Donor-Specific HLA Antibodies in a Cohort Comparing Everolimus With Cyclosporine After Kidney Transplantation

American Journal of Transplantation 2012; 12: 1192 1198 Wiley Periodicals Inc. C Copyright 2012 The American Society of Transplantation and the American Society of Transplant Surgeons doi: 10.1111/j.1600-6143.2011.03961.x Donor-Specific HLA Antibodies in a Cohort Comparing Everolimus With Cyclosporine After Kidney Transplantation L. Liefeldt a,, *, S. Brakemeier a,, P. Glander a, J. Waiser a, N. Lachmann b,c.schönemann b, B. Zukunft a, P. Illigens a,d.schmidt a,k.wu a,c, B. Rudolph c, H.-H. Neumayer a and K. Budde a a Department of Nephrology, Charité Campus Mitte, Charité - Universitätsmedizin Berlin, Berlin, Germany b Center for Tumor Medicine, HLA-Laboratory, Charité - Universitätsmedizin Berlin, Berlin, Germany c Department of Pathology, Charité -Universitätsmedizin Berlin, Berlin, Germany *Corresponding author: Lutz Liefeldt, lutz.liefeldt@charite.de S.B. and L.L. contributed equally to the study. Donor-specific HLA antibodies (DSA) have a negative impact on kidney graft survival. Therefore, we analyzed the occurrence of DSA and antibody-mediated rejection (AMR) in patients from two prospective randomized trials in our center. At 3 4.5 months posttransplant 127 patients were randomized to continue cyclosporine or converted to everolimus therapy. The presence of DSA was prospectively assessed using Luminex assays. AMR was defined according to the Banff 2009 classification. Antibody screening was available in 126 patients with a median follow-up of 1059 days. Seven out of 65 (10.8%) patients on cyclosporine developed DSA after a median of 991 days. In comparison, 14/61 patients (23.0%) randomized to everolimus developed DSA after 551 days (log-rank: p = 0.048). Eight patients on everolimus compared to two patients on cyclosporine developed AMR (log-rank: p = 0.036). Four of 10 patients with AMR all in the everolimus group lost their graft. A multivariate regression model revealed everolimus, >3 mismatches and living donor as significant risk factors for DSA. Acute rejection within the first year, >3 mismatches, everolimus and living donor were independent risk factors for AMR. This single center analysis demonstrates for the first time that everolimus-based immunosuppression is associated with an increased risk for the development of DSA and AMR. Key words: Antibody-mediated rejection, cyclosporine, donor-specific antibodies, everolimus, kidney transplantation, immunosuppression Abbreviations: AMR, antibody-mediated rejection; AR, acute rejection; CD, deceased donor; CI, confidence interval; CNI, calcineurin-inhibitor; CyA, cyclosporine A; DSA, donor-specific HLA antibodies; EC-MPS, enteric-coated mycophenolate sodium; EVR, everolimus; FSGS, focal segmental glomerulosclerosis; GFR, glomerular filtration rate; GN, glomerulonephritis; HLA, human leucocyte antigen(s); HR, hazard ratio; IS, immunosuppression; ITT, intention to treat; KTX.kidney transplantation; LD, living donor; MDRD, modification in diet in renal disease; MPS, mycophenolate sodium; MP, methylprednisolone; mtori, mammalian target of rapamycin inhibitor; PRA, panel reactive antibodies; SD, standard deviation; Tac, tacrolimus; TCMR, T-cell-mediated rejection. Received 04 August 2011, revised 02 December 2011 and accepted for publication 16 December 2011 Introduction HLA antibodies in general and donor-specific HLA antibodies (DSA) in particular are significant risk factors for graft survival after renal transplantation (1 4). HLA antibodies may cause antibody-mediated rejection (AMR) which is associated with poor prognosis (5 9). Prominent histological features of acute AMR are glomerulitis, peritubular capillaritis, dilatation of peritubular capillaries and interstitial edema (5,8). In addition, subclinical AMR and persistence of DSA may progress to chronic AMR, now known as a major cause of graft dysfunction and late graft loss (8,10 13). Chronic AMR is characterized by basement membrane abnormalities of glomerular and peritubular capillaries leading to chronic transplant glomerulopathy and nephron loss (11,12,14). Clinical signs of chronic AMR include proteinuria as hallmark of glomerular injury and a slow deterioration of allograft function (15,16). The development of nonnephrotoxic immunosuppressive regimens is an important goal in transplantation and recently, several studies describing successful conversion to a mammalian target of rapamycin inhibitor (mtori)-based regimen have been published (17 20). Most of these studies reported a significant improvement of renal function after conversion to a CNI-free regimen, however, long-term data are limited (21). 1192

