Authors that have contributed equally.
Efficacy of Induction Therapy with ATG and Intravenous Immunoglobulins in Patients with Low-Level Donor-Specific HLA-Antibodies
Article first published online: 26 MAR 2010
©2010 The Authors Journal compilation©2010 The American Society of Transplantation and the American Society of Transplant Surgeons
American Journal of Transplantation
Volume 10, Issue 5, pages 1254–1262, May 2010
How to Cite
Bächler, K., Amico, P., Hönger, G., Bielmann, D., Hopfer, H., Mihatsch, M. J., Steiger, J. and Schaub, S. (2010), Efficacy of Induction Therapy with ATG and Intravenous Immunoglobulins in Patients with Low-Level Donor-Specific HLA-Antibodies. American Journal of Transplantation, 10: 1254–1262. doi: 10.1111/j.1600-6143.2010.03093.x
- Issue published online: 20 APR 2010
- Article first published online: 26 MAR 2010
- Received 09 December 2009, revised 19 January 2010 and accepted for publication 07 February 2010
- Antibody-mediated rejection;
- intravenous immunoglobulins
Low-level donor-specific HLA-antibodies (HLA-DSA) (i.e. detectable by single-antigen flow beads, but negative by complement-dependent cytotoxicity crossmatch) represent a risk factor for early allograft rejection. The short-term efficacy of an induction regimen consisting of polyclonal anti-T-lymphocyte globulin (ATG) and intravenous immunoglobulins (IvIg) in patients with low-level HLA-DSA is unknown. In this study, we compared 67 patients with low-level HLA-DSA not having received ATG/IvIg induction (historic control) with 37 patients, who received ATG/IvIg induction. The two groups were equal regarding retransplants, HLA-matches, number and class of HLA-DSA. The overall incidence of clinical/subclinical antibody-mediated rejection (AMR) was lower in the ATG/IvIg than in the historic control group (38% vs. 55%; p = 0.03). This was driven by a significantly lower rate of clinical AMR (11% vs. 46%; p = 0.0002). Clinical T-cell-mediated rejection (TCR) was significantly lower in the ATG/IvIg than in the historic control group (0% vs. 50%; p < 0.0001). Within the first year, allograft loss due to AMR occurred in 7.5% in the historic control and in 0% in the ATG/IvIg group. We conclude that in patients with low-level HLA-DSA, ATG/IvIg induction significantly reduces TCR and the severity of AMR, but the high rate of subclinical AMR suggests an insufficient control of the humoral immune response.
Preformed donor-specific HLA-antibodies (HLA-DSA) represent a major barrier for successful renal transplantation. High-level HLA-DSA defined by positive complement-dependent cytotoxicity crossmatches (CDC-XM) are associated with a substantial risk for early antibody-mediated rejection (AMR) and allograft loss, even if sophisticated desensitization regimens are used (1). Therefore, such high-level HLA-DSA are considered a contraindication in most kidney transplant centres.
HLA-DSA detectable only by more sensitive assays such as the flow-cytometric crossmatch (FC-XM) or single HLA-antigen flow beads (SAFB) can be classified arbitrarily as low-level HLA-DSA. In contrast to high-level HLA-DSA, low-level HLA-DSA are widely regarded as a risk factor for transplantation, but not an absolute contraindication (2). Indeed, several studies have reported that an induction therapy consisting of polyclonal anti-T-lymphocyte globulin (ATG) and intravenous immunoglobulins (IvIg) with/without additional plasmapheresis allows transplanting such patients with reasonably good short-term outcomes, while only few intermediate-to-long-term outcome data have been published yet (3–7).
Nevertheless, the therapeutic short-term efficacy of ATG/IvIg induction in patients with low-level HLA-DSA cannot be conclusively assessed in these studies because either no control group (3,4) or only patients without HLA-DSA (5–7) were included for comparison. Clearly, a prospective randomized trial would be the best way to determine the therapeutic efficacy of ATG/IvIg induction, but this must be regarded as unethical, because low-level HLA-DSA are associated with a significant risk for early rejection and inferior allograft survival using standard immunosuppression (8,9). Therefore, a historic population consisting of patients who have been transplanted in the presence of low-level HLA-DSA without ATG/IvIg induction is the most appropriate available control group.
