Résumé
BACKGROUND AND PURPOSE: Studies have shown higher stroke incidence in areas with higher levels of deprivation. We aimed to determine the pattern of association between various area socioeconomic status (SES) indicators and stroke incidence in specific sex and age groups. METHODS: Data are from the Dijon stroke registry for the period 1995 to 2003. The analyses included 1255 cases aged older than 40 (median age, 76.8). Poisson regression was used to model stroke incidence according to the SES level of 61 small areas. RESULTS: Among women, stroke incidence was higher in neighborhoods with large income inequality (incidence rate ratio, 1.34; P=0.003), higher proportions of unemployed (1.24; P=0.02), of non-French nationals (1.21, P=0.02), and of rented housing (1.31; P=0.03). Areas with a higher proportion of people aged older than 60 were associated with lower stroke incidence (incidence rate ratio, 0.72; P=0.01). Analysis by specific age-groups showed stronger effects among the 60- to 74-year-olds. Among men, no associations between SES and stroke incidence were identified overall but analysis by age groups showed significant effect among the 40- to 59-year-olds. In this age group, incidence rate ratios were 1.47 for unemployment (P=0.01), 1.86 for no car ownership (P=0.02), and 1.56 for income inequality (P=0.01). Among stroke cases, no trend in vascular risk factors prevalence according to area SES was identified. CONCLUSIONS: Variations of stroke incidence were more marked for the SES indicators of wealth and of income inequality. They were apparent at an earlier age in men and showed a stronger gradient among women.