Risk prediction for recipient outcomes following living donor kidney transplantation (LDKT) remains a challenge despite detailed predonation information available from living kidney donors. Interestingly, recipients of kidneys from living donors who subsequently develop end-stage renal disease (ESRD) also have higher graft failure,1 suggesting the 2 donor kidneys share risk factors that could inform recipient outcomes. Given that donor ESRD is rare,2,3 an earlier and more common postdonation outcome such as hypertension4 could serve as a surrogate to improve and better individualize counseling and management for recipients.
Hypertension is one of the most frequent proximal causes of donor ESRD;5,6 early postdonation hypertension might indicate preexisting donor subclinical renal disease,7 which could increase the risk of graft failure for LDKT recipients. For example, the current literature on renal hyperfiltration in patients with ESRD suggests that hyperfiltration is associated with progressive kidney disease, particularly in those with ESRD caused by hypertension8 or diabetes.9 Recipients with hypertension- or diabetes-caused ESRD might therefore be at higher risk for graft failure associated with underlying subclinical renal disease in the donor kidney. Similarly, because there is an association between hypertension and glomerular hyperfiltration in African Americans,10 African American LDKT recipients whose donors developed hypertension might be at higher risk for graft failure.
With robust early donor follow-up data becoming available through Organ Procurement and Transplantation Network (OPTN) mandates,11,12 we now have the opportunity to examine how early postdonation incident hypertension is associated with LDKT recipient graft failure. To do this, we used national registry data to compare LDKT outcomes in recipients whose donors did or did not develop early postdonation hypertension. We further examined how this association might be different in subgroups that might be more sensitive to donor kidney hyperfiltration, namely African American LDKT recipients and those with ESRD caused by hypertension or diabetes.
MATERIALS AND METHODS
This study used data from the Scientific Registry of Transplant Recipients (SRTR) standard analysis files, available as of 9 January 2018. The SRTR data system includes data on all donor, waitlisted candidates, and transplant recipients in the United States, submitted by the members of the OPTN, and has been described elsewhere.13 The Health Resources and Services Administration (HRSA), US Department of Health and Human Services provides oversight to the activities of the OPTN and SRTR contractors.
We studied 45 441 adult LDKT recipients between 1 January 2008 and 3 January 2016 with 2-year follow-up information for their living donors. We excluded recipients whose donors were missing follow-up on hypertension at all mandated follow-up reports (n = 2346, 5.2%) and recipients whose donors had baseline hypertension (n = 1358, 3.0%) or were missing data regarding baseline hypertension at the time of donation (n = 143, 0.3%) (Figure 1). We also excluded recipients missing baseline data on body mass index (BMI) or cause of ESRD (n = 3693, 8.8%); there were no other baseline characteristics with missing data.
Early Postdonation Incident Hypertension
We defined early postdonation incident hypertension as a donor reporting hypertension at one or more follow-up reports mandated at 6 months, 1 year, and 2 years postdonation in response to the form question “Donor developed hypertension requiring medication.” Characteristics of recipients whose donors did not develop hypertension were compared to those whose donors developed early postdonation incident hypertension using chi-square tests and Wilcoxon rank-sum tests, as appropriate.
Propensity Score Weighting
Because we were interested in isolating the effect of early postdonation hypertension on graft outcomes and because the donors who developed early postdonation incident hypertension and their recipients were quite different from recipients and donors who did not develop hypertension, we used inverse probability of treatment weighting using a propensity score, a method that seeks unbiased estimates of average treatment effects with observational data.14
We created a propensity score quantifying the probability of the LDKT recipient having a donor who developed early postdonation incident hypertension, using logistic regression. This model included recipient age as a linear spline with a knot at 40 years, sex, race, BMI, cause of ESRD, estimated glomerular filtration rate (eGFR) at hospital discharge after LDKT; and donor age, sex, race, BMI, smoking status, ABO incompatibility, and relationship with recipient. These factors were selected because they were associated with early incident postdonation hypertension and one or both of the graft failure outcomes described below.