The aim of this study was to determine the short-term efficacy of ATG/IvIg induction in patients with low-level HLA-DSA by comparison of a prospective cohort having received ATG/IvIg (n = 37) with a historic control group not having received ATG/IvIg (n = 67).
Materials and Methods
The whole patient population includes 104 patients with low-level HLA-DSA (i.e. positive by SAFB, but negative by T-cell and B-cell CDC-XM). All retrospective analyses and therapeutic protocols were performed with approval of the local Institutional Review Board.
The historic control cohort consists of 67 patients who had retrospectively determined HLA-DSA detected by SAFB. These patients were identified in a retrospective study investigating 334 consecutive renal transplantations performed at our centre between 1999 and 2004 (8). They were considered at low risk for rejection based on negative current and remote T-cell and B-cell CDC-XM. Various immunosuppressive regimens were used during this period, but no patient received an induction therapy with ATG/IvIg (Table 1).
|Historic control (n = 67)||Anti-T-lymphocyte globulin (ATG)/intravenous immunoglobulins (IvIg) (n = 37)||p-level|
|Gender, females (%)||38 (57)||19 (51)||0.68|
|Age, median (range)||47 (15–71)||56 (26–73)||0.0007|
|Deceased donor, n (%)||35 (52)||26 (70)||0.10|
|Gender, females (%)||39 (58)||17 (47)||0.31|
|Age, median (range)||48 (1–76)||54 (7–73)||0.24|
|A/B, n with 1/2/3/4 mismatches||7/23/24/13||4/14/8/11||0.43|
|DRβ1, n with 0/1/2 mismatches||3/39/25||6/23/8||0.06|
|Number of DSA, n with 1/2/3/4/5 DSA||38/20/5/2/2||20/10/5/1/1||0.91|
|Class of DSA|
|Class I, n (%)||21 (31)||19 (51)||0.13|
|Class II, n (%)||30 (45)||12 (33)|
|Class I + II, n (%)||16 (24)||6 (16)|
|Cumulative strength of DSA [MFI]*|
|Median (range)||6494 (524–36 715)||2287 (543–26 537)||<0.0001|
|Cumulative strength of DSA, grouped|
|<2000 MFI, n (%)||10 (15)||16 (43)|
|2000–5000 MFI, n (%)||15 (22)||12 32)||0.001|
|5000–10 000 MFI, n (%)||16 (24)||4 (11)|
|>10 000 MFI, n (%)||26 (39)||5 (14)|
|Known presensitizing events**|
|Prior transplants, n (%)||30 (45)||18 (49)||0.84|
|Blood transfusions, n (%)||22 (33)||21 (57)||0.02|
|Pregnancies, n (%)||25 (37)||17 (46)||0.41|
|None, n (%)||35 (52)||n/a|
|Basiliximab, n (%)||26 (39)||n/a|
|Daclizumab, n (%)||6 (9)||n/a|
|ATG/IvIg, n (%)||n/a||37 (100)|
|CyA-MMF-P, n (%)||30 (45)||n/a|
|Tac-Aza-P, n (%)||25 (37)||n/a|
|Sir-MMF-P, n (%)||12 (18)||n/a|
|Tac-MMF-P, n (%)||n/a||37 (100)|
The ATG/IvIg cohort consists of 37 consecutive patients who had prospectively determined HLA-DSA by SAFB. They were consecutively transplanted at our centre between 2005 and 2008 and were considered at high risk for rejection based on the presence of low-level HLA-DSA (10). They received an induction therapy consisting of ATG (ATG-Fresenius, Fresenius Medical Care, Switzerland) 9 mg/kg body weight prior to reperfusion of the allograft and 3 mg/kg body weight on day 1–4 as well as IvIg 0.4 g/kg body weight prior to reperfusion of the allograft and on day 1–4 (total dose 2 g/kg body weight). Maintenance immunosuppression consisted of tacrolimus (Tac) (Prograf, Astellas, Wallisellen, Switzerland), mycophenolate-mofetil (MMF) (CellCept, Roche, Basel, Switzerland) and prednisone (P). Target Tac trough levels were 10–15 ng/mL for the first month, 8–12 ng/mL for months 2–3, 6–10 ng/mL for months 4–6 and 4–8 ng/mL thereafter. Steroids were tapered to 0.1 mg/kg body weight by month 3 posttransplant.
