Long-Term Outcome after Early Mammalian Target of Rapamycin Inhibitor-Based Immunosuppression in Kidney Transplant Recipients

Background: The use of mammalian target of rapamycin inhibitors (mTORis) in kidney transplantation increases the risk of donor-specific human leukocyte antigen (HLA) antibody formation and rejection. Here, we investigated the long-term consequences of early mTORi treatment compared to calcineurin inhibitor (CNI) treatment. Methods: In this retrospective single-center analysis, key outcome parameters were compared between patients participating in randomized controlled immunosuppression trials between 1998 and 2011, with complete follow-up until 2018. The outcomes of eligible patients on a CNI-based regimen (n = 384) were compared with those of patients randomized to a CNI-free mTORi-based regimen (n = 81) and 76 patients randomized to a combination of CNI and mTORi treatments. All data were analyzed according to the intention-to-treat (ITT) principle. Results: Deviation from randomized immunosuppression for clinical reasons occurred significantly more often and much earlier in both mTORi-containing regimens than in the CNI treatment. Overall patient survival, graft survival, and death-censored graft survival did not differ between the treatment groups. Donor-specific HLA antibody formation and BPARs were significantly more common in both mTORi regimens than in the CNI-based immunosuppression. Conclusions: The tolerability and efficacy of the mTORi treatment in kidney graft recipients are inferior to those of CNI-based immunosuppression, while the long-term patient and graft survival rates were similar.


Introduction
The use of the calcineurin inhibitors (CNIs) cyclosporine or tacrolimus in combination with anti-IL-2 receptor antibodies (e.g., basiliximab) and mycophenolic acid (MPA; e.g., mycophenolate mofetil (MMF) or enteric-coated mycophenolate sodium (EC-MPS)), with or without steroids, has led to major advances in renal transplantation [1] with excellent shortterm outcomes, and they are recommended by current guidelines for immunologically low-risk patients [2].However, CNIs are nephrotoxic and are a frequent secondary cause of graft loss [3].The development of mammalian target of rapamycin inhibitors (mTORis) has led to the availability of less nephrotoxic immunosuppressive drug combinations.Sirolimus (rapamycin) is a macrocyclic triene antibiotic produced by the fermentation of Streptomyces hygroscopicus.Sirolimus was discovered in a soil sample collected from Rapa Nui, also known as Easter Island [4].Sirolimus has been approved as an initial immunosuppressive treatment after kidney transplantation in combination with cyclosporine and corticosteroids [5].After 2-3 months, the cyclosporine dose should be tapered and eventually withdrawn.Thereafter, the regimen should consist of sirolimus and corticosteroids [6].
Everolimus in combination with cyclosporine and corticosteroids has been approved as an initial immunosuppressive treatment after kidney and heart transplantations.Furthermore, it has been approved for use in liver allograft recipients in combination with tacrolimus and corticosteroids.Notably, the withdrawal of CNIs is not recommended in everolimus-containing regimens after solid organ transplantation.To avoid the renal toxicity of CNIs in combination with everolimus, it has been recommended to reduce cyclosporine or tacrolimus early after transplantation [7].
Twenty years after the approval of sirolimus (Europe, 2001; US, 2003) and its derivative everolimus (Europe, 2004; US, 2010), the mTORi principle is rarely used and is still seeking its place in clinical kidney transplantation.Side-effects are frequent, and KDIGO found no clear advantage in mTORis in comparison with other antiproliferative agents or as a replacement for or combination with CNIs [2].Our center has participated in many clinical trials of mTORis, and we observed an increased risk of de novo donor-specific HLA antibody (dnDSA) formation and antibody-mediated rejection in patients treated with everolimus compared to those treated with cyclosporine [8].More recently, studies from the US and Australia, as well as a meta-analysis, have described increased mortality among kidney transplant recipients treated with mTORi-based immunosuppression [9][10][11][12].These and other recent investigations on the use of everolimus in combination with reduced CNI exposure [13][14][15] or standard-dose CNIs [16,17] in renal transplantation led us to re-evaluate and extend our previous analyses.We aimed to investigate the long-term consequences of early mTORi use (either in combination with CNIs or in a CNI-free regimen) compared with standard CNI-based immunosuppression.

Methods
In this retrospective single-center analysis, we studied the outcomes of patients who participated in a number of clinical trials at our center and who were primarily randomized to CNIs, a CNI-free mTORi-based regimen, or a combination of CNIs and mTORis.Thus, the analysis benefits from the stringent exclusion criteria, the randomized use of immunosuppressive agents, and complete long-term follow-up in our database.