DSA Risk and Immunosuppression after KTX So far, the risk factors for the development of DSA are not completely defined, and only a few studies tried to identify risk factors for chronic AMR. De novo DSA formation was associated with early acute rejections, HLA-DR matching, nonadherance and pretransplant immunization in a number of retrospective observational studies (2,7,8,10,11,16). All these findings suggest a relationship between the intensity of immunosuppression and sensitization. In this context the role of different immunosuppressive regimens is still unclear, and led us to investigate a potential influence of two different immunosuppressive regimens on de novo DSA formation. Therefore, we compared the effect of conventional cyclosporine-based therapy with a calcineurininhibitor (CNI)-free, everolimus-based regimen on the formation of DSA in a single center analysis using patients from two prospective randomized controlled trials. Methods Patients Between June 2005 and March 2010 200 kidney transplant patients of our center participated in two randomized trials, comparing an early conversion to an everolimus-based regimen on renal function with a cyclosporine-based regimen. Overall 58 patients were enrolled in the ZEUStrial (ClinicalTrials.gov: NCT00154310) and 142 patients were enrolled in the CRAD001ADE13-trial (ClinicalTrials.gov: NCT00514514). Both trials had an identical initial immunosuppressive regimen, consisting of Basiliximab induction (20 mg pretransplant and on day 4), 1.44 g/day enteric-coated mycophenolate sodium (EC-MPS), methylprednisolone (500 mg preoperatively tapered to 4 mg on month 3) and cyclosporine (trough levels 150 220 ng/ml in first 3 months, tapered to 100 150 ng/ml starting on month 6). Both trials included low-to-moderate risk patients and had identical inclusion and exclusion criteria at study entry and at randomization (17). Adult recipients (18 70 years) of a first or second kidney transplant were eligible for enrollment. The key exclusion criteria were multiple organ transplantation, loss of a previous graft due to immunological reasons in the first year after transplantation, receipt of an organ donated after cardiac death, donor age <5 or >65 years, current or peak panel reactive antibodies (PRA) >25%, platelets <75,000/mm 3, an absolute neutrophil count of <1.500/mm 3 or leucocytes <2.500/mm 3, hemoglobin <6 g/dl, or severe liver disease. At randomization, patients were required to receive treatment with cyclosporine, EC-MPS ( 720 mg/day) and corticosteroids, with serum creatinine 265 lmol/l. Patients were excluded from randomization when they had experienced graft loss, previous changes to the immunosuppressive regimen due to immunologic reasons, severe rejection ( Banff grade II), recurrent acute rejection, steroid resistant acute rejection or if, at the time of randomization, they were dialysis-dependent, proteinuria >1 g/day, clinically significant infection or contravened study entry criteria for hematology or liver disease. Randomization was performed using a central automated, validated fax-system, with patients stratified according to living or deceased donor. The only difference in the study design between both studies was the time of randomization, which took place at month 4.5 for the ZEUS trial (17) and at month 3 for the CRAD001ADE13 trial. At these time points n = 50 patients in the ZEUS trial and n = 77 patients in the CRAD001ADE13 trial were randomized to either (1) continue cyclosporine (trough level after month 6: 100 150 ng/ml), EC-MPS ( 720 mg twice daily) and steroids or (2) stepwise conversion from cyclosporine to everolimus (target trough concentration 6 10 ng/ml), EC-MPS ( 720 mg twice daily) and steroids. Altogether 41 patients did not undergo randomization and 32 patients were