To avoid bias due to extreme values, we used stabilized weights.15 To see how well the population was balanced by weighting, we calculated standardized differences between LDKT recipients whose donors developed early postdonation incident hypertension and those whose donors did not, both before and after weighting. Standardized differences of <0.1 after weighting demonstrate a balance between groups.14
Recipient Graft Failure
To examine how LDKT recipient all-cause graft failure (ACGF) was associated with early postdonation incident hypertension, we estimated incidence of graft failure at 1, 5, and 10 years after transplantation using Kaplan–Meier methods with log-rank tests. We then used Cox proportional hazards regression models to quantify the adjusted hazard ratio (aHR) of early postdonation incident hypertension in the weighted population. We also created standardized survival curves adjusted for donor and recipient characteristics by using the weighted population. As a sensitivity analysis, we repeated this for recipient death-censored graft failure (DCGF).
To determine whether ACGF differed for recipients with hypertension- or diabetes-caused ESRD, and for African American recipients, we examined these subgroups separately and repeated the analysis using inverse probability of treatment weighting and Cox proportional hazards regression as described above.
We examined the association between ACGF and the median systolic blood pressure reported for each donor across all follow-up reports as a sensitivity analysis, scaled per 10 mm Hg. We used Cox proportional hazards regression adjusting for recipient age as a linear spline with a knot at 40 years, sex, race, BMI, cause of ESRD, education, and peak panel-reactive antibody; and donor age, sex, race, BMI, smoking status, ABO incompatibility, human leukocyte antigen mismatches, and relationship with recipient. Because there were 5211 donors (13.7%) who were missing systolic blood pressure at all follow-up reports, we used multiple imputation by chained equations to impute missing values using predictors of recipient age, sex, race, cause of ESRD, BMI, education, insurance, peak panel-reactive antibody, and ACGF; and donor age, sex, race, BMI, smoking status, ABO incompatibility, human leukocyte antigen mismatches, relationship with recipient, preoperative eGFR, preoperative systolic blood pressure, preoperative diastolic blood pressure, postoperative systolic blood pressure, postoperative diastolic blood pressure, and baseline hazard of early hypertension requiring antihypertensive medication.
Confidence intervals are reported as per the method of Louis and Zeger.16 An α of 0.05 was considered statistically significant. All analyses were performed using Stata 14.2/MP for Linux (College Station, TX).
Of 37 901 living kidney donors, 923 (2.4%) developed early postdonation incident hypertension. Donors who developed early postdonation incident hypertension were older (median 49 versus 42 y, P < 0.001), more frequently male (47.7% versus 37.3%, P < 0.001), less frequently Hispanic (9.2% versus 14.2%, P < 0.001), and more frequently overweight (46.2% versus 42.0%) or obese (30.0% versus 22.3%, P < 0.001) compared to donors who did not develop hypertension. Donors who developed hypertension were more frequently a first-degree relative (44.2% versus 42.4%) or spouse or partner (18.4% versus 13.3%, P < 0.001) to their recipient compared to donors who did not develop hypertension (Table 1).
Of donors who developed postdonation hypertension, the median systolic blood pressure reported across all follow-up reports was 131 mm Hg (interquartile range [IQR] 120.5–141, with 11.3% missing) and the median diastolic blood pressure reported across all follow-up reports was 81 mm Hg (IQR 75–88.5, with 11.4% missing) (Figure 2). Of donors who did not develop postdonation hypertension, the median systolic blood pressure reported across all follow-up reports was 119 mm Hg (IQR 111.5–126, with 14.0% missing) and the median diastolic blood pressure reported across all follow-up reports was 74 mm Hg (IQR 69.5–80, with 14.0% missing) (Figure 2).
LDKT recipients whose donors developed hypertension were older (median 53 versus 50 y, P < 0.001), less frequently Hispanic (10.6% versus 14.7%, P = 0.002), and had a lower eGFR at discharge after transplantation (median 53 versus 56 mL/min, P < 0.001) (Table 1). After applying inverse probability of treatment weights to the study population, recipient and donor characteristics between those who developed postdonation hypertension and those who did not were balanced as demonstrated by all standardized differences <0.1 (Table 2). In other words, estimates of the association between postdonation incident hypertension and recipient outcomes are unbiased by baseline characteristics because balance in baseline characteristics has been achieved between groups through weighting.14
All-cause Graft Failure
At 1, 5, and 10 years after transplantation, 2.9%, 17.0%, and 38.6% of LDKT recipients whose donors developed early postdonation incident hypertension had ACGF compared to 2.8%, 12.8%, and 30.5% of those whose donors did not develop hypertension (P = 0.02). LDKT recipients whose donors developed early postdonation incident hypertension were at 19% higher risk for ACGF (HR 1.031.191.38, P = 0.02). However, after weighting to control for donor and recipient characteristics, LDKT recipients whose donors developed early postdonation incident hypertension were not at higher risk for ACGF (aHR 0.851.031.25, P = 0.72) (Figure 3).
