Case Report: Clinical Consequences of Adjusting Estimated GFR for Black Race

CASE REPORT

A 72-year-old Black woman was seen in the general nephrology clinic at a large academic center. She had a 10-year history of CKD secondary to diabetes mellitus type II and hypertension. Her medical history was also notable for hyperlipidemia, obstructive sleep apnea, and diabetic retinopathy. As of October 2020, her CKD-EPI eGFR adjusted for Black race was 20 mL/min/1.73m², which met the criteria for transplant referral. Twelve months prior to her CKD-EPI eGFR, SCr adjusted for Black race was 23 mL/min/1.73m² (non-Black eGFR = 20). During that time, she had continued consulting with her nephrologist, receiving an eGFR assessment every 2–4 months that showed steady decline in eGFR (Table 1). Due to the reliance on race-adjusted eGFR in CKD-EPI, her referral to the kidney transplant clinic for an initial evaluation was delayed by approximately 12 months. Had she not been identified as Black, this delay would not have occurred. When her nephrologist explained the connection between the timing of the referral and the binary use of race in GFR estimating equations, she was visibly disappointed. She was referred to vascular surgery for placement of arteriovenous fistula for dialysis preparation, partly due to concern that waitlist times would necessitate dialysis as a bridge to a kidney transplant.

Table 1.

Outpatient assessment of eGFR in kidney clinic

Date	eGFR (mL/min per 1.73 m2) adjusted for Black race	eGFR (mL/min per 1.73 m2) not adjusted for Black race
6/19	26	23
8/19	25	22
9/19	27	23
10/19	23	20
2/20	22	19
6/20	21	18
10/20	20	17

DISCUSSION

The current standard for estimating glomerular filtration rate uses the Chronic Kidney Disease Epidiemiology Collaboration (CKD-EPI) formula, which was formed using multiple regression analysis of patient serum creatinine, age, sex, and race, in comparison with directly measured, iothalamate-GFR. ^1,2 Mean measured GFR of this cohort was 68 cc/min/1.73m² (SD 40), 12–14% of subjects were over the age of 65, and separate groups were used for development, internal validation, and external validation of the formula. CKD-EPI adjusts estimated GFR (eGFR) for Black versus non-Black people in binary fashion, assigning Black patients 15.9% higher eGFRs than non-Black patients with the same serum creatinine, age, and sex. ^1,2 Adjusting for Black race in these equations is based both on regression models from the cohort of patients in CKD-EPI, and on the assumption that Black people have consistently higher generation of creatinine than non-Black people due to greater muscle mass. However, race is a social construct that defines differences in appearance, not physiology. The biological admixture of ancestry is more variable than a binary decision of Black or non-Black. If one were to include the broad stroke of race in a clinical decision algorithm, the biological differences are both subtle and more nuanced than a binary decision.

Adjustment of eGFR in Black individuals has serious clinical consequences, as it can delay referral to specialist nephrology care, affect eligibility for clinical studies with inclusions or exclusions that are based on GFR, increase the use of medications that can have toxicity with lower GFR, affect surgical decision-making that is based on GFR, and, as with the patient from our clinic, delay waitlist time for kidney transplantation. Patients with declining kidney function must have an eGFR SCr of 20 mL/min/1.73m² or less before they can start accruing time on the kidney transplant waitlist. ³ Adjustment of eGFR for Black race assigns higher GFR estimations to Black patients, and thus delays their placement onto the kidney transplant waitlist. Many patients with kidney failure who are also on the transplant waitlist are unable to receive their transplant before requiring dialysis. Dialysis is associated with high symptom burden, healthcare utilization, and cardiovascular risk. ^4–6 Accordingly, every extra year on the transplant waitlist prior to starting dialysis means one less year of dialyisis hardship before receipt of a transplant. This case report joins other studies in demonstrating that Black individuals have been disadvantaged for transplantation by a systematic adjustment to GFR for all Black patients. ^7–9

