Higher degrees of human leukocyte antigen (HLA) matching are associated with improved acute rejection and graft survival rates (1, 2). As such, the current kidney allocation scheme set forth by the United Network of Organ Sharing (UNOS) gives priority to candidates with zero HLA mismatches or fewer degrees of mismatch at the DR loci.
In the United States, the majority of kidney transplant recipients receive antibody induction therapy (3). Its use has been associated with a lower incidence of acute rejection compared with no induction (4, 5). However, the potential downsides of induction therapy include an increase in cost, infectious complications, cytopenias, and malignancies (6–8). Although the incidence of acute rejection after kidney transplantation has decreased in recent years, a corresponding improvement in long-term graft survival has not been observed (9). Death with graft function remains the most common cause of graft loss in the United States, and infectious death is one of the leading causes of death among kidney transplant recipients (10). Given the potential complications associated with immunosuppressive therapy, the risk-benefit ratio of administered immunosuppression in kidney recipients must be carefully considered, particularly for those considered to be at low-immunological risk.
In this study, we sought to investigate the impact of induction therapy in HLA 0-mismatch deceased donor kidney recipients, comparing their outcomes to other recipients with varying degrees of HLA 0-mismatch. We also report patient and graft outcomes associated with individual induction agents in HLA 0-mismatch recipients.
RESULTS
Baseline Characteristics
Of the 44,008 patients in the total study population, 14.3% received kidney grafts from 0-mismatch donors (n=6274) and 85.7% received kidney grafts from HLA-mismatched donors (n=37,734). Within the HLA-mismatched group, 4604 recipients had no DR mismatches, 18,431 recipients had one DR mismatch, and 14,699 recipients had two DR mismatches. The median follow-up duration was 834 days (25th–75th range: 380–1433 days).
Table 1 shows baseline characteristics of 0-mismatch and HLA-mismatched recipients. The proportions of recipients given antibody induction in the HLA 0 mismatch group and HLA-mismatched group were 75% and 75.9%, respectively (P=0.12). Compared with HLA-mismatched recipients, more HLA 0-mismatch recipients received interleukin-2 receptor antagonist (IL-2RA) induction (30.3% vs. 25.4%, P<0.001) and less received thymoglobulin (35.7% vs. 41.7%, P<0.001). Among the HLA 0-mismatch group, fewer recipients of alemtuzumab induction were maintained on steroid- containing regimens at hospital discharge compared with recipients of thymoglobulin, IL-2RA, and no induction (alemtuzumab 24.7% vs. thymoglobulin 64.4% vs. IL-2RA 85.3% vs. no induction 85.4%, P<0.001 across the four subgroups). At 6 months posttransplant, fewer recipients of alemtuzumab induction were maintained on steroids (alemtuzumab 23.1% vs. thymoglobulin 61.0% vs. IL-2RA 76.4% vs. no induction 73.6%, P<0.001) and mycophenolate mofetil (MMF)/mycophenolic acid (MPA) (alemtuzumab 71.0% vs. thymoglobulin 80.6% vs. IL-2RA 80.2% vs. no induction 74.2%, P<0.001 across the four subgroups).
;)
TABLE 1: Baseline characteristics
Association of HLA Matching With Outcomes
Transplant outcomes according to the degree of HLA mismatch (0 HLA mismatch, 0 DR mismatch, 1 DR mismatch, and 2 DR mismatches) are shown in Figure 1. A greater incidence of acute rejection during the initial transplant hospitalization, at 6 months (from discharge to 1 year), and at 1 year posttransplant (from discharge to 1 year) was observed with higher degrees of HLA mismatch (Fig. 1A); there was an inverse correlation between the degree of HLA mismatch and graft (Fig. 1B) and patient survival (Fig. 1B). Compared with the entire HLA mismatched group, recipients in the HLA 0-mismatch group had lower rejection rates (initial hospitalization: 2.2% in 0 mismatch group vs. 4.4% in mismatched group, P<0.001; at 6 months: 5.6% vs. 9.8%, P<0.001; at 1 year: 7.4% vs. 12.8%, P<0.001), a lower death-censored graft failure rate (6.3% vs. 11.8%, P<0.001), and lower mortality rate over the follow-up period (9.0% vs. 10.8%, P<0.001).
