Significance Statement
Under current US national policy, determining a patient’s eligibility for kidney transplantation waitlist registration requires the patient’s GFR to be ≤20 ml/min. Because disease progression is faster for Black versus White patients, this policy may contribute to racial disparities in accruable time on the waitlist before dialysis initiation. The authors used models to determine the association between race and time to ESKD from an eGFR of ≤20 ml/min per 1.73 m2, finding this time was shorter for Black versus White patients. They then estimated that allowing registration of Black patients on the transplant waitlist at higher levels of kidney function (as early as an eGFR of 24–25 ml/min per 1.73 m2) had the potential to reduce the observed disparities in accruable wait time.
Keywords: chronic kidney disease, transplantation, dialysis, progression of chronic renal failure, glomerular filtration rate
Abstract
Background
Patients may accrue wait time for kidney transplantation when their eGFR is ≤20 ml/min. However, Black patients have faster progression of their kidney disease compared with White patients, which may lead to disparities in accruable time on the kidney transplant waitlist before dialysis initiation.
Methods
We compared differences in accruable wait time and transplant preparation by CKD-EPI estimating equations in Chronic Renal Insufficiency Cohort participants, on the basis of estimates of kidney function by creatinine (eGFRcr), cystatin C (eGFRcys), or both (eGFRcr-cys). We used Weibull accelerated failure time models to determine the association between race (non-Hispanic Black or non-Hispanic White) and time to ESKD from an eGFR of ≤20 ml/min per 1.73 m2. We then estimated how much higher the eGFR threshold for waitlisting would be required to achieve equity in accruable preemptive wait time for the two groups.
Results
By eGFRcr, 444 CRIC participants were eligible for waitlist registration, but the potential time between eGFR ≤20 ml/min per 1.73 m2 and ESKD was 32% shorter for Blacks versus Whites. By eGFRcys, 435 participants were eligible, and Blacks had 35% shorter potential wait time compared with Whites. By the eGFRcr-cys equation, 461 participants were eligible, and Blacks had a 31% shorter potential wait time than Whites. We estimated that registering Blacks on the waitlist as early as an eGFR of 24–25 ml/min per 1.73 m2 might improve racial equity in accruable wait time before ESKD onset.
Conclusions
Policies allowing for waitlist registration at higher GFR levels for Black patients compared with White patients could theoretically attenuate disparities in accruable wait time and improve racial equity in transplant access.
According to current United Network for Organ Sharing policy, patients are eligible to be registered on the kidney transplant waitlist when their GFR falls to ≤20 ml/min by any method of estimating kidney function. Additional wait time for deceased organ priority can be accumulated if patients are registered on the waitlist before dialysis initiation (or the wait time will automatically begin to accrue on the date of dialysis initiation). Currently, most transplant centers in the United States do not begin evaluating for kidney transplantation candidacy until the eGFR is close to 20 ml/min per 1.73 m2. However, the transplant candidacy evaluation may take several months between referral and waitlist registration, if appropriate. Thus, the point in the course of disease when an individual is identified as being eligible for transplant evaluation and the rapidity of progression of CKD between GFR of 20 ml/min per 1.73 m2 and ESKD may be critical factors that influence the time available for living donor identification, waitlist registration, and accumulation of additional (preemptive) wait time for transplantation.1–4
Some studies have suggested that CKD progresses more rapidly in the Black population, especially during the advanced stages of disease, which may disadvantage Blacks compared with Whites in terms of the time that could theoretically be used to prepare for transplant.5–8 Black patients are also referred later for transplant evaluation than White patients, which may additionally contribute to the disparities in access to transplantation.2–4,9
There has also been significant debate recently over the use of a term for Black race in the equation most commonly used to estimate kidney function, and whether such a term contributes to disparities in transplant access.5,10,11 Although the consideration of race in the CKD-Epidemiology Collaboration (CKD-EPI) equation accounts for non-GFR determinants of serum creatinine, functionally, a Black individual will be registered later on the transplant waitlist compared with a White individual with the same age, sex, and serum creatinine concentration.10,12
Thus, at least three factors may contribute to racial disparities in transplant access during the late stages of CKD: faster progression of CKD, systematically higher estimates of GFR using recommended estimating equations as is done in routine clinical practice, and referral at lower eGFR. One approach that could attenuate these sources of disparity would be to allow for waitlist registration of Black individuals on the kidney transplant waitlist at higher values of eGFR, which would theoretically improve the equity in time that could be accrued preemptively before onset of kidney replacement therapy, but would also require a change in national waitlisting policy.
In this study, we first aimed to quantify the disparities in accruable wait time before onset of ESKD on the basis of current national policies using different approaches to estimate GFR. We then estimated the eGFR at which Black patients would need to be referred and waitlisted to equalize the potential wait times among Black and White patients using different GFR estimating strategies in the context of a thought experiment.
Methods
Study Population
The Chronic Renal Insufficiency Cohort (CRIC) Study is a multicenter observational cohort study in the United States.13 Participants with an eGFR between 20 and 70 ml/min per 1.73 m2 were recruited between June 2003 and September 2008. Inclusion and exclusion criteria have been previously published.13,14 CRIC participants were followed annually at in-person study visits, during which medical history, medication use, comorbidity, and laboratory data were assessed, including serum creatinine, cystatin C, and proteinuria.15 Because we were interested in racial disparities in time spent in the advanced stages of CKD experienced by Black patients, and only Black race is included as a factor in GFR estimation (e.g., CKD-EPI equations) under routine clinical practice, we only included non-Hispanic White (henceforth referred to as White) and non-Hispanic Black (henceforth referred to as Black) CRIC participants.11,16 Race was self-reported by CRIC participants. We excluded visits where a participant was ≥75 years, given the lower likelihood that such individuals would be deemed eligible for kidney transplantation. Informed consent was obtained from participants at study enrollment. We used deidentified data from the National Institute of Diabetes and Digestive and Kidney Diseases Central Repository in this study. The University of California San Francisco Institutional Review Board considers this study exempt human subject research. All analyses were conducted in STATA 16 (Statacorp, LLC, TX).
