Introduction

Currently, waittime on the national deceased donor waitlist can be accrued starting from the date of dialysis initiation. However, additional time can be accrued starting from the date when a patient's GFR is ≤20 ml/min per 1.73 m2 if a recipient is registered on the waitlist before dialysis initiation.1 We previously demonstrated that use of the race-free Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) 2021 equations could provide better equity in the preemptive waittime that could theoretically be accrued compared with use of the CKD-EPI 2009 equations (which included a term for Black race for the creatinine-based equation).2–5 The differences in waittime that we observed by race were related at least in part to the faster progression of CKD in Black than White patients and to the later eligibility of Black patients for preemptive waitlisting (due to upward adjustment of their eGFR) when a Black race term was previously included in the GFR-estimating equations.2,3

However, our previous work included only Black and White patients followed longitudinally within a research cohort and excluded patients of other racial/ethnic groups.2,3 Furthermore, some experts have suggested using risk-based strategies to determine preemptive waitlist eligibility as an alternative option to using a single eGFR threshold to better prioritize patients with the highest risk of kidney failure for preemptive waitlisting.6–9 Given that our previous work was conducted in a research cohort that may not represent patients in real-world practice where follow-up and practice patterns are not uniformly protocolized, we sought to validate and extend our previous work within a large electronic health record (EHR) cohort using data from two health care systems. We compared use of an eGFR versus a risk-based threshold as strategies for preemptive waitlisting and determined which strategy provided the most equity in preemptive waittime (within a theoretical context) across different racial/ethnic groups.9–11

Methods Study Population

We included Black, White, Asian/Pacific Islander, and Hispanic patients from the University of California San Francisco and University of California Davis between age 18 and 75 years with ≥2 outpatient serum creatinine values (to avoid the incorporation of patients with AKI) that corresponded to an eGFR <30 ml/min per 1.73 m2 taken at least 90 days apart between 2012 and 2019 (Figure 1 and Supplemental Table 1). Patients were included for analysis only if their last available eGFR was lower than the first eGFR included for the study to ensure these were patients with progressive declines in eGFR. Multiracial patients or those with unknown race were excluded from our analysis (N=36). We also excluded patients whose laboratory measurements were only available after they had received a kidney transplant, were on dialysis at the time of laboratory measurements, or had a cancer diagnosis between their first eGFR measurement <30 ml/min per 1.73 m2 and their last predialysis eGFR measurement. This study was approved by the University of California San Francisco, which served as the single Institutional Review Board of record, with concurrence from the University of California Davis Institutional Review Board. All patients who were included for analysis were linked to the United States Renal Data System (USRDS) for ascertainment of the date of onset of ESKD (kidney failure) so that follow-up could occur for all included patients.

Figure 1:

Cohort derivation. KFRE, kidney failure risk equation; UACR, urine albumin–creatinine ratio; UCD, University of California Davis; UCSF, University of California San Francisco; UPCR, urine protein–creatinine ratio.

Data Parameters Used to Model the eGFR Trajectory in the Advanced Stages of CKD

To model the trajectory of eGFR decline, we included any serum creatinine values in the EHR systems that corresponded to an eGFR <45 ml/min per 1.73 m2 (to ensure the incorporation of sufficient datapoints to further refine the eGFR slope, Supplemental Figure 1). We included eGFR at the time of dialysis initiation or preemptive transplantation if available to construct our trajectories. Our primary analysis was conducted using the 2021 CKD-EPI GFR estimating equations, which are race free. We censored all data as of December 31, 2019, because these were the most recent data available through the USRDS.

Data Parameters Used to Model the Kidney Failure Risk Equation Trajectory in the Advanced Stages of CKD

The kidney failure risk equation (KFRE) has been validated in large, racially diverse populations as a predictor of 2- or 5-year risk of onset of kidney failure using patients' age, sex, eGFR, and degree of albuminuria (or proteinuria).10–12 We chose to use the 2-year KFRE score because 2 years is closer to the time frame relevant for preparation for kidney transplantation. A 2-year online calculator is readily available (kidneyfailurerisk.com). We chose to use the four-variable (rather than the eight-variable) KFRE score because there were substantial missing data for the other values required for the eight-variable KFRE score, such as serum albumin, and because previous studies have shown similar discrimination and calibration of the four- and eight-variable equations.10,13 If urine albumin–creatinine ratio (UACR) was not available but a urine protein–creatinine ratio (UPCR) or urine dipstick was, we converted the UPCR or urine dipstick values to UACR using a standardized and validated approach.14 We carried forward previous values of UPCR or UACR if these values were missing at the time when the eGFR was predicted to fall below 20 ml/min per 1.73 m2. KFRE scores were computed using the CKD-EPI 2021 race-free equations.

