Advertisement

The Use of Race in eGFR: Why Racial Justice Requires Accuracy

      Current best evidence for estimated glomerular filtration rate (eGFR) includes a correction factor for Black patients. The validity of this correction factor has been called into question as another form of racial bias and a number of critics have argued that it should be abolished. One type of argument for this conclusion is exemplified by Norris et al,
      • Norris KC
      • Eneanya ND
      • Boulware LE
      Removal of race from estimates of kidney function: first, do no harm.
      who conclude that the overall benefits of such a change would outweigh the costs. The supposed benefits of this reclassification include more Black patients being eligible for nephrology special care and kidney transplants while costs include more Black patients ineligible for certain medical treatments such as metformin. We argue to the contrary that simply dropping the race correction factor from eGFR estimates will cause foreseeable harm to Black patients by creating a situation where they regularly receive inappropriate medical treatment based on inaccurate and racially biased estimates of GFR.
      Whether a patient's reclassification is a good or a bad thing depends on whether the reclassification is more accurate than the original. A thought experiment demonstrates the importance of accuracy: if we imagined not just removing the correction factor, but instead, further lowering Black patients’ eGFR by an additional 16%, it would lead to even more Black patients being labeled as having chronic kidney disease (CKD) and receiving the “benefits” thereof. But arbitrarily lowering eGFR scores in this way would be grossly misguided.
      Inaccurate classification leads to harm. Imagine a patient with type 2 diabetes and stage 3 chronic kidney disease. Prescribing metformin could be either beneficial or dangerous so it is important that our estimation be based on the best evidence we have. Removing the race correction factor would systematically lower the eGFR score for Black patients and metformin would now be contraindicated more often. Norris et al acknowledge that this would be a negative consequence. However, this is only a negative consequence if metformin would actually be the appropriate treatment. If the critics of race correction were right, eGFR scores for Black patients were inflated and metformin use should have been more often avoided. But we know that this is not the case. Direct measurements indicate that Black individuals have higher mGFRs at every creatinine level
      • Levey AS
      • Stevens LA
      • Schmid CH
      • et al.
      CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration). A new equation to estimate glomerular filtration rate.
      . To ignore this leads to bad estimates and thus to inappropriate treatment. The reasoning in the metformin case applies generally–it is always important that treatment decisions be made based on the most accurate estimates available of a patient's physiologic status.
      While accuracy is clearly important, a different (though related) concern is about equality and fairness. It is morally inappropriate to treat people differently based on their race. But acknowledging racial differences is appropriate when this is necessary to ensure that patients are being treated equally. An instructive case that is parallel in some respects and strikingly different in others is the recent paper by Sjoding et al
      • Sjoding MW
      • et al.
      Racial bias in pulse oximetry measurement.
      titled “Racial Bias in Pulse Oximetry Measurement.” They found that pulse oximetry measurements systematically overestimated arterial oxygen saturation in patients with darker skin colors, potentially leading to crucially inappropriate medical decisions for many Black patients. Said another way, our estimate of blood oxygen levels is often racially biased. This is a racial injustice despite the fact that race appears nowhere in treatment protocols because many Black patients are being treated differently than white patients with the same blood oxygen levels. Proper treatment requires some sort of correction to pulse oximetry measurement. Whatever kind of “racial correction” is done, it is not appropriately thought of as treating people differently because of their race. Rather, it will be a way to ensure that they are being treated equally regardless of their race.
      By analogy, we should demand a similar thing in the case of eGFR. As in the example above with decisions about prescribing metformin, treating Black and non-Black patients with the same kidney function differently would be unacceptable. In fact, this is exactly what used to happen. Prior to 2016, FDA prescription guidelines stated that metformin was contraindicated when serum creatinine levels were above a sex-based threshold with no mention of race. Because Black individuals have higher creatinine levels relative to the same kidney functioning, they were being inappropriately denied access to metformin.
      • Shin J-I
      • Sang Y
      • Chang AR
      • et al.
      The FDA metformin label change and racial and sex disparities in metformin prescription among patients with CKD.
      The labeling was changed so that contraindication is now based on eGFR and eGFR is now calculated using the CKD-EPI equation, which corrects for exactly this bias. Prediction equations and treatment recommendations that are based on studies of mainly white patients are dangerous. But using a large, diverse sample and then ignoring the data that you get from it is no improvement. Removing the racial correction factor is a step backwards.
      Concerns about racial bias and differential treatment based on race are absolutely legitimate. Creatinine level and pulse oximetry are used to estimate true physiology, something crucial to making good medical decisions. This pushes the “racial bias” question back to how best to estimate GFR. If race is being used to improve estimates of GFR (which it is in this case as verified in Levey et al
      • Levey AS
      • Tighiouart H
      • Titan SM
      • et al.
      Estimation of glomerular filtration rate with vs without including patient race.
      ) then the use of race is actually preventing racial bias in an estimation procedure. There are good reasons to look for accurate ways of estimating GFR that do not rely on the race of the patient; research in this area is ongoing.
      • Diao JA
      • Inker LA
      • Levey AS
      • Tighiouart H
      • Powe NR
      • Manrai AK.
      In search of a better equation—performance and equity in estimates of kidney function.
      It is also important to be aware of the highly imprecise nature of eGFR estimates and a more holistic approach to diagnosis and management is often called for.
      • Sehgal AR.
      Race and the false precision of glomerular filtration rate estimates.
      But keeping the rest of the estimation procedure intact and arbitrarily dropping the race correction factor would actually be to introduce a racial bias into treatment by inappropriately relying on racially biased estimates. GFR estimation procedures should regularly be studied and updated as populations change, new technologies and tests are introduced, and more comprehensive and reliable data are available. But any recommendations for best practices in GFR estimation must be based on the accuracy of these estimates, as should all data that guide medical decisions. Racial equality depends on it.

