AbstractAimThe Framingham Risk Score (FRS), calculated by considering conventional risk factors of cardiovascular diseases, was developed to predict coronary heart disease in various populations. However, reverse epidemiology has been raised concerning these risk factors in predicting high cardiovascular mortality in hemodialysis patients. Our objectives are to determine whether FRS is associated with overall and cardiovascular mortality and the role of new risk markers when they were added to a FRS model in hemodialysis patients.MethodsThis study enrolled 201 hemodialysis patients aged 20–80 years old. The FRS is used to identify individuals categorized as low (<6% 10‐year risk), intermediate (6–20% risk) or high risk (>20% risk). Medical records were reviewed to collect clinical information. Data of ankle‐brachial index (ABI) and brachial‐ankle pulse wave velocity (baPWV) were obtained by an ABI‐form device.ResultsThe mean follow‐up period was 4.4 ± 1.5 years. Intermediate risk predicted overall hazard ratio (HR) (2.157, P = 0.039) and cardiovascular mortality (HR= 5.023; P = 0.004) versus low risk, but ‘high’ risk did not. High risk (vs low risk) predicted cardiovascular events (HR = 2.458, P = 0.05). Besides, the addition of ABI < 0.9 (P = 0.021) and baPWV (P = 0.014) to a FRS model significantly improved the predictive value for overall mortality.ConclusionIn hemodialysis patients, intermediate risk but not high risk categorization by FRS predicted overall and cardiovascular mortality, and high risk predicted cardiovascular events. ABI < 0.9 and baPWV provided additional predictive values for overall mortality. Future study is needed to develop hemodialysis‐specific equations and assess whether risk refinement using ABI < 0.9 and baPWV leads to a meaningful change in clinical outcomes.