Demographic Covariates and Comorbidity Ascertainment

Age, sex, race, and ethnicity were determined using data recorded in our EHR systems and supplemented by data from the USRDS. Insurance (Medicare, Medicaid, private, or other) was also obtained from our EHR systems. Patients' comorbidities, such as diabetes and systolic BP readings, were ascertained using the most recent data before the encounter when the eGFR was predicted to fall below 20 ml/min per 1.73 m2 (when time in our models began). Hypertension was determined on the basis of the use of at least one antihypertensive agent or two outpatient systolic BP values ≥130/80 mm Hg within a 90-day period before the eGFR was predicted to fall below 20 ml/min per 1.73 m2. Diabetes mellitus was ascertained on the basis of the use of insulin or oral hyperglyemic agents or an A1c ≥6.5% within 90 days of the encounter when the eGFR was predicted to fall below ≤20 ml/min per 1.73 m2.

Statistical Analysis

We used linear mixed models with Black race as a fixed effect, a fixed effect for time, a random intercept for patients, a random slope for time, and unstructured covariance to determine the estimated time that could be accrued by either an eGFR- or a KFRE-based approach to determine preemptive waitlist eligibility within a theoretical context. Thus, we estimated the time spent with an eGFR between 20 and 5 ml/min per 1.73 m2 or time spent between a 2-year KFRE predicted risk of kidney failure of 25% and an eGFR of 5 ml/min per 1.73 m2 (Supplemental Figure 1). The KFRE threshold of 25% was selected in a data-driven approach to provide comparable times to the analyses using an eGFR threshold of 20 ml/min per 1.73 m2 (which is what the current US organ allocation system uses as the threshold for the start of preemptive waittime accrual). In our linear mixed models, we estimated the trajectory of eGFR (or KFRE) over time by modeling time to a common eGFR (5 ml/min per 1.73 m2) because not all patients in the cohort started KRT and because we were interested in the maximal accrual time that would theoretically be available as in our previous work.15–18 Patients who died during follow-up contributed eGFR data until their time of death.

Thus, we created two separate models of decline in eGFR—one using eGFR by the 2021 CKD-EPI equation and a second using KFRE (incorporating the 2021 CKD-EPI equation, Supplemental Figure 1). We limited extrapolation of the eGFR trajectory to no more than four times the duration of the available data to increase confidence in the trajectories that we had developed using actual laboratory measurements that were available. From these models, we computed the time between eGFR of 20 and 5 ml/min per 1.73 m2, or KFRE of 25% and eGFR of 5 ml/min per 1.73 m2 on the basis of an extrapolation of when these thresholds were crossed from our linear mixed models for each racial/ethnic group.

Next, we tested for differences in the estimated time available in the advanced stages of CKD (between eGFR 20 ml/min per 1.73 m2 or KFRE of 25% and an eGFR of 5 ml/min per 1.73 m2) by race and ethnicity using unadjusted linear regression models and comparing the β coefficients (which represents preemptive waittime differences between different racial/ethnic groups in months) across these models. Bootstrapping for the 95% confidence interval (CI) of the differences in the β coefficients (with 1000 repetitions) was performed. The specific comparisons that we made were for differences in the β coefficients derived in different models, where we used the CKD-EPI 2021 equation versus the KFRE to derive estimated preemptive waittimes. In sensitivity analysis, we also repeated analyses using a KFRE threshold of 20%, which some studies have suggested as the threshold to begin preparations for KRT.6

Our primary analysis was unadjusted because the current referral and waitlist policy is on the basis of an absolute eGFR threshold, which does not allow clinicians to adjust for additional considerations aside from eGFR. Furthermore, the KFRE does already account for other covariates that may influence rate of progression of CKD, including age, sex, and albuminuria. However, in sensitivity analyses, we additionally adjusted models further for age, sex, diabetes status, median neighborhood income by zip code, insurance type, and systolic BP ascertained at the closest time point before the eGFR dropped to 20 ml/min per 1.73 m2.