FIGURE 1: Impact of human leukocyte antigen (HLA) matching on transplant outcomes: (A) acute rejection rates during the initial hospitalization, form discharge to 6 months, and from discharge to 1 year posttransplant; (B) Kaplan-Meier survival curves for death-censored graft survival and patient survival.
Impact of Antibody Induction on Transplant Outcomes
The impact of antibody induction on transplant outcomes was analyzed in HLA 0-mismatch and HLA-mismatched recipients (Fig. 2).
FIGURE 2: Impact of antibody induction (ind) on transplant outcomes in human leukocyte antigen (HLA) 0-mismatch group and mismatched group: (A) acute rejection rates (overall and antibody-treated) during the initial hospital stay for transplant, from discharge to 6 month, and from discharge to 1 year posttransplant; (B) Kaplan-Meier survival curves for death-censored graft survival and patient survival.
Acute Rejection
In the HLA 0-mismatch group, recipients of antibody induction had significantly lower acute rejection rates during the initial hospitalization and at 6 months posttransplant, but not at 1 year posttransplant when compared with those not receiving induction (Fig. 2A). In comparison, those receiving induction in the HLA-mismatched group had lower acute rejection rates during the initial hospitalization, at 6 months, and at 1 year posttransplant compared with those not receiving induction (Fig. 2A).
In multivariate logistic regression, induction was associated with decreased risks of acute rejection during the initial hospitalization (odds ratio [OR]=0.45 [0.31–0.64], P<0.001) and at 6 months (OR=0.71 [0.57–0.89], P=0.003), but not at 1 year posttransplant (OR=0.88 [0.71–1.09], P=0.25) in the HLA 0-mismatch group. There was a decreased risk of antibody-treated rejection in HLA 0-mismatch recipients associated with induction during the initial hospitalization (OR=0.37 [0.24–0.57], P<0.001), but this effect was not observed at 6 months (OR=0.88 [0.62–1.25], P=0.47) and 1 year posttransplant (OR=1.05 [0.68–1.61], P=0.84). In HLA-mismatched recipients, induction was associated with decreased risks of rejection during the initial hospitalization (OR=0.45 [0.40–0.50], P<0.001), at 6 months (OR=0.78 [0.72–0.84], P<0.001), and at 1 year (OR=0.87 [0.81–0.93], P<0.001). Induction was associated with a decreased risk of antibody-treated rejection during the initial hospitalization OR=0.42 [0.37–0.48], P<0.001) in HLA-mismatched recipients, but this effect was not observed at 6 months (OR=0.97 [0.85–1.10], P=0.61) and at 1 year posttransplant (OR=1.09 [0.95–1.25], P=0.24).
Death-Censored Graft Failure
Kaplan-Meier survival curves for death-censored graft survival according to the usage of antibody induction for HLA 0-mismatch and HLA-mismatched recipients are shown in Figure 2(B). For both 0-mismatch and HLA-mismatched recipients, no significant differences in death-censored graft survival were observed between recipients with and without antibody induction (HLA 0-mismatch: survival rates at 1, 3, and 5 years posttransplant: 97.1%, 93.7%, and 89.9% in those with induction; 97.7%, 94.8%, and 89.4% in those without; P=0.57; HLA-mismatch: 95.1%, 88.6%, and 81.0% in those with induction; 95.1%, 88.2%, and 81.2% in those without; P=0.67).
In multivariate Cox regression analysis, antibody induction was not significantly associated with death-censored graft failure in the HLA 0-mismatch (HR=1.06 [0.83–1.35], P=0.65) and mismatched group (HR=1.04 [0.97–1.12], P=0.28).
Mortality
Kaplan-Meier survival curves for patient survival according to the usage of antibody induction in the two patient groups are shown in Figure 2(B). There were no significant differences in patient survival between those receiving and not receiving antibody induction both in the HLA 0-mismatch group (survival rates at 1, 3, and 5 years posttransplant: 96.8%, 91.9%, and 85.6% in those with induction; 96.6%, 89.8%, and 83.6% in those without; P=0.21) and HLA mismatched group (95.4%, 89.2%, and 81.9% in those with induction; 95.7%, 89.7%, and 82.2% in those without; P=0.20). In multivariate Cox regression analysis, antibody induction was not significantly associated with mortality in both the HLA 0 mismatch (HR=0.87 [0.72–1.05], P=0.16) and mismatched group (HR=1.05 [0.97–1.13], P=0.24).
