Mapping the modifiable exposome for ischemic stroke: association, causality, and preventive potential.

BACKGROUND AND AIMS: Previous hypothesis-driven studies focused on limited risk factors for ischemic stroke (IS) with inconsistent findings. We aimed to systematically identify modifiable factors, assess causality, quantify joint impact, and evaluate public health implications of conservative versus radical multi-domain IS intervention strategies. METHODS: We conducted a large prospective cohort study using UK Biobank and a bidirectional two-sample Mendelian randomization study. Exposures included 323 modifiable factors in seven domains: early life, health and medical history, lifestyle, local environment, physical measures, psychosocial factors, and sociodemographics. Exposome-wide association scan was used to investigate the associations between exposures and IS onset, and multivariable Cox models to assess their joint impact. Weighted population attributable fraction was estimated by partially eliminating exposures. RESULTS: 497,856 IS-free UK adults were enrolled during 2006-2010, and 10,593 experienced IS with a median follow-up of 14.53 years. Among 68 modifiable factors significantly associated with IS, 18 demonstrated robust causality. Controlling for the overlapping hazards of exposures, worsening risk profiles in each domain independently increased IS risk. Eradicating all modifiable risk factors would prevent 72.21% (95% CI 69.37%-74.83%) of IS, whereas a one-third reduction in the exposure levels of modifiable factors would prevent 68.19% (95% CI 64.22%-71.81%). These findings remained consistent across populations stratified by age, sex, and polygenic risk score-based genetic risk. CONCLUSIONS: This hypothesis-free and triangulated study provided comprehensive information on the modifiable exposome for IS. Our findings suggested that a moderate reduction in risk factor exposure achieved a preventive impact comparable to total eradication, supporting the potential for efficient public health strategies.