Effect of Maintaining Immunosuppression After Kidney Allograft Failure on Mortality and Retransplantation : Transplantation Direct

Journal Logo

Kidney Transplantation

Effect of Maintaining Immunosuppression After Kidney Allograft Failure on Mortality and Retransplantation

Balakrishnan, Suryanarayanan MD1; Smith, Byron PhD, MS2; Bentall, Andrew MD3; Kukla, Aleksandra MD3; Merzkani, Massini MD4; Stegall, Mark MD3; Schinstock, Carrie MD3

Author Information
Transplantation Direct 9(1):p e1415, January 2023. | DOI: 10.1097/TXD.0000000000001415
  • Open



The number of patients with end-stage renal disease (ESRD) undergoing kidney transplantation continues to increase. Although short-term allograft survival has improved, long-term allograft survival remains largely unchanged.1,2 These factors have led to an increase in the number of renal allografts that fail each year.3 In 2019 alone, >7000 patients in the United States returned to dialysis after allograft failure (AF).4 In fact, >12 500 patients on the United Network of Organ Sharing list have failed allografts, and this number is expected to rise.5,6

It remains unclear how best to care for patients with failed allografts who return to dialysis.7 These patients are medically complicated and often return to dialysis burdened with years of ESRD, multiple comorbidities, and immunosuppression side effects, putting them at high risk for death. Understanding the management of immunosuppression in this context is particularly challenging. Discontinuation of immunosuppression can increase sensitization and potentially reduce the opportunity for retransplantation, whereas immunosuppression maintenance may increase the risk of malignancy, infections, cardiovascular disease, and mortality.8,9 There is a paucity of evidence guiding the management of immunosuppression in patients with AF; thus, clinical practice patterns vary, affecting any conclusions drawn from the retrospective data review.

The overall goal of this study was to understand the effect of immunosuppression maintenance on the outcomes after AF. We were in a unique position to conduct this study because of our long-term follow-up program after transplantation and the integrated dialysis center. We aimed to determine the association between immunosuppression maintenance and sensitization, as measured by calculated panel reactive antibody (cPRA), retransplantation, and mortality after AF. To this end, we conducted a single-center retrospective study of consecutive patients who returned to dialysis after AF.


Study Population

This was a retrospective cohort study of all kidney transplant recipients who received a solitary kidney transplant at the Mayo Clinic in Rochester, Minnesota, between October 2007 and May 2017, had AF before February 2020, and returned to dialysis. We chose October 2007 as a starting date because it was when single antigen bead assays for HLA antibody detection became standard of care at our center. Patients who died within 30 d of AF (n = 15) were excluded from the analysis because of the inability to distinguish between the acute kidney injury and AF. This study was approved by our institutional review board (identification: 18-008184). Because this was a retrospective study, our institutional review board waived the need for patient consent.

The data were obtained from our transplant database to identify patients with AF, as defined by the date of dialysis initiation or preemptive (before dialysis initiation) retransplantation. Only patients who returned to dialysis were included in the immunosuppression-related analyses. Information collected from electronic medical records included basic demographic information, comorbid conditions, serial cPRA, and information regarding transplant reevaluation, transplantation, dialysis access, and death date.

Diabetes was identified via a chart review. Peripheral vascular disease was defined as surgery for a vascular obstruction (angioplasty, stenting, or amputation) or claudication. Chronic infection was defined as infection that necessitated chronic antibiotic suppressive therapy. Clinical notes and reviews of echocardiograms and procedures were used to diagnose cardiac arrhythmia, cardiac valvular disease, and/or coronary artery disease. Cardiovascular disease was defined as at least one of the following: ejection fraction <55% m coronary artery disease (myocardial infection, cardiac stent placement, coronary artery bypass graft surgery, arrhythmia, or valvular disease detected on echocardiogram that required intervention). Causes of AF were determined by chart review. All rejections were biopsy confirmed.


We collected information regarding immunosuppression before and after AF. Medication names, doses, dates of reduction, and date withdrawals were collected. Given the heterogeneity in immunosuppression adjustments after AF, we chose to use complete immunosuppression withdrawal or continuation as the variables for the study.

Assessment of Sensitization

We collected the available cPRA information before and after AF. To determine cPRA, a mean fluorescence intensity of 2000 was used as the threshold for antibody positivity to assign unacceptable antigens. When determining the association between immunosuppression withdrawal and sensitization, we ensured that the withdrawal occurred before the last cPRA.

