Topiwala · BMJ (Clinical research ed.) 2017 · prospective cohort study · n=550

Moderate alcohol consumption as risk factor for adverse brain outcomes and cognitive decline: longitudinal cohort study.

Cited 476 times in the scientific literature.

Level 3 - non-randomized controlled study

Prospective observational cohort study with repeated exposure measurements and follow-up neuroimaging

PubMed 28588063 · doi:10.1136/bmj.j2353 · record verified 2026-08-29

What was done

An observational cohort study evaluated 550 community-dwelling UK adults from the Whitehall II imaging substudy (mean baseline age 43.0 years, SD 5.4; non-alcohol dependent via CAGE screening). Weekly alcohol intake and cognitive performance were assessed repeatedly over 30 years (1985–2015). Multimodal brain MRI was conducted at study endpoint (2012–2015). Twenty-three participants were excluded due to poor imaging quality, gross structural abnormalities, or missing clinical data. Measured outcomes included hippocampal atrophy, grey matter density, white matter microstructure, longitudinal cognitive decline, and cross-sectional cognitive performance at scan time.

What was found

Higher alcohol consumption over 30 years was associated with increased odds of hippocampal atrophy in a dose-dependent manner. Consuming >30 units/week had the highest risk versus abstainers (odds ratio 5.8, 95% CI 1.8 to 18.6; P ≤ 0.001). Moderate intake (14–21 units/week) was associated with higher odds of right-sided hippocampal atrophy (OR 3.4, 95% CI 1.4 to 8.1; P = 0.007). Light drinking (1 to <7 units/week) showed no protective effect compared to abstinence. Higher intake was also associated with corpus callosum microstructural differences and faster lexical fluency decline, but showed no association with cross-sectional cognitive scores, semantic fluency change, or word recall change.

Why it matters

This study challenges the notion that moderate alcohol consumption is neuroprotective, providing evidence of dose-dependent adverse structural brain outcomes even at commonly accepted intake levels.

Limits

MRI was acquired only at study completion rather than at baseline, limiting structural conclusions to cross-sectional differences at endpoint. Alcohol intake was self-reported over time. The cohort consisted of British civil servants, potentially limiting generalizability.

Cited by