% (ARR 3.2% over 4 years, NNT ~30). All the early events were in the stenting arm (7 strokes, 1 death vs none)
- CEA + medical vs medical alone: 3.7% vs 5.3%, not significant
Medical arms treated to SBP <130 (<140 before 2018) and LDL <1.8 mmol/L; a TIA or stenosis progression on follow-up triggered revascularisation. Event rate on medical therapy alone was 5–6% over 4 years. One trial: the stenting benefit is small and in absolute terms depends on operator complication rates. CREST-2 used only credentialed, high-volume operators (surgeons <3% stroke/death), so results may not transfer. Guidelines predate it. Default stays intensive medical therapy; selected patients wanting to lower long-term risk → discuss with vascular/stroke at a high-volume centre.
after revascularisation
- after stenting → DAPT (ASA + clopidogrel) for at least 30 days (ESC I), then single antiplatelet lifelong. Beyond ~3 months, observational data show no ischaemic gain and possible excess bleeding — no RCT
- after CEA → single antiplatelet lifelong; duplex within the first month, then yearly until stable (ESC)
- control BP before and after either procedure; hyperperfusion syndrome (headache, seizure, ICH) can occur despite BP control → urgent neuroimaging to exclude ICH and strict BP lowering
- statin lifelong
- most mortality is cardiac, not stroke → aggressive lipid, BP and smoking control matters as much as the neck
key trials
| Study | Population | Result |
|---|---|---|
| NASCET / ECST pooled (Rothwell, Lancet. 2004), n=5,893 | symptomatic carotid stenosis | 5-year ipsilateral stroke ARR 16.0% at ≥70% (no near-occlusion) and 4.6% at 50–69% (AHA); no benefit at 30–49%, harm at <30%; benefit front-loaded (NNT ~5 ≤2 weeks vs ~125 >12 weeks), larger in men and age ≥75 |
| CREST-2 (2026) | asymptomatic ≥70%, intensive medical therapy | stenting ↓ 4-year events (6.0 → 2.8%); CEA not significant |
| Yoo, Stroke. 2025 (observational, n≈12,000) | post-CAS DAPT: 3–6 mo vs longer | no difference in stroke or bleeding |
| Kishi, J Stroke Cerebrovasc Dis. 2025 (observational, n=160 matched) | post-CAS DAPT: ≤3 mo vs >3 mo | >3 months → more bleeding, similar ischaemic events |
Prepared 2026-11-19. Not yet verified — check every number against the sources.