Statistical Analysis

Data are presented as median (interquartile range [IQR]). Continuous variables were compared using the t test or Kruskal-Wallis test based on distribution, and categorical variables were compared using the chi-square test. All statistical tests were 2-sided, and P < 0.05 was considered statistically significant. Immunosuppression status (maintenance versus withdrawal) was considered a time-dependent variable. Cox regression models were used to model the factors associated with mortality and retransplantation. Interactions between immunosuppression status and age ≥50 y were also analyzed as variables. Schoenfeld residuals were used to assess the assumptions of constant proportionality‚ and splines were used to assess nonlinearity. To determine the cutoff point of 50 y of age at transplant for the interaction term, linearity was first assessed within the main effect of age. Therefore, the linearity of the interaction between age and immunosuppression usage was assessed, including both main effects in the model. Based on the term plots of the resultant spline, we identified 50 as a reasonable cutoff point. Incidence curves were generated using death and retransplantation as competing risks. The Cox models were fitted using cause-specific endpoints. Incidence curves based on these predictions were adjusted for competing risks using the Aalen-Johansen estimate. All analyses were performed using BlueSky Statistics, version 7.2, and R Statistical Program, version 3.6.2.


Baseline Demographics

From October 2007 to May 2017, 1354 solitary ABO-compatible kidney transplants were performed. In this group, 97 patients experienced kidney AF and survived for at least 30 d. During this time, 35% of the patients (34/97) died after starting dialysis, 30% (29/97) were on dialysis at the last follow-up, 25% (24/97) were on dialysis before receiving another kidney transplant, and 10% (10/97) received a preemptive kidney transplant (Figure 1). Among the patients who returned to dialysis after AF, the median (IQR) follow-up time after AF was 2.4 y (1.0–4.4 y).

Patients studied.

The demographics at the initial transplantation among patients who did not receive preemptive transplantation are shown in Table 1. The median (IQR) age at transplant among all patients was 45 y (28–58 y), but patients who died after AF were older. The median (IQR) age at transplant among those who died was 57 (43–67). The majority of patients were male (60% [52/87]), 87% (76/87) were Caucasian, 25% (22/87) had received a deceased donor transplant, and 31% (27/87) had a prior failed renal allograft. The main cause of ESRD was glomerulonephritis (60% [52/87])‚ and most patients received antithymocyte globulin for immunosuppression induction (78% [68/87]). Notably, the standard induction regimen at our center was antithymocyte globulin or alemtuzumab (depending on drug availability) for patients <age 65 or with HLA antibodies at the time of transplant. The median (IQR) cPRA was 0 (0–94), but 44% (38/87) had pretransplant donor-specific antibodies (DSAs), and 24% (21/87) had a positive flow cytometric crossmatch with their donor.

TABLE 1. - Demographics at initial transplantation among patients who were not preemptively transplanted after allograft failure
Demographic Total (N = 87) Death (N = 34), 39% Dialysis (N = 29), 33% Nonpreemptive transplant (N = 24), 28% P
Age at transplant (median [IQR]) 45 (28–58) 57 (43–67) 41 (25–51) 36 (26–51) <0.001
Sex male, n (%) 52 (60) 23 (68) 14 (48) 15 (63) 0.29
Race/ethnicity, n (%) 0.41
 Caucasian 76 (87) 33 (97) 22 (76) 21 (88)
 Asian 4 (5) 0 (0) 3 (10) 1 (4)
 Black/African American 4 (5) 1 (3) 2 (7) 1 (4)
 Unknown/not reported 1 (1) 0 (0) 1 (3) 0 (0)
 Hispanic 2 (2) 0 (0) 1 (3) 1 (4)
Deceased donor transplant, n (%) 22 (25) 10 (30) 7 (24) 5 (21) 0.75
Prior failed allograft, n (%) 27 (31) 13 (38) 7 (24) 7 (29) 0.48
Cause of ESRD, n (%) 0.39
 Diabetes 10 (11) 7 (21) 2 (7) 1 (4)
 Cystic kidney disease 6 (7) 3 (9) 1 (3) 2 (8)
 Hypertension 3 (3) 0 (0) 1 (3) 2 (8)
 Congenital uropathy 5 (6) 1 (3) 1 (3) 3 (13)
 Glomerulonephritis 52 (60) 17 (50) 21 (72) 14 (58)
 Stone disease 2 (2) 1 (3) 0 (0) 1 (4)
 Other 9 (10) 5 (15) 3 (10) 1 (4)
Induction, n (%)
 Thymoglobulin 68 (78) 26 (76) 24 (83) 18 (75)
 Alemtuzumab 10 (11) 1 (3) 4 (14) 5 (21)
 Basiliximab 9 (10) 7 (21) 1 (3) 1 (4)
HLA mismatch, median (IQR) 4 (3–5) 4 (3–5) 3 (3–5) 4 (3–5) 0.96
cPRA (median [IQR]) 0 (0–94) 0 (0–96) 0 (0–72) 0 (0–50) 0.53
DSA, n (%) 38 (44) 15 (44) 11 (38) 12 (50) 0.57
Positive crossmatch, n (%) 21 (24) 8 (24) 6 (21) 7 (29) 0.70
cPRA, calculated panel reactive antibody; DSA, donor-specific antibody; ESRD, end-stage renal disease; IQR‚ interquartile range.