Identification of CRIC Participants Eligible for Waitlist Registration by Three Different Estimating Equations under Current National Policy
We examined time from onset of waitlist eligibility to need for kidney replacement therapy using three different criteria, two approaches that included the consideration of race and one that did not (Supplemental Figure 1). Specifically, we constructed three different cohorts of individuals who were eligible for wait time accrual after evaluation by a transplant center (assuming timely referral and no contraindications to transplantation in a theoretical context) using three strategies of estimation of kidney function. First, we used the 2012 CKD-EPI creatinine-based equation (eGFRcr),15 applying an adjustment term for Black race when applicable, to construct a cohort that would be eligible for waitlist registration.11 We then used the 2012 CKD-EPI cystatin-C based equation (eGFRcys), which does not apply any adjustment for race (Supplemental Figure 1).11 Finally, we repeated our estimates of eGFR using the 2012 eGFRcr-cys combined equation, which also applies an adjustment term for Black race.
We included participants who had at least one eGFR <25 ml/min per 1.73 m2 and a subsequent eGFR ≤20 ml/min per 1.73 m2 to construct our three cohorts. This requirement for a downward eGFR trajectory was to ensure that participants had progressive loss of kidney function and would reasonably be referred by a provider for transplant candidacy consideration.
Primary Predictor and Outcome
The primary predictor was race (non-Hispanic White or non-Hispanic Black) obtained via self-report. The outcome was time to onset of ESKD (dialysis or transplant) from the qualifying visit where eGFR first fell to ≤20 ml/min per 1.73 m2. ESKD was ascertained by participant self-report or their named contact to determine the date of dialysis initiation. Deaths were identified through report from next of kin, retrieval of death certificates or obituaries, review of hospital records, and linkage with the Social Security Death Index.
Confirming the Presence of Racial Disparities in Time Spent in the Zone of Eligibility under Current Waitlisting Policies
We examined baseline characteristics of participants included in each of the three cohorts we constructed by race using chi-squared and Kruskal-Wallis tests as appropriate.
We then sought to confirm the presence of racial differences in time spent in the advanced stages of CKD by using Weibull accelerated failure time models, with race as the predictor and time to ESKD as the outcome in unadjusted analyses. Weibull models are advantageous because they describe differences between groups by proportional reductions in time rather than hazard ratios. Time began in these models when eGFR fell to the qualifying eGFR (≤20 ml/min per 1.73 m2), and participants were followed until ESKD, death, or loss to follow-up. We considered unadjusted models to be our primary analyses given that we were interested in disparities in time to onset of ESKD by race on the basis of estimations of GFR using the guideline-recommend estimating equations (CKD-EPI). At the majority of transplant centers and according to the United Network for Organ Sharing, other risk factors for CKD progression, such as age or proteinuria, are not considered in policies surrounding eligibility for waitlist registration before dialysis initiation. Hence, we considered our unadjusted models to more closely mimic current practice and policies surrounding waitlisting.
In sensitivity analyses, we adjusted our Weibull accelerated failure models for age at the visit with the qualifying eGFR and sex (Model 1), and in additional sensitivity analysis (Model 2), other socioeconomic factors (annual household income, education, insurance), logarithm of urine protein/creatinine ratio, history of diabetes, peripheral vascular disease, coronary artery disease, heart failure, and receipt of angiotensin-receptor blockers or angiotensin-converting enzyme inhibitors (as a marker of care quality). We did not adjust these models for eGFR as the eGFR at time of entry serves as a marker and mediator of the rate of progression and time spent in the “zone of eligibility” (defined as the time between when eGFR fell below 20 ml/min per 1.73 m2 at the qualifying visit and onset of ESKD) for the accrual of wait time before ESKD. We also computed the absolute difference in time spent in the zone of eligibility for each of the three cohorts above.
Exploring Alternative Options for Waitlist Eligibility Determination that Could Provide Improved Equity in Time Spent in the Zone of Eligibility Across Different Racial Groups
Next, in a data-driven approach, we applied an algorithm to solve for how much higher the waitlist threshold for kidney function (eGFR) would need to be in Blacks compared with Whites to achieve an unadjusted time ratio of 1 in Weibull models (equity in time spent in the zone of eligibility by race) for each of the three approaches to estimating eGFR as described above in separate analyses. Once we identified the adjustment factor that was needed to achieve a theoretical time ratio of 1, we used bootstrapping with 1000 repetitions to determine the 95% confidence intervals (95% CIs) for each of the three adjustment factors separately. We then applied this adjustment factor to the eGFR of Black participants when determining entry into the zone of eligibility and re-examined time to ESKD using Weibull and Cox models in unadjusted and adjusted analyses to confirm this approach achieved theoretical equity in time spent in the zone of eligibility for Blacks and Whites.
Finally, because the adjustment factor that we derived in our algorithm (1.20–1.25) was close to the coefficient applied to account for Black race in CKD-EPI equations (e.g., a 16% upwards adjustment is applied to the eGFRcr estimation of kidney function for Blacks), we repeated our Weibull and Cox models (unadjusted and adjusted) including only participants who qualified for inclusion by having at least one eGFR <25 ml/min per 1.73 m2 and one eGFR ≤20 ml/min per 1.73 m2 at any subsequent visit, but computing estimates of kidney function using eGFRcr or eGFRcr-cys without consideration of race (i.e., dropping the race term). We did not perform this analysis using eGFRcys equations because these equations do not include a race term.
