Race Adjustment in eGFR Equations Does Not Improve Estimation of Acute Kidney Injury Events in Patients with Cirrhosis

Abstract

Background: Accuracy of glomerular filtration rate estimating (eGFR) equations has significant implications in cirrhosis, potentially guiding simultaneous liver kidney allocation and drug dosing. Most equations adjust for Black race, partially accounted for by reported differences in muscle mass by race. Patients with cirrhosis, however, are prone to sarcopenia which may mitigate such differences. We evaluated the association between baseline eGFR and incident acute kidney injury (AKI) in patients with cirrhosis with and without race adjustment.

Methods: We conducted a retrospective national cohort study of veterans with cirrhosis. Baseline eGFR was calculated using multiple eGFR equations including Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI), both with and without race adjustment. Poisson regression was used to investigate the association between baseline eGFR and incident AKI events per International Club of Ascites criteria.

Results: We identified 72,267 patients with cirrhosis, who were 97.3% male, 57.8% white, and 19.7% Black. Over median follow-up 2.78 years (interquartile range 1.22-5.16), lower baseline eGFR by CKD-EPI was significantly associated with higher rates of AKI in adjusted models. For all equations this association was minimally impacted when race adjustment was removed. For example, removal of race adjustment from CKD-EPI resulted in a 0.1% increase in the association between lower eGFR and higher rate of AKI events per 15 mL/min/1.73 m² change (p < 0.001).

Conclusions: Race adjustment in eGFR equations did not enhance AKI risk estimation in patients with cirrhosis. Further study is warranted to assess the impacts of removing race from eGFR equations on clinical outcomes and policy.

Keywords: Acute kidney injury; End-stage liver disease; Glomerular filtration rate; Race adjustment; Renal function.

Figure 1 –. Number of Acute Kidney Injury (AKI) Events During Follow-up, Stratified by Baseline CKD-EPI Estimated GFR (with Race Adjustment)*

Abbreviations: CKD-EPI = Chronic Kidney Disease Epidemiology Collaboration; eGFR = estimated glomerular filtration rate (mL/min/1.73m²) * The Kruskal-Wallis test was used to test for significant differences in number of AKI events by baseline eGFR. Lower baseline eGFR was significantly associated with higher numbers of AKI events during follow-up (p<0.001). Data are presented for the entire cohort.

Figure 2 –. Association between Baseline eGFR and Expected AKI Events over 10 Years* for eGFR Equations with and without Race-Adjustment (eGFR-NR): (A) CKD-EPI, (B) MDRD-6, (C) MDRD-4; and (D) Cumulative Difference in Events between eGFR-NR and eGFR across the Baseline eGFR Spectrum

Abbreviations: eGFR = estimated glomerular filtration rate (mL/min/1.73m²); CKD-EPI = Chronic Kidney Disease-Epi; NR = no race term (i.e., Black race coefficient removed from the equation); MDRD = Modification of Diet in Renal Disease * Each model is adjusted for age, sex, race (in the case of the all patients models), baseline model for end-stage liver disease (MELD) score, diabetes mellitus, hypertension, etiology of liver disease, and history of decompensated cirrhosis. Data shown are for the entire cohort. Caption: shaded regions correspond to 95% confidence bands. In the entire cohort, the association between GFR estimating equations and estimated AKI incidence was mildly strengthened by removing race adjustment for CKD-EPI, and mildly weakened for MDRD-4. This is more apparent in panel D, where cumulative difference in AKI events was calculated by integrating the difference in eGFR and eGFR-NR curves for each of panels A, B, and C, respectively. Removing race adjustment from CKD-EPI would predict an additional ~1.5 AKI events across the eGFR spectrum, whereas it would predict ~2 fewer events for MDRD-4.

Figure 3 –. Cox-Adjusted† Survival Curves for the Combined Renal Endpoint,* Stratified by CKD-EPI Baseline eGFR (A) with Race Adjustment and (B) without Race Adjustment

Abbreviations: CKD-EPI = Chronic Kidney Disease-Epi; eGFR = estimated glomerular filtration rate (mL/min/1.73m²) * Combined renal endpoint defined by stage 5 chronic kidney disease, dialysis status, or kidney transplant (see Supplemental Table 2) † Model is adjusted for age, sex, race (in the case of the all patients models), baseline model for end-stage liver disease (MELD) score, diabetes mellitus, hypertension, etiology of liver disease, and history of decompensated cirrhosis.