Opening a narrowed but symptom-free carotid artery did not improve thinking
In the CREST-2 trials, stenting or surgery added to intensive medical treatment left cognition no better at four years than medical treatment alone in people with severe asymptomatic carotid stenosis.
| Group | Value (value) |
|---|---|
| CAS + IMM | 0.22 (0.13 to 0.3) |
| IMM alone (CAS trial) | 0.18 (0.09 to 0.27) |
| CEA + IMM | 0.25 (0.17 to 0.33) |
| IMM alone (CEA trial) | 0.23 (0.15 to 0.32) |
Repairing a severely narrowed carotid artery that had not yet caused a stroke did not improve patients' thinking and memory, according to the prespecified cognitive results of the CREST-2 randomised trials published in JAMA [s1]. Adding a stent or surgery to intensive medical treatment left cognition no better at four years than intensive medical treatment alone [s1].
The question matters because asymptomatic carotid stenosis — plaque narrowing the neck artery in people who have had no stroke or warning symptoms — is linked in observational studies to poorer cognition, and that link has been used to argue for opening the artery to protect the brain [s1]. CREST-2 tested the argument directly, and it did not hold.
What the trials did
CREST-2 is actually two parallel randomised trials in patients aged 35 or older with asymptomatic carotid stenosis of 70% or greater [s1]. One trial compared carotid stenting plus intensive medical management (CAS + IMM) against intensive medical management alone; the other compared carotid endarterectomy, the surgical option, plus intensive medical management (CEA + IMM) against medical management alone [s1]. Enrolment ran from December 2014 through July 2024, with follow-up ending in July 2025 [s1].
Cognition was a prespecified secondary outcome. Five tests covering learning, memory, executive function, attention and processing speed were given before treatment and then annually for up to four years, scored centrally by assessors blinded to which treatment each patient had received [s1]. The prespecified primary cognitive analysis was the fully adjusted composite score at four years, and the researchers set 0.50 standard deviations as the minimal difference that would count as clinically meaningful [s1].
What it found
Among 2,366 US patients, 2,165 (91.5%) took part in the cognitive assessments — 1,078 in the stenting trial (529 CAS + IMM, 549 medical management) and 1,087 in the surgery trial (543 CEA + IMM, 544 medical management) [s1]. Mean age was 70.1 years and 37.8% were women [s1].
At four years the adjusted composite cognitive scores were essentially identical between groups. In the stenting trial the score was 0.22 with CAS + IMM versus 0.18 with medical management alone, a between-group difference of 0.03 (95% CI −0.09 to 0.15) [s1]. In the surgery trial it was 0.25 with CEA + IMM versus 0.23 with medical management, a difference of 0.02 (95% CI −0.09 to 0.14) [s1]. Both are a fraction of the 0.50 threshold set as clinically meaningful, and no individual cognitive test showed a significant advantage for revascularisation either [s1].
A companion substudy, CREST-H, went further and asked whether patients whose brains showed reduced blood flow on perfusion imaging — the group most plausibly helped by restoring flow — benefited [s2]. Among 291 participants with complete data, including 51 who already had reduced cognition at baseline, there was no differential effect of revascularisation by blood-flow status over one year (adjusted difference −0.15 SD; 95% CI −0.54 to 0.24; interaction P=0.40) [s2]. That undercuts the most mechanistically appealing version of the hypothesis.
How to read it
The cleanest reading is that revascularisation does not fix cognition in this population, even in the subgroup where restoring blood flow should matter most [s1][s2]. That is a genuinely useful negative: it removes "protecting the brain" as a reason to open an asymptomatic carotid artery, a procedure that carries its own upfront stroke and complication risk.
Two caveats keep the finding honest. Cognition was a secondary outcome; CREST-2's primary purpose is to compare stroke and death across these strategies, and those results speak to a different question [s1]. And the trials enrolled people healthy enough to be randomised to surgery, so they do not describe every patient with carotid disease. The accompanying editorial's framing captures it: revascularisation of an asymptomatic carotid artery may still be justified for some patients — but not to improve cognition.
Readers weighing the upstream question of whether to look for this narrowing at all can see the evidence on screening asymptomatic adults for carotid stenosis, which independent guideline bodies have generally not recommended.
Why it matters
Procedures justified by a plausible mechanism, rather than by trial evidence that they help patients, are a recurring problem in medicine. CREST-2 took a mechanism widely assumed to be true — narrowed artery, starved brain, restore flow, restore function — and showed the last two steps do not follow [s1][s2]. For patients, it means a cognitive complaint is not, by itself, a reason to undergo carotid surgery or stenting.
What to watch
The trials' main stroke-and-death results, comparing each intervention with modern medical therapy, are the outcome that will actually shape practice, and are the number to watch when they report in full.
This article describes trial results and is not medical advice. Decisions about carotid procedures are for patients and their treating clinicians.
Sources
- Revascularization of Asymptomatic Carotid Artery Stenosis: Cognitive Results of the CREST-2 Randomized Trials — JAMA, 16 September 2026 (prespecified cognitive outcomes)
- Revascularization in Asymptomatic Carotid Artery Stenosis With Hemodynamic Impairment and Cognitive Outcomes: The CREST-H Substudy — JAMA, 16 September 2026 (blood-flow subgroup analysis)
Sources
- Revascularization of Asymptomatic Carotid Artery Stenosis: Cognitive Results of the CREST-2 Randomized Trials — JAMA , September 16, 2026
- Revascularization in Asymptomatic Carotid Artery Stenosis With Hemodynamic Impairment and Cognitive Outcomes: The CREST-H Substudy of the CREST-2 Randomized Clinical Trial — JAMA , September 16, 2026
More on
Should healthy people be screened for narrowed neck arteries?
For adults with no symptoms, the US Preventive Services Task Force recommends against carotid ultrasound screening — a rare 'D' grade, because finding the narrowing does more harm than good.
A neuroprotective drug improved recovery after ischemic stroke in a Chinese trial
In the LAIS trial, more patients given intravenous loberamisal recovered with little or no disability at 90 days — a rare positive result in a field littered with failed neuroprotectants.
A tau-targeting Alzheimer's drug failed to slow decline in a phase 2 trial
Ceperognastat, an oral OGA inhibitor, showed no benefit on the primary cognitive endpoint at 100 weeks. The higher dose progressed 32% faster than placebo and had more serious adverse events.
Tirofiban after stroke thrombectomy improved recovery in a large China trial
In ATTRACTION, tirofiban given after successful thrombectomy left more stroke patients independent at 90 days than placebo, though symptomatic brain bleeds trended higher and were not significantly different.