randomized in the CRAD001ADE13 trial to a third regimen, consisting of low dose cyclosporine in combination with everolimus and steroids. Due to the low number and the shorter observation time the latter group was not included in the analysis. According to the protocol, all patients received 4 mg methylprednisolone for at least 12 months. According to the center s practice steroid withdrawal was allowed after the first year. In the case of side effects, EC-MPS was reduced, and the immunosuppressive protocol was adjusted to the individual needs. HLA-antibody screening Serum samples were prospectively collected at least once a year and at the time of indication graft biopsies for HLA-antibody screening starting in 2006. All serum samples were qualitatively screened for HLA antibodies by the Luminex-based bead assay LABScreen Mixed (One Lambda, Canoga Park, CA, USA). Donor specificity of HLA antibodies was determined by LABScreen Single Antigen beads (One Lambda). All tests were performed according to the manufacturer s guidelines (4). Graft biopsies Graft biopsies were performed in the case of impaired kidney function and/or development of proteinuria. All histological findings were recategorized in 2010 according to the Banff 09 classification (12) by two experienced pathologists (BR, KW). The pathologists were not aware of the treatment allocation. Follow-up, compliance All patients received a regular follow-up in our center (during the core phase and the prespecified 5-year observational follow-up) and all clinical data were recovered from our web-based electronic patient record system, TBase (22,23). In addition three nephrologists involved in the follow-up of the patients (SB, BZ, PI) independently assessed the treatment adherence using three categories: 1- excellent, 3 moderate, 5- bad. Statistics Endpoints were the incidence of de novo DSA, time to first detection of DSA, first biopsy-proven AMR and graft survival. In addition serum creatinine and proteinuria were assessed. All analyses were by intention to treat (ITT). Comparisons between treatment groups were performed with the chi-squared test for categorical data, Wilcoxon Mann Whitney test for continuous data, and cumulative incidence plots with a log-rank test for time-to-event data. A p-value <0.05 was considered to be statistically significant. Time to first detection of DSA and AMR was estimated with censoring at date of death, graft failure or end of observation on March 31, 2011. In univariate regression analyses several factors were tested on their impact on the developmentof de novo DSA and AMR: underlying renal disease (GN vs. other), recipient age (>55 vs. 55 years), gender, waiting time (>24 vs. 24 months), immunosuppressive regimen (everolimus vs. cyclosporine), reduced MPS-dose (<1.44 g/day), number of mismatches (4 6 vs. 0 3), donor type (living vs. deceased), treated biopsy-proven acute rejection(s) within the first year, number of KTX (second vs. first), patient s compliance (graded into 3 categories). American Journal of Transplantation 2012; 12: 1192 1198 1193

Liefeldt et al. The proportional hazards assumption for categorical covariates was graphically tested by plotting log[-log(s(t))] versus time for strata of each covariate and valid if the curves were approximately parallel and did not cross. All variables with a p<0.20 in the univariate analysis were included simultaneously into a Cox proportional hazard model and backward elimination strategy (p>0.05) was applied to identify significant risk factors for development of DSA and occurrence of AMR. All statistical analyses were performed using SPSS for Windows release 16.0.0 (SPSS Inc., Chicago, IL, USA). Results Patient population From 200 patients enrolled 159 patients were eligible for randomization, and a total of 127 patients were randomized to continue cyclosporine (n = 66) or to conversion to everolimus (n = 61). Demographic characteristics and details of the immunosuppressive treatment are summarized in Table 1. Results of antibody screening were available in 126 patients (99.2%). On average 4.3 ± 2.8 samples per patient were analyzed during a follow-up period of 33.8 ± 20.6 months. Samples were analyzed with a similar frequency in both groups with 2.6 and 3.0 