To the Editor:
Estimated glomerular filtration rate (eGFR) has been commonly used as a measure of kidney function in clinical practice and research. The widely used CKD-EPI creatinine equation uses sex, age, race (Black or non-Black), and serum creatinine level (Scr) to calculate eGFR, assigning a 16% higher eGFR for individuals identified as Black despite the same age, sex, and Scr.1 Because race is a social, nonbiological construct,2,3 to provide a unifying approach for GFR estimation3 a new CKD-EPI creatinine equation that does not use the race variable was recently developed4 and is currently recommended by the National Kidney Foundation and American Society of Nephrology Task Force for US adults.5,6 To understand the potential impact of the new equation on estimating racial disparities in clinical outcomes, we assessed Black-White disparities in kidney replacement therapy (KRT) and death as outcomes after incident CKD defined using the 2009 and 2021 CKD-EPI creatinine equations (2009-CKD-EPI and 2021-CKD-EPI, respectively). Our hypothesis is that 2021-CKD-EPI, developed without using the race variable, would reduce estimates of racial disparities in CKD outcomes.
The US Veterans Health Administration (VHA) provided a single-source population from which veterans, either non-Hispanic White or non-Hispanic Black, were determined to have incident CKD GFR category 3 or higher based on outpatient Scr using (1) 2009-CKD-EPI with its race term included1 to construct one cohort, and (2) 2021-CKD-EPI developed without the race variable4 to construct an alternative cohort. Race and ethnicity were self-reported. We defined incident CKD by the first occurrence of 2 consecutive eGFRs <60 mL/min/1.73 m2 at ≥91 days apart but within 18 months.7 Follow-up time started from the second, confirmatory, eGFR value in each cohort to avoid immortal time bias and continued up to 10 years or May 31, 2018. For each cohort, we obtained hazard ratios of death (including death after KRT) for Black versus White veterans using Cox proportional hazards regression and for KRT using cause-specific hazards regression censored for death. Additional detail is given in Item S1. This study was approved by Institutional Review Boards.
Between January 2007 and December 2016, 84,090 Black and 507,303 White veterans had incident CKD defined by 2009-CKD-EPI, and 101,693 Black (including 67,038 [66%] who were in both cohorts and 34,655 [34%] patients new to this cohort) and 449,802 White veterans (366,118 [81%] in both cohorts and 83,684 [19%] new patients) had incident CKD by 2021-CKD-EPI. Some patients identified by 2009-CKD-EPI were not identified as incident CKD by 2021-CKD-EPI (Item S2). The Black incident CKD group defined by 2021-CKD-EPI was younger and had fewer comorbid conditions compared with its counterpart defined by 2009-CKD-EPI (Table 1), primarily owing to inclusion of younger and healthier Black veterans identified by 2021-CKD-EPI (Table S1). Conversely, the White group by 2021-CKD-EPI was slightly older, with more comorbid conditions (Table 1), primarily owing to removal of some healthier White veterans from its incident CKD group (Table S1). Rates of KRT (10.5 vs 15.2 per 1,000 patient-years) and death (51.6 vs 62.9 per 1,000 patient-years) were both substantially reduced in Black veterans using 2021-CKD-EPI versus 2009-CKD-EPI (Table 2). Conversely, rates of KRT and death were both increased in White veterans. After adjustment for basic covariates, Black veterans had a 37% greater hazard of KRT than White veterans under 2021-CKD-EPI, a marked reduction from the 172% greater Black-versus-White hazard under 2009-CKD-EPI. Also, Black veterans had a 9% lower adjusted hazard of death than White veterans under 2021-CKD-EPI, in contrast to a 10% greater adjusted hazard of death with 2009-CKD-EPI. These disparities were similar after further adjustment for comorbidities.
Table 1.