Diagnosis of rejection and definition of AMR
All reported rejection episodes were biopsy-proven. Clinically indicated allograft biopsies were preformed when serum creatinine deteriorated by more than 20%. As about half of the patients in the historic control group were transplanted before protocol biopsies at month 3 and 6 posttransplant became part of the clinical routine, only 45% of patients in the historic control group had protocol biopsies compared to 95% of patients in the ATG/IvIg group. Biopsy specimens (two cores obtained with a 16 gauge needle) were evaluated by light microscopy and immunofluorescence for C4d as previously reported (11,12). Findings were graded according to the Banff 2007 classification (13). C4d positivity was defined as either focal (10–50% of peritubular capillaries [PTC]) or diffuse (>50% of PTC) staining for C4d by immunofluorescence. Because reported AMR episodes occurred within the first 6 months posttransplant (median 23 days, interquartile range 7–90) and all patients had HLA-DSA in the day-of-transplant sera, we presumed that circulating HLA-DSA were also present at the time of histological AMR diagnosis. AMR was defined and graded according to histological findings by light microscopy, C4d staining in PTC and allograft function as follows:
- 1Clinical/subclinical C4d positive acute AMR: clinically indicated or protocol biopsy demonstrating C4d positivity in PTC and transplant glomerulitis and/or peritubular capillaritis and/or thrombotic microangiopathy in glomeruli and/or arteritis. This group is consistent with Banff acute AMR type II or III (13).
- 2Clinical/subclinical C4d negative acute AMR: clinically indicated or protocol biopsy demonstrating negative C4d staining in PTC, but transplant glomerulitis and peritubular capillaritis ± arteritis. By the current Banff criteria, these cases would be classified as suspicious for acute AMR.
- 3Clinical/subclinical C4d positivity only: clinically indicated or protocol biopsy demonstrating positive C4d staining in PTC, but no morphological features of AMR. This phenotype was recently integrated into the Banff classification for AMR (13).
Treatment of rejection
In the historic control group, clinical AMR and TCR >IA was treated with a 7 day course of ATG and steroid pulses (3–5 × 500 mg i.v.), clinical TCR Banff grad ≤IA with steroid pulses only (3 × 500 mg i.v.). Subclinical AMR and TCR were mostly treated with steroid pulses and rarely with ATG, depending on the severity. In the ATG/IvIg group, clinical AMR was treated with steroid pulses (3–5 × 500 mg i.v.) and IvIg (5 × 0.4 g/kg body weight); plasmapheresis and rituximab (Mabthera, Roche) were added depending on the severity and clinical response to the treatment. Subclinical AMR was treated in most cases with steroid pulses (3 × 500 mg i.v.) and IvIg ± rituximab depending on the severity. Clinical C4d positivity only, which was observed exclusively in the historic control group, was mostly treated with steroid pulses; subclinical C4d positivity only was not treated in most cases.
Detection of HLA-antibodies and assignment as HLA-DSA
All sera were tested for class I (i.e. HLA-A/B/Cw) and class II (i.e. HLA-DR/DQ/DP) HLA-antibodies using SAFB on a Luminex platform (LabScreen LS1A04 Lot 002 or 003; LS2A01 Lot 005 or 006; OneLambda, Canoga Park, CA). A positive result was defined as a baseline normalized mean fluorescence intensity (MFI) >500. Donor specificity of HLA-antibodies was determined by comparison of the HLA-antibody specificities with the HLA-typing of the donor as previously reported (8). For every individual HLA-DSA, the reported strength is based on the MFI of one SAFB. In case of more than one HLA-DSA against different HLA-antigens, the cumulative strength was calculated by adding the individual MFI values.
T and B cells were isolated using immunomagnetic beads (Dynabeads, Dynal Biotech, Oslo, Norway). One μL of donor T and B cells was incubated with 1 μL of recipient sera for 30 and 40 min, respectively. Five μL rabbit complement and staining solution were added and incubated for 45 min. T- and B-cell CDC-XM were considered positive when the observed cell death exceeded 10% above background. No prospective FC-XM were performed.