samples per year of follow-up in the cyclosporine and everolimus arm, respectively (p = 0.4). Table 1: Characteristics of the study population Cyclosporine Everolimus group group Variable (n = 66) (n = 61) Recipient age (years) 49.5 ± 11.7 45.5 ± 13.3 mean±sd Male gender number (%) 43 (65%) 30 (49%) Donor age (years) mean±sd 49.9 ± 13.2 47.3 ± 12.7 Waiting time (months) median 58.3 (0 122) 41.0 (0 387) (range) Mismatches median (range) 2.5 (0 6) 3 (0 5) Renal disease (GN vs. other) 32 vs. 34 30 vs. 31 Living donor number (%) 20 (30.3%) 20 (32.8%) Number of KTX (first vs. 60 vs. 6 55 vs. 6 second) Patients on randomized 58/65 (89%) 54/60 (90%) treatment at 12 months number (%) Patients on randomized treatment at end of observation number (%) 42/66 (64%) 40/61 (66%) Mean trough level (ng/ml) CyA: 122 ± 12 EVR: 7.1 ± 1.2 Mean MPS-dose (mg/day) 1232 ± 278 1210 ± 259 Steroid-free immunosuppression at end of observation number (%) 41/66 (62%) 36/61 (59%) GN = glomerulonephritis; SD = standard deviation; CyA = cyclosporine A; EVR = everolimus; MPS = mycophenolate sodium. Figure 1: Cumulative incidence plot of DSA-detection in 61 patients (red) with everolimus-based immunosuppression compared to 66 patients (black) with cyclosporine-treatment (log-rank: p = 0.048). Formation of DSA Seven out of 65 (10.8%) cyclosporine-treated patients developed de novo DSA during a median follow-up of 1273 (542 2107) days. DSA were detected on median 991 (12 1188) days after transplantation. A higher incidence of DSA was observed in patients randomized to everolimus (14/61, 23.0%, Figure 1). Despite a similar follow-up of 1173 (542 2103) days the time to first occurrence of DSA was shorter compared to the cyclosporine group (551, range 10 1176 days). All CyA-patients with de novo DSA formation developed DSA against HLA class II antigens. One patient additionally developed a class I antibody (Cw7). Class II DSA were directed against DR in 4 patients, DQ in 2 patients, and both DR and DQ in 1 patient. All DSA were detected sustained with mean fluorescence intensities (MFI) ranging from 805 MFI to 1593 MFI and 1480 MFI to 11 175 MFI for class I- and class II-DSA, respectively. De novo DSA of everolimus-treated patients were directed exclusively against class I antigens in 3 (2 against HLA-A/B, 1 against -B), class II antigens in 8 (5 against HLA-DQ, 2 against -DR and 1 against -DP), and both class I- and IIantigens in 3 patients (one against HLA-A/DR, -A/DQ and -A/DR/DQ each). DSA of everolimus-treated patients were detected sustained with MFI values of 492 to 15 116 and 727 to 6559 for class I- and class II-DSA, respectively. Biopsy-proven AMR, graft losses and outcome Ten out of 21 de novo DSA-positive patients developed biopsy-proven AMR (Table 2). Eight patients randomized 1194 American Journal of Transplantation 2012; 12: 1192 1198

DSA Risk and Immunosuppression after KTX Table 2: Cases of biopsy-proven AMR in patients randomized to receive cyclosporine- or everolimus-based immunosuppression Ageat ISatfirst Timeto Graft Patient KTX Renal Type Number Mismatches detection DSA DSA Time to function # Gender (years) disease of KTX of KTX A/B/DR IS (ITT) of DSA (days) specificities IS at AMR AMR (days) (GFR) TCMR 1 M 51 Scleroderma CD 1 0/2/2 CyA Tac+MPS+MP 307 DQ2, DR17 Tac+MPS+MP 306 18 ml/min Prior to AMR 2 F 34 SLE LD 1 0/1/1 CyA CyA+MPS+MP 12 B7, DR14 CyA+MPS+MP 14 22 ml/min Prior to AMR 3 M 35 Polycystic CD 1 0/0/0 EVR EVR+MPS+MP 380 DQ2 EVR+MPS+MP 406 lost Prior to AMR 4 M 41 GN CD 1 1/1/1 EVR EVR+red.MPS 752 DP3 EVR+red.MPS 1150 lost None +lowmp +lowmp 5 M 24 GN LD 1 1/1/1 EVR EVR+red.MPS 257 A24, DQ7 EVR+red.MPS 820 lost Simultaneous with +MP AMR 6 M 21 Dysplasia CD 2 1/2/1 EVR CyA+MPS+MP 10 B62 EVR+red.MPS+MP 135 lost Prior to AMR 7 F 46 GN LD 1 2/2/1 EVR EVR+MPS+MP 805 A1, DR7 EVR+red.MPS+MP 757 12 ml/min None 8 M 35 Alport LD 1 2/2/1 EVR EVR+MPS+MP 173 DQ8 EVR+MPS 623 59 ml/min Simultaneous with AMR 9 M 33 GN LD 1 1/0/1 EVR CyA+MPS+MP 1027 DR1 EVR+red.MPS 898 35 ml/min None 10 M 44 Polycystic LD 1 2/2/1 EVR EVR+MPS 1176 A33 EVR+MPS 1218 34 ml/min None KTX = kidney transplantation; IS = immunosuppression; ITT = intention to treat; DSA = donor-specific HLA antibodies; AMR = antibody-mediated rejection; LD = living