Impact of Individual Induction Agents on Transplant Outcomes in HLA 0 Mismatch and Mismatched Recipients
Acute Rejection
In HLA 0-mismatch recipients, compared with no induction, thymoglobulin and IL-2RA were both associated with a decreased risk of overall rejection during the initial hospitalization and at 6 months, but not at 1 year posttransplant. The two agents were associated with a decreased risk of antibody-treated rejection during the initial hospitalization, but not at 6 months and 1 year. In multivariate Cox regression, neither thymoglobulin nor IL-2RA was associated with a risk of acute rejection at 1 year posttransplant (Table 2).
TABLE 2: Effects of individual induction agent on outcomes in HLA 0 mismatch and mismatched recipients
In HLA-mismatched recipients, compared with no induction, IL-2RA was associated with a decreased risk of overall rejection during the initial hospitalization, but not at 6 months and 1 year posttransplant. Thymoglobulin was associated with a decreased risk of overall rejection over the first year posttransplant. The two agents were associated with a decreased risk of antibody-treated rejection during the initial hospitalization, but not at 6 months and 1 year.
Graft Failure/Mortality
Kaplan-Meier survival curves for graft and patient survival according to the selection of induction agent in 0-mismatch recipients are shown in Figure 3. There were no significant differences in graft and patient survival over the follow-up period (834 days in median) among non-induction, thymoglobulin, IL-2RA, and alemtuzumab recipients. In multivariate Cox regression analysis, neither thymoglobulin nor IL-2RA were associated with a difference in the risk of death- censored graft survival or patient death over the follow-up period when compared with no induction, in 0-mismatch and mismatched recipients (Table 2).
FIGURE 3: Kaplan-Meier survival curves for (A) graft survival and (B) patient survival in HLA 0-mismatch recipients according to the selection of antibody induction.
Subgroup Analyses
African Americans
In HLA 0-mismatch African Americans recipients (n=721), antibody induction was associated with a decreased risk of rejection during the initial hospitalization (OR=0.23[0.08–0.65], P=0.006), and a trend toward decreased risk of rejection at 6 months (OR=0.63 [0.35–1.15], P=0.14) and 1 year posttransplant (OR=0.64 [0.37–1.14], P=0.13) in the multivariate logistic regression model. The small patient number may have contributed to the non-significant findings. In HLA-mismatched African Americans recipients (n=13,209), antibody induction was associated with decreased risks of rejection during the initial hospitalization (OR=0.41 [0.35–0.48], P<0.001), at 6 months (OR=0.64 [0.57–0.72], P<0.001) and 1 year posttransplant (OR=0.72 [0.65–0.80], P<0.001). Induction was not significantly associated with a risk of graft loss in both HLA 0-mismatch (HR=0.82 [0.46–1.47], P=0.71) and mismatched recipients (HR=0.94 [0.84–1.04], P=0.24). The use of induction was not significantly associated with mortality in both HLA 0-mismatch (HR=1.10 [0.59–2.0], P=0.76) and mismatched recipients (HR=0.98 [0.85–1.11], P=0.71).
Recipients of Triple Immunosuppression
To minimize the confounding effects of different maintenance immunosuppression regimens, we performed a subgroup analysis of recipients who received a triple-maintenance immunosuppression regimen consisting of tacrolimus, MMF/MPA, and steroids at discharge.
In HLA 0-mismatched recipients, induction (yes vs. no) was associated with a decreased risk of acute rejection during the initial hospitalization (OR=0.46 [0.30–0.73], P=0.001) and a trend toward decreased acute rejection at 6 months (OR=0.75 [0.55–1.01], P=0.06) but no difference at 1 year (OR=0.89 [0.66–1.19], P=0.44). In HLA-mismatched recipients, the protective effect of induction on acute rejection was observed during the initial hospitalization (OR=0.44 [0.39–0.51], P<0.001), at 6 months (OR=0.75 [0.67–0.83], P<0.001), and at 1 year posttransplant (OR=0.81 [0.74–0.90], P<0.001).