Results
Differences in Potential Wait Time by eGFRcr Estimating Equation
When we used the eGFRcr equation to identify individuals who were theoretically eligible for accumulation of time on the waitlist before ESKD, we found that 444 participants were eligible for inclusion. The median age was 62 years and eGFR was 16.6 ml/min per 1.73 m2, with the eGFR being lower for Black (15.8 ml/min per 1.73 m2) compared with White participants (17.6 ml/min per 1.73 m2) at the qualifying visit (Table 1). The protein/creatinine ratio was higher for Blacks compared with Whites.
Table 1.
Overall characteristics of the study cohorts by various CKD-EPI estimating equations and by race
| Characteristic | ||||
|---|---|---|---|---|
| Cohort using eGFRcr as inclusion criteria | ||||
| Overall, n=444 | White, n=157 | Black, n=287 | P value | |
| Median age (IQR) yrs | 62 (55–68) | 63 (56–68) | 62 (54–68) | 0.15 |
| Female, n (%) | 211 (48) | 63 (40) | 148 (52) | 0.02 |
| Median eGFR by CKD-EPI (IQR) in ml/min per 1.73 m2 | 16.6 (13.9–18.4) | 17.6 (15.4–18.8) | 15.8 (13.4–17.9) | <0.001 |
| Median urine protein/creatinine ratio (g/g) (IQR) | 1.29 (0.42–3.02) | 1.11 (0.27–2.25) | 1.36 (0.50–3.36) | 0.03 |
| Diabetes, n (%) | 268 (60.4) | 84 (53.5) | 184 (64.1) | 0.03 |
| Heart failure, n (%) | 71 (16) | 14 (8.9) | 57 (19.9) | 0.003 |
| MI, n (%) | 121 (27.3) | 41 (26.1) | 80 (27.9) | 0.74 |
| PVD, n (%) | 48 (10.8) | 15 (9.6) | 33 (11.5) | 0.63 |
| ESKD, n (%) | 313 (71) | 96 (61) | 217 (76) | 0.002 |
| Death, n (%) | 34 (7.7) | 11 (7.0) | 23 (8.0) | 0.85 |
| Cohort using eGFRcys as inclusion criteria | ||||
| Overall, n=435 | White, n=169 | Black, n=266 | P value | |
| Median age (IQR) yrs | 63 (56–68) | 64 (58–70) | 62 (56–68) | 0.06 |
| Female, n (%) | 219 (50) | 73 (43) | 146 (55) | 0.02 |
| Median eGFR by CKD-EPI (IQR) in ml/min per 1.73 m2 | 16.8 (15.0–18.4) | 17.4 (15.1–18.6) | 16.6 (14.6–18.3) | 0.04 |
| Median urine protein/creatinine ratio (g/g) (IQR) | 1.08 (0.36–2.52) | 0.84 (0.26–2.19) | 1.17 (0.42–2.71) | 0.06 |
| Diabetes, n (%) | 284 (65.3) | 106 (62.7) | 178 (66.9) | 0.41 |
| Heart failure, n (%) | 93 (21.4) | 25 (14.8) | 68 (25.6) | 0.008 |
| MI, n (%) | 147 (33.8) | 62 (36.7) | 85 (32) | 0.35 |
| PVD, n (%) | 57 (13.1) | 25 (14.8) | 32 (12) | 0.47 |
| ESKD, n (%) | 291 (67) | 102 (60) | 189 (71) | 0.02 |
| Death, n (%) | 60 (14) | 26 (15) | 34 (13) | 0.48 |
| Cohort using eGFRcr-cys as inclusion criteria | ||||
| Overall, n=461 | White, n=177 | Black, n=284 | P value | |
| Median age (IQR) yrs | 62 (55–68) | 64 (55–69) | 62 (54–68) | 0.15 |
| Female, n (%) | 229 (49.7) | 76 (42.9) | 153 (53.9) | 0.03 |
| Median eGFR by CKD-EPI (IQR) in ml/min per 1.73 m2 | 17 (14.6–18.5) | 17.4 (15.4–18.5) | 16.5 (14.0–18.4) | 0.01 |
| Median urine protein/creatinine ratio (g/g) (IQR) | 1.15 (0.37–2.59) | 0.90 (0.27–2.18) | 1.33 (0.45–2.89) | 0.009 |
| Diabetes, n (%) | 291 (63.1) | 105 (59.3) | 186 (65.5) | 0.20 |
| Heart failure, n (%) | 83 (18) | 23 (13) | 60 (21.1) | 0.03 |
| MI, n (%) | 136 (29.5) | 54 (30.5) | 82 (28.9) | 0.75 |
| PVD, n (%) | 53 (11.5) | 24 (13.6) | 29 (10.2) | 0.30 |
| ESKD, n (%) | 323 (70.1) | 111 (62.7) | 212 (74.7) | 0.009 |
| Death, n (%) | 47 (10.2) | 20 (11.3) | 27 (9.5) | 0.53 |
MI, myocardial infarction; PVD, peripheral vascular disease.
Using Weibull models, the time to ESKD was 32% shorter in Blacks compared with Whites (95% CI, 13% to 47%) in unadjusted analyses (Table 2). The median time spent in the zone of eligibility for wait time accrual was approximately 9 months shorter for Blacks compared with Whites using eGFRcr (27, interquartile range [IQR], 11–55 versus 18 months, IQR, 7–38) (Table 3). When using Cox models, Blacks were 1.44 times (95% CI, 1.13 to 1.84) more likely to progress to ESKD than Whites, which was consistent with observations from Weibull models (Table 2).
