Comparison of Outcomes after Delayed Graft Function: Sirolimus-Based Versus Other Calcineurin-Inhibitor Sparing Induction Immunosuppression Regimens

Abstract

Background. 

Sirolimus (SRL) may increase the incidence of or prolong delayed graft function (DGF) after cadaveric renal transplantation. This study compares transplant outcomes of SRL-based induction immunosuppression (IS) with other calcineurin-inhibitor (CNI) sparing regimens in the DGF setting.

Methods. 

Adult cadaveric renal-transplant recipients who received transplants between January 1, 1997 and June 30, 2001 and experienced DGF (n=132) were divided into three groups by induction IS: A, depleting antibody (n=41); B, SRL (n=49); and C, neither (n=42). All recipients also received steroids and mycophenolate mofetil with delayed initiation of CNIs when good renal function returned. Patient survival, graft survival, and time to rejection within 1 year of transplantation were assessed by Kaplan-Meier analysis. One-year graft function was compared using Kruskal-Wallis and Fisher’s exact tests.

Results. 

The SRL group had longer DGF duration (P=0.01). The three groups had comparable patient (P=0.27) and graft survival (P=0.69), but the depleting antibody group experienced less rejection (P=0.004). There were no clinically significant differences in 1-year graft function.

Conclusions. 

In our analysis of a large and modern cohort of adult cadaveric transplant recipients with DGF, induction immunosuppression with a depleting antibody preparation reduced rejection, whereas SRL prolonged DGF duration. All three CNI-sparing induction IS regimens resulted in comparable patient survival, graft survival, and graft function.

Delayed graft function (DGF), most commonly defined as the requirement for dialysis during the first week after transplantation, is thought to result from injury incurred by a kidney during and immediately after donation and transplantation (¹). Many risk factors for the development of DGF have been identified and are well accepted. It has been suggested that the more unfavorable the factors, the more severe the injury, the more likely DGF will occur, and the more prolonged the DGF. Although the vast majority of grafts recover, it is believed that the occurrence of DGF compromises outcomes for both the recipient and the graft (^1–10).

Data has recently emerged regarding the potential impact of sirolimus (SRL), a recently available immunosuppressive agent, on DGF. SRL is a bacterial macrolide that, in complex with its cellular receptor, FK506 binding protein, potently inhibits downstream signaling by the mammalian target of rapamycin (mTOR) (^{11, 12}). mTOR is a Ser/Thr kinase widely distributed among multiple signaling pathways in many cells, including renal tubular epithelia. Animal models, both in vitro and in vivo, have suggested that mTOR inhibition results in decreased cellular proliferation, which exacerbates or perpetuates renal ischemic injury (^{13, 14}). Because a transplanted kidney inevitably incurs injury, it is not surprising that comparable effects of SRL on human kidney-transplant recipients have recently been observed. Smith et al. (¹⁵) have reported an increased risk of developing DGF with perioperative SRL administration. Our group has shown that full exposure to SRL significantly prolongs the duration of DGF (¹⁶). Because potent risk factors for developing DGF such as donor age and cold ischemia time did not correlate with time to graft function in our analysis, we postulated that SRL compromises the allograft’s ability to recover from injury. In light of these recent reports, we have undertaken the present study to compare transplant outcomes associated with SRL-based versus other calcineurin-inhibitor (CNI)-sparing regimens in the DGF setting.

MATERIALS AND METHODS

Patients and Data Collection

Retrospective chart review was performed to collect information regarding recipient, donor, transplant, and posttransplant parameters for the 132 of 563 (23.4%) adults who experienced DGF after single-organ, deceased kidney transplantation at University of California San Francisco (UCSF) between January 1, 1997 and June 30, 2001. DGF was defined as any dialysis requirement within 1 week of transplantation.

Immunosuppression

The UCSF IS protocol for recipients with DGF evolved during the study period, predominantly in response to the availability of new agents during the 4.5-year study period. However, the basic philosophy guiding protocol design, avoidance of CNIs to optimize renal recovery, remained coherent.