Typing of HLA-antigens
HLA-A/B/DR antigens were determined by serology (Biotest, Rockaway, NJ) and confirmed by SSP DNA typing (Protrans, Indianapolis, IN). In the historic control group, HLA-DQ antigens were primarily inferred from the HLA-DR antigens, which are in strong linkage disequilibrium (14). If a potential donor-specific HLA-DQ antibody was present, HLA-DQβ antigens of the donor were verified by SSP DNA typing (Protrans). In the ATG/IvIg group, HLA-DQβ antigens were determined by serology and confirmed by SSP DNA typing. In both groups, HLA-Cw and/or HLA-DPβ antigens of the donor were assigned by DNA typing, if the recipient had HLA-Cw and/or HLA-DP antibodies.
We used JMP software version 7.0 (SAS Institute Inc., Cary, NC) for statistical analysis. For categorical data, Fisher's exact test or Pearson's chi-square test were used. Parametric continuous data were analyzed by Student's t-tests. For nonparametric continuous data, the Wilcoxon rank-sum test was used. Survival analysis was performed by the Kaplan–Meier method and groups compared using the log-rank test. Nominal logistic regression was used to determine independent predictors for clinical AMR. A p-value <0.05 was considered to indicate statistical significance.
Baseline patient characteristics of the two groups are detailed in Table 1. The two groups were not different regarding donor parameters, HLA-matching, prior renal transplants and pregnancies, as well as number and class of HLA-DSA. However, patients in the ATG/IvIg group had lower total MFI of HLA-DSA in the day-of-transplant sera than patients in the historic control group (median 2287 [range 543–26 537] vs. median 6494 (range 524–36 715); p < 0.0001). Median follow-up time in the ATG/IvIg and the historic control group was 2 years (range 1–4.7) and 8.5 years (range 5–11), respectively.
Rejection episodes and allograft function
The cumulative incidence of clinical/subclinical AMR within the first 6 months posttransplant was lower in the ATG/IvIg than in the historic control group (38% vs. 55%; p = 0.03) (Figure 1A). Most strikingly, clinical AMR occurred only in 4/37 patients (11%) in the ATG/IvIg group compared to 31/67 patients (46%) in the historic control group (p = 0.0002). The four patients experiencing clinical AMR in the ATG/IvIg group had total MFI of 1363, 7732, 12 265 and 26 537 at the time of transplantation. The prevalence of subclinical AMR at 3 and 6 months posttransplant was not different between the ATG/IvIg and the historic control group (29% vs. 32% and 33% vs. 46%, respectively; p ≥ 0.43). The phenotypes of subclinical AMR in the two groups were not different (Table 2).
|Historic control (n = 67)||Anti-T-lymphocyte globulin (ATG)/ Intravenous immunoglobulins (IvIg) (n = 37)||p-level|
|Patients with any biopsies, n (%)||64 (96)||36 (97)||1.0|
|Patients with clinical biopsies, n (%)||54 (81)||8 (22)||<0.0001|
|One, n (%)||17 (26)||7 (19)|
|Two or more, n (%)||37 (55)||1 (3)|
|Clinical rejection episodes*|
|Antibody-mediated rejection (AMR), n (%)||31 (46)||4 (11)||0.0002|
|C4d positive acute AMR, n||18||4|
|C4d negative acute AMR, n||7||0|
|C4d positivity only, n||6||0|
|T-cell-mediated rejection (TCR), n (%)||34 (50)||0||<0.0001|
|IA and IB, n||14|
|Protocol biopsy at 3 months|
|Biopsies performed/patients at risk, n/n (%)||28/63 (44)||35/36 (97)||<0.0001|
|AMR, n (%)||9 (32)||10 (29)||0.79|
|C4d positive acute AMR, n||3||5|
|C4d negative acute AMR, n||2||2|
|C4d positivity only, n||4||3|
|TCR, n (%)||12 (43)||4 (11)||0.008|
|IA and IB, n||4||0|
|Protocol biopsy at 6 months|
|Biopsies performed/patients at risk, n/n (%)||28/62 (45)||33/36 (92)||<0.0001|
|AMR, n (%)||13 (46)||11 (33)||0.43|
|C4d positive acute AMR, n||3||5|
|C4d negative acute AMR, n||2||2|
|C4d positivity only, n||8||4|
|TCR, n (%)||13 (46)||6 (18)||0.03|
|IA and IB, n||4||0|
The cumulative incidence of clinical/subclinical AMR or TCR was significantly lower in the ATG/IvIg group than in the historic control group (38% vs. 72%; p = 0.0002) (Figure 1B). Clinical TCR was not observed in the ATG/IvIg group, while 34/67 patients (50%) in the historic control group experienced clinical TCR (p < 0.0001). The prevalence of subclinical TCR at 3 and 6 months posttransplant was lower in the ATG/IvIg than in the historic control group (11% vs. 43%, and 18% vs. 46%, respectively; p ≤ 0.03). All subclinical TCR in the ATG/IvIg group were borderline tubulitis, while more severe grades were observed in the historic control group (Table 2).