Table 2.
Time to ESKD starting from eGFR of 20 ml/min per 1.73 m2 by Weibull accelerated failure and Cox models
| Entry into Cohort when eGFR is ≤20 ml/min per 1.73 m2 | Entry into Cohort when eGFR is ≤20 ml/min per 1.73 m2 with Application of an Adjustment Factor to eGFRs for Blacksa | |||
|---|---|---|---|---|
| Blacks Compared with Whites | Weibull Model Time Ratio (95% CI) | Cox model Hazard Ratio (95% CI) | Weibull Model Time Ratio (95% CI) | Cox model Hazard Ratio (95% CI) |
| Univariable model | ||||
| eGFRcr | 0.68 (0.53 to 0.87) | 1.44 (1.13 to 1.84) | 1.00 (0.79 to 1.27) | 1.00 (0.79 to 1.26) |
| eGFRcys | 0.65 (0.49 to 0.84) | 1.46 (1.15 to 1.86) | 1.00 (0.79 to 1.27) | 1.00 (0.80 to 1.25) |
| eGFRcr-cys | 0.69 (0.54 to 0.87) | 1.41 (1.12 to 1.78) | 1.00 (0.80 to 1.25) | 1.00 (0.81 to 1.25) |
| Multivariable model 1b | ||||
| eGFRcr | 0.70 (0.55 to 0.90) | 1.41 (1.10 to 1.80) | 1.05 (0.83 to 1.34) | 0.95 (0.75 to 1.20) |
| eGFRcys | 0.67 (0.51 to 0.88) | 1.42 (1.11 to 1.81) | 1.04 (0.82 to 1.31) | 0.97 (0.77 to 1.22) |
| eGFRcr-cys | 0.69 (0.54 to 0.88) | 1.41 (1.12 to 1.78) | 1.00 (0.80 to 1.25) | 1.00 (0.81 to 1.25) |
eGFR of Black individuals was computed by standard CKD-EPI estimating equations and subsequently multiplied by an adjustment factor that we searched for to equalize time spent in the zone of eligibility by race; thus, Blacks would be eligible for entry into the cohort earlier than under current national waitlisting policy.
Includes age at entry and sex as covariates.
Table 3.
Median time between eGFR of ≤20 ml/min per 1.73 m2 and ESKD by different CKD-EPI estimating equations
| Median mo to ESKD (IQR) from Visit where Qualifying eGFR Was ≤20 ml/min per 1.73 m2 | White | Black |
|---|---|---|
| eGFRcr | 27 (11–55) | 18 (7–38) |
| eGFRcys | 32 (12–69) | 21 (8–44) |
| eGFRcr-cys | 29 (12–61) | 20 (8–42) |
Differences in Potential Wait Time by eGFRcys
When we used the eGFRcys equation to determine the cohort of patients eligible for waitlist registration, we found 435 eligible individuals, but these individuals were not necessarily the same individuals as the ones identified using eGFRcr equation and did not necessarily enter our study at the same visit. The eGFRcys was statistically significantly lower for Blacks (16.6 ml/min per 1.73 m2) compared with Whites (17.4 ml/min per 1.73 m2), but the difference was not as great as when using eGFRcr equation with adjustment for race (Table 1). The degree of proteinuria was higher for Blacks compared with Whites, although this did not achieve statistical significance.
Using Weibull models, the time to ESKD was 35% shorter in Blacks compared with Whites (95% CI, 16% to 51%) in unadjusted analyses (Table 2). There was a median 11-month difference in potentially accruable wait time for Blacks compared with Whites (Table 3). When using Cox models, Blacks were also more likely to progress to ESKD than Whites (hazard ratio, 1.46; 95% CI, 1.15 to 1.86, Table 2).
Differences in Potential Wait Time by eGFRcr-cys
When we used the eGFRcr-cys equation to determine the cohort of patients eligible for waitlist registration, we found 461 eligible individuals, but these individuals were not necessarily the same as the ones identified using other approaches and did not necessarily enter our study at the same visit. The eGFRcr-cys was statistically significantly lower for Blacks (16.5 ml/min per 1.73 m2) compared with Whites (17.4 ml/min per 1.73 m2), and degree of proteinuria was higher for Blacks compared with Whites (Table 1).
Using Weibull models, the time to ESKD was 31% shorter in Blacks compared with Whites (95% CI, 13% to 46%) in unadjusted analyses (Table 2). There was a median 9-month difference in potentially accruable wait time for Blacks compared with Whites (Table 3). When using Cox models, Blacks were also more likely to progress to ESKD than Whites (Table 2, Supplemental Table 1).
Alternative Approaches to Determining Waitlist Eligibility
When we searched for the “adjustment factor” needed to equalize time spent in the advanced stages of CKD, we found the factor that needed to be applied to the current threshold for waitlist eligibility (eGFR ≤20 ml/min per 1.73 m2) was 1.19 (95% CI, 0.80 to 2.32) if determining kidney function by eGFRcr, 1.25 (95% CI, 0.84 to 2.44) if using eGFRcys, and 1.25 (95% CI, 0.85 to 2.33) if using eGFRcr-cys equations. These adjustment factors correspond to allowing Blacks to be registered on the waitlist when their eGFR is ≤24 to 25 ml/min per 1.73 m2 (although continuing to waitlist Whites when their eGFR is ≤20 ml/min per 1.73 m2).
The baseline characteristics of cohorts identified after applying this adjustment factor to the eGFR threshold for entry into the cohort by various CKD-EPI equations are shown in Table 4. Of note, after applying this adjustment factor, differences in proteinuria between Blacks and Whites were attenuated across all cohorts.