All recipients received steroids and mycophenolate mofetil (2–3 g orally per day adjusted for hematologic parameters or clinical symptoms). Maintenance CNI initiation, either cyclosporine or tacrolimus, was typically delayed until resolution of DGF. Sensitized recipients (those with history of previous transplant or peak panel reactive antibody (Ab)>30%) were often induced with a depleting Ab preparation; OKT3 was used in 1997, whereas rabbit thymoglobulin was used from 1998 to the end of the study period. Unsensitized recipients were induced with horse antithymocyte globulin in 1997. Beginning at the end of 1997, unsensitized recipients received anti-interleukin-2R receptor (IL-2R) Abs instead. A final protocol change occurred in December 1999 when SRL was added to the anti-IL-2R Ab protocol for primary transplant recipients with peak panel reactive Ab less than 30%. A loading dose was given followed by maintenance doses targeting for trough blood levels of 10 to 15 ng/mL.

Maintenance IS was steroids, mycophenolate mofetil, and CNI for all but six patients (1 died 7 days after transplantation; 3 were enrolled in CNI-free studies and 2 remained on SRL). Tacrolimus was the predominant CNI choice, used for 67% of all recipients.

Rejection

Transplant recipients with DGF typically underwent allograft biopsies at 5- to 10-day intervals until clinical evidence of improving renal function. Two cores were obtained for histologic assessment. All rejection episodes were biopsy-proven and treated with either OKT3 or rabbit thymoglobulin.

Statistical Analysis

Induction groups were compared for all recipient, donor, transplant, and posttransplant parameters using the Kruskal-Wallis test for continuous variables and Fisher’s exact test for discontinuous variables. Kaplan-Meier analysis was used to determine patient survival, graft survival, and time to rejection. Differences in survival distribution were assessed using both Wilcoxon and log-rank statistics. All analyses were performed using SAS, version 8.2 (SAS Institute, Cary, NC).

RESULTS

Recipient, Donor, and Transplant Characteristics

Our study cohort of 132 adult cadaver kidney transplant recipients with DGF were divided into three groups by induction IS regimen (Table 1). Group A comprised 42 (32%) recipients who received depleting Ab (rabbit or horse antithymocyte globulin or OKT3); group B comprised 49 (37%) recipients who received SRL, with (44, 90%) or without (5, 10%) IL-2R Ab; group C comprised 41 (31%) recipients who received neither depleting Ab nor SRL; the majority (32, 78%) received IL-2R Ab. The mean (median) months of follow-up were 42.9 (33), 22.0 (30), and 36.0 (29) months for groups A, B, and C.

TABLE 1: Donor, recipient, and transplant characteristics by induction immunosuppression group.

The three DGF patient groups were similar in all recipient, donor, and transplant parameters except for the distribution of sensitized recipients (P=0.0009) and the degree of human leukocyte antigen (HLA) matching (P=0.03) (Table 1). There was a trend toward longer cold ischemia time in group A and shorter warm ischemia time in group B.

Duration of DGF

DGF was defined as any dialysis requirement during the first week after transplantation, and the duration of DGF was defined as the day of the last dialysis treatment. Dialysis requirement terminated on the transplant day (last dialysis day 0) or 1 day after transplantation (last dialysis day 1) for 11 (8%) recipients distributed evenly among the three groups (4, 4, and 3 in groups A, B, and C). The duration of DGF differed significantly among the three groups (Table 1); both mean and median days of DGF were greatest for the SRL group (P=0.01).

Patient Survival

During the study period, there were 12 (9%) deaths: 2, 3, and 7 in groups A, B, and C (Fig. 1A). Nine of the 12 deaths occurred within 1 year of transplantation (1, 3, and 5 in groups A, B, and C). Overall patient survival rates at 1 year (97.6%, 91.8%, and 87.7%, groups A, B, and C), 3 years (97.6%, 91.8%, and 84.8%, groups A, B, and C), and 5 years (93.0% and 81.4%, groups A and C) did not appear to differ.

