Larsson · American journal of epidemiology 2011 · dose-response meta-analysis of prospective cohort studies · n=11 studies (479,689 participants, 10,003 cases)

Coffee consumption and risk of stroke: a dose-response meta-analysis of prospective studies.

Cited 198 times in the scientific literature.

Level 3 - non-randomized controlled study

Meta-analysis of non-randomized prospective cohort studies

PubMed 21920945 · doi:10.1093/aje/kwr226 · record verified 2026-08-28

What was done

A dose-response meta-analysis of prospective studies examined the association between coffee consumption and stroke risk. PubMed and Embase (January 1966 to May 2011) and reference lists were searched for prospective studies reporting relative risks across 3 or more intake categories. Study-specific relative risks were pooled using a random-effects model, and spline transformations were used to model non-linear dose-response curves.

What was found

Across 11 prospective studies with 10,003 stroke cases and 479,689 participants, a non-linear relationship was identified (P for nonlinearity = 0.005). Compared with no coffee consumption, stroke relative risks were 0.86 (95% CI: 0.78, 0.94) for 2 cups/day, 0.83 (95% CI: 0.74, 0.92) for 3-4 cups/day, 0.87 (95% CI: 0.77, 0.97) for 6 cups/day, and 0.93 (95% CI: 0.79, 1.08) for 8 cups/day. Heterogeneity between study trends was low (I² = 12% and 20% across spline transformations).

Why it matters

This meta-analysis demonstrates that moderate coffee consumption (2 to 6 cups per day, peaking at 3-4 cups) is associated with a modest reduction in stroke risk, whereas very high intake showed no clear association.

Limits

All included studies were observational cohorts, meaning residual confounding cannot be ruled out. The abstract does not stratify by stroke subtype (ischemic vs. hemorrhagic), coffee preparation type, or caffeine content.

Cited by