Death-censored Graft Failure
Our sensitivity analysis examining DCGF confirmed inferences. At 1, 5, and 10 years after transplantation, 1.8%, 8.7%, and 25.4% of LDKT recipients whose donors developed early postdonation incident hypertension had DCGF compared to 1.7%, 7.2%, and 15.5% of those whose donors did not develop hypertension (P = 0.14). LDKT recipients whose donors developed early postdonation incident hypertension were not at higher risk for DCGF (HR 0.941.161.44, P = 0.1). After weighting, LDKT recipients whose donors developed early postdonation incident hypertension were not at higher risk for DCGF (aHR 0.811.051.37, P = 0.66) (Figure 4).
The association between early postdonation incident hypertension and recipient ACGF did not vary among LDKT recipients with ESRD from hypertension (n = 6251), those with ESRD from diabetes (n = 8563), or African American recipients (n = 4967). After weighting to control for donor and recipient characteristics, there was no statistically significant association between donor incident hypertension and recipient ACGF in any subgroup analysis (recipients with hypertension-caused ESRD: aHR 0.651.101.85, P = 0.73; recipients with diabetes-caused ESRD: aHR 0.560.801.13, P = 0.20; African American recipients: aHR 0.701.101.73, P = 0.68).
Higher postdonation systolic blood pressure was associated with higher risk of ACGF. For each increment of 10 mm Hg of postdonation systolic blood pressure, there was a 5% higher risk of ACGF (aHR 1.031.051.08, P < 0.001).
In this national study of 37 901 living donor kidney transplant recipients, we used propensity score weighting to isolate the association between early postdonation incident hypertension in donors and the graft outcomes of their recipients. In adjusted models, we found no association between early postdonation incident hypertension and recipient ACGF (aHR 1.03, P = 0.72) or DCGF (aHR 1.05, P = 0.66). We then studied subgroups of recipients who might be at highest risk for graft failure—African American recipients and those with ESRD caused by hypertension and diabetes—and also found no higher risk of graft failure associated with early postdonation incident hypertension in these subgroups.
Our finding of no association between early incident postdonation hypertension and recipient graft failure is surprising, given the known link between donor hypertension and ESRD5,6 and association of recipient graft failure with donor ESRD.1 However, we did find an association between donor postdonation systolic blood pressure and graft failure. It may be that kidneys from donors whose postdonation hypertension was more easily controlled by antihypertensive medication were not at higher risk for graft failure, while those from donors whose hypertension was more difficult to control did carry a higher risk of graft failure. It also might be that the reporting of diagnosed hypertension in the registry is less reliable than the reporting of measured blood pressure. Regardless, our findings suggest that there are differences in renal physiology between donors and the recipients of their kidneys. Prior literature does suggest differences in renal physiology in donors versus those with kidney disease. Although Lenihan et al17 found that donor hyperfiltration was a benign adaptive process not associated with albuminuria in their study of 21 living kidney donors at median 6 years postdonation,18 Steiner19 cites evidence that hyperfiltration causes mechanical injury and progressive loss of GFR in rat studies and suggests that long-term outcomes deserve further consideration. Our results contrast with the study from Miller et al20 of 115 124 kidney transplant recipients, which found that recipients were at higher risk of graft failure due to “nephron underdosing” when their donors were ≥10 kg lighter than they were. Although our epidemiologic study found no association between the diagnosis of postdonation incident hypertension and recipient graft loss, we found that the reported postdonation systolic blood pressure may be more indicative of a linked physiology between the donor’s remaining kidney and transplanted kidney. Further work should continue to clarify the physiologic impact of subclinical kidney disease and glomerular hyperfiltration on both kidney donors and transplant recipients.
Our study has several limitations inherent to its use of national registry data. We are limited by missing data and the clinical granularity of the available data. We do not have information regarding donor albuminuria, detailed family history, or presence of apolipoprotein L1 high-risk variants,21,22 all of which impact donor risk for later ESRD.23 Although we use a propensity score method for balancing baseline characteristics between recipients whose donors developed early postdonation incident hypertension and those whose donors did not, we cannot account for potential unmeasured confounders that differ between groups.