Induction was not associated with a risk of death-censored graft failure in HLA 0-mismatch (HR=1.09 [0.76–1.57], P=0.64) and mismatched recipients (HR=1.07 [0.96–1.19], P=0.20). Induction was not significantly associated with a risk of mortality in HLA 0-mismatch (HR=1.02 [0.77–1.37], P=0.88) and mismatched recipients (HR=1.07, P=0.21). These results are comparable with those in the overall cohort analysis.
The impact of individual induction agents on outcomes in the triple-immunosuppression subgroup are summarized in the supplementary table (see SDC 1, https://links.lww.com/TP/A594). In HLA 0-mismatch recipients, thymoglobulin and IL-2RA were both associated with a decreased risk of overall rejection during the initial hospitalization, but not at 6 months and at 1 year posttransplant compared with no induction. The two agents were associated with a decreased risk of antibody-treated rejection during the initial hospitalization, but not at 6 months and 1 year. In HLA-mismatched recipients, IL-2RA was associated with a decreased risk of overall rejection during the initial hospitalization, but not at 6 months and at 1 year posttransplant compared with no induction. Thymoglobulin was associated with a decreased risk of overall rejection over the first year posttransplant. The two agents were associated with a decreased risk of antibody-treated rejection during the initial hospitalization, but not at 6 months and 1 year. These results are comparable with those in the overall cohort analysis.
DISCUSSION
The use of antibody induction therapy has increased in the last decade such that the majority of kidney transplant recipients in the United States now receive antibody induction (3). Over the same period, acute rejection rates have decreased. The reduction in rejection incidence has been attributed in part to the advent of newer immunosuppressive agents in the late 1990s (9). In our study, we found that the cumulative rate of rejection at 1 year in HLA 0-mismatch recipients was 8.2%; among HLA 0-mismatch recipients, 7.8% of induction therapy recipients had lower acute rejection rates in the first year compared with 9.5% of those not receiving induction (P=0.03). The difference in rejection rates between induction therapy recipients and non-recipients mainly arose from a reduction in early rejection during the initial hospitalization (1.6% vs. 4.1%, P<0.001) and to a lesser extent from a lower rate of rejection from hospital discharge to 6 months (5.2% vs. 6.9%, P=0.049). There was no difference in the rate of rejection from discharge to 1 year (7.1% vs. 8.2%, P=0.4). The reduction in rejection incidence did not lead to an improvement in graft or patient survival among induction recipients with follow-up to 5 years.
In our previous study using the Organ Procurement and Transplant Network/UNOS database of an earlier cohort of kidney recipients from 1999 to 2001, 51% of deceased donor recipients received induction therapy (thymoglobulin or IL-2RA) and those with HLA mismatches were more likely to receive antibody induction (11). Induction was associated with a reduced risk of acute rejection in the first year after transplant for deceased donor recipients regardless of the level of HLA mismatch. In the more recent era, the usage of antibody induction has been steadily increasing, even in recipients with low immunological risk. Of the current study cohort (adult recipients between 2003 and 2008), approximately 75% received antibody induction, with similar utilization of induction in HLA 0-mismatch and HLA-mismatched recipients. Our data show that the impact of induction therapy on acute rejection incidence is greater for HLA-mismatched recipients compared with HLA 0-mismatch recipients. Less rejection was observed during the initial transplant hospitalization, at 6 months, and at 1 year among HLA-mismatched induction therapy recipients compared with non-recipients.
The most common induction agents used in the United States are thymoglobulin and IL-2RA (basiliximab and daclizumab); evidence for their efficacy comes from prospective randomized trials and registry database series. In a prospective randomized trial comparing thymoglobulin and basiliximab induction among deceased donor recipients considered to be at high risk for acute rejection or delayed graft function, there was a lower incidence of acute rejection overall and antibody-treated rejection among thymoglobulin recipients compared with basiliximab (7). These results were substantiated by Noel et al. (12) in a randomized trial comparing daclizumab and thymoglobulin induction in high-immunologic risk recipients (current panel reactive antibody [PRA] ≥30%, peak PRA ≥50%, or re-transplanted recipients). However, despite a decrease in acute rejection associated with thymoglobulin in these two trials, no differences in graft survival were observed in either study.
