>
>
Competing Risks Survival Analysis of Time to Dialysis Initiation and Pre-Dialysis Mortality in Chronic Kidney Disease Stage 4-5: A Registry Cohort of 28,400 Patients
Competing Risks Survival Analysis of Time to Dialysis Initiation and Pre-Dialysis Mortality in Chronic Kidney Disease Stage 4-5: A Registry Cohort of 28,400 Patients
Publisher : PJPCR
Author(s)
Yuki A. Shimizu; Conrad P. Muller; Blessing N. Adeyemi
Abstract
This study investigates competing risks analysis of time to dialysis initiation, pre-dialysis death, and kidney transplant in CKD stage 4-5 patients using cause-specific hazards and subdistribution hazard models within the context of nephrology and clinical biostatistics, an area of growing scientific importance given its implications for CKD prognosis tools, shared decision-making for renal replacement therapy planning, and clinical trial design in late CKD. Using Fine-Gray subdistribution hazard model for competing risk endpoints (dialysis, pre-ESRD death, kidney transplant) with cause-specific hazard comparison in 28,400 CKD 4-5 patients, we examine rapid eGFR decline increasing cause-specific hazard for dialysis initiation while comorbidity burden and age increasing competing mortality risk, with subdistribution hazard model appropriately accounting for competing outcomes in real-world prognosis in 28,400 CKD stage 4-5 patients (eGFR 10-29) from multi-state renal registry (2010-2018) with median follow-up 3.8 years drawn from 12-state U.S. renal registry database linked to Medicare claims and ESRD Network data for outcome ascertainment. Results indicate that 5-year cumulative incidence of dialysis is 42.4% by cause-specific analysis versus 38.4% by competing risks (4.0 pp overestimation), with diabetes as strongest dialysis subdistribution hazard predictor (SHR 1.84) and age >75 years predicting pre-dialysis mortality (cause-specific HR 2.84) (p < 0.001), with 4.0 pp dialysis probability overestimation without competing risks; SHR 1.84 for diabetes as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to nephrology and clinical biostatistics and carry actionable implications for the design of programs and policies targeting CKD prognosis tools, shared decision-making for renal replacement therapy planning, and clinical trial design in late CKD.
