Objectives: Risk of death is high for hemodialysis (HD) patients but it varies considerably among individuals. There is few clinical tool to predict long-term survival rates for HD patients yet. The aim of this study was to develop and validate a easy-to-use nomogram for prediction of 1-, 5-, and 10-year survival among HD patients.
Methods: This study retrospectively enrolled 643 adult HD patients who was randomly assigned to two cohorts: the training cohort (n = 438) and validation cohort (n = 205), univariate survival analyses were performed using Kaplan-Meier's curve with log-rank test and multivariate Cox regression analyses were performed to identify predictive factors, and a easy-to-use nomogram was established. The performance was assessed using the area under the curve (AUC), calibration plots, and decision curve analysis.
Results: The score included seven commonly available predictors: age, diabetes, use of arteriovenous fistula (AVF), history of emergency temporary dialysis catheter placement, cardiovascular disease (CVD), hemoglobin (Hgl), and no caregiver. The score revealed good discrimination in the training and validation cohort (AUC 0.779 and 0.758, respectively) and the calibration plots showed well calibration, indicating suitable performance of the nomogram model. Decision curve analysis showed that the nomogram added more net benefit compared with the treat-all strategy or treat-none strategy with a threshold probability of 10% or greater.
Conclusions: This easy-to-use nomogram can accurately predict 1-, 5-, and 10-year survival for HD patients, which could be used in clinical decision-making and clinical care.
Keywords: Hemodialysis; all-cause mortality; nomogram; prediction model; survival.