FIGURE 1.: Kaplan-Meier curves of patient survival (A) and death censored graft survival (B) for delayed graft function (DGF) groups defined by induction immunosuppression. Group A, lymphocyte depleting antibody (Ab); group B, sirolimus (SRL); group C, neither.

Graft Survival

Of the 132 grafts with DGF, 3 (2%) grafts never functioned and were excluded from all analyses of graft survival: 1 graft still in DGF was lost to patient death on posttransplant day 7; 1 graft was lost to refractory rejection; and 1 graft never functioned for unknown reasons. Of the 129 transplants which functioned, 29 (23%) were lost during the follow-up period. Twelve grafts were lost within 1 year of transplantation, the majority (⁹) secondary to recipient death (1, 3, and 5 in groups A, B, and C). Seventeen (13%) grafts were lost after the first year (7, 5, and 5 in groups A, B, and C). Overall graft survival rates at 1 year (95.2%, 91.5%, and 82.5%, groups A, B, and C), 3 years (82.7%, 87.2%, and 74.3%, groups A, B, and C), and 5 years (70.5% and 71.3%, groups A and C) did not appear to differ among the three groups (data not shown). Rates of graft survival with death censored at 1 year (97.6%, 95.6%, and 97.2%, groups A, B, and C), 3 years (84.7%, 91.2%, and 88.0%, groups A, B, and C), and 5 years (75.8% and 88.0%, groups A and C) also appear comparable among the three groups (Fig. 1B).

Graft Function

Three parameters were used to assess graft function 1 year after transplantation: creatinine (Cr), Cr clearance, and delta (Δ)Cr (1-year Cr to 6-month Cr) (Table 2). This analysis excluded 15 of 129 (12%) grafts: 12 grafts were lost during the first year, and 3 grafts were missing data. There were no apparent differences among the three groups with respect to 1-year Cr (P=0.57) or Cr clearance (P=0.33). Group B, however, did show a more favorable profile of ΔCr (P=0.05). However, all three groups had similar number and percentage of transplants with ΔCr ≥ 0.3 or more mg/dL (P=0.74), the threshold associated with a significant decline in long-term graft survival (¹⁷).

TABLE 2: One-year graft function by induction immunosuppression group

Rejection

In our cohort of 132 DGF patients, 45 (34%) recipients had at least one episode of biopsy-proven rejection within 1 year of transplantation. The incidence of rejection at 1 year was 14.3%, 44.9%, and 41.5% for groups A, B, and C (Fig. 2). The majority of recipients (35 of 45) who rejected did so within the first month after transplantation; 32 recipients rejected before or concurrent with the resolution of DGF. The graphs diverge within the first month and then remain essentially parallel for the follow-up period. Overall, the incidence of rejection for black versus non-black recipients was similar (37.1% vs. 33.0%). However, among the three IS groups, non-black recipients had similar rejection rates (16.1%, 35.3%, and 46.9% for groups A, B, and C), but black recipients had dramatically different rejection rates (9.1%, 66.7%, and 22.2% for groups A, B, and C). We investigated the frequency of allograft biopsy during DGF to determine whether differences in biopsy practice may account for different incidences of rejection among the IS groups. Limiting our analysis to recipients whose DGF duration exceeded 3 days, the three induction groups differed in the number of patients without biopsy (Table 3). However, the DGF duration of these nonbiopsied recipients were very comparable, suggesting a coherent biopsy practice. Similarly, although the groups differed in the number of patients with allograft biopsies, the number of biopsies normalized by the number of patients and the mean DGF duration was also very comparable.

FIGURE 2.: Rejection for DGF groups defined by induction immunosuppression. Group A, lymphocyte depleting Ab; group B, SRL; group C, neither.

TABLE 3: Biopsy frequency during DGF by induction immunosuppression group

To explore the high incidence of rejection in the SRL group, we examined SRL levels. For nearly half (23 of 49) of group-B recipients, no SRL levels were sent: 6 of 15 rejectors (all 6 rejected within 7 days after transplantation) and 17 of 34 nonrejectors with short-lived DGF (5.0 days mean duration) who were quickly transitioned to CNIs. For the 26 recipients with available SRL levels, the first SRL level was obtained 6.9 (6) mean (median) days after transplantation. Interestingly, there were no significant differences between the first SRL level of rejectors and nonrejectors: 10.3 (9.8) mean (median) ng/mL for rejectors and 8.3 (7.4) mean (median) ng/mL for nonrejectors.

