Risk of ESRD in Prior Living Kidney Donors

Jennifer L. Wainright; Amanda M. Robinson; Amber R. Wilk; David K. Klassen; Wida S. Cherikh; Darren E. Stewart


American Journal of Transplantation. 2018;18(5):1129-1139. 

In This Article


Between April 1, 1994 and September 30, 2016 a total of 123 526 LKDs donated in the United States. Of those, 218 LKDs developed ESRD; 131 were added to the OPTN kidney waiting list, of which 97 received deceased donor kidney transplants. Eleven received living donor kidney transplants (8 of 11 were listed), and 26 were still on the waiting list at the end of the study period. Sixty–nine listings and 75 transplants occurred before initiation of dialysis, but 84 LKDs with ESRD were neither listed nor transplanted in the study period. Renal failure diagnoses included conditions related to hypertension (33.5%), glomerulonephritis (23.9%), and diabetes (7.8%) (see Supplemental Materials). Median follow–up time was 10.3 years. Median time between donation and onset of ESRD was 11.1 years (interquartile range: 8.3–14.6).

Univariate Analyses

Compared with LKDs who did not develop ESRD, LKDs with ESRD were younger at donation (38.0 vs 41.0 years), more likely to be black (32.1% vs 12.3%), male (58.3% vs 40.2%), obese (BMI > 30; 40.0% vs 22.8%), related to the recipient (85.8% vs 63.2%) and first–degree relatives (79.8% vs 54.9%), and less likely to live in a wealthy neighborhood (median: $57.2k vs $65.1k; Table 1).

The overall cumulative incidence of ESRD 20 years postdonation was 49.0 per 10 000 LKDs (Figure 1A). Cumulative incidence at 20 years for black (104.2) and other race (77.8) donors was notably higher than for white (38.0) and Hispanic donors (40.7; Figure 1B). First–degree relatives of the recipient had a higher cumulative incidence of ESRD at 20 years (full sibling: 56.3, parent: 64.8, child: 53.7) than more distant relatives (40.7) or unrelated donors (25.2; Figure 1C). The 20–year cumulative incidence for donors in the lowest quartile of neighborhood incomes (Q1: 83.4) was notably higher than for donors in wealthier neighborhoods (Q2: 43.5, Q3: 47.2, Q4: 30.5; Figure 1D). This univariate association between neighborhood income and risk of ESRD was found for white, black, and other race donors, but not for Hispanic donors (Figure 2). Risk increased exponentially over time, especially for the groups at highest risk.

Figure 1.

Cumulative incidence of ESRD among living kidney donors, per death–censored Kaplan–Meier analysis overall (A), by donor race (B), by donor relationship to the recipient (C), and by donor neighborhood income quartile at donation (D). ESRD, end–stage renal disease

Figure 2.

Cumulative incidence of ESRD in living kidney donors by income quartile for white (A), black (B), and Hispanic donors (C). Other ethnic groups excluded due to fewer than 10 events per quartile in accordance with Centers for Medicare and Medicaid Services (CMS) policy. ESRD, end–stage renal disease

Multivariable Analyses

The C–statistic for the model was 0.74. As expected, we found a number of statistically significant predictors of ESRD within the first 20 years after donation (Table 2; Figure 3). Compared with female donors, male LKDs had a higher risk (aHR: 1.75; 95% confidence interval [95%CI]: 1.33–2.31), and donors with higher BMI were at higher risk (aHR: 1.34 per 5 kg/m2 increase; 95%CI: 1.10–1.64). Black donors (compared at the median age of 41) had a higher risk than white donors (aHR: 2.79; 95%CI: 1.92–4.06).

Figure 3.