In conclusion, we found that early postdonation hypertension did not portend a higher risk of recipient graft failure in the same way that eventual postdonation ESRD does. These findings held when limiting to recipient subsets at highest risk for progressive kidney disease due to graft hyperfiltration: African American recipients and those with ESRD caused by hypertension and diabetes. While these findings do not improve risk prediction beyond current models, they suggest some insights into the physiology of LDKT grafts in the milieu of the recipient with ESRD.
The data reported here have been supplied by the Hennepin Healthcare Research Institute (HHRI) as the contractor for the Scientific Registry of Transplant Recipients (SRTR). The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an official policy of or interpretation by the SRTR or the US Government.
1. Muzaale AD, Massie AB, Anjum S, et al. Recipient outcomes following transplantation of allografts from live kidney donors who subsequently developed end-stage renal disease.Am J Transplant201616123532–3539
2. Muzaale AD, Massie AB, Wang MC, et al. Risk of end-stage renal disease following live kidney donation.JAMA20143116579–586
3. Mjøen G, Hallan S, Hartmann A, et al. Long-term risks for kidney donors.Kidney Int2014861162–167
4. Holscher CM, Kmd SB, Thomas AG, et al. Early hypertension and diabetes after living kidney donation: a national cohort study.Transplantation2018
5. Anjum S, Muzaale AD, Massie AB, et al. Patterns of end-stage renal disease caused by diabetes, hypertension, and glomerulonephritis in live kidney donors.Am J Transplant201616123540–3547
6. Matas AJ, Berglund DM, Vock DM, et al. Causes and timing of end-stage renal disease after living kidney donation.Am J Transplant20181851140–1150
7. Fahmy LM, Massie AB, Muzaale AD, et al. Long-term renal function in living kidney donors who had histological abnormalities at donation.Transplantation201610061294–1298
8. Palatini P, Dorigatti F, Saladini F, et al. Factors associated with glomerular hyperfiltration in the early stage of hypertension.Am J Hypertens20122591011–1016
9. Tonneijck L, Muskiet MH, Smits MM, et al. Glomerular hyperfiltration in diabetes: mechanisms, clinical significance, and treatment.J Am Soc Nephrol20172841023–1039
10. Kotchen TA, Piering AW, Cowley AW, et al. Glomerular hyperfiltration in hypertensive African Americans.Hypertension2000353822–826
11. Henderson ML, Thomas AG, Shaffer A, et al. The national landscape of living kidney donor follow-up in the United States.Am J Transplant201717123131–3140
12. Reed RD, Shelton BA, MacLennan PA, et al. Living kidney donor phenotype and likelihood of postdonation follow-up.Transplantation20181021135–139
13. Massie AB, Kucirka LM, Kuricka LM, et al. Big data in organ transplantation: registries and administrative claims.Am J Transplant20141481723–1730
14. Austin PC, Stuart EA. Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies.Stat Med201534283661–3679
15. Cole SR, Hernán MA. Constructing inverse probability weights for marginal structural models.Am J Epidemiol20081686656–664
16. Louis TA, Zeger SL. Effective communication of standard errors and confidence intervals.Biostatistics20091011–2
17. Lenihan CR, Busque S, Derby G, et al. Longitudinal study of living kidney donor glomerular dynamics after nephrectomy.J Clin Invest201512531311–1318
18. Blantz RC, Steiner RW. Benign hyperfiltration after living kidney donation.J Clin Invest20151253972–974
19. Steiner RW. Invited response to “hyperfiltration after donation and living kidney donor risk.”Am J Transplant201717112992
20. Miller AJ, Kiberd BA, Alwayn IP, et al. Donor-recipient weight and sex mismatch and the risk of graft loss in renal transplantation.Clin J Am Soc Nephrol2017124669–676
21. Gaillard F, Gribouval O, Courbebaisse M, et al. Comparison of postdonation kidney function between Caucasian donors and low-risk APOL1 genotype living kidney donors of African ancestry.Transplantation201810211e462–e463
22. Ma L, Divers J, Freedman BI. Mechanisms of injury in APOL1-associated kidney disease.Transplantation20191033487–492
23. Slinin Y, Brasure M, Eidman K, et al. Long-term outcomes of living kidney donation.Transplantation201610061371–1386