Kidney Medicine · Aug 25, 2026

Frailty as a Predictor of Mortality in Dialysis: A Multicenter Validation of the Clinical Frailty Scale

By Ingwiller M, Chantrel F, Krummel T et al.

RATIONALE & OBJECTIVE: Frailty is highly prevalent in dialysis and is associated with poor outcomes, yet scalable assessment in routine care remains challenging. We evaluated the feasibility and prognostic performance of the clinical frailty scale (CFS) applied by dialysis nurses and compared it with the REIN predictive score (RPS), a comorbidity-based mortality tool routinely available in French dialysis care. STUDY DESIGN: A prospective multicenter observational cohort. SETTING & PARTICIPANTS: Adults receiving maintenance dialysis at 6 centers in northeastern France (n = 792 prevalent patients). PREDICTORS: Nurse-rated 9-point CFS (analyzed both continuously and in 3 categories: 1-3 well, 4-5 vulnerable, and 6-9 frail) and RPS calculated at the time of CFS assessment. OUTCOMES: All-cause mortality at 12 and 24 months; kidney transplant treated as a competing event. ANALYTICAL APPROACH: Frailty ratings were obtained by trained dialysis nurses; interrater reliability was assessed in a subset with duplicate ratings. Cumulative incidence functions for death and transplantation were estimated using competing-risk methods. Associations with mortality were evaluated using Fine-Gray subdistribution hazard models adjusted for age, sex, diabetes, and central venous catheter use, plus CFS and RPS. Discrimination was assessed with C-statistics at 12 and 24 months. RESULTS: Two independent CFS ratings were available for 509 patients; agreement was moderate when grouped into 3 categories (κ = 0.60). Twelve-month and 24-month mortality increased stepwise across CFS categories (well: 3.56% and 9.78%; vulnerable: 10.49% and 21.30%; frail: 27.27% and 45.87%). In multivariable models, both higher CFS (sHR per point 1.33; 95% CI, 1.19-1.49) and higher RPS (sHR per point 1.05; 95% CI, 1.02-1.09) were independently associated with mortality. Discrimination was good: at 12 months, C-statistics were 0.73 (95% CI, 0.68-0.79) for CFS and 0.69 (95% CI, 0.63-0.75) for RPS; at 24 months, 0.71 (95% CI, 0.66-0.75) and 0.67 (95% CI, 0.62-0.72), respectively. A CFS score ≥6 identified an exploratory high-risk group for 12-month mortality. LIMITATIONS: Prevalent cohort from a single region; nonstandardized timing of CFS assessments; no systematic ethnicity data. CONCLUSIONS: In routine dialysis practice, nurse-applied CFS is feasible, reproducible when grouped into 3 categories, and predicts mortality over 36 months with performance comparable to, and independent of, a comorbidity-based mortality score. These findings support integrating the CFS into longitudinal risk assessment in dialysis, with external validation needed before using specific cut points to guide treatment decisions.

doi.org/10.1016/j.xkme.2026.101510Open accessPubMed 42889815Biomarkers