Table 4.
Overall characteristics of the study cohorts by various CKD-EPI estimating equations and by race, allowing Black patients to enter at higher levels of eGFR on the basis of applied adjustments
| Characteristic | ||||
|---|---|---|---|---|
| Cohort using eGFRcr as inclusion criteria | ||||
| Overall, n=562 | White, n=157 | Black, n=405 | P value | |
| Median age (IQR) yrs | 62 (55–68) | 63 (56–68) | 62 (54–68) | 0.11 |
| Female, n (%) | 270 (48.0) | 63 (40.1) | 207 (51.1) | 0.02 |
| Median eGFRa by CKD-EPI (IQR) in ml/min per 1.73 m2 | 16.4 (13.5–18.4) | 17.6 (15.4–18.8) | 15.9 (12.9–18.0) | <0.001 |
| Median urine protein/creatinine ratio (g/g) (IQR) | 1.11 (0.35–2.71) | 1.11 (0.27–2.25) | 1.12 (0.36–2.78) | 0.46 |
| Diabetes, n (%) | 341 (60.7) | 84 (53.5) | 257 (63.5) | 0.03 |
| Heart failure, n (%) | 91 (16.2) | 14 (8.9) | 77 (19.0) | 0.003 |
| MI, n (%) | 152 (27.1) | 41 (26.1) | 111 (27.4) | 0.83 |
| PVD, n (%) | 61 (10.9) | 15 (9.6) | 46 (11.4) | 0.65 |
| ESKD, n (%) | 373 (66.4) | 96 (61.2) | 277 (68.4) | 0.11 |
| Death, n (%) | 50 (8.9) | 11 (7.0) | 39 (9.6) | 0.41 |
| Cohort using eGFRcys as inclusion criteria | ||||
| Overall, n=598 | White, n=169 | Black, n=429 | P value | |
| Median age (IQR) yrs | 63 (56–69) | 64 (58–70) | 62 (55–68) | 0.12 |
| Female, n (%) | 298 (49.8) | 73 (43.2) | 225 (52.5) | 0.05 |
| Median eGFRa by CKD-EPI (IQR) in ml/min per 1.73 m2 | 16.6 (14.0–18.2) | 17.4 (15.1–18.6) | 16.4 (13.5–18.1) | <0.001 |
| Median urine protein/creatinine ratio (g/g) (IQR) | 0.95 (0.29–2.48) | 0.84 (0.26–2.19) | 0.99 (0.30–2.61) | 0.41 |
| Diabetes, n (%) | 392 (65.6) | 106 (62.7) | 286 (66.7) | 0.39 |
| Heart failure, n (%) | 129 (21.6) | 25 (14.8) | 104 (24.2) | 0.01 |
| MI, n (%) | 207 (34.6) | 62 (36.7) | 145 (33.8) | 0.51 |
| PVD, n (%) | 74 (12.4) | 25 (14.8) | 49 (11.4) | 0.27 |
| ESKD, n (%) | 379 (63.4) | 102 (60.4) | 277 (64.6) | 0.35 |
| Death, n (%) | 82 (13.7) | 26 (15.4) | 56 (13.1) | 0.51 |
| Cohort using eGFRcr-cys as inclusion criteria | ||||
| Overall, n=632 | White, n=177 | Black, n=455 | P value | |
| Median age (IQR) yrs | 62 (55–68) | 64 (55–69) | 62 (54–68) | 0.14 |
| Female, n (%) | 308 (48.7) | 76 (42.9) | 232 (51.0) | 0.08 |
| Median eGFRa by CKD-EPI (IQR) in ml/min per 1.73 m2a | 16.3 (13.5–18.2) | 17.4 (15.4–18.5) | 15.7 (13.0–18.0) | <0.001 |
| Median urine protein/creatinine ratio (g/g) (IQR) | 0.98 (0.30–2.69) | 0.90 (0.27–2.18) | 0.98 (0.33–2.77) | 0.27 |
| Diabetes, n (%) | 402 (63.6) | 105 (59.3) | 297 (65.3) | 0.17 |
| Heart failure, n (%) | 125 (19.8) | 23 (13.0) | 102 (22.4) | 0.008 |
| MI, n (%) | 193 (30.5) | 54 (30.5) | 139 (30.6) | 1.000 |
| PVD, n (%) | 81 (12.8) | 24 (13.6) | 57 (12.5) | 0.79 |
| ESKD, n (%) | 414 (65.5) | 111 (62.7) | 303 (66.6) | 0.40 |
| Death, n (%) | 77 (12.2) | 20 (11.3) | 57 (12.5) | 0.79 |
MI, myocardial infarction; PVD, peripheral vascular disease.
Downward adjusted eGFR with application of our adjustment factor to simulate entry of Blacks into the cohort at a higher eGFR threshold.
Using this approach to identify our cohort of waitlist-eligible individuals, the time to ESKD was equivalent in Blacks compared with Whites in unadjusted analyses as might be expected, but was also very similar for Blacks compared with Whites in adjusted analyses (Table 2, Supplemental Table 1A and B) across all CKD-EPI estimating equations.
Differences in Potential Wait Time by CKD-EPI without Consideration of Race (Sensitivity Analysis)
When we used an approach of dropping the race term from the eGFRcr equation, differences in proteinuria were also attenuated when comparing Blacks with Whites (Supplemental Table 2A). In Weibull models, the time to ESKD among Blacks was 0.94 times that of Whites (95% CI, 0.74 to 1.19) in unadjusted analyses and no longer statistically significant (Supplemental Table 2B). When using Cox models, there was also no difference in time to ESKD by race (Supplemental Table 2B).