Finally, the impact of rejection was assessed by comparing patient and graft survival of our DGF cohort according to rejection status. Rejecters and nonrejectors enjoyed similar patient survival, graft survival (Fig. 3), and 1-year graft function (data not shown).

FIGURE 3.: Patient and graft survival for DGF patients defined by rejection status.

DISCUSSION

The choice of optimal IS for the DGF setting continues to present a challenge. We have undertaken the present study to assess whether three different CNI-sparing IS regimens produce different outcomes for cadaveric grafts with DGF. This was particularly motivated by recent reports that SRL, a non-nephrotoxic immunosuppressant, may nevertheless exacerbate or perpetuate graft injury (^{15, 16}). It is important to recognize that although this study is retrospective, it is also one of the largest published single-center cohorts of DGF recipients. We certainly acknowledge that patient distribution into induction IS groups was not random. However, the primary difference among the groups was that group A had a disproportionate number of sensitized recipients. All of the analyses, when repeated with unsensitized patients alone, yielded comparable results (data not shown). Although the groups also differed in the degree of HLA matching, we would argue that the small difference was clinically insignificant. Otherwise, the three groups were similar in all donor, recipient, and transplant characteristics. We attribute this parity to our center’s consistent approach to IS for DGF. Therefore, the majority of recipients and, particularly, the unsensitized recipients were effectively randomized to an IS strategy by transplant date.

After transplantation, the three groups had similar patient survival, graft survival, and 1-year graft function but differed in the duration of dialysis and the incidence of rejection. Recipients who received SRL experienced longer DGF than those who did not, as we have previously reported (¹⁶). Our current analysis suggests that prolongation of DGF by SRL does not appear to compromise transplant outcome. Historically, because the duration of DGF often reflected the degree of graft injury, prolonged DGF correlated with inferior transplant outcomes (^18–21). We have alternatively postulated that SRL exposure prolonged DGF not by increasing injury severity but rather by preventing recovery. This model of SRL action would predict our current results of apparent outcome parity. Although we expected that group A would have the lowest incidence of rejection, we were surprised that groups B and C were entirely comparable. We suspected but did not find subtherapeutic SRL levels for group-B rejectors. We did find a striking 67% incidence of rejection for blacks in group B. Increased rejection rates for black recipients on SRL-based, CNI-free regimens have been previously reported (^{12, 22}) and may speak to differences in immunologic risk factors or drug-metabolism profiles.

Perhaps our most curious finding is that, in spite of significant differences in rejection incidence, the three DGF patient groups had similar outcomes. Although there may be disagreement as to the impact of DGF alone, there is strong consensus that rejection superimposed upon DGF portends poorly for graft survival (^{1, 4, 7, 8, 10}). Our study, although large as a single-center cohort of DGF recipients, may be underpowered to detect differences with the current follow-up period. Alternatively, expeditious diagnosis (typically by protocol biopsies during DGF) and aggressive treatment of rejection may have attenuated potentially deleterious effects on transplant outcomes.

The apparent similarity of outcomes among the three groups still begs the question as to the optimal IS regimen in the DGF setting. For recipients of high immunologic risk (sensitized, retransplant, possibly black), a depleting Ab appears wise. However, for recipients of low immunologic risk (unsensitized, first transplants, non-black), such potent IS may be neither necessary nor prudent. If a SRL-based regimen is selected, the transplant center and physician should be prepared for a longer period of DGF with the concomitant need for vigilant monitoring and frequent protocol biopsies to ensure expeditious diagnosis and treatment of any underlying rejection. Because there is no single ideal IS regimen for DGF, the choice must made on an individual basis, guided by a balanced compromise between optimization of renal recovery and efficacy against rejection without unnecessary over-immunosuppression.