Clinical Impact of the Refit CKD-EPI 2021 Creatinine-Based eGFR Equation

Background The National Kidney Foundation recently endorsed the refit Chronic Kidney Disease Collaboration (CKD-EPI) equation for estimated glomerular filtration rate (eGFR) using creatinine, age and sex [2021 eGFR(Cr)(AS)] without a coefficient for race. We evaluated the impact of adopting the 2021 eGFR(Cr)(AS) equation or a variation of the 2009 CKD-EPI eGFR equation without race [2009 CKD-EPI eGFR(Cr)(ASR-NB)] compared to the original CKD-EPI eGFR [2009 eGFR(Cr)(ASR)]. Methods The studied population included patients with a clinically ordered iothalamate clearance (n = 33 889). Bias was assessed as the difference between measured and estimated GFR, P30 was defined as the percentage of estimates within 30% of measured GFR, and concordance was determined according to relevant clinical thresholds. Results Among Black patients, the median bias for 2009 eGFR(Cr)(ASR), 2009 eGFR(Cr)(ASR-NB), and 2021 eGFR(Cr)(AS) was -1.32 mL min(-1) (1.73 m(2))(-1) (95CI -2.46 to -0.26), -8.81 mL min(-1) (1.73 m(2))(-1) (95CI -9.93 to -7.58), and -6.08 mL min(-1) (1.73 m(2))(-1) (95CI -7.18 to -4.92), respectively. The median bias among non-Black patients was -0.15 m min(-1) (1.73 m(2))(-1) (95CI -0.84 to -0.08) for 2021 eGFR(cr)(AS) compared to -3.09 mL min(-1) (1.73 m(2))(-1) (95CI -3.17 to -3.03) for the 2009 eGFR(Cr)(ASR). P30 and concordance were not significantly different in either racial group. The net reclassification improvement at a measured GFR <20 mL min(-1) (1.73 m(2))(-1) was 6.4% (95CI 0.36 to 12.4) for Black patients and -5.1% (95CI -6.0 to -4.1) for non-Black patients using the 2021 eGFR(Cr)(AS) equation. Conclusions Overall, the change in reported eGFR was minimal. However, these changes led to significant reclassification improvements at lower eGFR, which will indirectly improve equitable access to CKD resources.

Automated summary

This study evaluated the impact of adopting a new eGFR equation without race on CKD patients. The results showed that the new equation had different effects on bias among Black and non-Black patients, but there were no significant differences in P30 and concordance between the two racial groups.