Sensitivity Analyses

Because there were variations in the frequency of serum creatinine measurements that were available across different racial and ethnic groups, we also repeated our analysis limiting the modeling of our trajectories to the inclusion of only one randomly selected serum creatinine measurement per 90-day period of follow-up, or using the median of all eGFR values per 90-day period to build our trajectories, thus including an equal number of measurements for our trajectory modeling. In additional sensitivity analysis, we modeled our trajectories using the actual eGFR if a patient developed kidney failure; for patients who did not develop kidney failure, we modeled their time to an eGFR of 5 ml/min per 1.73 m2. Finally, we also excluded from our analyses individuals who died and repeated our primary analyses to ensure that results were consistent.

Results Characteristics of the Cohort

We included 1290 adults who met our inclusion criteria for analysis using CKD-EPI 2021 race-free equations. Characteristics of these individuals are shown in Table 1. The median age of the cohort was 58 (interquartile range [IQR], 48–66) years, 43% were women, 21% were of non-Hispanic Black race, 21% Asian/Pacific Islander, and 19% of Hispanic ethnicity. Approximately 36% of individuals had commercial insurance. Most of this cohort had hypertension and approximately 34% had diabetes. The median albuminuria at entry into the study was 469 (IQR, 252–1386) mg/g.

Table 1 - Characteristics of patients by race and ethnicity Cohort Characteristics and Outcomes Asian/Pacific Islander Black Hispanic Non-Hispanic White Total n=265 (21%) n=276 (21%) n=242 (19%) n=507 (39%) n=1290 (100%) Age at cohort entry, yr, median (IQR) 60 (50–66) 56 (48–63) 54 (44–64) 60 (50–66) 58 (48–66) Sex, n (%)  Male 149 (56) 149 (54) 137 (57) 295 (58) 730 (57)  Female 116 (44) 127 (46) 105 (43) 212 (42) 560 (43) Hypertension, n (%) 211 (80) 232 (84) 197 (81) 426 (84) 1066 (83) Systolic BP, mm Hg, median (IQR) 140 (126–154) 145 (132–161) 139 (127–157) 135 (122–148) 139 (126–154) Diabetes, n (%) 103 (39) 99 (36) 88 (36) 148 (29) 438 (34) Insurance type, n (%)  Medicare 55 (21) 97 (35) 63 (26) 113 (22) 328 (25)  Medicaid 56 (21) 87 (32) 84 (35) 95 (19) 322 (25)  Commercial 110 (42) 59 (21) 70 (29) 225 (44) 464 (36)  Other 44 (17) 33 (12) 25 (10) 74 (15) 176 (14) UACR, mg/g, median (IQR) 840 (337–1964) 478 (262–1229) 1010 (337–1762) 337 (99–1229) 469 (252–1386) Years from first to last eGFR, median (IQR) 1.4 (0.7–2.8) 1.8 (0.9–3.4) 1.4 (0.7–2.7) 1.8 (0.9–3.3) 1.6 (0.8–3.1) First eGFR <30 ml/min per 1.73 m2, median (IQR) 25 (18–28) 24 (18–28) 25 (19–28) 25 (20–28) 25 (19–28) Last eGFR available, ml/min per 1.73 m2, median (IQR) 11 (7–16) 10 (6–14) 11 (8–16) 12 (9–18) 11 (8–17) eGFR at kidney failure (if applicable) in ml/min per 1.73 m2, median (IQR) 8 (6–11) 8 (6–11) 10 (8–12) 10 (8–13) 9 (7–12) Site, n (%)  University of California San Francisco 172 (65) 118 (43) 100 (41) 224 (44) 614 (48)  University of California Davis 93 (35) 158 (57) 142 (59) 283 (56) 676 (52) Outcomes  Estimated years from eGFR of 20 to 5 ml/min per 1.73 m2    Mean±SD 2.9±2.2 3.2±2.3 2.9±2.5 3.7±3.4 3.3±2.8    Median (IQR) 2.3 (1.6–3.1) 2.5 (1.7–3.8) 2.1 (1.4–3.3) 2.6 (1.8–4.2) 2.4 (1.7–3.7)  Estimated years from KFRE of 25% to eGFR of 5 ml/min per 1.73 m2    Mean±SD 3.4±2.3 3.6±2.6 3.4±2.8 3.8±3.5 3.6±3.0    Median (IQR) 2.9 (2.0–3.8) 2.9 (2.1–4.2) 2.6 (1.8–4.0) 2.8 (1.8–4.6) 2.8 (1.9–4.2)  Kidney failure (dialysis or transplant) 187 (71) 194 (70) 162 (67) 319 (63) 862 (67)  Death before reaching eGFR of 5 ml/min per 1.73 m2 32 (12) 40 (14) 33 (14) 104 (21) 209 (16)

IQR, interquartile range; KFRE, kidney failure risk equation; UACR, urine albumin–creatinine ratio.