Characteristics at AF

The patient and graft characteristics after AF are shown in Table 2. The median (IQR) recipient age when the allograft failed was 49 y (32–61 y), but the median (IQR) age among patients who eventually died was 59 y (46–71 y). The median (IQR) graft survival was 3.2 y (1.3–5.1 y) among the cohort.

TABLE 2. - Characteristics after allograft failure
Characteristic Total (N = 87) Death (N = 34) Dialysis (N = 29) Nonpreemptive transplant (N = 24) P
Age at failure, median (IQR), y 49 (32–61) 59 (46–71) 42 (29–55) 38 (30–52) <0.001
Graft survival, median (IQR), y 3.2 (1.3–5.1) 3.0 (1.7–5.3) 3.6 (1.8–4.5) 2.7 (1.1–5.1) 1.0
Cause of allograft failure, n (%) 0.05
 Antibody-mediated rejection 31 (36) 11 (32) 9 (31) 11 (46)
 Mixed or T-cell–mediated rejection 9 (10) 2 (6) 4 (14) 3 (13)
 Acute kidney injury 5 (6) 4 (12) 1 (3) 0 (0)
 Surgical complication 4 (5) 1 (3) 0 (0) 3 (13)
 Early graft thrombosis 3 (3) 1 (3) 0 (0) 2 (8)
 Recurrent glomerulonephritis 15 (17) 4 (12) 10 (34) 1 (4)
 Unknown 3 (3) 2 (6) 1 (3) 0 (0.0)
 Multifactorial 13 (15) 8 (24) 3 (10) 2 (8)
 Other 4 (5) 1 (3) 1 (3) 2 (8)
Follow-up time after graft failure, median (IQR), y 2.4 (1.0–4.4) Time to death: 1.7 (0.9–3.8) Time on dialysis after failure: 3.6 (2.4–7.4) Time to transplant: 1.7 (0.9–2.7) <0.001
Deemed ineligible for retransplant, n (%) 26 (30) 15 (44) 11 (38) NA <0.001
Comorbidities, n (%)
 Cardiovascular disease all a 32 (37) 21 (62) 6 (21) 5 (21) <0.001
 Ejection fraction <55% 7 (8) 6 (18) 1 (3) 0 (0) 0.03
 Coronary artery disease b 8 (9) 7 (21) 1 (3) 0 (0) 0.01
 Diabetes 20 (23) 11 (32) 6 (21) 3 (13) 0.22
 Peripheral vascular disease 6 (7) 5 (15) 1 (3) 0 (0) 0.07
 Active malignancy 4 (5) 2 (6) 1 (3) 1 (4) 0.49
 Chronic infection c 7 (8) 3 (9) 1 (3) 3 (13) 0.66
Dialysis access, n (%) <0.01
 Hemodialysis (n = 71) 82%
  Catheter 44/71 (62) 17 (50) 14 (48) 13 (54)
  AV fistula 27/71 (38) 12 (35) 10 (34) 5 (21)
 Peritoneal dialysis catheter 8 (9) 0 (0) 4 (14) 4 (17)
 Unknown 8 (9) 5 (15) 1 (3) 2 (8)
aTo qualify for cardiovascular disease, the patient had to have at least one of the following: documented ejection fraction <55%; coronary artery disease, arrhythmia, or valvular disease detected on echocardiogram that required intervention.
bCoronary artery disease was defined as having documented myocardial infarction, cardiac stent placement, or coronary artery bypass graft surgery.
cRequires suppressive antibiotics.
AV, arteriovenous; IQR‚ interquartile range; NA, not available.