Baseline Characteristics of Black and White Veterans at Incidence of CKD Defined by the 2021 and 2009 CKD-EPI Equations
| Characteristic | Identified by the 2021 CKD-EPI (n = 551,495) |
Identified by the 2009 CKD-EPI (n = 591,393) |
||
|---|---|---|---|---|
| Black | White | Black | White | |
| No. of participants | 101,693 | 449,802 | 84,090 | 507,303 |
| Age, y | 65.3 ± 10.2 | 74.1 ± 9.6 | 67.0 ± 10.3 | 73.2 ± 9.6 |
| Age <65 years | 52,656 (51.8%) | 80,685 (1 7.9%) | 37,875 (45.0%) | 103,429 (20.4%) |
| Male sex | 96,095 (94.5%) | 437,603 (97.3%) | 80,234 (95.4%) | 492,891 (972%) |
| BMI, kg/m2 | 29.8 ± 6.3 | 29.9 ± 6.0 | 29.8 ± 6.5 | 29.9 ± 5.9 |
| Systolic BP, mm Hg | 133.0 ± 19.8 | 130.2 ± 18.2 | 133.4 ± 20.4 | 130.3 ± 17.9 |
| Diastolic BP, mm Hg | 77.1 ± 12.3 | 71.2 ± 11.1 | 76.3 ± 12.6 | 71.8 ± 11.1 |
| eGFR, mL/min/1.73 m2 | 51.2 ± 8.4 | 51.4 ± 7.8 | 50.6 ± 8.9 | 51.4 ± 7.8 |
| Hypertension | 94,644 (93.1%) | 416,629 (92.6%) | 80,320 (95.5%) | 461,841 (91.0%) |
| Diabetes | 54,467 (53.6%) | 230,136 (51.2%) | 49,053 (58.3%) | 246,490 (48.6%) |
| Heart failure | 21,653 (21.3%) | 130,674 (29.1%) | 21,235 (25.3%) | 132,569 (26.1%) |
| Coronary artery disease | 33,652 (33.1%) | 242,851 (54.0%) | 31,835 (37.9%) | 259,190 (51.1%) |
| Cardiac dysrhythmia | 27,377 (26.9%) | 1 98,826 (44.2%) | 25,425 (30.2%) | 211,704 (41.7%) |
| Other cardiac diseases | 29,208 (28.7%) | 185,218 (41.2%) | 26,507 (31.5%) | 198,232 (39.1%) |
| CVA/TIA | 19,497 (19.2%) | 129,699 (28.8%) | 18,179 (21.6%) | 137,146 (27.0%) |
| PVD | 24,310 (23.9%) | 1 64,778 (36.6%) | 22,826 (27.1%) | 174,148 (34.3%) |
| COPD | 26,390 (26.0%) | 165,477 (36.8%) | 23,168 (27.6%) | 181,023 (35.7%) |
| Anemia | 38,656 (38.0%) | 1 79,468 (39.9%) | 36,180 (43.0%) | 185,669 (36.6%) |
| Cancer | 23,490 (23.1%) | 124,088 (27.6%) | 21,364 (25.4%) | 134,297 (26.5%) |
| GI bleeding disorders | 13,880 (13.6%) | 64,031 (14.2%) | 12,310 (14.6%) | 68,880 (13.6%) |
| Liver disease | 6,581 (6.5%) | 24,696 (5.5%) | 5,656 (6.7%) | 27,478 (5.4%) |
| UACR | ||||
| <30 mg/g | 25,698 (25.3%) | 117,343 (26.1%) | 21,127 (25.1%) | 132,422 (26.1%) |
| 30–300 mg/g | 12,596 (12.4%) | 59,265 (13.2%) | 11,946 (14.2%) | 60,357 (11.9%) |
| >300 mg/g | 4,946 (4.9%) | 16,282 (3.6%) | 5,336 (6.3%) | 15,216 (3.0%) |
| Missing | 58,453 (57.5%) | 256,912 (57.1%) | 45,681 (54.3%) | 299,308 (59.0%) |
| Incident years 2007–2010 | 33,672 (33.1%) | 1 72,355 (38.3%) | 27,715 (33.0%) | 197,405 (38.9%) |
| Incident years 2011–2013 | 28,171 (27.7%) | 121,379 (27.0%) | 23,542 (28.0%) | 134,849 (26.6%) |
| Incident years 2014–2016 | 39,850 (39.2%) | 156,068 (34.7%) | 32,833 (39.0%) | 1 75,049 (34.5%) |
Values for continuous variables given as mean ± standard deviation. Categories may not sum to 100% due to rounding. Abbreviations: CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration; BMI, body mass index; BP, blood pressure; CVA/TIA, cerebrovascular accident/transient ischemic attack; PVD, peripheral vascular disease; COPD, chronic obstructive pulmonary disease; GI, gastrointestinal; UACR, urinary albumin-creatinine ratio.
Table 2.
Hazard Ratios of Outcomes for Black Veterans Compared toWhite VeteransWith Incident CKD Defined by Each Equation
| Event Rate (per 1,000 Patient-Years) |
HR (95% CI) | |||
|---|---|---|---|---|
| Unadjusted | Adjusted for Basic Covariatesa | Additionally Adjusted for Comorbiditiesb | ||
| KRT | ||||
| 2021-CKD-EPIc | ||||
| Black | 10.5 | 2.39 (2.31–2.47) | 1.37 (1.32–1.42) | 1.27 (1.22–1.32) |
| White | 4.4 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| 2009-CKD-EPIc | ||||
| Black | 15.2 | 4.56 (4.40–4.71) | 2.72 (2.62–2.82) | 2.21 (2.13–2.29) |
| White | 3.4 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| Death | ||||
| 2021-CKD-EPIc | ||||
| Black | 51.6 | 0.62 (0.61–0.62) | 0.91 (0.90–0.92) | 0.96 (0.94–0.97) |
| White | 83.5 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| 2009-CKD-EPIc | ||||
| Black | 62.9 | 0.85 (0.84–0.86) | 1.10 (1.09–1.12) | 1.10 (1.08–1.11) |
| White | 74.3 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
Abbreviation: HR, hazard ratio.
Basic covariates included age, sex, eGFR, and incident-year period.
