# Performance of the 2021 Estimated Glomerular Filtration Rate CKD-EPI Refit and the European Kidney Function Consortium (EKFC) Formulas

## TL;DR

This study compares kidney function estimation formulas in Black and White populations in the U.S., finding that both new formulas underestimate kidney function at lower levels.

## Contribution

The study validates the EKFC and REFIT formulas in U.S. populations, highlighting racial differences in GFR estimation accuracy.

## Key findings

- REFIT equation underestimates measured GFR in Black individuals at eGFR < 80 and in White individuals at eGFR > 20.
- EKFC underestimates measured GFR at eGFR > 20 in both Black and White individuals.
- REFIT had the least absolute median bias compared to EKFC and CKD-EPI in both racial groups.

## Abstract

Background: The glomerular filtration rate (GFR) is a universal clinical measure central to assessing kidney function and to the management of kidney disorders. Several formulas for the estimation of GFR are in use. The European Kidney Function Consortium (EKFC) formula has been reported to more accurately estimate the GFR as compared to the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula and its recent version (REFIT equation) in European and African populations. However, validation of the EKFC equation in diverse U.S. populations, especially the Black subpopulation, is needed. Methods: Data from the electronic medical records of 75,442 individuals presenting to a large safety net county hospital with measurements of serum creatinine and/or iohexol clearance studies were used to calculate the estimated GFR (eGFR) and to determine CKD stage using the various reported eGFR formulas. The correlation between eGFR and measured GFR was determined for each equation. Results: The median eGFR for Black participants using the CKD-EPI, REFIT, and EKFC formulas was 130.6 mL/min/1.73 m2, 82.0 mL/min/1.73 m2, and 80.6 mL/min/1.73 m2 (p < 0.001), respectively. For White participants, the median eGFR using the CKD-EPI, REFIT, and EKFC formulas was 145.3 mL/min/1.73 m2, 105.6 mL/min/1.73 m2, and 99.2 mL/min/1.73 m2, respectively (p < 0.001). The REFIT equation underestimates the mGFR in Black individuals at eGFR < 80 mL/min per 1.73 m2 and in White individuals at eGFR > 20 mL/min per 1.73 m2. In comparison, the EKFC equation underestimates the mGFR at eGFR > 20 mL/min per 1.73 m2 in both Black and White individuals. The REFIT equation had the least absolute median bias as compared to EKFC and CKD-EPI in both Black and White participants (p < 0.0001). The P30 of the REFIT and EKFC equations was not statistically different for either Black or White participants (p = 0.16, p = 0.37). Conclusions: Although the accuracies (P30) of the EKFC and REFIT equations are not statistically significant (p = 0.16 and 0.37, Black and White individuals, respectively), adopting the EKFC formula in Americans requires the evaluation of each subpopulation. Both the EKFC and REFIT formulas underestimate the mGFR at a lower eGFR, which may have a direct impact on CKD classification for Black and White patients, with potentially significant implications for clinical management.