By race and ethnicity, the median age at entry into our study was younger for Black and Hispanic patients compared with non-Hispanic White patients, and the prevalence of commercial insurance was lower in Black and Hispanic patients (Table 1). Non-Hispanic White patients also had a lower prevalence of diabetes and lower median systolic BP at the time of entry into our cohort compared with all other racial and ethnic groups. The median albuminuria was highest among Hispanic patients (1010 mg/g; IQR, 337–1762) and Asian/Pacific Islander patients (840 mg/g; IQR, 337–1964).

About 67% of individuals initiated KRT during the study follow-up. However, there was variability in the prevalence of onset of KRT with 70% of non-Hispanic Black patients, 63% of non-Hispanic White patients, 67% of Hispanic patients, and 71% of Asian/Pacific Islander patients reaching the need for KRT.

Modeled Time between eGFR of 20 and 5 ml/min per 1.73 m2

When using the CKD-EPI 2021 equation, non-Hispanic White patients spent a mean of 3.7±3.4 years between an eGFR of 20 and 5 ml/min per 1.73 m2. This mean time was shorter for Black patients (3.2±2.3 years), Hispanic patients (2.9±2.5 years), and Asian/Pacific Islander patients (2.9±2.2 years, Figure 2 and Table 1). The time spent between an eGFR of 20 and 5 ml/min per 1.73 m2 was on average 6.8 months shorter for those who were non-Hispanic Black (95% CI, −11.7 to −1.9 months), 10.2 months shorter for those of Hispanic ethnicity (95% CI, −15.3 to −5.1 months), and 10.3 months shorter for Asian/Pacific Islander patients (95% CI, −15.3 to −5.4 months) relative to non-Hispanic White patients (Table 2).

Figure 2:

Mean time (in years) available for preemptive waittime accrual by race and ethnicity. CI, confidence interval.

Table 2 - Predicted months to eGFR of 5 ml/min per 1.73 m2 from either an eGFR of 20 ml/min per 1.73 m2 or a kidney failure risk equation estimated risk of 25% by race and ethnicity and comparisons of differences across these two models Race and Ethnicity eGFR 20 to eGFR 5 ml/min per 1.73 m2 KFRE 25% to eGFR 5 ml/min per 1.73 m2 Differences across the eGFR versus KFRE Models (with Bootstrapped CI)a β (mo) 95% CI β (mo) 95% CI Difference (mo) 95% CI Asian/Pacific Islander −10.3 −15.3 to −5.4 −5.4 −10.7 to −0.1 4.9 1.7 to 8.1 Black −6.8 −11.7 to −1.9 −2.5 −7.8 to 2.7 4.3 0.8 to 7.7 Hispanic −10.2 −15.3 to −5.1 −4.8 −10.3 to 0.6 5.3 1.9 to 8.8 Non-Hispanic White Reference Reference

CI, confidence interval; KFRE, kidney failure risk equation.

aIn this model, we first determined the difference in the β coefficients across the two separate models by eGFR and by kidney failure risk equation thresholds. We then used a bootstrapping approach where the analysis was repeated 1000 times to estimate confidence intervals around the difference in time (which are not easily obtained otherwise because we are comparing differences in time across two separate models).

Time between KFRE of 25% and Modeled eGFR of 5 ml/min per 1.73 m2

When we repeated our models to estimate time between KFRE of 25% and eGFR of 5 ml/min per 1.73 m2 (incorporating the 2021 CKD-EPI equation into the KFRE), the mean waittime available for accrual was 3.8 years for non-Hispanic White patients, 3.6 years for Black patients, 3.4 years for Hispanic patients, and 3.4 years for Asian/Pacific Islander patients (Figure 2 and Table 1). Use of a KFRE threshold incorporating the 2021 eGFR equations to determine waittime yielded nonstatistically significant differences in time for Black (−2.5 months; 95% CI, −7.8 to 2.7) and Hispanic (−4.8 months; 95% CI, −10.3 to 0.6) patients compared with non-Hispanic White patients. However, statistically significant differences in time were still noted in Asian/Pacific Islander patients (−5.4 months; −10.7 to −0.1) compared with non-Hispanic White patients (Table 2).