In low-immunologic risk recipients, no prospective randomized trial has evaluated the impact of induction therapy on graft outcomes and the best available evidence arises from retrospective registry series. In an analysis of the Scientific Registry of Transplant Recipients database of an older cohort of kidney recipients (1998–2003), Patlolla et al. (13) reported that induction therapy with IL-2RA or lymphocyte-depleting antibodies was associated with a lower adjusted risk of rejection at 6 months and 1 year compared with no induction. Furthermore, IL-2RA was associated with a small decreased risk for death-censored graft failure compared with no induction. A stronger effect of IL-2RA induction on acute rejection incidence at 6 months with increasing degrees of HLA mismatch was observed (P for interaction between IL-2RA and HLA-mismatch=0.007); however, there was no differential effect of lymphocyte-depleting induction on acute rejection at 6 months by degree of HLA mismatch (P for interaction=0.11). The study may not be applicable to more contemporary kidney transplant recipients, however. Reflecting immunosuppression usage from an earlier era, a greater proportion of patients in the study by Patlolla et al. was maintained on cyclosporine-based immunosuppression rather than tacrolimus and approximately one third of recipients receiving lymphocyte-depleting induction did not receive thymoglobulin.
A more recent analysis performed by Lim et al. (14) evaluated outcomes associated with the use of IL-2RA induction using the Australia and New Zealand Dialysis and Transplant Registry. In low immunologic-risk recipients, defined by less than or equal to two HLA mismatches and PRA less than 10%, there was no association between the use of IL-2RA induction and acute rejection at 6 months, despite greater utilization of cyclosporine and azathioprine-based maintenance immunosuppression in recipients not receiving induction therapy.
To control for the influence of maintenance immunosuppression on graft outcomes, we performed a subgroup analysis of recipients maintained on tacrolimus, MMF, and prednisone, the most common maintenance immunosuppression regimen used in the United States. Antibody induction recipients had a reduction in acute rejection during the initial transplant hospitalization, a trend toward decreased rejection at 6 months, and no difference in acute rejection or at 1 year. The benefit of antibody induction on acute rejection beyond the initial transplant hospitalization in those maintained on tacrolimus, MMF, and prednisone was seen only in HLA-mismatched recipients. These findings are comparable with that observed in the overall cohort analysis. It should be noted that we could not account for immunosuppression changes occurring after the initial transplant hospitalization nor did we have information on dosages and drug levels. In our study, the overall incidence of acute rejection was low among HLA 0-mismatch recipients throughout the first year. Although there was a statistically significant reduction in acute rejection during the initial transplant hospitalization and at 6 months, the absolute risk reduction was small (2.5% risk reduction during the initial hospitalization and 1.7% reduction over the first 6 months). Furthermore, it may be that the impact of early acute rejection episodes may not be as profound as those developing late. It has been reported that long-term graft survival is markedly worse among those developing a late episode of rejection (≥3 months posttransplant) compared with those occurring within the first 3 months (15). Therefore, given the inherently low incidence of acute rejection in HLA 0-mismatch recipients coupled with the modest benefit observed with induction therapy, our findings cast uncertainty about the benefit of induction therapy in HLA 0-mismatch recipients.
This retrospective analysis is subject to limitations inherent to the nature of registry studies, such as reporting bias or error. We adjusted for known recipient, donor, and transplant-related factors in our multivariate models. However, it is possible that other unidentified confounders may not be adjusted for in retrospective analysis. Furthermore, we were unable to evaluate for indicators of drug toxicity, such as number of re-hospitalizations, infectious complications, and adverse effects. It is also important to note that the number of alemtuzumab recipients in this study was small, and the observations made about antibody induction were thus largely driven by outcomes associated with thymoglobulin and IL-2RA administration. Finally, to fully impact the relative risk/benefit of induction in HLA 0-mismatch recipients, a cost-benefit analysis should be considered.
