BACKGROUND:
Epidemiological studies of stroke outcomes frequently assume proportional and age-invariant effects of risk factors. However, such assumptions may obscure meaningful age-related heterogeneity in long-term mortality risk.

METHODS:
We applied a dual age-stratification framework to a cohort of first-ever ischemic stroke patients enrolled between 2003 and 2024. Mortality associations were examined using Cox proportional hazards models stratified by baseline age and Poisson regression models stratified by attained age during follow-up, enabling complementary assessment of age at exposure and age at risk.

RESULTS:
Several established risk factors, including hypertension and hyperlipidemia, exhibited stable associations across age strata. In contrast, diabetes, hematological indices, inflammatory markers, and health-related behaviors showed strong age-dependent effects. Notably, these factors were associated with substantially higher mortality risks at younger ages, with effect sizes diminishing at older attained ages.

CONCLUSIONS:
Our findings demonstrate that post-stroke mortality risk is highly age-dependent and sensitive to the analytic definition of age. Incorporating life-course–informed and attained-age approaches can reveal clinically and epidemiologically relevant heterogeneity that is masked in conventional models. These results support the routine integration of age-sensitive modeling strategies in stroke outcome research and broader NCD epidemiology.