Rejection was the leading cause of graft loss. Antibody-mediated rejection was the presumed cause of graft loss in 36% (31/87) of patients, and 10% (9/87) of patients lost their graft from a mixed or T-cell–mediated rejection. The median IQR follow-up time on dialysis after AF to the endpoint of interest (nonpreemptive transplant, death, or last follow-up) was 2.4 y (1.0–4.4 y). The time to death was 1.7 y (0.9–3.8 y), and the time to retransplantation (excluding preemptive transplantation) was 1.7 y (0.9–2.7 y).

Comorbidities at the time of AF are shown in Table 2. The overall prevalence of cardiovascular disease was 37%, but it was disproportionally higher (62%; P < 0.01) among patients who died. Diabetes was present in 23% of the patients, whereas peripheral vascular disease, malignancy, and chronic infection were documented in <10% of the patients (Table 2). After AF, most patients (82% [71/87]) started hemodialysis, and 62% (44/71) of these started with a catheter. Only 9% (8/87) of patients immediately initiated peritoneal dialysis after AF, and dialysis information was not available for 9% (8/87) of patients. None of the patients in the peritoneal dialysis group died during follow-up.

Immunosuppression, Severe Infection, and Graft Intolerance After AF

Immunosuppression information was available for 90% (78/87) of the patients who started dialysis after AF. At the time of AF, the most common immunosuppression regimen included a calcineurin inhibitor, mycophenolate mofetil, and prednisone (74% [64/87]), and the majority (83% [72/87]) of patients had some immunosuppression reduction (Table 3). This was consistent with the standard maintenance immunosuppression regimen used for maintenance among patients with a functioning during the time the study was conducted. Nearly half, 46% (40/87)‚ of the patients stopped immunosuppression at a median (IQR) time of 0.9 y (0.2–2.2 y) after AF. The time from AF until complete discontinuation of immunosuppression was shorter among patients who later received another transplant and longest among patients who remained on dialysis after AF during follow-up (P = 0.04). The median (IQR) time to complete immunosuppression withdrawal among patients who received a transplant after starting dialysis was 0.2 y (0.1–0.6 y) compared with 1.0 y (0.6–1.9 y) among patients who died and 1.8 y (0.6–2.7 y) among patients who remained on dialysis at last follow-up. After AF, 26% (23/87) of the patients had a transplant nephrectomy‚ and 51% (44/87) had a blood transfusion.

TABLE 3. - Immunosuppression and sensitizing events after allograft failure
Total (N = 87) Death (N = 34) Dialysis (N = 29) Transplant (N = 24) P
Immunosuppression regimen at failure, n (%) 0.33
 Calcineurin inhibitor, mycophenolate, and prednisone 64 (74) 22 (65) 26 (90) 16 (67)
 Calcineurin inhibitor, mycophenolate, and steroid free 3 (3) 1 (3) 0 (0) 2 (8)
 Calcineurin inhibitor and prednisone 9 (10) 5 (15) 2 (7) 2 (8)
 Other 14 (16) 7 (21) 1 (3) 6 (25)
Immunosuppression reduction, a n (%) 72 (83) 26 (77) 26 (90) 20 (83) 0.56
Immunosuppression withdrawal, n (%) 0.03
 Yes 40 (46) 12 (35) 18 (62) 10 (42)
 No 38 (44) 15 (44) 9 (31) 14 (58)
 Unknown 9 (10) 7 (21) 2 (7) 0 (0)
Time to all immunosuppression withdrawal, median (IQR), y 0.9 (0.2–2.2) 1.0 (0.6–1.9) 1.8 (0.6–2.7) 0.2 (0.1–0.6) 0.04
Transplant nephrectomy, n (%) 23 (26) 8 (24) 7 (24) 8 (33) 0.05
Blood transfusion after failure, n (%) 44 (51) 18 (53) 18 (62) 8 (33) <0.001
aReduction = decrease from before failure or discontinuation.
To our knowledge, no pregnancies occurred after graft failure. IQR‚ interquartile range.

Only 35% (8/23) of the nephrectomies were done for graft intolerance after some degree of immunosuppression reduction; 4 of these patients had been on at least a low level of calcineurin inhibitor and mycophenolate when they developed intolerance. The rest of the nephrectomies were done for early graft failure (48% [11/23]) or miscellaneous reasons (17% [4/23]). The overall incidence of graft intolerance after immunosuppression was low.