We bootstrapped and identified statistically significant differences in the β coefficients (representing theoretical preemptive waittime differences between different racial/ethnic groups) comparing use of an eGFR versus KFRE threshold for Black, Hispanic, and Asian/Pacific Islander patients showing superiority of the KFRE threshold for improving equity in time using the KFRE (all P < 0.05) across all subgroups (Table 2).

When we examined the time difference by race and ethnicity in adjusted models (Supplemental Table 2), we noted differences in potential waittime accrual in Asian/Pacific Islander and Hispanic patients compared with non-Hispanic White patients, but this finding did not achieve statistical significance in Black patients, although the time differences were generally consistent with those observed in primary models. Use of the KFRE threshold yielded smaller absolute differences in waittime accrual (in months) across all racial and ethnic groups that were not statistically significantly different compared with non-Hispanic White patients for all groups in adjusted models, although these models may be over adjusted because of their inclusion of factors already in the KFRE.

Sensitivity Analysis

When we repeated our analyses using a KFRE threshold of 20% instead of the 25% used in our primary analysis, results were qualitatively similar with those found in our primary analysis, although the difference in waittime was no longer statistically significant for Black individuals when comparing the KFRE with an eGFR threshold across the two models (P = 0.09, Supplemental Table 3). However, in general, use of the KFRE still provided more equity in waittime than use of an eGFR threshold.

When we repeated our analyses comparing time between an eGFR 20 and 5 ml/min per 1.73 m2 using only one randomly selected eGFR within each 90-day period or the median of all eGFRs within each 90-day period of follow-up, findings were generally consistent with those in our primary analysis (Supplemental Table 4). When we computed time using the date of actual kidney failure onset (or predicted onset of eGFR of 5 ml/min per 1.73 m2 if patient had not yet reached KRT), findings were similar to our primary analysis, except that use of the KFRE also attenuated differences in potential waittime accrual for Asian/Pacific Islander individuals (Supplemental Table 5). Finally, when we excluded patients who died before their estimated date of an eGFR of 5 ml/min per 1.73 m2 (N=209), results were qualitatively similar (Supplemental Table 6).

Discussion

There has been significant concern over inequities in access to kidney transplantation in non-White patients over the past two decades. In this study, we compared different strategies of determining preemptive waittime accrual using real-world EHR data from two large health care systems. When we used the CKD-EPI 2021 race-free equation to estimate potential waittime accrual in accordance with current national waitlisting policies, we observed a statistically significant shorter waittime for Black, Hispanic, and Asian/Pacific Islander patients that could potentially be accrued in comparison to non-Hispanic White patients. When we used a KFRE threshold as the criteria for entry into our study, we observed attenuation in the inequities for Black and Hispanic, but not Asian/Pacific Islander individuals.

A number of observational studies have suggested that late (or non) referral for transplant evaluation and longer time to the completion of recipient evaluations both contribute to the lower access of non-White patients to kidney transplantation.19–24 We had hypothesized that structural racism may be one factor that contributed to the inequities in access to preemptive waitlisting across different racial and ethnic groups given the current national policy for waitlisting is on the basis of a single GFR value and does not allow one to adjust for other considerations. Although additional waittime can be accrued starting from the date when the GFR is ≤20 ml/min if a candidate was registered on the waitlist before dialysis initiation (preemptively),25–27 the use of a single eGFR threshold to determine eligibility for waittime accrual ignores differences in the rate of progression of CKD (which may be related to presence or absence of diabetes and albuminuria) that are known to be more rapid in non-White patients.28 Patients with more rapidly progressive CKD (observed in the Black and Hispanic population)28,29 would thus have less preemptive waittime that could be accrued before dialysis initiation, even if their disease is recognized and they are appropriately referred for kidney transplantation.

Consistent with our hypothesis, when we used a risk-based strategy to determine waittime that could be theoretically accrued using the KFRE, we found that the inequities in time spent in the advanced stages of CKD across different racial/ethnic groups were improved over the use of a single eGFR threshold, particularly for Black and Hispanic individuals. Because the KFRE includes age, sex, eGFR, and albuminuria in its computation,10,11 we believe the KFRE provides a more equitable strategy for the determination of the risk for needing KRT than using a single eGFR threshold and could be considered as a strategy for determining waitlist eligibility.

