Estimated pulse wave velocity is associated with all-cause and cardiovascular mortality in individuals with stroke A national-based prospective cohort study

Extensive evidence underscores the potential of estimated pulse wave velocity (ePWV) as a robust tool for predicting disease prevalence and mortality. However, its comparative effectiveness in forecasting all-cause and cardiovascular disease (CVD) mortality, particularly among stroke populations, remains inadequately characterized in relation to the traditional Framingham Risk Score (FRS) model. This prospective study included 1202 individuals with stroke from the National Health and Nutrition Examination Survey conducted between 1999 and 2014, with comprehensive follow-up data. Survey-weighted Cox regression models were employed to examine the association between ePWV and the risks of all-cause and CVD mortality. Subgroup analyses were performed to evaluate the stability of ePWV in predicting these outcomes. A generalized additive model was utilized to explore the dose-response relationship between ePWV and mortality risk. Receiver operating characteristic curves were then used to assess and compare the prognostic capabilities of ePWV and FRS models for 10-year all-cause and CVD mortality. After adjustment for relevant covariates, each 1 m/s increase in ePWV was associated with a 44% and 65% increase in all-cause and CVD mortality, respectively. ePWV demonstrated consistent prognostic performance across the majority of stroke subpopulations. Notably, ePWV exhibited a nonlinear relationship with all-cause mortality (P for nonlinearity =.045) while maintaining a linear association with CVD mortality (P for nonlinearity =.293). Furthermore, ePWV outperformed the FRS model in predicting 10-year all-cause (Integrated Discrimination Improvement = 0.061, 95% confidence interval: 0.031-0.095, P =.007) and CVD mortality (95% confidence interval: 0.005-0.083, P =.02). ePWV is an independent risk factor for both all-cause and CVD mortality in individuals with stroke, demonstrating superior predictive value compared to the traditional FRS model for forecasting these outcomes.