We also examined the proportion of patients who had at least 1 severe infection requiring hospitalization based on whether immunosuppression was maintained. We found that a similar proportion of patients had severe infection requiring hospitalization among the immunosuppression withdrawal and maintenance groups. Specifically, 65% (26/40) of patients who completely withdrew immunosuppression were hospitalized at least once with a severe infection versus 52.6% (20/38) who did not completely withdrawal immunosuppression (P = 0.35).

Change in cPRA After AF

The median (IQR) cPRA among all patients before AF was 40 (0–87) and increased to 93 (37–100) after AF (Table 4). Furthermore, cPRA increased to 100% in 31% (15/49) of patients. This increase in cPRA appeared to be driven by immunosuppression withdrawal because the cPRA remained unchanged (the median [IQR] change in cPRA was 0 [–0.5 to 0.5]) when immunosuppression was maintained but increased from a median (IQR) of 18 (0–99) to 96 (88.5–100.0; P = 0.003) when immunosuppression was withdrawn.

TABLE 4. - Change in cPRA after allograft failure
Total (N = 87) Death (N = 34) Dialysis (N = 29) Transplant (N = 24) P
Before failure
 Last cPRA, median (IQR) 40 (0–87) 48 (0–97) 29 (0–77) 18 (0–75) 0.57
 N-missing 3 3 0 0
After failure
 Last cPRA, median (IQR) 93 (37–100) 83 (7–99) 89 (36–100) 94 (48–100) 0.61
 Patients with cPRA ≥100 31% (15/49) 21% (3/14) 38% (6/16) 31% (6/19) 0.63
 N-missing 38 20 13 5
Change in cPRA from before to after failure
 median (IQR) 0 (0–45) 0 (0–0) 0 (0–31) 37 (0–81) 0.006
 Range –29 to 100 –17 to 2 –29 to 99 –7 to 100
The median (IQR) change in cPRA was 0 (–0.5 to 0.5) when immunosuppression was maintained but increased from a median (IQR) of 18 (0–99) up to 96 (88.5–100.0; P = 0.003) among patients whose immunosuppression was withdrawn.
cPRA, calculated panel reactive antibody; IQR, interquartile range.

Factors Associated with Death and Retransplantation After AF

Using univariate and multivariate analyses and excluding patients who received a preemptive transplant, we examined the factors associated with the competing outcomes of death and retransplantation. With univariate analysis, the factors associated with death included older age at AF (hazard ratio [HR], 1.1; confidence interval [CI], 1.0-1.1), peripheral vascular disease (HR, 4.6; CI, 1.5-13.4), and all-cause cardiovascular disease (HR, 3.8; CI, 1.7-8.2) (Table 5). Maintaining at least some immunosuppression was not associated with death (HR, 0.9; CI, 0.4-2.1). In the multivariate analysis, only age at AF (HR, 1.1; CI, 1.0–1.1) and all-cause cardiovascular disease remained associated with increased death (HR, 2.9; CI, 1.2-7.0). Importantly, the second-order term that included immunosuppression and age >50 y was associated with a reduced hazard of death (HR, 0.1; CI, 0.0-0.7). In other words, older patients who remained on immunosuppression were less likely to die. Figure 2A shows the predicted incidence of death and retransplant based on Cox proportional models (adjusted for competing risks) among patients aged 30 and 65 y on and off maintenance immunosuppression. The specific causes of death were not identified in nearly half of the patients (47% had an unknown cause of death); however, among those with an identifiable cause of death, 33% (6/18) died of infection, 28% (5/18) died of cardiovascular disease, 28% (5/18) died of miscellaneous reasons, 6% (1/18) died of malignancy, and 6% (1/18) died of neurological disease.

