ABSTRACT

Stroke is a major public health challenge worldwide; yet, the impact of habitual coffee and caffeine consumption on stroke risk remains unclear, with conflicting evidence suggesting both protective and harmful effects. In this study, genetic variants linked to coffee and caffeine consumption were identified from prior genome-wide meta-analyses and used as instrumental variables. Summary-level data for stroke, including ischemic and hemorrhagic subtypes, were obtained from genome-wide association meta-analyses involving 1,913,565, 1,020,314, and 567,056 participants, respectively. Bidirectional 2-sample Mendelian randomization analyses were performed to assess the causal relationships between coffee/caffeine intake and stroke. Novel genetic loci, key genes, and pathways identified in our genome-wide association studies meta-analysis validated the reliability of genome-wide association studies summary statistics as instrumental variables. Forward Mendelian randomization analyses revealed that genetically-predicted coffee and caffeine consumption was associated with a reduced risk of stroke, with pooled odds ratios for stroke-related traits of 0.927 (95% CI, 0.877-0.979; P = 0.007), 0.898 (95% CI, 0.794-1.015; P = 0.085), and 0.954 (95% CI, 0.646-1.408; P = 0.812) for coffee consumption, and 0.831 (95% CI, 0.711-0.972; P = 0.0202), 0.897 (95% CI, 0.799-1.007; P = 0.0656), and 0.924 (95% CI, 0.834-1.023; P = 0.1275) for caffeine consumption. Reverse Mendelian randomization analyses found no evidence of a causal effect of stroke on coffee or caffeine consumption, and no significant heterogeneity or pleiotropy was detected. These findings suggest a potential protective role of coffee and caffeine against stroke and highlight the importance of integrating dietary habits and genetic determinants into future stroke prevention strategies.

‍