The practice of factoring Black versus non-Black race into estimates of GFR began with the MDRD cohort in 1999 ¹ which reported that age, sex, and race accounted for most of the variation in non-GFR determinants of serum creatinine (MDRD Eq. 1). Their stepwise multiple regression model found that non-Black ethnicity predicted a significantly lower GFR. Therefore, an adjustment of 1.18 was used for GFR calculation in Black participants, who comprised 12% of the cohort in the MDRD study. At the time, it was believed that the adjustment for people identified as Black in the cohort could be extrapolated to all Black people. ¹ However, the studies cited by MDRD to support the assumption that Black people have consistently more skeletal muscle with corresponding greater creatinine generation than non-Black individuals ^10–12 were not generalizable for several reasons. First, Cohn et al. reported Black individuals had higher total-body potassium than sex- and age-matched White individuals, but did not measure muscle mass. Harsha et al. compared body fat content in Black and White volunteer children from the socioeconomically disadvantaged city of Bogalusa, LA. The cohort did not include adults, and nutritional intake of children was not measured. Lastly, Worrall et al. reported that racial variation in creatine kinase was independent of lean body mass, but their cohort was small and not from the USA (30 Afro-Carribbeans and 30 White British hospital workers).

$$\text{eGFR}\left(\text{mL}/\text{min}/\mathrm{1.73}{\text{m}}^{\mathrm{2}}\right)=\mathrm{175}\times {\left({\text{S}}_{\text{cr}}\right)}^{-\mathrm{1.154}}\times {\left(\text{Age}\right)}^{-\mathrm{0.203}}\times \left(\mathrm{0.742}\text{if female}\right)\times \left(\mathrm{1.212}\text{if African American}\right)$$ 1

When CKD-EPI improved upon and supplanted MDRD, the correction factor for Black versus non-Black race remained a dichotomous variable, with a correction factor of 1.159 for eGFR if Black, leading to a 15.9% higher estimate of GFR (CKD-EPI Eq. 2). Compared with MDRD, the CKD-EPI cohort included a higher proportion of people identified as Black in the development and validation cohorts (32% and 31% Black, respectively). ² Designation of Black race in the study was determined either by the patient or by study investigators (the latter, presumably based on appearance); a prescient 2011 editorial recounted how some patients of color asked their physician, “How did they define African American?” ¹³.

$$\text{eGFR}\left(\text{mL}/\text{min}/\mathrm{1.73}{\text{m}}^{\mathrm{2}}\right)=\mathrm{141}\times {\left({\text{S}}_{\text{cr}}/\kappa ,\mathrm{1}\right)}^{\alpha }\times \text{max}{\left({\text{S}}_{\text{cr}}/\kappa ,\mathrm{1}\right)}^{-\mathrm{1.209}}\times {\mathrm{0.993}}^{\text{Age}}\times \mathrm{1.018}\left(\text{if female}\right)\times \mathrm{1.159}\left(\text{if Black}\right)$$ 2

While the mathematics in the CKD-EPI study cannot be disputed, it is possible that the difference in Black versus non-Black creatinine generation was a factor of that particular cohort of patients. If the Black patients who were recruited were more physically active, drove less, walked more, or were of a lower socio-econonic scale that required more manual labor, then creatinine generation may have been higher due to lifestyle, not Black race. If muscle mass and creatinine generation were an African-ancestral predisposition, then Black patients from Africa (with less population admixture) would also require the Black correction factor for accurate use of CKD-EPI, but this is not the case. In South Africa, the Black correction factor of 1.159 increased the bias compared to using CKD-EPI without adjustment for race (bias of − 12.4 mL/min vs − 4.9 mL/min, respectively) using iohexol clearance as the gold standard for GFR measurement. ¹⁴ A similar analysis in the Congo found that CKD-EPI with the Black adjustment overestimated GFR with a bias of 17.9 mL/min/1.73m² (95% CI 13.3; 21.2), compared to a bias of 2.3 mL/min/1.73m² (95% CI − 1.3; 5.8) without use of the Black adjustment using iothalamate as the gold standard for GFR measurement. ¹⁵