Patients
To identify as many patients eligible for this study as possible, all adult recipients of a kidney allograft transplanted between 10 September 1998 and 8 February 2011 were screened.In 1998, the first patient was randomized to an mTORi in a clinical trial.By 8 February 2011, the last transplantation of a patient randomized to receive mTORi-based immunosuppression was performed.To exclude patients with early intolerability, we included only those with at least 4 weeks of randomized treatment with CNIs (intentionto-treat (ITT) group 1), mTORis (ITT group 2), or a combined treatment with CNIs and mTORis (ITT group 3).The prospective randomized controlled trials comparing different immunosuppressant regimens are summarized in Table 1 .The decision for transplantation was based on a negative CDC crossmatch and the absence of DSAs at the time of transplantation.The dosing of CNIs and mTORis was based on trough levels and in accordance with study protocols.Initially, standard doses of mycophenolic acid were administered, and later individualized on the basis of side-effects and tolerability.Most trials allowed for a steroid treatment according to center policy.Our center policy is to administer steroids for at least 12 months.Thereafter, the withdrawal of steroids is based on individual decision making, considering the immunological risk (namely, previous rejections), side-effects, underlying renal disease, and patient preferences.

HLA Antibody Screening
HLA antibody screening was performed on the discretion of the renal physician responsible for the follow-up and at the time of indication of graft biopsies.In addition, serum samples were prospectively collected at least once a year for the HLA antibody screening starting in 2006.All of the serum samples were qualitatively screened for HLA antibodies using the Luminex-based bead assay LABScreen Mixed (One Lambda; Canoga Park, CA, USA).Donor specificity of the HLA antibodies was determined using LABScreen Single Antigen Beads (One Lambda).All of the tests were performed in accordance with the manufacturer's guidelines.

Graft Biopsies
All of the graft biopsies were performed for clinical indications, such as impaired kidney function and/or proteinuria.Renal allograft pathology was performed by two experienced nephropathologists (B.R. and K.W.) according to the Banff classification.All of the histological findings were re-categorized in 2010 according to the Banff 09 classification [43], and later to the Banff 2017 classification [44].

Follow-Up
All of the patients underwent regular follow-up at our outpatient clinic.All of the clinical data were retrieved from the web-based electronic patient record system TBase [45].The end of the observation period was June 30th, 2018.
The Institutional Review Board of Charité-Universitätsmedizin Berlin approved the retrospective analysis of patient data, with a waiver of informed consent (EA1/048/14).

Statistics
We used IBM SPSS Statistics for Windows version 27.0.(IBM Corp., Armonk, NY, USA) for all of the statistical analyses.Descriptive data were summarized as medians with ranges.Categorical data were presented as proportions.
The endpoints were death, graft loss, de novo DSA formation, and first biopsy-proven acute rejection (BPAR) episodes after the first month.Events over time (Kaplan-Meier) were analyzed using the log-rank test.
Cox regression analyses were used to identify prognostic factors that contributed to the endpoint variability.Therefore, univariate analyses with a number of factors that potentially affect the outcomes were performed: recipient age (≥65 vs. <65 years), recipient sex (female vs. male), number of KTXs (repeated vs. first), waiting time (<60 vs. ≥60 months), renal disease (GN vs. other), PRAs pretransplant (positive vs. negative), donor age (≥65 vs. <65), donor sex (female vs. male), donor type (living vs. deceased), number of mismatches (4-6 vs. 0-3), KTX function (DGF vs. primary function), AR within the first month (yes vs. no), ITT group (mTORis, CNIs + mTORis vs. CNIs), steroid stop (yes vs. no), and KTX date (after March 31, 2005 vs. before March 31, 2005).Variables with a p < 0.20 in the univariate analysis were included in a multivariate Cox proportional hazard model, and a stepwise backward elimination strategy (p > 0.05) was applied to identify the risk factors for overall patient survival, death-censored graft loss, overall graft loss, dnDSA formation, and biopsy-proven acute rejection after the first month.
A probability of less than 0.05 was considered as statistically significant.

