Evaluation of a predictive model of end-stage kidney disease in a French-based cohort

Abstract

Background: The risk of ESKD is highly heterogeneous among renal diseases, and risk scores were developed to account for multiple progression factors. Kidney failure risk equation (KFRE) is the most widely accepted, although external validation is scarce. The objective of this study was to evaluate the usefulness of this score in a French case-control cohort and test the pertinence of the proposed thresholds.

Methods: A retrospective case-control study comparing a group of patients starting renal replacement therapy (RRT) to a group of patients with CKD stages 3-5. Multivariate analysis to assess the predictors of ESKD risk. Discrimination of 4-, 6- and 8-variable scores using ROC curves and compared with eGFR alone and albumin/creatinine ratio (ACR) alone.

Results: 314 patients with a ratio of 1 case for 1 control. In multivariate analysis, increasing age and higher eGFR were associated with a lower risk of ESKD (OR 0.62, 95% CI 0.48-0.79; and OR 0.72, 95% CI 0.59-0.86, respectively). The log-transformed ACR was associated with a higher risk of ESKD (OR 1.25 per log unit, 95% CI 1.02-1.55). The 4-variable score was significantly higher in the RRT group than in the CKD-ND group, and was more efficient than the eGFR (AUROC 0.66, 95% CI 0.60-0.72, p = 0.018) and the log-transformed ACR (AUROC 0.63 95% CI 0.60-0.72, p = 0.0087) to predict ESKD. The 6-variable score including BP metrics and diabetes was not more discriminant as the 4-variable score. The 8-variable score had similar performance compared with the 4-score (AUROC 8-variable score: 0.70, 95% CI 0.64-0.76, p = 0.526). A 40% and 20% score thresholds were not superior to eGFR < 15 and 20 mL/min/1.73 m², respectively. A 10% threshold was more specific than an eGFR < 30 mL/min/1.73 m².

Conclusion: KFRE was highly discriminant between patients progressing to ESKD vs those non-progressing. The 4-variable score may help stratify renal risk and referral in the numerous patients with stage 3 CKD. Conversely, the proposed thresholds for creating vascular access or preemptive transplantation were not superior to eGFR alone.

Keywords: Chronic kidney disease; End-stage kidney disease; Prediction model; Progression; Risk score.