GFR Estimations without Current Race-Based Equations

The use of indicators for Black race in equations commonly used to estimate glomerular filtration rate (GFR) from the level of serum creatinine has come under scrutiny and criticism of late.

Adults who identify as Black have, on average, higher levels of serum creatinine, independent of age, sex, and GFR, compared with those who do not identify as Black. Equations developed to estimate GFR from serum creatinine level have incorporated information on race, a practice that should, it has been argued, be eliminated. The inclusion of race suggests that race is a biologic rather than primarily a social construct. Nonetheless, concerns have been raised regarding possible misclassification of estimated GFR resulting from the removal of the race coefficient from current equations.

Chi-yuan Hsu, MD, and colleagues conducted an analysis of data from a large, national study involving adults with chronic kidney disease (CKD). The researchers sought to gain insights into the relationships among race, genetically derived ancestry, the serum creatinine level, and the serum cystatin C level to identify strategies for the accurate estimation of GFR without reliance on racial classifications. Results were reported online in the New England Journal of Medicine [doi:10.1056/NEJMoa2103753].

The study cohort was comprised of participants enrolled in the CRIC (Chronic Renal Insufficiency Cohort) study, a multicenter, prospective, observational study that included racially and ethnically diverse patients. The current analysis included a randomly selected subgroup of 1423 CRIC participants. Of those, 1248 participants had data available on race as reported by the participant, genetic ancestry markers, as well as serum creatinine, serum cystatin C, and 24-hour creatinine levels and were included in the final analysis.

Of the 1248 participants in the current study, 37% (n=458) identified as Black or Black and multiracial; the median percentage of African ancestry was 82.6% in those who identified as Black and 0.2% in those who identified as non-Black. Standardized differences between Black and non-Black participants in regard to age, sex, and measured eGFR were low. Mean serum creatinine levels were higher in Black participants; mean cystatin C levels were not higher in Black Participants.

In the validation data set, among the Black participants estimated GFR calculated on the basis of level of serum creatinine, age, and sex alone underestimated measured GFR by a median of 3.99 mL/min/1.73 m2 (95% confidence interval [CI], 2.17-56.2). Among non-Black participants, the median difference between measured and estimated GFR was –0.92 mL/min/1.73 m2 (95% CI, –2.29 to 0.55), an indication of statistical bias in Black participants when a race or ancestry term was not used.

There were no systematic differences between measured and estimated GFR when race as reported by participants (Black and non-Black) or percentage of African ancestry were included in the models. Models that included a race or ancestry term were correspondingly more accurate for Black participants with respect to P10 (proportion of estimated GFRs within 10% of measured GFR) (42%; 95% CI, 34 to 50) compared with models that did not include such a term (31%; 95% CI, 24 to 39).

Independent of age, sex, and measured GFR, there was an association between Black race and a 10.7% (95% CI, 8.8 to 12.7) higher serum creatinine level than among participants of non-Black race. Further, there was an association between every 10% increase in the percentage of African Ancestry and an increase of 1.3% (95% CI, 1.1 to 1.6) in the serum creatinine level in the full study sample.

The inclusion of non-GFR determinants of serum creatinine level such as metrics of body composition and urinary excretion of creatinine did not eliminate the misclassification introduced by the removal of race or ancestry from GFR estimating equations based on serum creatinine.  Non-GFR determinants of serum creatinine that differed according to Black (as opposed to non-Black) rage and higher percentage of African ancestry included higher body mass index, body-surface area, height , weight, bioelectrical impedance analysis phase angle, bioelectrical impedance analysis-quantified fat-free mass, and 24-hour urinary excretion of creatinine. There was no association between Black race or a higher percentage of African ancestry and tubular secretion of creatinine. There were associations between Black race or higher percentage of African ancestry and lower dietary protein as assessed by the Diet History Questionnaire.

In the final model that included several of the non-GFR determinants of serum creatinine, the race coefficients were not fully attenuated and there was 8.7% (95% CI, 5.8 to 11.7) higher measured GFR in Black participants compared with non-Black participants and 1.1% (95% CI, 0.8 to 1.5) higher measured GFR per 10% increase in the percentage of African ancestry.

Following adjustment for age, sex, and measured GFR, there was no association between Black race and cystatin C level. The difference comparing Black participants with non-Black participants in the full study sample was 0.03% (95%CI, –2.12 to 2.11). In addition, there was no independent association between African ancestry and cystatin C level (0.2% per 10% increase in the percentage of African ancestry; 95% CI, –0.25 to 0.28).

Models with cystatin C, age, and sex alone derived from the development data set resulted in estimates of GFR that were very close to the measured GFR in Black participants (median difference, 0.33 mL/min/1.73 m2; 95% CI, –1.43 to 1.92). The GFR estimates in Black participants when cystatin C-based equations were used were accurate (P10, 41%; 95% CI, 34% to 49%) as compared with equations using the serum creatinine level, age, sex, or race or percentage of African ancestry. The model had no meaningful improvement in the statistical bias or accuracy when a race term or an ancestry term was included in an equation based on cystatin-C.

Study limitations included the use of only research volunteers, small sample sizes, and the inability to generalize results to those with higher levels of GFR or populations outside the United States.

“Our study showed that the use of serum cystatin C rather than serum creatinine for GFR estimation produced estimates of similar validity while eliminating the negative consequences of race-based approaches,” the researchers said.

Takeaway Points

  1. Results of an analysis of data from a large national study to examine the relationships among race, ancestry, serum creatinine level, and serum cystatin C level to develop strategies for accurate estimation of glomerular filtration rate (GFR) without inclusion of racial classifications.
  2. The researchers considered three alternatives: replacement of race with a quantitative measure of ancestry in estimation of GFR; replacement of race with non-GFR determinants of serum creatinine that vary by race; and elimination of the need to consider race by use of cystatin C as the glomerular filtration marker.
  3. The use of cystatin C generated accurate results that eliminated the negative consequences of current race-based approaches.