Patient Population
Figure 1 illustrates the steps used to define the study cohort.Only patients with at least four weeks of randomized treatment with CNIs (ITT group 1, n = 384), mTORis (ITT group 2, n = 81), or simultaneous treatment with CNIs and mTORis (ITT group 3, n = 76) were included in the study.The baseline characteristics of the intention-to-treat groups are presented in Table 2.
A total of 268 of 384 (69.84%) patients who were randomized to receive a CNI-based regimen remained on CNI maintenance treatment until graft loss, death with a functioning graft, or the end of observation.Of the CNI sub-cohort, 17 patients were switched to mTORis during the first year, and 90 were switched later (after 12 months) on the discretion of their renal physicians.The reasons for the mTORi changeover included neoplasia (n = 23), CNI nephrotoxicity (n = 61), viral infections (n = 6), or other reasons, including participation in other clinical trials (n = 17).
Of the 81 patients who were randomized to receive an mTORi-based regimen, only 20 (24.7%) remained on mTORis until graft loss, death, or the end of observation.In total, 59 patients were switched to CNIs for clinical reasons (mTORi-specific toxicity, n = 22; biopsy-proven acute rejections (BPARs), n = 15; detection of dnDSA formation, n = 12; increased risk of dnDSA formation, n = 10).Two patients were converted to belatacept because of their participation in a clinical trial.
Of the 76 patients who were randomized to mTORis plus CNIs, only 4 (5.3%)remained on this combination until graft loss, death, or the end of observation.Twenty-three patients were switched to other regimens because of CNI toxicity.An mTORi-specific toxicity was mentioned in the records of 14 patients.Ten patients left the protocol after a BPAR diagnosis.Five patients were switched to prevent ABMR after dnDSA formation was detected.Combinations of the mentioned reasons for changeover or other reasons led to the end of the combined treatment in 20 patients.Deviations from randomized immunosuppression were significantly more common and occurred earlier in the mTORi groups than in the CNI-based treatment group (Figure 2).

Patient Survival
The risk of death in our cohort was not dependent on early immunosuppression (Figure 3).Sixty months after transplantation, the overall survival rates were 88.9%, 90.8%, and 89.1% in patients treated with mTORis, a combination, and CNI-treated patients, respectively.
Recipient age, underlying renal disease, donor sex, donor type, and decision to withdraw steroids were significant confounders in the multivariate model of patient survival, whereas early immunosuppression was not (Table 3).
formation was detected.Combinations of the mentioned reasons for changeover or other reasons led to the end of the combined treatment in 20 patients.
Deviations from randomized immunosuppression were significantly more common and occurred earlier in the mTORi groups than in the CNI-based treatment group (Figure 2).Patients on the combination of mTORis plus CNIs were more likely to be changed than patients on mTORis plus MPA (log-rank: p = 0.001).

Patient Survival
The risk of death in our cohort was not dependent on early immunosuppression (Figure 3).Sixty months after transplantation, the overall survival rates were 88.9%, 90.8%, and 89.1% in patients treated with mTORis, a combination, and CNI-treated patients, respectively.Recipient age, underlying renal disease, donor sex, donor type, and decision to withdraw steroids were significant confounders in the multivariate model of patient survival, whereas early immunosuppression was not (Table 3).

Graft Loss
The risk of death-censored graft loss was not dependent on early immunosuppression (Figure 4).Recipient age, number of KTXs, and occurrence of DGF increased the risk of death-censored graft loss in the multivariate model, whereas neither type of randomized early immunosuppression did (Table 3).The decision to discontinue steroids was associated with a decreased risk of death-censored graft loss.
Analysis of the overall graft survival did not show significant differences between the ITT groups (Figure 5).Recipient age, donor type, DGF, steroid withdrawal, renal disease, and year of transplantation (era) were significantly associated with overall graft survival (Table 3).

BPAR
Biopsy-proven acute rejection in indication biopsies ≥1 month after transplantation was significantly more common in mTORi-treated patients than in those treated with CNIbased immunosuppression (Figure 6).Sixty months after transplantation, the cumulative rates of the first BPAR more than 1 month after transplantation were 33.8%, 33.4%, and 17.3% in CNI-free mTORi-treated, mTORi + CNI, and CNI-treated patients, respectively.
Risk factors for BPAR in the multivariate analysis included a higher number of mismatches, the occurrence of DGF, and randomized treatment with mTORis.Remarkably, the decision to withdraw steroids from the immunosuppressive maintenance regimen was not associated with increased rejection rates after month 1 (Table 3).

dnDSA Formation
De novo DSA formation was significantly more common throughout the observation period in both mTORi-containing regimens than in the CNI group (Figure 7).Sixty months after transplantation, dnDSAs were detected in 35.1%, 30.4%, and 15.6% of CNI-free mTORitreated, CNI + mTORi, and CNI-treated patients, respectively.Analysis of the overall graft survival did not show significant differences between the ITT groups (Figure 5).Recipient age, donor type, DGF, steroid withdrawal, renal disease, and year of transplantation (era) were significantly associated with overall graft survival (Table 3).

