Across 97 studies, thicker carotid intima-media (HR 1.07 per standard deviation) and 50-100% asymptomatic carotid stenosis (HR 1.48, 95% CI 1.06-2.07) were linked to higher dementia risk, mostly in stroke-free people. Non-randomized studies suggested revascularization improved cognition (SMD 0.73), but all carried high bias risk and heterogeneity topped 89%. The large CREST-2 trial found essentially zero cognitive effect from stenting (0.03), and ACST-1 showed no long-term dementia difference (HR 0.98).
The split between association and intervention is the story. Carotid disease likely marks a vascular risk profile (hypertension, dyslipidemia, diabetes, smoking) that independently damages the brain, so fixing one artery does not rescue cognition. That points adults toward upstream prevention: blood pressure control, LDL lowering, glycemic management, and not smoking, which has far stronger evidence for protecting brain health than procedures. The pooled dementia associations are modest and rest on few studies (three for stenosis), and observational designs cannot establish causation. The apparent benefit in earlier small trials looks like the familiar pattern of promising early results shrinking in larger, better-controlled work. One caveat: CREST-2 and ACST-1 enrolled mostly asymptomatic patients, so effects in symptomatic disease remain less clear. This is a preprint that has not been peer-reviewed, and the conclusions may shift on review. Overall the work is confirmatory, valuable chiefly for tempering expectations that carotid procedures protect cognition.