The validity of using race to estimate GFR has been challenged. ^16–18 Key professional societies such as the American Association for Clinical Chemistry (AACC), the American Society of Nephrology (ASN), and the National Kidney Foundation (NKF) are assessing the use of race in diagnosing kidney disease ¹⁹. The NKF-ASN joint task force has stated that race modifiers should not be included in equations to estimate kidney function, and that current race-based equations should be replaced by a suitable approach that is accurate, inclusive, and standardized in every laboratory in the USA. ²⁰ The AACC has created a committee to examine the clinical impact and utility of the Black race correction factor and any follow-up plans AACC should consider. ⁸ Several clinical laboratories at academic hospitals including the University of Washington, Beth Israel Deaconess, San Francisco General, and Vanderbilt University have excluded an adjustment for being Black from estimated GFR. At the time of this submission, clinical laboratories at the University of Pittsburgh Medical Center are also in the process of removing the adjustment for Black race from GFR estimation.

Because human race taxonomy is not a reliable substitute for genetic differences, the use of racial taxons in medical practice and research is a source of debate in the scientific community. ^7,21 In this case, the use of Black race as a binary variable in estimating GFR does not account for those who identify as multi- or biracial and does not account for population admixture. ²² Moreover, the use of Black race as a surrogate for increased muscle mass is not based on solid scientific data and may have originated from historical bias. While there is fear that removing race from the estimation of GFR will lead to overdiagnosis and overtreatment of Black patients, ²³ the current use of the race adjustment also has the potential to divert resources and care away from Black patients and perpetuate disparities in kidney care. ²⁴

The practical effects of using race in the GFR estimation are beginning to be assessed clinically. Analysis of the Partners in Health Care Registry found 2,225 Black patients with CKD, of whom 64 had an eGFR of > 20 mL/min/1.73m² with the adjustment and < 20 mL/min/1.73m² without the adjustment for Black race. Of these 64 patients who might have been accruing waitlist time without the Black adjustment, none had a transplant referral or was on the transplant waitlist. ²⁵ By overestimating the eGFR, our patient had a similar delay in transplant referral. Overestimating eGFR in Black patients also risks using medications with toxicity at reduced kidney function, may affect entry into experimental trials that have GFR entry and exclusion criteria, affects clinical decision-making for surgical decisions such as the choice of radical versus partial nephrectomy for localized renal cancer, ²⁶ and could delay referral for fistula or graft placement, which risks having Black patients abruptly start emergent dialysis in the hospital without a mature access.

Efforts are being made in the scientific community to determine alternatives to race adjustment in CKD-EPI calculation. ²⁷ The first step is for clinicians to keep emphatically clarifying that estimates of GFR are only estimates. We know that malnourishment and decreased muscle mass overestimate eGFR, while patients with higher muscle mass have underestimated eGFR. ²⁸ Cystatin-C, a protein produced at a constant rate by nucleated cells and almost entirely eliminated in the proximal renal tubule, has shown promise in estimating GFR with minimal bias and may be a more accurate test than serum creatinine. ^29,30 Cystatin-C testing is commercially available on in vitro diagnostic analyzers.

In conclusion, we presented a patient who not only was disadvantaged by the Black race correction factor, but also became visibly disappointed when learning of the ramifications of the correction factor on her transplant status. Such revelations can affect patient trust in the medical system. In the occasional clinical situation where estimating GFR is insufficient, either because an individual has extremes of muscle mass (either high or low) or because a precise GFR is needed for clinical decision-making, we recommend not using the Black race correction factor to estimate GFR. Instead, we support the use of Cystatin-C eGFR without race correction, or direction measurement of GFR by iohexol or iothalamate.

Declarations

Conflict of Interest

Dr. Johnstone, Dr. Bansal, and Dr (can) Skiba report no conflicts of interest. Dr. Palmer has received funding from the NIH, Roche Diagnostics and Beckman Coulter, and consulted with and received an honorarium from Siemens Healthineers and Beckman Coulter.