Calculated GFR by the MDRD formula at 12 months posttransplant was 8 mL/min higher in the ATG/IvIg than in the historic control group (median 52 mL/min [range 24–87] vs. median 44 mL/min [range 13–102]; p = 0.20).
Graft and patient survival
Patient survival was not different between the two groups (p = 0.20), but this result has to be interpreted with caution because the follow-up time in the ATG/IvIg group was very short (Figure 2A). So far, two patients died in the ATG/IvIg group. One patient deceased 2 years posttransplant at the age of 70 years following bacterial sepsis, the other patient died 3 years posttransplant at the age of 63 years due to cardiac arrest. Both patients had well-functioning allograft.
Death-censored allograft survival was also not different between the two groups (p = 0.45), but again this has to be regarded as preliminary due to the short follow-up time in the ATG/IvIg group (Figure 2B). So far, three allografts were lost in the ATG/IvIg group. One was related to a donor-transmitted infection of the allograft with aspergillus fumigatus requiring transplant nephrectomy 6 weeks posttransplant (15). Two occurred at 2 and 4 years posttransplant because of ongoing AMR (the first patient had 3 HLA-DSA with cumulative 12 265 MFI at the time of transplantation; the second patient had 1 HLA-DSA with 590 MFI). In the historic control group 5/67 allografts (7.5%) were lost within the first year, all related to refractory early AMR.
Predictive factors for early clinical AMR
Thirty-five of the combined 104 patients of both groups (34%) developed clinical AMR (22 C4d positive acute AMR, 7 C4d negative acute AMR, 6 C4d positivity only). The multivariate analysis to determine independent predictors for clinical AMR included all significant parameters in the univariate analysis (i.e. recipient gender, HLA-DR mismatches, number of HLA-DSA, class of HLA-DSA, total MFI of HLA-DSA, prior transplants, prior blood transfusions, prior pregnancies and induction with ATG/IvIg). Absence of ATG/IvIg induction therapy was the only independent predictor for clinical AMR (p = 0.001) (Table 3). When all patients with clinical/subclinical C4d positivity only were excluded from the analysis (i.e. 6 clinical and 6 subclinical C4d positivity only), 29 of 92 patients (32%) developed clinical AMR. Multivariate analysis in this population revealed the absence of ATG/IvIg induction therapy (p = 0.01) and the number of HLA-DSA (p = 0.007) as independent predictors for clinical AMR.
