Removal of Race Adjustment Improved Bias, Accuracy in eGFR Equations for Black Adults

A systematic review found that alternative approaches to calculating estimated glomerular filtration rate (eGFR) were worth looking into to properly account for diversity in populations.

A review published in PLoS One found that removal of race adjustments in measurement of estimated glomerular filtration rate (eGFR) improved bias, accuracy, and precision in eGFR equations, prompting the researchers to conclude that alternative methods of calculating eGFR were worth looking into.

Accurate diagnosis of chronic kidney disease (CKD) is important for the management of the disease and relies on proper assessments of kidney function. eGFR, which measures serum filtration markers to assess kidney function, is used most often for this purpose. An increasing focus on health equity and the lack of evidence supporting race adjustment has led to the omission of race from severa; eGFR equations, and the National Kidney Foundation and the American Society of Nephrology now recommend use of an equation without the race variable. This review aimed to determine how accurate eGFR is with and without race adjustment.

The researchers used PubMed, Embase, ScienceDirect, Web of Science, and Google Scholar to find applicable studies. Studies were included if they had validation and/or comparison studies that included Black adults 18 years and older; compared an eGFR formula with measured GFR (mGFR) or with another eGFR formula; and were published in English, French, Spanish, or Portuguese. A study was excluded if it did not use mGFR as a reference test, did not use either a Modification of Diet in Renal Disease (MDRD) or Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) eGFR formula, did not use isotope dilution mass spectrometry traceable creatinine assays, or did not report outcomes by Black race.

There were 12 studies that evaluated eGFR equations with and without race adjustment vs mGFR and were included in this study; 7 were cross-sectional, 3 were retrospective, and 1 used pooled data from previous studies. There were 3 studies that focused on participants with renal dysfunction and 5 that included a mixed population of patients with and without kidney disease.

All included studies that evaluated CKD-EPI used the 2009 creatinine-based equation (CKD-EPICr). There were 3 studies from the United States, 6 studies from African countries, 1 study from France, 1 study from the United Kingdom, and 1 study from Brazil.

Removal of race adjustment was found to improve bias in 10 of the 11 studies that evaluated the bias of the CKD-EPICr equations. A study based in the United States that included patients with and without CKD found that bias was worse after the removal of race adjustment. Bias improved in the 7 studies that evaluated the bias of the 4-variable MDRD equation.

Removal of the race coefficient was found to improve accuracy in all 8 studies that evaluated accuracy in the CKD-EPICr equation. Removal of the race coefficient in the MDRD equation also improved accuracy in the 6 studies that assessed it.

In the 6 studies evaluating the precision of CKD-EPICr with and without race adjustment, 4 of them found an improvement in precision. The 5 studies that evaluated the precision of the 4-variable MDRD equation also found improvement in precision when race adjustment was removed.

The assessment of correlation between CKD-EPICr and mGFR was mixed in the 6 studies reporting it. Increasing GFR values for CKD-EPICr saw the difference between eGFR and mGFR decrease with and without race adjustment. There was an increased correlation between eGFR and mGFR in a large cohort from the Democratic Republic of Congo and the Ivory Coast when the race coefficient was removed in 1 study. The 3 studies that assessed the correlation between MDRD and mGFR found an increased correlation when eGFR race adjustment was removed in 2 studies.

There were some limitations to this review. The limitations in eGFR from variations in non-GFR determinants could be relevant. eGFRCr likely overestimates kidney function in sicker patient populations, which can bring kidney function closer to an expected value after removal of the race coefficient. The studies included had suboptimal accuracy measures, which indicates the need for more studies on the topic.

The researchers concluded that this review “provides the highest level of evidence against race adjustment when estimating GFR. Race is an inappropriate proxy for genetics all over the world, and efforts to eliminate its use in estimating GFR should be global.”


Umeukeje EM, Koonce TY, Kusnoor SV, et al. Systematic review of international studies evaluating MDRD and CKD-EPI estimated glomerular filtration rate (eGFR) equations in Black adults. Plos One. 2022;17(10):e0276252. doi:10.1371/journal.pone.0276252

Related Videos
Kirsten Johansen, MD.
Jennifer Green, MD.
Related Content
CH LogoCenter for Biosimilars Logo