The finding of inequities in theoretical preemptive waittime accrual in the Asian and Pacific Islander population is additive to the previous literature. Disparities in access of Asian and Pacific Islander individuals to kidney transplantation have been previously described and associated with factors such as neighborhood poverty and linguistic isolation.23,30 Although much attention has been focused on the inclusion of a Black race term in GFR-estimating equations, in a real-world setting, use of the CKD-EPI 2021 equation still led to theoretical preemptive waittime inequities for Black, Hispanic, and Asian individuals. Further efforts are needed to improve equity in the amount of waittime that could be accrued when patients are referred, evaluated, and preemptively waitlisted, regardless of race and ethnicity.

Our study has several strengths. First, we used a large, diverse cohort of patients seen for usual care in two large academic centers and were able to extend our previous work to Hispanic and Asian/Pacific Islander populations. Second, the cohort experienced a high rate of kidney failure events. Third, we were able to ascertain eGFR trajectories over time, which national databases do not collect data on. Fourth, we compared a well-validated risk estimating equation (the KFRE) that is widely available to the use of a single eGFR threshold (in accordance with current national policy) to evaluate alternative approaches that may improve equity in preemptive waiting time accrual.

We acknowledge several limitations to our study. Our study was conducted within a theoretical context and does not address whether using any alternative strategy that we tested to determine preemptive waittime accrual would lead to actual improvements in transplant rates for non-White individuals. We are unable to determine to what extent the disparities observed in preemptive waittime are due to differential access to providers who are able to efficiently guide patients through the evaluation process with expedience and leverage opportunities to have the patient waitlisted at the earliest time point possible. Furthermore, factors related to social determinants of health or other medical comorbidities may prolong time from referral to waitlisting, although we emphasize that our study was not designed to identify the mediators of the inequities but to examine different policy-based strategies that could be applied when determining preemptive waittime accrual to attenuate any observed racial and ethnic inequities. Finally, practice patterns in our two large health care systems may not generalize to other settings, our data have not been validated externally, and additional solutions beyond changing the threshold for preemptive waitlisting may be needed to further address the inequities we observed.

In conclusion, although the KFRE attenuated differences in theoretically accruable waittime for Black and Hispanic patients that were observed when using a CKD-EPI 2021 eGFR threshold for waittime accrual, persistent inequities were observed for Asian/Pacific Islander individuals. Our study highlights the importance of recognizing that disparities in access to transplantation persist despite omission of a Black race term from GFR-estimating equations, particularly for those whose race and ethnicity were not accounted for in the previous equations and who may be underrepresented in most research cohorts. Further studies are needed to confirm our findings and to determine whether alternative risk-based approaches in guiding preemptive waitlisting and waittime accrual may associate with improvement in inequities in access to kidney transplantation across all racial and ethnic groups.

Disclosures

S. Amaral reports consultancy for Bristol Myers Squibb-DSMB; research funding from Laffey-McHugh Foundation, NIAID, NICHD, and NIH-NIDDK; advisory or leadership roles for Ad Hoc MOT Committee, Dialysis Patient Citizens Advisory Council, Education Center, and POC and VCA committees—UNOS/OPTN; and other interests or relationships with ASPN Executive Council. L.-X. Chen reports research funding from Allovir, Astellas, CSL Behring, Dexcom, Memo Therapeutics, TruGraf, and Veloxis. K.L. Johansen reports employment with Hennepin Healthcare; consultancy for Akebia, GSK, and Vifor; an advisory or leadership role for GSK; and role as an Associate Editor of JASN. E. Ku reports ownership interest in Edison Company; research funding from CareDX, Natera, and NIH; advisory or leadership roles for American Journal of Kidney Diseases (Associate Editor) and American Kidney Fund Health Equity Coalition; and other interests or relationships with Fidelity Trust and John Andrew Lang Philanthropic Fund. C.E. McCulloch reports consultancy for Acuta Capital Partners and Amgen. M.R. Weir reports consultancy for Akebia, AstraZeneca, Bayer, Boehringer-Ingelheim, CareDx, Janssen, Merck, Novo Nordisk, and Vifor Pharma—all are modest (less than $10000) and honoraria for ad hoc advisory board meetings. All remaining authors have nothing to disclose.