|Clinical AMR (n = 35)||No clinical AMR (n = 69)||Univariate p-level||Multivariate p-level|
|Gender, females (%)||24 (69)||33 (48)||0.06||0.99|
|Age, median (range)||51 (15–68)||53 (18–73)||0.43|
|Deceased donor, n (%)||20 (57)||41 (59)||0.84|
|Gender, females (%)||20 (57)||36 (53)||0.83|
|Age, median (range)||52 (1–76)||48 (4–73)||0.90|
|A/B, mean ± std||2.6 ± 1||2.7 ± 0.9||0.48|
|DRβ1, mean ± std||1.4 ± 0.6||1.1 ± 0.6||0.01||0.24|
|Number of DSA, mean ± std||2 ± 1.1||1.5 ± 0.9||0.01||0.07|
|Class of DSA|
|Class I, n (%)||9 (26)||31 (45)||0.08||0.63|
|Class II, n (%)||15 (43)||27 (39)|
|Class I+II, n (%)||11 (31)||11 (16)|
|Cumulative strength of DSA [mean fluorescence intensity(MFI)]*|
|Median (range)||7471 (524–34 929)||3189 (543–36 715)||0.01||0.86|
|Prior transplants, n (%)||11 (31)||37 (54)||0.04||0.24|
|Blood transfusions, n (%)||8 (23)||35 (51)||0.007||0.13|
|Pregnancies, n (%)||19 (54)||23 (33)||0.06||0.48|
|ATG/IvIg FK-MMF-P, n (%)||4 (11)||33 (48)||0.0002||0.001|
To the best of our knowledge, this is the first study that investigated the short-term efficacy of ATG/IvIg induction in patients with low-level HLA-DSA by comparison with an immunologically well-matched historic control group. The main observation in this comparative cohort study was that ATG/IvIg induction significantly reduces the incidence of TCR and the severity of AMR compared to standard immunosuppression without ATG/IvIg. Specifically, the overall incidence of clinical rejection (AMR and TCR) dropped from 64% in the historic control group to 11% in patients receiving ATG/IvIg induction (p < 0.0001). In addition, calculated GFR at one year was 8 mL/min higher in the ATG/IvIg than in the historic control group (52 vs. 44 mL/min). Whether these encouraging short-term results will translate into improved long-term allograft survival, requires an extended observation time. In this regard, the still-observed subclinical AMR rate of 30% at 3 and 6 months posttransplant is worrisome.
The long-term impact of subclinical AMR on allograft survival is currently still unknown. However, accumulating data demonstrate that subclinical AMR can progress to transplant glomerulopathy as well as interstitial fibrosis and tubular atrophy, which are among the three top-ranked specific causes for allograft losses (16–18). It is tempting to speculate that successful treatment of subclinical AMR might be beneficial to improve long-term allograft survival. The key question is, how ongoing subclinical humoral rejection can be stopped given the fact that it developed and/or persisted despite induction therapy with ATG/IvIg and maintenance immunosuppression with Tac/MMF/P. Recently, Perry et al. have shown in vitro that ATG, IvIg and rituximab failed to inhibit alloantibody production by plasma cells, whereas bortezomib induced plasma cell apoptosis (19). Limited in vivo data suggest that bortezomib might indeed become a valuable treatment option for (sub)clinical AMR (19–21). Clearly, prospective randomized studies with serial allograft biopsies are necessary to better delineate the efficacy of bortezomib for prevention/treatment of (sub)clinical AMR.
This study might also help to explain the recently reported contradictive results regarding the clinical relevance of low-level HLA-DSA detected by SAFB. Our data demonstrate that ATG/IvIg induction and maintenance immunosuppression with Tac/MMF/P reduces the overall clinical rejection rate to 11% in patients with low-level HLA-DSA and that this regimen is the strongest independent predictor for prevention of clinical AMR. Interestingly, three studies that found no or only a limited short-term clinical impact of low-level HLA-DSA defined by SAFB used ATG induction and Tac/MMF/P maintenance immunosuppression in all (22,23) or the majority of patients (24). In contrast, without ATG induction low-level HLA-DSA defined by SAFB were associated with a high risk for early clinical rejection and/or a lower allograft survival at 5 years posttransplant (8,9,25). Therefore, the clinical relevance of low-level HLA-DSA has to be interpreted in the context of the used immunosuppression.
Several groups have reported clinical AMR rates from 7% to 44% in patients with low-level HLA-DSA using induction regimens with ATG ± IvIg ± plasmapheresis (3–7). Although it is difficult to directly compare these cohorts with our data due to differences in patient characteristics, HLA-DSA detection methods and the immunosuppressive regimens, more than 56% of all these patients did not experience AMR. This suggests that some low-level HLA-DSA appear to be clinically irrelevant per se or are controllable by these induction regimens–-at least in the short-term—while others are not. Clearly, defining pathogenic factors of HLA-DSA that are predictive for the occurrence of AMR and decreased long-term allograft survival will be essential for better risk stratification (26).