donor; CD = deceased donor; CyA = cyclosporine A; EVR = everolimus; Tac = tacrolimus; MPS = mycophenolate sodium; red.mps = daily MPS-dose less than 1440 mg; MP = methylprednisolone; lowmp = daily MP-dose less than 4 mg; GFR = glomerular filtration rate (MDRD); TCMR = T-cell-mediated rejection. Figure 2: Cumulative incidence plot of first antibodymediated rejection in 61 patients (red) with everolimus-based immunosuppression compared to 66 patients (black) with cyclosporine-treatment (log-rank: p = 0.036). to the everolimus-based immunosuppression, and two with continued use of cyclosporine (log-rank test: p = 0,036; Figure 2). Four of the eight everolimus patients with AMR lost their graft despite intensive efforts to maintain kidney function. The kidney function of the other patients with AMR is shown in Table 2. Three additional graft losses were observed in patients without AMR: two patients in the cyclosporine group (1 graft loss due to multiple infectious complications, and 1 graft loss due to extensive interstitial fibrosis). One patient randomized to everolimus lost the graft due to recurrence of FSGS. Serum creatinine, proteinuria and graft- and patient-survival of the cohort are summarized in Table 3. Risk factors for DSA formation In order to further define risk factors for the formation of DSA we performed univariate regression analyses. Waiting time, immunosuppressive regimen (by ITT), >3 mismatches, living donors and treated acute rejections during the first year were identified as potential confounders (p < 0.20) in univariate regression analyses. Using these potential risk factors in a multivariate backward elimination (p>0.05) Cox regression model, only everolimus-based immunosuppressive regimen, >3 mismatches and graft from a living donor remained significant risk factors associated with de novo DSA formation (Table 4). Risk factors for AMR Glomerulonephritis as renal disease, recipient age, gender, everolimus-based immunosuppressive regimen (by ITT), >3 mismatches, living donors and treated acute rejections American Journal of Transplantation 2012; 12: 1192 1198 1195

Liefeldt et al. Table 3: Outcome after KTX in patients randomized to receive cyclosporine- or everolimus-based immunosuppression Parameter Cyclosporine (n = 66) Everolimus (n = 61) Patient survival number (%) Month 12 65 (98.5%) 61 (100.0%) Month 24 64 (97.0%) 60 (98.4%) Last observation 1 63 (95.4%) 60 (98.4%) Death censored graft Month 12 65/65 (100.0%) 60/61 (98.4%) survival number (%) Month 24 63/64 (98.4%) 57/60 (95.0%) Last observation 2 61/63 (96.8%) 55/60 (91.7%) Creatinine (mg/dl) Month 12 1.46 ± 0.61 (n = 65) 1.21 ± 0.57 (n = 60) median ± interquartile range Month 24 1.69 ± 0.57 (n = 49) 1.31 ± 0.84 (n = 44) >24 Months 1.73 ± 0.70 (n = 40) 1.45 ± 1.06 (n = 36) # Proteinuria (mg/day) Month 12 179 ± 172 (n = 60) 258 ± 183 (n = 46) median±interquartile range Proteinuria >1 g/day:n = 0 Proteinuria >1 g/day:n = 2 Month 24 150 ± 97 (n = 39) 380 ± 608 (n = 34) Proteinuria >1 g/day:n = 0 Proteinuria >1 g/day:n = 6 >24 Months 132 ± 185 (n = 35) 278 ± 608 (n = 33) Proteinuria >1 g/day:n = 1 Proteinuria >1 g/day:n = 7 p < 0.05, p < 0.01, p < 0.001, #p = 0.075 (Wilcoxon Mann Whitney test). 1 median follow-up: 1094 (372 2107) days, 2 median follow-up: 1108 (372 2107) days within the first year were identified as potential risk factors for AMR in univariate regression analyses (p < 0.20). The multivariate regression model identified everolimusbased immunosuppression, living donor, >3 mismatches and treated acute rejections during the first year as risk factors for AMR (Table 4). Discussion The development of DSA posttransplant is an important risk factor for graft loss. Data on DSA formation under mtori-based immunosuppression are limited to a small pilot study in combination with belatacept (24). Thus, the present analysis from two prospective randomized trials represents the first systematic investigation on this important topic. In our analysis we clearly observed an increased risk for DSA formation and AMR under an everolimusbased, CNI-free immunosuppressive protocol. Our control group with cyclosporine and EC-MPS had a similar frequency of DSA (10.8%) compared with the recent report by Larsen et al. (25), who