Tijdsgewogen CAVI voorspelt sterfte bij dialysepatiënten

INTRODUCTION: Chronic kidney disease and end-stage renal disease are burdened by extremely high cardiovascular mortality. The Cardio-Ankle Vascular Index (CAVI) is a blood-pressure-independent marker of arterial stiffness, but a single measurement may not reflect the time-dependent nature of vascular remodelling in haemodialysis. We hypothesized that a time-weighted CAVI (twCAVI), integrating arterial stiffness with dialysis vintage, would better predict all-cause mortality. METHODS: In this observational cohort study, 94 maintenance haemodialysis patients were followed from February 2016 to December 2024. Arterial stiffness was assessed by CAVI, and twCAVI was defined as mean CAVI × dialysis vintage (months). After preprocessing and feature selection, the association between arterial stiffness and overall survival over approximately 8 years was examined using a Cox proportional hazards model. Internal validity was assessed by 1,000 bootstrap iterations and expressed via Harrell's concordance index (C-index) and time-dependent area under the curve (AUC). RESULTS: The final model retained six predictors: sex, vascular access, creatinine, alkaline phosphatase (ALP), ferritin, and twCAVI. In the multivariate model, male sex (HR 2.56, p = 0.0018), ALP (HR 1.01, p = 0.0082), ferritin (HR 1.002, p = 0.0067) and twCAVI (HR 1.0021, p = 0.0464) were independently associated with higher mortality, while higher creatinine was protective (HR 0.87, p = 0.0445). Central venous catheter use showed a borderline trend towards increased risk compared with arteriovenous fistula (HR 1.68, p = 0.0509). Bootstrap validation yielded a median C-index of 0.75 (95% confidence interval [CI] 0.68-0.83), with time-dependent AUCs generally ≥0.80 up to 102 months. Notably, exploratory survival analyses revealed an age-dependent inverse association between phosphorus levels and mortality, supporting the reverse epidemiology hypothesis in the elderly. CONCLUSION: twCAVI, capturing the cumulative vascular burden risk, is a predictor of mortality in haemodialysis patients and improves long-term risk discrimination. A proof-of-concept web-based calculator was developed to facilitate the educational exploration of this metric.