In conclusion, acute rejection was infrequent among HLA 0 mismatch deceased donor kidney transplants during 2003 to 2008 in the United States. Antibody induction was associated with a modest salutary effect on acute rejection early posttransplant (during the initial transplant hospitalization and at 6 months) but not at 1 year; among HLA 0-mismatch recipients maintained on tacrolimus, MMF/MPA, and corticosteroids at hospital discharge, antibody induction was associated with a reduced risk of acute rejection during the initial transplant hospitalization alone. Compared with no induction, thymoglobulin and IL-2RA were associated with a lower risk of acute rejection over the first 6 months but not at 1 year. There was no association between the use of induction therapy and death-censored graft and patient survival over the follow-up period (834 days in median) among 0-mismatch recipients. Our findings may be of benefit when considering the use of induction therapy for 0-mismatch recipients.
MATERIALS AND METHODS
Study Population
Using the Organ Procurement and Transplant Network/UNOS database as of November 27, 2009, we identified all adult recipients (≥18 years) who received a deceased donor kidney transplant alone between 2003 and 2008 (n=47,854). Those who had missing information on recipient/donor HLA phenotypes (n=16) or induction therapy (n=891), received anti-thymocyte globulin/muromonab-CD3 induction (n=667), or received multiple induction agents (n=2272) were excluded. A total of 44,008 recipients were selected into the total study population. Among them, 17,980 recipients (40.9%) received thymoglobulin induction, 11,480 recipients (26.1%) received an IL-2RA, 3882 recipients (8.8%) received alemtuzumab, and 10,666 recipients (24.2%) received no antibody induction. The impact of induction (thymoglobulin, IL-2RA, or alemtuzumab) on acute rejection during the initial transplant hospitalization, between discharge and 6 months posttransplant, and between discharge and 1 year posttransplant; death-censored graft survival; and patient survival were compared among HLA 0-mismatch and HLA-mismatched recipients. The effect of individual induction agents on transplant outcomes was further analyzed.
Statistical Analysis
Recipient, donor, transplant, and immunosuppressive characteristics were described using means (±standard deviation) or frequencies as appropriate. Continuous variables, including age, body mass index, duration of pretransplant dialysis, and cold ischemia time were categorized because their associations with the study outcomes were not linear. Comparisons among groups were made using the Wilcoxon rank-sum test for continuous variables and chi square test for categorical variables. Survival rates were estimated using the Kaplan-Meier product limit method. The log-rank test was used for comparison of the unadjusted survival curves.
Logistic regression models were used to estimate the ORs of acute rejection (overall rejection and antibody-treated rejection) during the initial hospital stay for transplant, at 6 months posttransplant(from hospital discharge to 6 months), and by1 year posttransplant (from hospital discharge to 1 year). Cox regression models were used to estimate the hazard ratios (HR) of death-censored graft failure and mortality during follow-up. The analyses were made in the overall cohort, a sub-group of African Americans, and a subgroup of recipients who received a triple-maintenance immunosuppression regimen consisting of tacrolimus, MMF/MPA, and steroids at hospital discharge.
Major recipient, donor, transplant, and immunosuppression-related factors were adjusted for in the multivariate analysis. Recipient-related factors included gender, age, pretransplant body mass index (kg/m2), race, primary payment (private insurance, Medicare, and others), presence of hepatitis C virus antibody, duration of pretransplant dialysis, and peak PRA. Donor-related factors included gender, expanded criteria donor, and donor/recipient cytomegalovirus serostatus. Transplant-related factors included cold ischemia time, delayed graft function, and transplant year. Immunosuppression-related factors included selection of tacrolimus, MMF/MPA, mammalian target of rapamycin inhibitors, and maintenance steroid use at hospital discharge (for rejections during transplant and by 6 months posttransplant) and at 6 months posttransplant (for rejection at 1 year posttransplant and graft/patient survivals over the follow-up period. Potential confounding factors in univariate analysis (P<0.1) and variables of immunosuppression at discharge were adjusted for in the multivariate analysis.
All P values were two-tailed and a P less than 0.05 was considered statistically significant. All analyses were performed using Stata Statistical Software, Release version 10 (StataCorp LP, College Station, TX).
ACKNOWLEDGMENT
The authors dedicate this paper to the late Dr. Alan H. Wilkinson, who died during the revision of this manuscript. He was a dedicated mentor and clinician, and his passing was a great loss to UCLA and the transplant community.