When estimating kidney function using eGFRcr-cys equations without consideration of race, differences in proteinuria by race persisted when comparing Blacks with Whites (Supplemental Table 2A). In addition, time differences spent in the “zone of eligibility” persisted in unadjusted and minimally adjusted models (Supplemental Table 2B) even when we dropped the race term during the computation of eGFR using eGFRcr-cys equations.
Discussion
The Kidney Disease Improving Global Outcomes (KDIGO) guidelines currently recommend using creatinine-based equations to estimate kidney function, and most electronic health record systems have switched to the use of the eGFRcr equation to automate estimates of eGFR.16,17 However, renewed controversy over the routine consideration of race in the estimation of kidney function has surfaced, given race is a social rather than a biologic construct and medicine is increasing its attention on systemic racism.10,18–21 These debates over the role of race in medicine are not limited to nephrology, but extend to multiple domains of medicine and have led to the decision by some institutions to change automated reporting of kidney function in electronic health records to not consider race.10,20,22,23 In this study, we confirmed that kidney progression is faster in Black compared with White patients during the late stages of CKD, and examined disparities in access to the kidney transplant waitlist within the context of using various different approaches to determine eligibility for preemptive waitlisting or to estimate kidney function.
Our study provides a number of important observations surrounding disparities in access to deceased donor transplantation and time available for donor identification before the need for dialysis. First, we confirmed the presence of racial disparities in potential time spent in the advanced stages of CKD using the most common clinical approach to estimating kidney function by serum creatinine.11,17 Specifically, we found that Blacks had less time than Whites between eGFR ≤20 ml/min per 1.73 m2 and onset of ESKD by the eGFRcr equation. When we used alternative equations such as the eGFRcys (which does not consider race) or eGFRcr-cys combined equation (which does consider race), the disparities in time spent in the advanced stages of CKD persisted. However, we found that use of a higher eGFR threshold of 24–25 ml/min per 1.73 m2 (depending on the specific CKD-EPI equation used), and therefore earlier eligibility for waitlisting in Blacks could reduce the disparity in time spent in the advanced stages of CKD. This increase in the threshold for Blacks is of a larger magnitude than the adjustment for Black race made by current CKD-EPI equations. Thus, merely eliminating the Black race term in standard CKD-EPI equations may be insufficient to achieve true equity in accruable wait time. We did confirm that estimating kidney function using an eGFRcr equation without the consideration of race did reduce the racial disparity in time spent in the zone of eligibility, but this approach was not as effective when using eGFRcr-cys equations.
One possible explanation for our observations is that there are inaccuracies in the estimation of kidney function by the CKD-EPI creatinine- or cystatin C- based equations in patients with very low eGFR that differ by race. Because these equations were derived from amalgamated data from a number of cohorts whose members mostly had mild to moderate stages of CKD,11,24 it is possible the performance of the CKD-EPI equations in the very late stages of CKD are not as robust as in earlier stages of CKD, and may overestimate kidney function in the Black relative to the White population.
In contrast, prior studies have shown that loss of kidney function by measured GFR is faster for Blacks compared with Whites, and hence errors in eGFR estimates are unlikely to explain this disparity.8,25 Thus, a second possible and more likely explanation for our findings is that progression of CKD in the advanced stages is faster among the Black population than the White population, which would be consistent with the higher level of proteinuria in the Blacks across all cohorts. If this explanation is correct, then allowing entry of Blacks into the zone of eligibility for preemptive wait time accrual earlier (i.e., at eGFR of 24–25 ml/min per 1.73 m2) than Whites could provide enough lead time to offset the more rapid progression of their CKD and potentially improve equity in time spent in the late stages of CKD. Such a strategy could be adopted into national policy if found to be effective at increasing access of Blacks to waitlist registration and kidney transplantation. Black patients may also need more time to complete the evaluation process,26 so that earlier referral and eligibility for waitlist registration even beyond the level needed to account for faster progression of kidney disease could help improve equity in access to transplantation.27 Although we acknowledge that some of the observed racial disparity in access to kidney transplantation may be related to socioeconomic status, these factors did not fully explain the difference in potential waitlist time that we observed, but may be operational at other steps in the complex transplant evaluation process. Thus, continual education to emphasize the importance of early referral of Black patients for transplant evaluation and advocacy for policy changes that promote access to the kidney transplant waitlist in this population is warranted.
We highlight that current waitlisting policies do not specify which method of measuring or estimating kidney function needs to be used for waitlist registration. Clinicians caring for patients and contemplating referral for transplant evaluation or waitlist registration have the option of using confirmatory testing (i.e., more than one estimate of kidney function) to enhance confidence in and potential accuracy of the estimates of kidney function as recommended by KDIGO. Some institutions have considered adopting the CKD-EPI cystatin C equation to avoid consideration of race when estimating eGFR and promote use of cystatin C over serum creatinine.28 The derivation of the eGFRcys was based primarily on cohorts where more individuals had mild to moderate rather than advanced CKD,11,24 so it is unclear whether cystatin-based estimates outperform creatinine-based estimates at very low levels of eGFR, which may be of concern. On the basis of our findings, the cystatin C–based CKD-EPI estimating equation does not appear to improve racial equity in accruable transplant wait time in the very advanced stages of CKD. Further studies are needed to corroborate our findings.