described the development of DSA in 7 12% of cyclosporine-treated patients. An important confounder for the formation of DSA is histo(in)compatibility. It is obvious that more mismatches increase the probability for the development of DSA. As a consequence the use of more potent immunosuppression in patients with suboptimal HLA match may be advisable. Our results indicate that conversion to a CNI-free, everolimus-based regimen needs careful selection of immunological low risk patients, preferentially with a wellmatched organ. Given the fact that most of our patients with AMR developed HLA class II DSA, an alternative prophylactic approach would be to increase the importance of the HLA-DR match in the allocation process. Table 4: Cox proportional hazard modeling for risk for de novo DSA formation and antibody-mediated rejection. Covariates with p > 0.20 in univariate analyses were not reported in the table (a) De novo DSA formation Univariate analyses Multivariate analyses covariates HR 95% CI p HR 95% CI p Time on dialysis (>24 vs. 24 months) 0.43 0.18 1.02 0.047 Donor type (living vs. deceased) 2.39 1.01 5.65 0.040 2.39 1.01 5.66 0.048 Number of mismatches (4 6 vs. 0 3) 3.13 1.32 7.44 0.006 3.26 1.37 7.75 0.008 Regimen (everolimus vs. cyclosporine) 2.43 0.98 6.04 0.048 2.67 1.07 6.66 0.035 Treated AR during the first year (yes vs. no) 3.11 1.03 9.39 0.034 (b) Antibody-mediated rejection Renal disease (GN vs. other) 2.42 0.63 9.35 0.187 Recipient age (>55 vs. 55 years) 0.03 0.00 6.20 0.024 Gender (female vs. male) 0.34 0.07 1.61 0.153 Donor type (living vs. deceased) 3.80 1.07 13.53 0.027 5.78 1.44 23.16 0.013 Number of mismatches (4 6 vs. 0 3) 3.84 1.11 13.30 0.022 5.10 1.39 18.72 0.014 Regimen (everolimus vs. cyclosporine) 4.53 0.96 21.38 0.036 5.35 1.11 25.70 0.036 Treated AR during the first year (yes vs. no) 8.46 2.33 30.75 0.000 10.22 2.56 40.87 0.001 The second category in parentheses is the reference category. DSA = donor-specific HLA antibody; HR = hazard ratio; CI = confidence interval; GN = glomerulonephritis; AR = acute rejection. 1196 American Journal of Transplantation 2012; 12: 1192 1198

DSA Risk and Immunosuppression after KTX Compared with CNI, the use of mtor inhibitors seems to be less effective with respect to the overall prevention of rejections. In the ZEUS study a 6% higher rate of acute rejections was observed in the everolimus group. Interestingly, most rejections occurred early after conversion, were treatable and patients with rejections had similar renal function at 1 year in this study (17). In the present multivariate analysis, we found that previous T-cell-mediated acute rejections were associated with the occurrence of AMR during long-term follow-up. The higher rate of rejections and AMR may reflect underimmunosuppression with the everolimus-based regimen and suggests that the combination of two antiproliferative immunosuppressants (everolimus and MPS) is not sufficient for an adequate rejection prophylaxis and seems not to sufficiently downregulate T- and B-cell-mediated immunity. The course in individual cases led us to speculate that daily MPS doses below the standard-dose might further increase the risk of DSA formation as well as the AMR risk. However, the average MPS doses were similar between groups and the univariate analysis did not support this hypothesis. Unfortunately mycophenolate exposure was not determined in this population because it was not requested by the study protocol. Further studies are needed to exclude potential underexposure with either everolimus or mycophenolate, and to investigate a potential class effect of mtori. Furthermore, our analysis was not suitable for the detection of deleterious effects of steroid withdrawal, because only patients with an uneventful first year (i.e. without rejections) were eligible for steroid withdrawal. Proteinuria is a known side effect of mtori (26). On the other hand, proteinuria has been shown to be an early marker of AMR (15). In the light of the results of our study, development of proteinuria under immunosuppression with mtori should prompt further diagnostics rather than routinely attributing this side effect to mtori medication. The formation of DSA suggests inadequate immunosuppression and the presence