BPAR
Biopsy-proven acute rejection in indication biopsies ≥1 month after transplantation was significantly more common in mTORi-treated patients than in those treated with CNIbased immunosuppression (Figure 6).Sixty months after transplantation, the cumulative rates of the first BPAR more than 1 month after transplantation were 33.8%, 33.4%, and 17.3% in CNI-free mTORi-treated, mTORi + CNI, and CNI-treated patients, respectively.
Risk factors for BPAR in the multivariate analysis included a higher number of mismatches, the occurrence of DGF, and randomized treatment with mTORis.Remarkably, the decision to withdraw steroids from the immunosuppressive maintenance regimen was not associated with increased rejection rates after month 1

dnDSA Formation
De novo DSA formation was significantly more common throughout the observation period in both mTORi-containing regimens than in the CNI group (Figure 7).Sixty months after transplantation, dnDSAs were detected in 35.1%, 30.4%, and 15.6% of CNI-free mTORi-treated, CNI + mTORi, and CNI-treated patients, respectively.Table 3 summarizes the contributions of different variables to the formation of de novo DSAs in the multivariate model.In the final model, the randomized use of mTORis was associated with an increased risk of dnDSA formation, such as pretransplant PRAs, the number of mismatches, and the occurrence of DGF.Remarkably, steroid withdrawal from immunosuppression was associated with less dnDSA formation in the multivariate model.
In secondary confirmatory analyses, we studied whether clinically initiated conversion from CNIs to mTORis was associated with a higher frequency of dnDSAs.The switch from CNIs (ITT group 1) to mTORis for clinical reasons was also associated with an increased risk of dnDSA formation, similar to the ITT analysis (Figure 8).Patients who Table 3 summarizes the contributions of different variables to the formation of de novo DSAs in the multivariate model.In the final model, the randomized use of mTORis was associated with an increased risk of dnDSA formation, such as pretransplant PRAs, the number of mismatches, and the occurrence of DGF.Remarkably, steroid withdrawal from immunosuppression was associated with less dnDSA formation in the multivariate model.
In secondary confirmatory analyses, we studied whether clinically initiated conversion from CNIs to mTORis was associated with a higher frequency of dnDSAs.The switch from CNIs (ITT group 1) to mTORis for clinical reasons was also associated with an increased risk of dnDSA formation, similar to the ITT analysis (Figure 8).Patients who continued CNIs over time had the lowest risk of dnDSAs (after 60 months: 11.2% with unchanged CNI treatment vs. 24.4% after conversion to mTORis).

Discussion
In general, long-term data from randomized trials are sparse.The long-term followup of patients who participated in several randomized clinical trials at our center enabled us to perform a detailed retrospective ITT analysis of the long-term outcomes of different

