Background: As the population ages, factors affecting prognosis after dialysis initiation in very elderly patients remain unclear. We aimed to develop a prognostic score to predict short-term survival in this group.
Methods: This retrospective cohort study included 116 patients aged 85 or older who initiated hemodialysis at our hospital between 2010 and 2023, out of a total of 2279 patients. Using a Cox regression model, we analyzed factors affecting 3-year all-cause mortality. Based on these analyses, a prognostic scoring system was developed to predict patient outcomes.
Results: Baseline factors independently associated with mortality (hazard ratio (HR), 95% confidence interval (CI)) included body mass index (BMI) < 18.5 (2.14, 1.17-3.91), cardiovascular disease (2.16, 1.10-4.22), advanced cancer (4.65, 1.73-12.5), impaired walking (2.33, 1.17-4.64), emergency dialysis without fistula (1.88, 1.03-3.44), fluid overload (2.70, 1.43-5.07), serum albumin < 3.0 g/dL (2.32, 1.25-4.32), and C reactive protein (CRP) > 0.5 mg/dL (3.87, 1.85-8.10). A scoring system was developed, assigning 1 point each for BMI < 18.5, cardiovascular disease, impaired ambulation, emergency dialysis without fistula, fluid overload, and serum albumin < 3.0 g/dL, and 2 points for advanced cancer and CRP > 0.5 mg/dL. The score demonstrated an area under the ROC curve of 0.83. The hazard ratios for 3-year mortality for scores of 4-5 and 6 or higher, compared to 0-3, were 2.55 (95% CI: 1.04-6.23) and 7.10 (3.36-15.0), respectively (P for trend < 0.001). Subgroup analyses confirmed the score's robustness.
Conclusion: The prognostic score for 3-year mortality in very elderly patients who initiate dialysis should be validated in different populations and is expected to inform decisions regarding dialysis initiation.
Keywords: Hemodialysis; Prognosis; Very elderly.
© 2025. The Author(s) under exclusive licence to Italian Society of Nephrology.