TABLE 5. - Univariate and multivariate factors associated with death or retransplantation after allograft failure
Death Univariate analysis Multivariate analysis
Age at failure (y) 1.1 (1.0-1.1) <0.001 1.1 (1.0-1.1) 0.0005
Male sex 1.7 (0.7-3.7) 0.22
On immunosuppression 0.9 (0.4-2.1) 0.49 0.7 (0.5-7.3) 0.34
Diabetes 1.8 (0.8-4.0) 0.18
Peripheral vascular disease 4.6 (1.5-13.4) 0.006 2.7 (0.9-8.6) 0.08
Cardiovascular disease (all) 3.8 (1.7-8.2) <0.001 2.9 (1.2-7.0) 0.02
Prior allograft survival (y) 1.1 (0.9-1.3) 0.41
AV fistula at the time of failure 0.9 (0.4-2.1) 0.86
Peritoneal dialysis at the time of failure 0.0 (0.0-infinity) 1.0
Nephrectomy after graft failure 0.7 (0.3-1.8) 0.60
On immunosuppression × age >50 a 0.2 (0.0-0.8) 0.03 0.1 (0.0-0.7) 0.02
Retransplant (nonpreemptive)
 Age at failure (y) 1.0 (1.0-1.0) 0.58
 Male sex 1.4 (0.6-3.2) 0.55
 On immunosuppression 0.8 (0.3-2.0) 0.63
 Diabetes 0.6 (0.2-2.0) 0.38
 Peripheral vascular disease 0.0 (0.0-infinity) 1.0
 Cardiovascular disease (all) 0.7 (0.2-1.7) 0.39
 Prior allograft survival (y) 1.1 (0.9-1.3) 0.51
 AV fistula at the time of failure 0.6 (0.2-1.6) 0.31
 Peritoneal dialysis at the time of failure 1.3 (0.4-4.1) 0.56
 Nephrectomy after graft failure 0.9 (0.4-2.1) 0.49
 On immunosuppression × age >50 a 0.3 (0.1-1.35) 0.12
aUnivariate results include adjustment for main effects (age and immunosuppression status).
AV, arteriovenous; CI, confidence interval; HR‚ hazard ratio.

Predicted incidence of death and retransplantation stratified by age and IS based on Cox regression modeling. The predicted incidence of death (A) and retransplant (B) based on Cox proportional models (adjusted for competing risks) among patients at age 30 and 65 on and off maintenance IS. With multivariate analysis (Table 5), age at failure (HR, 1.1; CI, 1.01-1.01; P = 0.0005), cardiovascular disease (HR, 2.9; CI, 1.2-7.0; P = 0.02), and the second-order term IS × age >50 (HR, 0.1; CI, 0.0-0.7; P = 0.02) were associated with death. These factors were not associated with retransplantation. CI, confidence interval; HR‚ hazard ratio; IS, immunosuppression; y/o, year old.

On univariate analysis, age at AF, male sex, immunosuppression maintenance, diabetes, peripheral vascular disease, all-cause cardiovascular disease, prior allograft survival (y), arteriovenous fistula, peritoneal dialysis, and nephrectomy were not associated with eventual (nonpreemptive) retransplantation. We also did not find a statistically significant interaction between age and immunosuppression maintenance on retransplantation (Figure 2B).

Of the 24 retransplants performed after AF and temporary dialysis, 46% (11/24) were from deceased donors, and 54% (12/24) were from living donors. Fifty-eight percent (7/12) of the living donor transplants were facilitated through paired kidney donations. Additionally, 25% (6/24) of the transplants were DSA-positive, and 13% (3/24) were positive crossmatches. Transplant recipients with a cPRA ≥90% and 100% were numerically more likely to receive deceased donor transplants than less sensitized recipients, but this did not reach statistical significance. The proportion of transplants that were DSA-positive, positive crossmatch, or facilitated by kidney paired donation was similar among patients with high cPRA ≥90% and 100% and less sensitized candidates.


Outcomes after kidney AF remain poor. Within a median of 2 y after AF and return to dialysis, 35% of the patients died, 30% remained on dialysis, and only 25% received another kidney transplant. Older patients with cardiovascular disease have the highest risk of mortality. Sensitization is also a major problem after AF, as nearly 30% of patients reached a cPRA of 100%. Complete immunosuppression withdrawal was clearly associated with sensitization and a greater change in cPRA. However, we did not find evidence that complete immunosuppression withdrawal after AF had a substantial impact on retransplantation. We also did not find an association between the maintenance of immunosuppression and death. In fact, patients who maintained immunosuppressive therapy were less likely to die. We are doubtful that this suggests that immunosuppression is protective; instead, we believe that older and sicker patients may have taken off immunosuppression for safety. Together, our findings suggest that‚ in the current era of robust kidney paired donation programs and waiting list prioritization for sensitized patients, it may be appropriate to discontinue immunosuppression after AF despite the risk of sensitization if there is a delay in retransplantation. Validation of these findings in a large prospective multicenter randomized trial with a standardized immunosuppression withdrawal approach would be of value to the transplant community.

Our results validate findings from multiple prior studies but offer a new perspective.10,11 A heightened risk of death after AF is known,12 and the risk of sensitization after AF has been studied.13–15 The novel aspect of this study was the incorporation of the impact of ongoing immunosuppression on these important outcomes.