of DSA may indicate a higher risk for complications later in the course even in patients with actually stable graft function. Therefore, conversion to everolimus as well as reduction of other immunosuppressive agents should be considered very carefully in these patients. The fluctuating titers of DSA and the latency between the formation of DSA and the histological confirmation of AMR underline that AMR is a long-term problem in kidney transplantation and necessarily requires long-term follow-up (14). As mtori impacts B-cell proliferation and potentially antibody formation, we did not expect an increased risk of de novo DSA formation. Contrary, CNI inhibitors, which are thought to act primarily through inhibition of T-cell activation prevented DSA formation more effectively, suggesting an interplay between T and B cells for complete humoral immune activation. Whether the increased risk of de novo DSA development persists over time and is equally high later in the course after transplantation should be investigated in further studies. Limitations of our study are the single center nature, the limited number of patients and the relatively small number of events. That is why confidence intervals are wide and we can give only an approximation of the true risk. At current our observation is limited to everolimus at trough levels of 5 8 ng/ml in combination with MPS. It is obvious, that a higher number of patients, a more frequent screening and probably protocol biopsies would have helped for the early detection of subclinical AMR. However, up-to-date, no treatment guidelines for such cases do exist. Advantages of our study are the homogeneous treatment protocols, the prospective routine HLA-antibody screening and the long and complete follow-up. In conclusion, our study emphasizes the need and the utility of HLA antibody monitoring in immunosuppressive trials, especially in CNI-minimization or elimination regimens. In our cohort of a well-characterized and randomized lowto-moderate risk population the conversion to the CNI-free, everolimus-based regimen was associated with the occurrence of DSA and AMR. Our data clearly need confirmation in larger prospective trials, but given the poor prognosis associated with AMR we think it is important to highlight this serious long-term problem for more rigorous prospective investigations in the future. Until those data are available, the conversion from a CNI-based therapy to an everolimusbased therapy needs to be carefully discussed, especially in patients with an increased immunological risk or preexisting DSA. The potential benefit of better renal function in the majority of patients has to be balanced with safety concerns associated with the development of DSA in a significant proportion of everolimus-treated patients. Acknowledgment The authors gratefully acknowledge the statistical advice provided by Andreas Walter, Statistician, at PAREXEL Int. GmbH. Disclosure The following authors have no conflicts of interest to disclose as described by the American Journal of Transplantation: S.B., C.S., B.Z., P.I., D.S., K.W. and B.R. The following authors of this manuscript have conflicts of interest to disclose as described by the American Journal of Transplantation: L.L., P.G. and J.W. received honoraria from Novartis. N.L. is advisor for One Lamda s German distributor BmT. H.H.N. received research funds and honoraria from Novartis. K.B. received research funds and/or honoraria from Pfizer, Novartis, Astellas, Roche, Hexal, Bristol- Myers Squibb, LCP Pharma, TCL Pharma and Siemens. American Journal of Transplantation 2012; 12: 1192 1198 1197

