Abstract Number: Esoc2026a1678 Structural Sexism and Stroke Incidence Among Black and White Women and Men: Evidence From the Reasons for Geographic and Racial Differences in Stroke (Regards) Study

Abstract Background and aims Individual-level social factors are disproportionately associated with stroke for women, but the association between structural sexism (SS) and stroke is unclear. Methods Individuals from the nationwide REGARDS study without stroke were included (n=27,881) during 2003-2007 and followed for stroke or death until September 30, 2022. Using data listed in Table 1, SS was calculated as a sum score of inequalities between women and men across multiple social determinants of health at the county or state level and linked to REGARDS participants. Associations between SS and its subdomains with stroke for each sex, accounting for the competing risk of death and clustering of individuals within counties, were assessed by sex-stratified Fine-Gray models with robust sandwich estimators. Models were adjusted for covariates listed in Table 2. An interaction between SS and sex was added to an unstratified model. Results Over a median 12.5-year follow-up, 1,793 individuals (mean age=64.6 years, 55.5% women) had incident stroke. SS was associated with increased stroke rates (HR=1.05, 95% CI 0.97, 1.13 per one standard deviation increase in SS) among women and decreased stroke rates (HR=0.92, 95% CI 0.85, 0.99) among men (interaction p-value=0.01). For subdomains of SS (Table 2), socioeconomic inequality was associated with increased stroke rates (HR=1.15, 95% CI 1.04, 1.28) among women, and conservative religion was associated with decreased stroke rates (HR=0.86, 95% CI 0.76, 0.97) among men. Conclusions Findings suggest the potentially harmful impact of SS on stroke risk among women. Future studies should aim to unravel the underlying mechanisms of this association. Conflict of interest Chen Chen. nothing to disclose. Virginia Howard. nothing to disclose. Natalie Colabianchi. nothing to disclose. Debora Kamin Mukaz. nothing to disclose. Lynda Lisabeth. nothing to disclose. Table 1 - belongs to Methods Table 2 - belongs to Results