CREST-H: restoring carotid flow does not restore cognition

Design

prespecified substudy within a multicentre phase 3 randomised trial, 61 North American sites

Population

392 adults aged 35–86 without dementia with unilateral asymptomatic internal carotid stenosis ≥70%; 291 with complete data, 51 with reduced baseline cognition

Primary outcome

difference in composite cognitive z score over 1 year, tested for interaction with baseline haemodynamic impairment

Effect

adjusted difference −0.15 SD (95% CI −0.54 to 0.24); interaction β 0.018 (95% CI −0.025 to 0.061), p=0.40

Asymptomatic carotid stenosis with measurable cerebral haemodynamic impairment is associated with worse cognition, and the obvious inference is that the reduced flow is causing it — so opening the artery should help. CREST-H tested the inference inside CREST-2, enrolling 392 adults without dementia who had unilateral asymptomatic internal carotid stenosis of 70% or more, with perfusion imaging to quantify impairment by interhemispheric time-to-peak delay.

Among the 51 participants who had reduced cognition at baseline, revascularisation plus intensive medical management produced no cognitive advantage over medical management alone at one year: adjusted difference in composite cognitive z score −0.15 SD (95% CI −0.54 to 0.24). The interaction between treatment and degree of haemodynamic impairment — the actual question — was not significant (β 0.018, 95% CI −0.025 to 0.061, p=0.40). Both groups improved slightly, which is what repeated cognitive testing produces.

The authors' conclusion is the right one: in chronic asymptomatic carotid stenosis, cognitive dysfunction is not a flow problem that revascularisation fixes. It is more plausibly small-vessel disease and shared vascular risk, which is treated medically or not at all.

• Do not offer carotid revascularisation for cognitive indications — there is now randomised evidence against it.
• Perfusion imaging showing delayed time-to-peak identifies impaired haemodynamics; it does not identify a patient who will benefit cognitively.
• The stroke-prevention question is separate and is what CREST-2 itself addresses; nothing here changes that discussion.
• Where a patient or family raises 'improving blood flow to the brain', this is the study to cite.

Why it matters

It closes off a rationale that has been used to justify intervening on asymptomatic stenosis when the stroke-prevention case was weak.

Don't overread it

The primary analysis rested on 51 participants with reduced baseline cognition — this rules out a large cognitive benefit, not a small one.

The statistics, in plain English

The interaction test is the whole analysis here: it asks whether the treatment effect changes as haemodynamic impairment worsens, and at p=0.40 with an interval straddling zero, it does not. The primary estimate of −0.15 SD sits with an interval from −0.54 to 0.24, so a moderate benefit is not formally excluded — but only 51 participants had reduced baseline cognition, out of 392 enrolled and 291 analysed, which is a thin base for a subgroup-defined primary outcome.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.