Kidney Week

Abstract: FR-PO936

Kidney Failure Risk Equation Prediction in a Real-World Population with CKD

Session Information

• CKD Epidemiology, Risk Factors, Prevention - II
  November 03, 2023 | Location: Exhibit Hall, Pennsylvania Convention Center
  Abstract Time: 10:00 AM - 12:00 PM

Category: CKD (Non-Dialysis)

• 2301 CKD (Non-Dialysis): Epidemiology, Risk Factors, and Prevention

Authors

• Shpaner, Leonid, University of California Los Angeles David Geffen School of Medicine, Los Angeles, California, United States

Leonid Shpaner,

• Petousis, Panayiotis, University of California Los Angeles David Geffen School of Medicine, Los Angeles, California, United States

Panayiotis Petousis,

• Duru, Obidiugwu, University of California Los Angeles David Geffen School of Medicine, Los Angeles, California, United States

Obidiugwu Duru,

• Daratha, Kenn B., Providence Medical Research Center, Spokane, Washington, United States

Kenn B. Daratha,

• Norris, Keith C., University of California Los Angeles David Geffen School of Medicine, Los Angeles, California, United States

Keith C. Norris,

• Tuttle, Katherine R., Providence Medical Research Center, Spokane, Washington, United States

Katherine R. Tuttle,

• Nicholas, Susanne B., University of California Los Angeles David Geffen School of Medicine, Los Angeles, California, United States

Susanne B. Nicholas,

• Bui, Alex, University of California Los Angeles David Geffen School of Medicine, Los Angeles, California, United States

Alex Bui,

Group or Team Name

• Center for Kidney Disease Research, Education, and Hope (CURE-CKD).

Background

Risk prediction helps to identify patients with chronic kidney disease (CKD) who may benefit from awareness, detection, and intervention to preserve kidney function. The purpose of this study was to externally validate the Kidney Failure Risk Equation (KFRE) using the 4-variable (4-KFRE) and 6-variable (6-KFRE) equations to predict end stage renal disease (ESRD; eGFR <15 mL/min/1.73 m²) over 2- and 5-year periods in a real-world population with moderate-to-severe CKD.

Methods

Patients with CKD stages 3 and 4 (N=20,947) were selected from electronic health records data in the CURE-CKD Registry from the UCLA (N=3,599) and Providence (N=17,348) Health systems. Patient demographics (age and sex), estimated glomerular filtration rate (eGFR), and log-normalized urine albumin/creatinine ratio were used for the 4-KFRE. The 6-KFRE was applied to patients who also had diabetes and hypertension.

Results

The prevalence of ESRD was 7% for the 2-year period and 11% for the 5-year period. The performance of the 6-KFRE and the 4-KFRE models was similar (4-KFRE is shown in Table 1). Specificity (0.994) reflects the model's ability to predict non-ESRD patients, while sensitivity (0.135) signifies the ability to predict ESRD patients. The area under the precision-recall curve (PR-AUC) was 0.466, and the area under the receiver operating curve (AUC ROC) was 0.853, suggesting accurate KFRE predictions. The model performance improved with lower eGFR.

Conclusion

The KFRE equations depicted high performance when applied to a real-world population with moderate-to-severe CKD. While highly specific for ESRD, it lacked sensitivity, but improved for more advanced CKD stages.

Results from the Four Variable KFRE (UCLA + Providence)

KFRE: Kidney Failure Risk Equation; ESRD: end stage renal disease; PPV: positive predictive value; PR AUC: precision recall area under the receiver operating curve