Performance of race-neutral eGFR equations in patients with decompensated cirrhosis

The 2021 Chronic Kidney Disease Epidemiology Collaboration equation [CKD-EPI 2021] is a race-neutral equation recently developed and rapidly implemented as a reference standard to estimate glomerular filtration rate(GFR). However, its role in cirrhosis has not been examined especially in low GFR. We analyzed the performance of CKD-EPI 2021 compared to other equations with protocol-measured GFR (mGFR) in cirrhosis. We analyzed 2090 unique adult patients with cirrhosis undergoing protocol GFR measurements using iothalamate clearance from 1985 to 2015 when listed for liver transplantation at Baylor University in Dallas and Fort Worth, Texas. Using mGFR as a reference standard, the CKD-EPI 2021 was compared to CKD-EPI 2012, Modification of Diet in Renal Disease-4, Modification of Diet in Renal Disease-6, Royal Free Hospital, and GFR Assessment in Liver disease overall and in certain subgroups (ascites, mGFR ≤ 30 mL/min/1.73 m2, diagnosis, Model for End-Stage Liver Disease and gender). We examined bias (difference between eGFR and mGFR), accuracy (p30: eGFR within ± 30% of mGFR) and agreement between eGFR and mGFR categories. CKD-EPI 2021 had the second lowest bias across the entire range of GFR after GFR Assessment in Liver disease (6.6 vs. 4.6 mL/min/1.73 m2, respectively, p < 0.001). The accuracy of CKD-EPI 2021 was similar to CKD-EPI 2012 (p30 = 67.8% vs. 67.9%, respectively) which was higher than the other equations (p < 0.001). It had a similar performance in patients with ascites, by diagnoses, Model for End-Stage Liver Disease subgroups, by gender, and in non-Black patients. However, it had a relatively higher overestimation in mGFR ≤ 30 mL/min/1.73 m2 than most equations (18.5 mL/min/1.73m2, p < 0.001). Specifically, 64% of patients with mGFR ≤ 30 mL/min/1.73m2 were incorrectly classified as a less severe CKD stage by CKD-EPI 2021. In Blacks, CKD-EPI 2021 underestimated eGFR by 17.9 mL/min/1.73 m2, which was higher than the alternate equations except for Royal Free Hospital (p < 0.001). The novel race-neutral eGFR equation, CKD-EPI 2021, improves the GFR estimation overall but may not accurately capture true kidney function in cirrhosis, specifically at low GFR. There is an urgent need for a race-neutral equation in liver disease reflecting the complexity of kidney function physiology unique to cirrhosis, given implications for organ allocation and dual organ transplant.