Very few recent studies have examined the mortality risks associated with immunosuppression after AF and the start of dialysis. It is known that uremia alone is associated with immune dysfunction; however, the additive risk of immunosuppression is unclear.16,17 A small nonrandomized study showed that immunosuppression after AF was associated with a 3-times-higher incidence of infection and related complications (odds ratio, 3.4).18 This high infection risk translated into reduced patient survival, as shown by a follow-up study from the same group that demonstrated higher mortality among patients who continued immunosuppression after AF.8 These previous studies were small and not performed in the modern era with current dialytic techniques and immunosuppressive regimens. Furthermore, HLA antibody detection techniques were less sophisticated; therefore, these older studies did not address the effect of immunosuppression on sensitization and retransplantation.

Several studies have examined the role of immunosuppression withdrawal and sensitization, but these studies did not thoroughly explore retransplantation.9,13–15,19 A single-center cohort study of 49 patients referred for retransplantation showed that immunosuppression withdrawal within 3 mo of AF was associated with sensitization and a higher PRA than immunosuppression withdrawal after this time point.14 Another single-center nonrandomized study showed that immunosuppression withdrawal was associated with sensitization and a PRA of >80%.13 Undergoing a transplant nephrectomy appeared to increased sensitization further.13

Similar to other studies, we validated the risk of sensitization with immunosuppression withdrawal but did not find an impact on subsequent transplantation. This is likely due to the changing landscape of kidney transplantation. In the last decade, kidney paired donation programs20 have expanded, and changes in the United Network of Organ Sharing allocation have led to increased transplantation among all sensitized patients, except for those in the 100% group with a precise cPRA >99.9%.21,22 Although it is possible that we were underpowered to detect important trends, it is remarkable that we did not even detect a signal for reduced retransplantation among patients whose cPRA increased up to 100%. Further supporting this notion, we did not find an increase in transplants with DSA or positive crossmatch among patients with the highest cPRA.

The strength of our study is that we used a well-characterized cohort. Given our emphasis on long-term follow-up combined with our large integrated health system and integrated dialysis program, we were able to obtain detailed follow-up data that were not readily available elsewhere. The major weakness of our study was the small single-center population and the nonstandardized and immunosuppression withdrawal regimen. Therefore, we may have been underpowered to detect the immunosuppressive effects. We are also a large transplant center with access to kidney paired donation and expertise in HLA-incompatible transplantation, which may have increased our patients’ access to retransplantation despite sensitization. Ideally, a large multicenter study of immunosuppression withdrawal versus maintenance among patients suitable for retransplantation should be performed to answer complex questions regarding the impact of immunosuppression on patient survival after AF. However, a study of that magnitude may not be feasible, given the varied baseline immunosuppression used in AF, the subjective nature of transplant candidacy, and the unfortunate separation among many dialysis providers and transplant centers.

However, many questions remain in this area. First, it remains unclear how best to withdraw immunosuppression in general. It is unclear how the speed of immunosuppression withdrawal influences the development of sensitization or graft intolerance. Interestingly, we found that the incidence of graft intolerance and need for subsequent nephrectomy were low, but we did not have a large enough cohort to understand the risk factors for graft intolerance in general. We do not know how the pace and degree of immunosuppression reduction can affect sensitization, and the relationship between immunosuppression reduction and retaining residual allograft function has been largely unexplored.

Although there are unanswered questions, these data can be applied to clinical practice. If a retransplant candidate has malignancy or life-threatening infection after graft failure, immunosuppression should be withdrawn despite the risk of sensitization. Sensitization may occur, but this is unlikely to have a major impact on retransplantation. Additionally, immunosuppression can be continued at least temporarily and safely if a patient has potential living donors and timely retransplantation is possible. It also appears that the maintenance of immunosuppression to preserve residual renal function is reasonable. This has not been directly studied, but we did not find a clear signal that ongoing immunosuppression is a major risk factor for death. We do not observe the benefit of long-term immunosuppression after AF purely to prevent sensitization and increase cPRA.

In conclusion, patients are vulnerable to AF, and their overall risk of short-term mortality is high. Maintenance of immunosuppression does not appear to increase the risk of death but prevents sensitization. However, the increased sensitization that occurs with immunosuppression does not translate to reduced transplantation in the current era. Therefore, it appears reasonable to discontinue immunosuppression after AF despite the risk of sensitization if retransplantation is delayed. Larger prospective randomized trials are needed to validate this conclusion.