Funding

E. Ku and S. Amaral: National Institutes of Health (R01 DK120886); K. Johansen: National Institutes of Health (R01 DK115629).

Acknowledgments

The data reported here have been supplied by the United States Renal Data System. The interpretation and reporting of the data presented here are the responsibility of the authors and in no way should be seen as an official policy or interpretation of the US Government. This manuscript was made possible by the Multidisciplinary Advancement of Transplant-Centered Health Outcomes Research Center. The funding sources had no input in the design, analysis, interpretation of the data, or drafting of the manuscript.

Author Contributions

Conceptualization: Sandra Amaral, Kirsten L. Johansen, Elaine Ku, Charles E. McCulloch, Matthew R. Weir.

Data curation: Ling-Xin Chen, Naeem Goussous, Elaine Ku, Isabelle Lopez.

Formal analysis: Timothy Copeland, Charles E. McCulloch.

Funding acquisition: Elaine Ku, Matthew R. Weir.

Investigation: Sandra Amaral, Timothy Copeland, Kirsten L. Johansen, Elaine Ku, Charles E. McCulloch, Jonathan D. Savant.

Methodology: Sandra Amaral, Timothy Copeland, Kirsten L. Johansen, Elaine Ku, Charles E. McCulloch.

Project administration: Naeem Goussous, Isabelle Lopez, Jonathan D. Savant.

Resources: Elaine Ku, Charles E. McCulloch.

Supervision: Sandra Amaral, Ling-Xin Chen, Elaine Ku, Charles E. McCulloch.

Writing – original draft: Elaine Ku.

Writing – review & editing: Sandra Amaral, Ling-Xin Chen, Isabelle Lopez, Charles E. McCulloch, Jonathan D. Savant.

Data Sharing Statement

The data are from institutions’ clinical health data records, which are not allowed to be shared under our data use agreements within our consortium and when we applied for use of the EHR data. In addition, the data are linked to the USRDS, and the USRDS do not allow for re-release of data by individual investigators.

Supplemental Material

This article contains the following supplemental material online at https://links.lww.com/CJN/B827.

Supplemental Figure 1. Summary of methods.

Supplemental Table 1. Characteristics of patients included and excluded from study.

Supplemental Table 2. Multivariable model of predicted months to an eGFR of 5 ml/min per 1.73 m2 from either an eGFR of 20 ml/min per 1.73 m2 or KFRE score of 25% by race/ethnicity.

Supplemental Table 3. Predicted months to eGFR of 5 ml/min per 1.73 m2 from either an eGFR of 20 ml/min per 1.73 m2 or a KFRE score of 20% by race/ethnicity with bootstrapped comparisons across models.

Supplemental Table 4. Predicted time difference between an eGFR of 20 and 5 ml/min per 1.73 m2 constructing trajectories using all available eGFRs, one random eGFR per 3 month period, or the median eGFR per 3 month period by race/ethnicity.

Supplemental Table 5. Comparison of time (in months) by race/ethnicity using actual eGFR at ESKD onset when available in sensitivity analysis.

Supplemental Table 6. Predicted months to eGFR of 5 ml/min per 1.73 m2 from either an eGFR of 20 ml/min per 1.73 m2 or a KFRE score of 25% by race/ethnicity with bootstrapped comparisons across models, excluding 209 patients who died during follow-up.

References 1. Organ Procurement & Transplantation Network, U. S. Department of Health & Human Services, Health Resources & Services Administration. Notice of OPTN Policy Change: Establish OPTN Requirement for Race-Neutral Estimated Glomerular Filtration Rate (eGFR) Calculations, 2022. Accessed March 8, 2023. https://optn.transplant.hrsa.gov/media/xn3nhhjr/policy-notice_establish-optn-req-for-race-neutral-egfr-calcls_mac.pdf 2. Ku E, McCulloch CE, Adey DB, Li L, Johansen KL. Racial disparities in eligibility for preemptive waitlisting for kidney transplantation and modification of eGFR thresholds to equalize waitlist time. J Am Soc Nephrol. 2021;32(3):677–685. doi:10.1681/ASN.2020081144 3. Ku E, Amaral S, McCulloch CE, Adey DB, Li L, Johansen KL. Comparison of 2021 CKD-EPI equations for estimating racial differences in preemptive waitlisting for kidney transplantation. Clin J Am Soc Ne