Liefeldt et al. References 1. Dunn TB, Noreen H, Gillingham K, et al. Revisiting traditional risk factors for rejection and graft loss after kidney transplantation. Am J Transplant 2011; 11: 2132 2143. 2. Hourmant M, Cesbron-Gautier A, Terasaki PI, et al. Frequency and clinical implications of development of donor-specific and nondonor-specific HLA antibodies after kidney transplantation. J Am Soc Nephrol 2005; 16: 2804 2812. 3. Mao Q, Terasaki PI, Cai J, et al. Extremely high association between appearance of HLA antibodies and failure of kidney grafts in a five year longitudinal study. Am J Transplant 2007; 7: 864 887. 4. Lachmann N, Terasaki PI, Budde K, et al. Anti-human leukocyte antigen and donor-specific antibodies detected by luminex posttransplant serve as biomarkers for chronic rejection of renal allografts. Transplantation 2009; 87: 1505 1513. 5. Lefaucheur C, Nochy D, Hill GS, et al. Determinants of poor graft outcome in patients with antibody-mediated acute rejection. Am J Transplant 2007; 7: 832 841. 6. Gloor J, Cosio F, Lager DJ, et al. The spectrum of antibody mediated renal allograft injury: Implications for treatment. Am J Transplant 2008; 8: 1367 1373. 7. Everly MJ, Everly JJ, Arend LJ, et al. Reducing de novo donorspecific antibody levels during acute rejection diminishes renal allograft loss. Am J Transplant 2009; 9: 1063 1071. 8. Loupy A, Suberbielle-Boissel C, Hill GS, et al. Outcome of subclinical antibody-mediated rejection in kidney transplant recipients with preformed donor-specific antibodies. Am J Transplant 2009; 9: 2561 2570. 9. Lefaucheur C, Loupy A, Hill GS, et al. Preexisting donor-specific HLA antibodies predict outcome in kidney transplantation. J Am Soc Nephrol 2010; 21: 1398 1406. 10. Gloor JM, Sethi S, Stegall MD, et al. Transplant glomerulopathy: Subclinical incidence and association with alloantibody. Am J Transplant 2007; 7: 2124 2132. 11. Cosio FG, Gloor JM, Sethi S, et al. Transplant glomerulopathy. Am J Transplant 2008; 8: 492 496. 12. Sis B, Mengel M, Haas M, et al. Banff 09 meeting report: Antibody mediated graft deterioration and implementation of Banff working groups. Am J Transplant 2010; 10: 464 471. 13. Hill GS, Nochy D, Bruneval P, et al. Donor-specific antibodies accelerate arteriosclerosis after kidney transplantation. J Am Soc Nephrol 2011; 22: 975 983. 14. Archdeacon P, Chan M, Neuland C, et al. Summary of FDA antibody-mediated rejection workshop. Am J Transplant 2011; 11: 896 906. 15. Fotheringham J, Angel C, Goodwin J, et al. Natural history of proteinuria in renal transplant recipients developing de novo human leukocyte antigen antibodies. Transplantation 2011; 91: 991 996. 16. Sánchez-Fructuoso AI, Santiago JL, Pérez-Flores I, et al. De novo anti-hla antibodies in renal allograft recipients: A cross-section study. Transplant Proc 2010; 42: 2874 2876. 17. Budde K, Becker T, Arns W et al., ZEUS Study Investigators. Everolimus-based, calcineurin-inhibitor-free regimen in recipients of de novo kidney transplants: An open-label, randomised, controlled trial. Lancet 2011; 377: 837 847. 18. Lebranchu Y, Thierry A, Toupance O, et al. Efficacy on renal function of early conversion from cyclosporine to sirolimus 3 months after renal transplantation: Concept study. Am J Transplant 2009; 9: 1115 1123. 19. Guba M, Pratschke J, Hugo C et al. SMART-Study Group. Renal function, efficacy, and safety of sirolimus and mycophenolate mofetil after short-term calcineurin inhibitor-based quadruple therapy in de novo renal transplant patients: One-year analysis of a randomized multicenter trial. Transplantation 2010; 90: 175 183. 20. Weir MR, Mulgaonkar S, Chan L, et al. Mycophenolate mofetilbased immunosuppression with sirolimus in renal transplantation: A randomized, controlled Spare-the-Nephron trial. Kidney Int 2011; 79: 897 907. 21. Lebranchu Y. Can we eliminate both calcineurin inhibitors and steroids? Transplant Proc. 2010; 42(Suppl 9): S25 S28. 22. Schröter K, Lindemann G, Fritsche L. TBase2 a web-based electronic patient record. Fundamenta Informaticae 2000; 43: 343 353. 23. Glander P, Budde K, Schmidt D, et al. The blood group O problem in kidney transplantation time to change? Nephrol Dial Transplant 2010; 25: 1998 2004. 24. Ferguson R, Grinyó J, Vincenti F, et al. Immunosuppression with belatacept-based, corticosteroid-avoiding regimens in de novo kidney transplant recipients. Am J Transplant 2011; 11: 66 76. 25. Larsen CP, Grinyó J, Medina-Pestana J, et al. Belatacept-based regimens versus a cyclosporine A-based regimen in kidney transplant recipients: Two-year results from the BENEFIT and BENEFIT- EXT studies. Transplantation 2010; 90: 1528 1535. 26. Letavernier E, Pe raldi MN, Pariente A, et al. Proteinuria following a switch from calcineurin inhibitors to sirolimus. Transplantation 2005; 80: 1198 1203. 1198 American Journal of Transplantation 2012; 12: 1192 1198