Our study has several strengths, including the inclusion of participants of a well-characterized cohort of patients with CKD followed longitudinally and simultaneous measurement of serum creatinine and cystatin C at prespecified intervals during research visits. This avoids bias that may be present in electronic health record data, where more frequent measurements of kidney function in patients with better access to care or greater severity of illness may increase the chances of a patient being recognized as eligible for transplant referral, evaluation, and waitlist registration. However, our study is a theoretical thought experiment and does not aim to examine whether patients were referred, underwent waitlist registration, or were deemed candidates for transplantation. Our study is also not designed to examine other aspects of estimating kidney function outside of the question of wait time accrual in the setting of advanced CKD. Our data do not address the potential consequences of ignoring the race term when estimating GFR, as such an approach may affect the accuracy of kidney function estimates, which may have adverse consequences in other important contexts (such as drug dosing, living donor eligibility determination, or clinical trial participation).19,25 We note that our cohort construction identifies patients who have progression of their CKD (because two eGFRs are required to meet our criteria before a patient enters into the study), and our study results do not necessarily apply to those without evidence of progression of their CKD over time. Given the observational nature of our study, residual confounding may be present.
In summary, our study supports the potential utility of using different thresholds for waitlist registration by race as a means of reducing disparities that could occur in wait time accrual for deceased donor transplantation and predialysis transplant preparation time. Our data do not support the routine use of cystatin C estimates, at least not via the 2012 estimating equation for the purposes of improving equity in access to kidney transplantation despite the lack of consideration of race in these eGFR estimates. Finally, we believe racial disparities in access to kidney transplantation are likely multifactorial, and additional factors may need to be addressed including differences in social determinants of health, referring provider practices, racism or perceived racism in the evaluation process, and other factors beyond the estimation of kidney function. Consideration of changes to national policies in the future that allow for earlier waitlisting of Black individuals for kidney transplantation could be one strategy to reduce disparities in access to transplantation if further studies can confirm our findings and demonstrate such a strategy would translate into improved parity in actual receipt of kidney transplantation. Additional studies and strategies are needed to improve equity in clinical outcomes of patients with CKD at all stages of disease.
Disclosures
D. Adey reports receiving research funding from Hansa Pharmaceuticals; honoraria from Natera; and being a scientific advisor or member of the American Society of Transplantation on the Governance Board, American Board of Internal Medicine on the Subspecialty Governance Board and Transplant International Journal on the Editorial Board. K. Johansen reports being a scientific advisor or member of the Steering Committee for GlaxoSmithKline prolyl hydroxylase inhibitor clinical trials program; and other interests/relationships as Associate Editor of JASN. Because K. Johansen is an editor of the JASN, they were not involved in the peer review process for this manuscript. A guest editor oversaw the peer review and decision-making process for this manuscript. E. Ku reports consultancy agreements with Tricida; and reports receiving research funding from CareDX. L. Li reports having an ownership interest in Anthem Inc., Johnson & Johnson, Pfizer Inc., Peloton Interactive Inc., Teladoc Health Inc., and UnitedHealth Group Inc.. All remaining authors have nothing to disclose.
Funding
E. Ku and K.L. Johansen, and C.E. McCulloch were funded by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) grant R01 DK115629. This work was also supported by NIDDK grant K24 DK85153 (to K.L. Johansen).
Acknowledgments
The CRIC study was conducted by the CRIC Investigators and supported by NIDDK. The data from the CRIC study reported here were supplied by NIDDK Central Repositories. This manuscript was not prepared in collaboration with Investigators of the CRIC study and does not necessarily reflect the opinions or views of the CRIC study, the NIDDK Central Repositories, or the NIDDK.
Footnotes
Published online ahead of print. Publication date available at www.jasn.org.
See related editorial, “Promoting Equity in Eligibility for Registration on the Kidney Transplantation Waiting List: Looking beyond eGFRcr,” on pages 523–525.
Supplemental Material
This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2020081144/-/DCSupplemental.
Supplemental Figure 1. Cohort derivation.
Supplemental Table 1. (A) Risk of ESKD by race using cohorts defined by different inclusion criteria in fully adjusted Weibull accelerated failure and Cox models. (B) Median time in the zone of eligibility after application of the adjustment factor designed to equalize time spent in the zone of eligibility to determine cohort entry.
Supplemental Table 2. (A) Characteristics of the study cohort by computing eGFR without consideration of race term to determine eligibility for entry into the cohort. (B) Time to ESKD starting from eGFR of ≤20 ml/min per 1.73 m2 using an approach that does not consider the race when computing estimates of eGFR using Weibull accelerated failure and Cox models.
