LAUSR

C linical laboratories in the United States and world wide have adopted the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equations for estimating glomerular filtration rate (GFR) in the clinical management of patients with CKD. Among their uses, the CKD-EPI equations are a clinical decision tool for preemptive placement on the kidney transplant waitlist before the initiation of maintenance dialysis, medication dosing, and referral for dialysis or a kidney transplant. These equations were developed and rigorously validated using isotopic GFR measurements as the comparator standard. In the developmental statistical modeling, the inclusion of a term for “Black race” improved model accuracy by 14% to 17% compared with models that did not include an adjustment for this variable (1). Recently, controversy has developed about the inclusion of race in the CKD-EPI equations. Those challenging its inclusion cite race as a social construct rather than a biological factor and assert that overestimation of estimated GFR (eGFR) by the CKD-EPI equations that include race harms Black patients with CKD (2, 3). In September 2021, the CKD-EPI published the CKDEPI equation refit without a race term (4). Simultaneously, the American Society of Nephrology and National Kidney Foundation jointly recommended that all laboratories immediately implement the CKD-EPI creatinine equation refit without the race variable (5). The CKD-EPI reported that the CKD-EPI creatinine equation with race adjustment overestimated measured GFR in Black persons by a median of 3.7 mL/min/1.73 m (95% CI, 1.8 to 5.4 mL/min/1.73 m), whereas the new CKD-EPI creatinine refit equation underestimated measured GFR in Black persons by a median of 3.6 mL/min/1.73 m (CI, 1.8 to 5.5 mL/min/1.73m) (4). In effect, the equation refit substitutes imprecision in one direction (overestimation of measured eGFR by 3.7 mL/min/1.73 m) with a nearidentical bias in the opposite direction (underestimation of measured eGFR by 3.6mL/min/1.73 m). Bundy and colleagues examined the prediction performance of 5 CKD-EPI eGFR equations (including the new equation refit without race and the older equations with race) and compared their discrimination and calibration properties with those of the Kidney Failure Risk Equation (KFRE) score in predicting 2-year risk for endstage kidney disease (ESKD) (6). The KFRE score was calculated using 4 variables (age, sex, eGFR, and estimated urinary albumin–creatinine ratio), thus utilizing more information in the prediction of 2-year ESKD risk than eGFR alone. Prediction performance between the CKD-EPI equations and KFRE score was compared in a cohort of 3786 participants aged 21 to 74 years enrolled in the CRIC (Chronic Renal Insufficiency Cohort) Study between 2003 and 2008. Bundy and colleagues' study (6) concluded that in both Black and non-Black participants, a KFRE score greater than 20% has the same specificity for predicting 2-year risk for ESKD (that is, 95%) as an eGFR cutoff of less than 20 mL/min/1.73 m (in any of the 5 CKD-EPI equations). However, in both Black and non-Black participants, a KFRE score greater than 20% also has at least a 10% greater sensitivity for predicting 2-year risk for ESKD than eGFR alone. Of note, Bundy and colleagues state that eGFR alone from any of the CKD-EPI equations “is an excellent predictor of ESKD” (6). Thus, when more information was available for ESKD risk prediction modeling— as in the KFRE score—precision of measurement was not materially significantly affected regardless of whether the CKD-EPI equation (embedded in the KFRE score) included race adjustment. The large amount of effort that went into eliminating race from the GFR estimating equations was a product of tunnel vision, with experts choosing to ignore both history and science. Before the MDRD (Modification of Diet in Renal Disease) eGFR equations were published in 1999, measurement of serum creatinine concentrations and calculation of the Cockcroft–Gault equation for creatinine clearance were the most commonly used tools for estimating kidney function in the clinical setting. The MDRD eGFR equations were based on a study population that was poorly representative of patients in the United States and subsumed by the more robust CKDEPI equations, which were based on more representative patient populations and widely accepted as more accurate clinically useful tools. However, relatively recently some proclaimed these equations as nothing more than racialized medicine and turned a clinically useful tool that benefits patients—including Black patients—into an emblem of systemic racism in clinical medicine. The race adjustment in the CKD-EPI eGFR equation has wrongly and unfairly been ascribed to higher serum creatinine concentrations resulting from larger muscle mass in Black persons, a concept that amounts to a blatantly racist inference. However, a racist connotation was never intended in the development of the CKD-EPI equations. The race adjustment in these equations arose from statistical modeling and was consistent with the average higher measured isotopic GFR in Black persons in Levey and associates' original publication (1). Moreover, the developers of the CKD-EPI equations never contended that larger muscle mass in Black patients was the rationale for race adjustment and have consistently advocated for both research to develop a more precise eGFR estimating equation without a race term and simpler and more clinically useful race-neutral measures of GFR (7). There are potential harms from removing race from the GFR estimating equations without heeding this injunction (8–10).