1. Meier-Kriesche HU, Schold JD, Kaplan B. Long-term renal allograft survival: have we made significant progress or is it time to rethink our analytic and therapeutic strategies? Am J Transplant. 2004;4:1289–1295.
2. Lamb KE, Lodhi S, Meier-Kriesche HU. Long-term renal allograft survival in the United States: a critical reappraisal. Am J Transplant. 2011;11:450–462.
3. Singh P, Filippone EJ, Colombe BW, et al. Sensitization trends after renal allograft failure: the role of DQ eplet mismatches in becoming highly sensitized. Clin Transplant. 2016;30:71–80.
4. Kochar GS, Langone AJ. How should we manage renal transplant patients with failed allografts who return to dialysis? Blood Purif. 2020;1:4.
5. Gaston RS, Fieberg A, Hunsicker L, et al. Late graft failure after kidney transplantation as the consequence of late versus early events. Am J Transplant. 2018;18:1158–1167.
6. Agrawal N, Pavlakis M. Caring for the patient with a failing allograft: challenges and opportunities. Curr Opin Org Transplant. 2019;24:416–423.
7. Davis S, Mohan S. Managing patients with failing kidney allograft: many questions remain. Clin J Am Soc Nephrol. 2022;17:444–451.
8. Smak Gregoor PJ, Zietse R, van Saase JL, et al. Immunosuppression should be stopped in patients with renal allograft failure. Clin Transplant. 2001;15:397–401.
9. Nimmo A, McIntyre S, Turner DM, et al. The impact of withdrawal of maintenance immunosuppression and graft nephrectomy on HLA sensitization and calculated chance of future transplant. Transplant Direct. 2018;4:e409.
10. Gill JS, Abichandani R, Kausz AT, et al. Mortality after kidney transplant failure: the impact of non-immunologic factors. Kidney Int. 2002;62:1875–1883.
11. Rao PS, Schaubel DE, Jia X, et al. Survival on dialysis post-kidney transplant failure: results from the Scientific Registry of Transplant Recipients. Am J Kidney Dis. 2007;49:294–300.
12. Kaplan B, Meier-Kriesche HU. Death after graft loss: an important late study endpoint in kidney transplantation. Am J Transplant. 2002;2:970–974.
13. Augustine JJ, Woodside KJ, Padiyar A, et al. Independent of nephrectomy, weaning immunosuppression leads to late sensitization after kidney transplant failure. Transplantation. 2012;94:738–743.
14. Casey MJ, Wen X, Kayler LK, et al. Prolonged immunosuppression preserves nonsensitization status after kidney transplant failure. Transplantation. 2014;98:306–311.
15. Del Bello A, Congy-Jolivet N, Sallusto F, et al. Donor-specific antibodies after ceasing immunosuppressive therapy, with or without an allograft nephrectomy. Clin J Am Soc Nephrol. 2012;7:1310–1319.
16. Kato S, Chmielewski M, Honda H, et al. Aspects of immune dysfunction in end-stage renal disease. Clin J Am Soc Nephrol. 2008;3:1526–1533.
17. Kimmel PL, Phillips TM, Simmens SJ, et al. Immunologic function and survival in hemodialysis patients. Kidney Int. 1998;54:236–244.
18. Gregoor PJ, Kramer P, Weimar W, et al. Infections after renal allograft failure in patients with or without low-dose maintenance immunosuppression. Transplantation. 1997;63:1528–1530.
19. Kosmoliaptsis V, Gjorgjimajkoska O, Sharples LD, et al. Impact of donor mismatches at individual HLA-A, -B, -C, -DR, and -DQ loci on the development of HLA-specific antibodies in patients listed for repeat renal transplantation. Kidney Int. 2014;86:1039–1048.
20. Flechner SM, Thomas AG, Ronin M, et al. The first 9 years of kidney paired donation through the National Kidney Registry: characteristics of donors and recipients compared with National Live Donor Transplant Registries. Am J Transplant. 2018;18:2730–2738.
21. Schinstock CA, Smith BH, Montgomery RA, et al. Managing highly sensitized renal transplant candidates in the era of kidney paired donation and the new kidney allocation system: is there still a role for desensitization? Clin Transplant. 2019;33:e13751.
22. Stewart DE, Kucheryavaya AY, Klassen DK, et al. Changes in deceased donor kidney transplantation one year after KAS implementation. Am J Transplant. 2016;16:1834–1847.
Copyright © 2022 The Author(s). Transplantation Direct. Published by Wolters Kluwer Health, Inc.