References
- 1.Farouk SS, Atallah S, Campbell KN, Vassalotti JA, Uribarri J: Implementation of a quality improvement strategy to increase outpatient kidney transplant referrals. BMC Nephrol 21: 192, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Basu M, Petgrave-Nelson L, Smith KD, Perryman JP, Clark K, Pastan SO, et al.: Transplant center patient navigator and access to transplantation among high-risk population: A randomized, controlled trial. Clin J Am Soc Nephrol 13: 620–627, 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Peev V: We wait too long to refer patients for transplantation. Semin Dial 29: 318–319, 2016 [DOI] [PubMed] [Google Scholar]
- 4.Knight RJ, Teeter LD, Graviss EA, Patel SJ, DeVos JM, Moore LW, et al.: Barriers to preemptive renal transplantation: A single center questionnaire study. Transplantation 99: 576–579, 2015 [DOI] [PubMed] [Google Scholar]
- 5.van den Beukel TO, de Goeij MC, Dekker FW, Siegert CE, Halbesma N; PREPARE Study Group: Differences in progression to ESRD between black and white patients receiving predialysis care in a universal health care system. Clin J Am Soc Nephrol 8: 1540–1547, 2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Derose SF, Rutkowski MP, Crooks PW, Shi JM, Wang JQ, Kalantar-Zadeh K, et al.: Racial differences in estimated GFR decline, ESRD, and mortality in an integrated health system. Am J Kidney Dis 62: 236–244, 2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Parsa A, Kao WH, Xie D, Astor BC, Li M, Hsu CY, et al.; AASK Study Investigators; CRIC Study Investigators: APOL1 risk variants, race, and progression of chronic kidney disease. N Engl J Med 369: 2183–2196, 2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Hunsicker LG, Adler S, Caggiula A, England BK, Greene T, Kusek JW, et al.: Predictors of the progression of renal disease in the modification of diet in renal disease study. Kidney Int 51: 1908–1919, 1997 [DOI] [PubMed] [Google Scholar]
- 9.Epstein AM, Ayanian JZ, Keogh JH, Noonan SJ, Armistead N, Cleary PD, et al.: Racial disparities in access to renal transplantation–clinically appropriate or due to underuse or overuse? N Engl J Med 343: 1537–1544; 1532 p preceding 1537, 2000 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Eneanya ND, Yang W, Reese PP: Reconsidering the consequences of using race to estimate kidney function. JAMA 322: 113–114, 2019 [DOI] [PubMed] [Google Scholar]
- 11.Inker LA, Schmid CH, Tighiouart H, Eckfeldt JH, Feldman HI, Greene T, et al.; CKD-EPI Investigators: Estimating glomerular filtration rate from serum creatinine and cystatin C [published correction appears in N Engl J Med 367: 2060, 2012]. N Engl J Med 367: 20–29, 2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hsu J, Johansen KL, Hsu CY, Kaysen GA, Chertow GM: Higher serum creatinine concentrations in black patients with chronic kidney disease: Beyond nutritional status and body composition. Clin J Am Soc Nephrol 3: 992–997, 2008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Feldman HI, Appel LJ, Chertow GM, Cifelli D, Cizman B, Daugirdas J, et al.; Chronic Renal Insufficiency Cohort (CRIC) Study Investigators: The Chronic Renal Insufficiency Cohort (CRIC) study: Design and methods. J Am Soc Nephrol 14[Suppl 2]: S148–S153, 2003 [DOI] [PubMed] [Google Scholar]
- 14.Lash JP, Go AS, Appel LJ, He J, Ojo A, Rahman M, et al.; Chronic Renal Insufficiency Cohort (CRIC) Study Group: Chronic Renal Insufficiency Cohort (CRIC) study: Baseline characteristics and associations with kidney function [published correction appears in Clin J Am Soc Nephrol 6: 2548–2553, 2011]. Clin J Am Soc Nephrol 4: 1302–1311, 2009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Anderson AH, Yang W, Hsu CY, Joffe MM, Leonard MB, Xie D, et al.: Estimating GFR among participants in the Chronic Renal Insufficiency Cohort (CRIC) Study. Am J Kidney Dis 60: 250–261, 2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Inker LA, Astor BC, Fox CH, Isakova T, Lash JP, Peralta CA, et al.: KDOQI US commentary on the 2012 KDIGO clinical practice guideline for the evaluation and management of CKD. Am J Kidney Dis 63: 713–735, 2014 [DOI] [PubMed] [Google Scholar]
- 17.Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group: KDIGO 2012. clinical practice guideline for the evaluation and management of chronic kidney disease. Available at: https://kdigo.org/wp-content/uploads/2017/02/KDIGO_2012_CKD_GL.pdf. Accessed April 5, 2019
- 18.Levey AS, Titan SM, Powe NR, Coresh J, Inker LA: Kidney disease, race, and GFR estimation. Clin J Am Soc Nephrol 15: 1203–1212, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Powe NR: Black kidney function matters: Use or misuse of race? JAMA 324: 737–738, 2020 [DOI] [PubMed] [Google Scholar]
- 20.Vyas DA, Eisenstein LG, Jones DS: Hidden in plain sight - reconsidering the use of race correction in clinical algorithms. N Engl J Med 383: 874–882, 2020 [DOI] [PubMed] [Google Scholar]
- 21.Evans MK, Rosenbaum L, Malina D, Morrissey S, Rubin EJ: Diagnosing and treating systemic racism. N Engl J Med 383: 274–276, 2020 [DOI] [PubMed] [Google Scholar]
- 22.Fontanarosa PB, Bauchner H: Race, ancestry, and medical research. JAMA 320: 1539–1540, 2018 [DOI] [PubMed] [Google Scholar]
- 23.University of Washington , Department of Medicine: UW Medicine to exclude race from calculation of eGFR (measure of kidney function). News and Events; May 29, 2020. Available at: https://medicine.uw.edu/news/uw-medicine-exclude-race-calculation-egfr-measure-kidney-function. Accessed July 5, 2020
- 24.Levey AS, Coresh J, Tighiouart H, Greene T, Inker LA: Strengths and limitations of estimated and measured GFR. Nat Rev Nephrol 15: 784, 2019 [DOI] [PubMed] [Google Scholar]
- 25.Levey AS, Tighiouart H, Titan SM, Inker LA: Estimation of glomerular filtration rate with vs without including patient race. JAMA Intern Med 180: 793–795, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Weng FL, Reese PP, Mulgaonkar S, Patel AM: Barriers to living donor kidney transplantation among black or older transplant candidates. Clin J Am Soc Nephrol 5: 2338–2347, 2010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Nonterah CW, Gardiner HM: Pre-transplant evaluation completion for Black/African American renal patients: Two theoretical frameworks. Patient Educ Couns 103: 988–998, 2020 [DOI] [PubMed] [Google Scholar]
- 28.Zoler M: Dropping race-based eGFR adjustment gains traction in US. Medscape. 2020. Available at: https://www.medscape.com/viewarticle/933418. Accessed July 10, 2020