Discussion
In general, long-term data from randomized trials are sparse.The long-term follow-up of patients who participated in several randomized clinical trials at our center enabled us to perform a detailed retrospective ITT analysis of the long-term outcomes of different immunosuppressive protocols following kidney transplantation.
High rates of deviations from randomized immunosuppression, on the one hand, and the desire to correlate early immunosuppression with late events were the arguments against a per-protocol analysis.
Patients randomized to mTORis after kidney transplantation, either as a replacement for CNIs or in combination with CNIs as a replacement for mycophenolic acid, clearly had inferior tolerability and efficacy, as evidenced by significantly higher discontinuation rates, rejection rates, and the frequent development of dnDSAs.Only a minority of patients were maintained on an originally randomized regimen containing mTORis until graft loss, death, or the end of observation.Patients randomized to receive standard CNI-based immunosuppression experienced lower rejection rates and reduced dnDSA formation, contributing to the overall good tolerability of this regimen.
High dropout rates of patients on mTORis were seen in virtually all trials with this class of immunosuppressants, last reported in the TRANSFORM and Athena studies [13][14][15][16][17].The adverse effects of sirolimus and everolimus are diverse [46,47] and summarized by Kaplan et al. [48].Importantly, we observed several serious side-effects, such as 12 pulmonary adverse events (e.g., pneumonitis and interstitial lung disease) as well as acute rejections and the development of dnDSAs, which might have serious consequences for long-term graft survival [3].
Even with a rather short exposure to mTORis, the rate of dnDSA formation significantly increased and was twice as high as that in the comparator arm.In combination with the increased rate of BPARs, this study clearly supports and extends our previous observation [8], which, in 2012, initiated a controversial debate on the efficacy of mTORis after kidney transplantation [49].One argument against our concerns was the relatively high rate of steroid-free patients at our center, including the speculation that steroid withdrawal in a CNI-free regimen could have contributed to alloimmunization.Our current analysis revealed the opposite effect: steroid withdrawal was associated with a significantly lower rate of dnDSA formation and BPARs.Steroid-free immunosuppression might reflect clinically driven individualization in low-risk patients, but, in combination with better patient and graft survival, it might also reflect an effective reduction in cardiovascular risk and steroid-associated side-effects.
Analyses of graft and patient survival did not support our hypothesis that early randomized immunosuppression significantly determines graft and patient survival.Nevertheless, our data again support the association between dnDSA formation and graft loss (log-rank: p < 0.001), and between BPARs and graft loss (log-rank: p < 0.001), independent of early immunosuppression.While both mTORi groups had a higher risk of dnDSAs and BPARs, this did not translate into more graft losses.Besides the limited statistical power, a potential reason for this might be that the mTORi treatment was frequently discontinued, and the duration of the initial randomized immunosuppression with mTORis in relation to the total observation time was short.In addition, routine dnDSA screening in our center increased awareness, and, in cases of dnDSA detection, clinicians intervened to prevent ABMR and graft loss [50][51][52].
In part, the unfavorable effects of mTORis on graft survival may also be counterbalanced by CNI nephrotoxicity [53].However, based on our long-term experience, CNI-free mTORi-based regimens are not a solution for the problem of CNI toxicity.Unfortunately, the combined use of mTORis and CNIs resulted in 30% (23/76 patients) withdrawals due to aggravated nephrotoxicity in the long term.Aggravation of nephrotoxicity with the combination of mTORis and CNIs was observed decades ago [54], and the combined use of tacrolimus and sirolimus [55][56][57] was associated with a high rate of CNI toxicity.Together with mTORi-associated side-effects, this combination had the lowest tolerability and partially explained the high dropout rates of everolimus arms in the TRANSFORM [15] and Athena [17] studies.In summary, the standard CNI regimen with mycophenolic acid clearly provided better tolerability and lower rates of dnDSAs and BPARs.
Recently, five years data from a Brazilian single-center study were published, which suggested a high degree of safety and efficacy in kidney transplant recipients receiving tacrolimus in combination with everolimus [58].The remarkably low rates of dnD-SAs after 5 years in this cohort contrasted with our own data and the results of other groups [59].An explanation for this surprising observation could be the high rates of zero HLA-DR mismatches (64-76%; the highest rate in the favored rabbit antithymocyte globulin/everolimus/low-tacrolimus/prednisone group), which were, in part, responsible for the good efficacy and low rates of dnDSAs.In our cohort, the rate of zero HLA-DR mismatches was approximately 31%, which was less than half of that in the Brazilian study.There is an increasing body of evidence for the importance of HLA-DR and epitope matching in the development of dnDSAs and rejection [60][61][62].
Not surprisingly, the high recipient age in our cohort was associated with patient survival.Other factors associated with impaired overall patient survival in the multivariate model were non-GN renal disease, male donor sex, and deceased donor type.Importantly, steroid-free immunosuppression reduced the risk of patient death.The detection of dnDSAs per se did not affect patient survival (log-rank: n.s.); however, BPARs significantly decreased overall patient survival (log-rank: p = 0.032), regardless of the initial immunosuppressive regimen.The lack of a significant influence of early immunosuppression on overall patient survival contrasts with previous studies from the US registry [11], Australia [12], and a meta-analysis [10], which consistently demonstrated increased mortality in mTORitreated kidney allograft recipients.In a prospective observational study of 993 prevalent kidney transplant recipients without a history of malignancy, long-term exposure to mTOR inhibitors was associated with significantly increased mortality [9].Again, patients in our cohort received follow-up in a center with longstanding mTORi experience, with close follow-up and timely interventions in case of efficacy failure and side-effects.This and the impact of competing risks over the long observation period were likely reasons for the lack of a significant influence of early randomized immunosuppression on patient survival.
Subgroups of patients who could potentially benefit from mTORi-based immunosuppression after kidney transplantation have yet to be defined.Univariate analysis of patients from ITT group 2 (mTORi) and ITT group 3 (mTORi + CNI) with above-average duration of randomized immunosuppression in the respective groups revealed that younger, female candidates of a first kidney graft after a longer waiting time might be such a subgroup if they receive an organ from a relatively young donor with a good HLA match (≤three mismatches).Unfortunately, these patients may want to become pregnant and are prone to ovarian cysts; therefore, they are not ideal candidates for mTORi treatment.The incidence of ovarian cysts in our center has been found to be 20.5% and 4.9% in mTORi treatment and control groups, respectively [47].
The present study has some limitations.The patients were recruited from heterogeneous clinical trials at a single center over a 13-year period.Practice patterns have changed with respect to the choice of CNIs, with respect to the target trough levels, and with respect to the monitoring of mycophenolic acid treatment, but the classes of immunosuppressive drugs and their combination have not changed.A significant number of patients did not receive an interleukin-2 receptor induction.Despite the accumulation of data from different immunosuppression trials, the number of mTORi patients was relatively small, especially at the end of the observation period.
Complete long-term follow-up with rather uniform treatment algorithms, the classification of biopsy results according to the most recent Banff classification, and the comparison of randomized and clinically driven mTOR inhibitor use with a CNI-based treatment are the advantages of this study.
From our long-term data, we conclude that the general tolerability and immunosuppressive potential of mTORis in kidney graft recipients is inferior to that of immunosup-pression based on the combination of CNIs and mycophenolic acid.The monitoring of mycophenolic acid by measurements of inosine 5 ′ -monophosphate dehydrogenase activity and the definition of individual target CNI trough levels, along with a strict follow-up policy, provide tools for the precise individualization of immunosuppression and good long-term outcomes.