In our study of 104 patients, readily available parameters pretransplant (i.e. characteristics of HLA-DSA and the route of sensitization) had no or not clinically useful predictive value for the occurrence of clinical AMR (Table 3). In agreement with our observation, another large study reported that the route of sensitization (i.e. prior transplantation versus other) and the pretransplant strength of low-level HLA-DSA determined by FC-XM were not predictive for the occurrence and severity of AMR (27). In contrast, three smaller studies found a correlation between pretransplant amounts of low-level HLA-DSA determined by solid-phase assays or FC-XM and the development of AMR (4,6,28). Again, consistent with our observation, Lefaucheur et al. found no association of the class of HLA-DSA (i.e. class I vs. class II) and the occurrence of AMR (6). Therefore, other than these readily available parameters seem to critically influence the clinical impact of low-level HLA-DSA. Interestingly, Burns et al. found that increased HLA-DSA production posttransplant was associated with the development of AMR suggesting that the occurrence and magnitude of a humoral memory response is a key element for subsequent outcomes (27). In addition, factors regulating complement activation/inhibition, recruitment of inflammation cells and protective mechanisms of donor endothelial cells might further modulate the clinical impact of HLA-DSA (29–31).
Our definition of AMR was rather broad, including biopsies with the morphological phenotype of AMR but negative C4d staining in PTC, and biopsies demonstrating C4d positivity without morphological features of AMR. Clearly, if more strict criteria would have been applied, the incidence of AMR would be lower. However, recent data from sequential protocol biopsies in patients with HLA-DSA indicate that there are frequent fluctuations in the extent of glomerulitis, peritubular capillaritis and C4d deposition in PTC over time. The authors concluded that ‘the entity of humoral lesions without C4d-deposition may represent a milder but progressive form of AMR’ (17). Furthermore, chronic active AMR without C4d deposition has recently been described and is associated with allograft failure (32). Isolated C4d positivity could correspond to a state of accommodation in AB0-incompatible transplantation, but this has likely a different significance in transplantations across HLA-DSA (33,34). In summary, these emerging data suggest that AMR is a dynamic process with fluctuating intensity not always fulfilling all criteria required for diagnosis of definitive AMR according to the Banff classification. The used assignment as AMR in our study is consistent with such a concept.
This study has several important limitations. Inherent to comparative cohort studies, imbalances of pretransplant patient characteristics may confound the observed treatment effect. All donor and immunological parameters were comparable between the two groups except the total amount of pretransplant HLA-DSA defined by MFI, which was significantly lower in the ATG/IvIg group. However, it is rather unlikely that this imbalance has led to an overestimation of the treatment effect of ATG/IvIg induction because the total MFI of HLA-DSA was not predictive for clinical AMR in the multivariate analysis.
Furthermore, it is difficult to determine whether the observed treatment effect is only related to the ATG/IvIg induction and not additionally to the maintenance immunosuppression with Tac/MMF/P, because Tac/MMF/P was not used in the historic control group. Notably, in the historic control group, we found no association of the used immunosuppressive regimens (i.e. Tac/azathioprine (Aza)/P, cyclosporine (CyA)/MMF/P, sirolimus (Sir)/MMF/P) and the occurrence of AMR (8) suggesting that maintenance immunosuppression is likely less important than ATG/IvIg induction for prevention of early rejection episodes.
Finally, detection and assignment of HLA-DSA was only based on SAFB and no FC-XM were performed. Therefore, we can not confirm by a cell-based assay that all detected HLA-DSA were in fact directed against the ‘true’ HLA-molecules of the donor. Although several recent studies reported concordance rates >85% between HLA-DSA defined by SAFB and FC-XM (10,35–37), technically related false positive SAFB results and thus false assignments as HLA-DSA might have occurred in our study (38,39). However, these are likely equally distributed within the two groups, because the detection of HLA-antibodies (i.e. use of only two different lots of SAFB with minor differences) and the assignment as HLA-DSA were identical.
In conclusion, ATG/IvIg induction significantly reduced the incidence of TCR and the severity of AMR in patients with low-level HLA-DSA. However, the observed clinical/subclinical AMR rate of 38% suggests an insufficient control of the humoral immune response. Adaptation/supplementation of this regimen with drugs that might target the humoral immune response more specifically (e.g. bortezomib) could be beneficial, but need to be studied in prospective randomized trials.
SS is supported by Swiss National Foundation grant 32473B_125482/1.