Risks factors for ESRD among living kidney donors: Adjusted hazard ratios and 95% confidence intervals. BP, blood pressure; eGFR, estimated glomerular filtration rate; ESRD, end–stage renal disease [Color figure can be viewed at wileyonlinelibrary.com]

There were also associations between age at donation and 20–year risk of ESRD, with a statistically significant interaction between age and race (P = .003). There was a significant positive association between age at donation and risk of ESRD (ie, higher risk with older age at donation) for white donors (aHR: 1.26 per decade of life; 95%CI: 1.04–1.54; Figure 4A). In contrast, we found a significant negative association between age at donation and risk (ie, higher risk with younger age at donation) for black donors (aHR: 0.75 per decade; 95%CI: 0.57–0.99; Figure 4B). Higher risk was associated with older age at donation for Hispanic donors (aHR: 1.18 per decade; 95%CI: 0.80–1.73; Figure 4C) and with younger age for donors of "other" racial groups (aHR: 0.67 per decade; 95%CI: 0.38–1.20; Figure 4D), but these 2 effects did not reach statistical significance.

Figure 4.

Risk of living kidney donors' ESRD relative to median donor age (41) and 95% confidence intervals for white (A), black (B), Hispanic (C), and other race donors (D). Donor age was restricted to the 5th and 95th percentiles. aHR, adjusted hazard ratio; ESRD, end–stage renal disease [Color figure can be viewed at wileyonlinelibrary.com]

There was an association between neighborhood income and risk of ESRD for donors in neighborhoods with below median income (aHR: 0.87 per $10k increase; 95%CI: 0.77–0.99), though this effect was not significant for donors living above median income (aHR: 1.01; 95%CI: 0.93–1.09). Compared with donors in median income neighborhoods ($65.1k), donors in less wealthy neighborhoods had a higher risk (Figure 5). Of note, this effect remained after adjusting for race.

Figure 5.

Living kidney donors' risk of ESRD relative to median neighborhood income ($65,129) and 95% confidence intervals. Neighborhood income was restricted to the 5th and 9th percentiles. aHR, adjusted hazard ratio; ESRD, end–stage renal disease [Color figure can be viewed at wileyonlinelibrary.com]

Donors who were first–degree relatives of their recipient had a higher risk of ESRD than unrelated donors (parent [aHR: 2.01; 95%CI: 1.26–3.21], full sibling [aHR: 1.87; 95%CI: 1.23–2.84], and child of the recipient [aHR: 1.60; 95%CI: 0.96–2.64]). Being the identical twin of the recipient was strongly associated with risk of ESRD (aHR: 19.79; 95%CI: 7.65–51.24), but there was a small number of events (n < 10, exact n suppressed per CMS policy) among a small number of identical twins. Being an "other relative" (eg, niece or grandparent) was not significantly associated with risk (aHR: 1.13; 95%CI: 0.59–2.18).

We found lower ESRD risk among donors with a higher eGFR at donation (aHR: 0.89 per 10 mL/min increase; 95%CI: 0.80–0.99). Systolic blood pressure was associated with ESRD risk (aHR: 1.15 per 10 mmHg increase; 95%CI: 0.99–1.35), but the effect did not reach statistical significance (P = .069).

Absolute Risk

We calculated the predicted absolute risk over 5–, 10–, 15–, and 20–year time horizons for LKDs based on age, race, and sex (Figure 6). Risk varies widely among subgroups of LKDs. With other factors set to the median or reference value, a 20–year–old white female LKD has a predicted 20–year ESRD risk of 8 per 10 000, compared to 111 per 10 000 for a 20–year–old black male LKD. We found higher absolute 20–year risk of ESRD at younger ages of donation for black LKDs and higher risk at older ages of donation for white LKDs. Risk increased exponentially over time after donation. When they donated at age 60, risks were nearly identical between black and white male LKDs and between black and white female LKDs.

Figure 6.

Heat map with predicted absolute risk of ESRD (per 10 000 donations) for living kidney donors by race, sex, and age, with 95% confidence intervals. All other variables were set to the median or reference value. ESRD, end–stage renal disease