Figure 1 .
Figure 1.Description of steps towards the definition of a study cohort for the comparison of differen immunosuppressant regimens early after kidney transplantation.

Figure 1 .
Figure 1.Description of steps towards the definition of a study cohort for the comparison of different immunosuppressant regimens early after kidney transplantation.

Figure 2 .
Figure 2. Deviation from randomized immunosuppression after transplantation according to the ITT group (blue-CNI, red-mTORi, purple-CNI + mTORi).Patients of both mTORi groups were more often switched to other immunosuppressive regimens than CNI patients (log-rank: p < 0.001 each).Patients on the combination of mTORis plus CNIs were more likely to be changed than patients on mTORis plus MPA (log-rank: p = 0.001).

Figure 2 .
Figure 2. Deviation from randomized immunosuppression after transplantation according to the ITT group (blue-CNI, red-mTORi, purple-CNI + mTORi).Patients of both mTORi groups were more often switched to other immunosuppressive regimens than CNI patients (log-rank: p < 0.001 each).Patients on the combination of mTORis plus CNIs were more likely to be changed than patients on mTORis plus MPA (log-rank: p = 0.001).

FOR PEER REVIEW 16 of 22 Figure 8 .
Figure 8. De novo DSA formation after kidney transplantation in patients from ITT group 1 with maintenance CNIs or conversion to a CNI-free, mTORi-based immunosuppression for clinical reasons (blue-ever CNI, red-mTORi-conversion).Maintenance of a CNI-based immunosuppression in comparison with a switch to mTORi (CNI-free) combinations protected from dnDSA formation (log-rank: p = 0.002).

Figure 8 .
Figure 8. De novo DSA formation after kidney transplantation in patients from ITT group 1 with maintenance CNIs or conversion to a CNI-free, mTORi-based immunosuppression for clinical reasons (blue-ever CNI, red-mTORi-conversion).Maintenance of a CNI-based immunosuppression in comparison with a switch to mTORi (CNI-free) combinations protected from dnDSA formation (log-rank: p = 0.002).

Table 1 .
Clinical immunosuppression trials used to study the long-term effects of early immunosuppression.

Table 2 .
Patient baseline characteristics (descriptive data are summarized as median and range).

Table 3 .
Cox proportional hazard modeling of the risks for patient death (a), death-censored graft loss (b), overall graft loss (c), acute rejections (d), and de novo DSA formation (e).Covariates with p > 0.20 in the univariate analyses are not reported in the table.

Table 3 .
Cox proportional hazard modeling of the risks for patient death (a), death-censored graft loss (b), overall graft loss (c), acute rejections (d), and de novo DSA formation (e).Covariates with p > 0.20 in the univariate analyses are not reported in the table.