Covariates were basic covariates plus BMI, systolic and diastolic BPs, and 13 comorbidities (hypertension, diabetes, heart failure, coronary artery disease, cardiac dysrhythmia, other cardiac diseases, CVA/TIA, PVD, COPD, anemia, cancer, GI bleeding disorders, and liver disease).
Analyses were performed separately for the 2 equations. For each equation, Black veterans were compared to the White veterans.
In this national veteran population, we found that 2021-CKD-EPI resulted in smaller estimates of racial disparities in clinical outcomes, but a greater disparity in age of CKD onset and many comorbidities at incident CKD. A main strength of the study is that the comparison was robust because all groups of veterans were anchored in new diagnosis of CKD according to the equation and were followed for an extended duration. The main limitation is the predominance of male participants in VHA.
These findings have important implications. Despite no change in actual patient outcomes, our disparity estimates of clinical outcomes between non-Hispanic Black and White veterans following incident CKD changed substantially using 2021-CKD-EPI because it changes eGFR values for both races, resulting in different incident CKD cohorts of both races. The research community should consider how to interpret prior and future estimates of disparities based on different eGFR equations. Our findings also suggest that the 2021 equation will identify more Black adults with CKD who are younger and relatively healthier. This will provide an important opportunity for health care providers to engage and intervene early to mitigate risks of progression and adverse outcomes in this high-risk group.
Supplementary Material
Acknowledgements:
Support for VA/CMS data was provided by the Department of Veterans Affairs, VA Health Services Research and Development Service, VA Information Resource Center (Project Numbers SDR 02-237 and 98-004). This work was also supported using resources and facilities at the Veterans Informatics and Computing Infrastructure (VINCI), VA HSR RES 13-457. Accessing VA national databases and computations was supported by the VINCI team in Salt Lake City, Utah.
Support:
Research reported herein was supported by NIH/NIDDK under award no. R01DK112008. Dr Norris was supported by NIH grants P30AG021684 and UL1TR000124. The funders had no role in study design, data collection, analysis, reporting, or the decision to submit for publication.
Footnotes
Financial Disclosure: The authors declare that they have no other relevant financial interests.
Disclaimer: The views expressed in this publication are those of the authors and do not reflect the official policy of the Department of the Army/Navy/Air Force, the Department of Defense, NIH, or the US government.
Peer Review: Received July 1, 2021. Evaluated by 2 external peer reviewers, with direct editorial input from a Statistics/Methods Editor and an Associate Editor, who served as Acting Editor-in-Chief. Accepted in revised form December 3, 2021. The involvement of an Acting Editor-in-Chief was to comply with AJKD’s procedures for potential conflicts of interest for editors, described in the Information for Authors & Journal Policies.
Contributor Information
Guofen Yan, Department of Public Health Sciences, University of Virginia School of Medicine, Charlottesville, Virginia.
Robert Nee, Nephrology Service, Walter Reed National Military Medical Center and Department of Medicine, Uniformed Services University, Bethesda, Maryland.
Julia J. Scialla, Department of Public Health Sciences, University of Virginia School of Medicine, Charlottesville, Virginia; Division of Nephrology, University of Virginia School of Medicine, Charlottesville, Virginia.
Tom Greene, Division of Biostatistics, University of Utah School of Medicine, Salt Lake City, Utah.
Wei Yu, Department of Public Health Sciences, University of Virginia School of Medicine, Charlottesville, Virginia.
Alfred K. Cheung, Division of Nephrology & Hypertension, Department of Internal Medicine, University of Utah, Salt Lake City, Utah; Renal Section, Veterans Affairs Salt Lake City Healthcare System, Salt Lake City, Utah.
Keith C. Norris, Department of Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California.
References
- 1.Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604–612. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Vyas DA, Eisenstein LG, Jones DS. Hidden in plain sight - reconsidering the use of race correction in clinical algorithms. N Engl J Med. 2020;383(9):874–882. [DOI] [PubMed] [Google Scholar]
- 3.Delgado C, Baweja M, Burrows NR, et al. Reassessing the inclusion of race in diagnosing kidney diseases: an interim report from the NKF-ASN Task Force. Am J Kidney Dis. 2021;78(1):103–115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Inker LA, Eneanya ND, Coresh J, et al. New creatinine- and cystatin C-based equations to estimate GFR without race. N Engl J Med. 2021;385(19):1737–1749. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Delgado C, Baweja M, Crews DC, et al. A unifying approach for GFR estimation: recommendations of the NKF-ASN Task Force on reassessing the inclusion of race in diagnosing kidney disease. Am J Kidney Dis. Published online September. 23, 2021. doi: 10.1053/j.ajkd.2021.08.003. [DOI] [PubMed] [Google Scholar]
- 6.Delgado C, Baweja M, Crews DC, et al. A unifying approach for GFR estimation: recommendations of the NKF-ASN Task Force on reassessing the inclusion of race in diagnosing kidney disease. J Am Soc Nephrol. 2021;32(12):2994–3015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Levey AS, Coresh J, Balk E, et al. National Kidney Foundation practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Ann Intern Med. 2003;139(2):137–147. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.