      Acknowledgement

      We thank Howard Curzer, PhD (Texas Tech), Anne C. Epstein, MD, (Texas Tech Health Sciences), Neil Kurtzman, MD (Texas Tech Health Sciences), and Kenneth Nugent, MD (Texas Tech Health Sciences) for valuable discussions on this topic.

      References

        • Norris KC
        • Eneanya ND
        • Boulware LE
        Removal of race from estimates of kidney function: first, do no harm.
        JAMA. 2021; 325: 135-137https://doi.org/10.1001/jama.2020.23373
        • Levey AS
        • Stevens LA
        • Schmid CH
        • et al.
        CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration). A new equation to estimate glomerular filtration rate.
        Ann Intern Med. 2009; 150: 604-612https://doi.org/10.7326/0003-4819-150-9-200905050-00006
        • Sjoding MW
        • et al.
        Racial bias in pulse oximetry measurement.
        N Engl J Med. 2020; 383: 2477-2478https://doi.org/10.1056/NEJMc2029240
        • Shin J-I
        • Sang Y
        • Chang AR
        • et al.
        The FDA metformin label change and racial and sex disparities in metformin prescription among patients with CKD.
        J Am Soc Nephrol. 2020; 31: 1847-1858https://doi.org/10.1681/ASN.2019101119
        • Levey AS
        • Tighiouart H
        • Titan SM
        • et al.
        Estimation of glomerular filtration rate with vs without including patient race.
        JAMA Intern Med. 2020; 180: 793-795https://doi.org/10.1001/jamainternmed.2020.0045
        • Diao JA
        • Inker LA
        • Levey AS
        • Tighiouart H
        • Powe NR
        • Manrai AK.
        In search of a better equation—performance and equity in estimates of kidney function.
        N Engl J Med. 2021; 384: 396-399https://doi.org/10.1056/NEJMp2028243
        • Sehgal AR.
        Race and the false precision of glomerular filtration rate estimates.
        Ann Intern Med. 2020; 173: 1008-1009https://doi.org/10.7326/M20-4951