- Email: [email protected]
Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation
Human Immunology xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
www.ashi-hla.org
journal homepage: www.elsevier.com/locate/humimm
Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation Igor Salvadé a,⇑, Vincent Aubert a, Jean-Pierre Venetz a, Dela Golshayan a, Anne-Catherine Saouli a, Maurice Matter b, Samuel Rotman c, Giuseppe Pantaleo a, Manuel Pascual a,⇑ a b c
Transplantation Center and Service of Immunology and Allergy, Lausanne University Hospital, 1011 Lausanne, Switzerland Service of Visceral Surgery, Lausanne University Hospital, 1011 Lausanne, Switzerland Institute of Pathology, Lausanne University Hospital, 1011 Lausanne, Switzerland
a r t i c l e
i n f o
Article history: Received 27 October 2015 Revised 8 April 2016 Accepted 8 April 2016 Available online xxxx Keywords: Kidney transplantation DSA cMFI Acute AMR
a b s t r a c t Background: Pretransplant anti-HLA donor-specific antibodies (DSA) are recognized as a risk factor for acute antibody-mediated rejection (AMR) in kidney transplantation. The predictive value of C4d-fixing capability by DSA or of IgG DSA subclasses for acute AMR in the pretransplant setting has been recently studied. In addition DSA strength assessed by mean fluorescence intensity (MFI) may improve risk stratification. We aimed to analyze the relevance of preformed DSA and of DSA MFI values. Methods: 280 consecutive patients with negative complement-dependent cytotoxicity crossmatches received a kidney transplant between 01/2008 and 03/2014. Sera were screened for the presence of DSA with a solid-phase assays on a Luminex flow analyzer, and the results were correlated with biopsy-proven acute AMR in the first year and survival. Results: Pretransplant anti-HLA antibodies were present in 72 patients (25.7%) and 24 (8.6%) had DSA. There were 46 (16.4%) acute rejection episodes, 32 (11.4%) being cellular and 14 (5.0%) AMR. The incidence of acute AMR was higher in patients with pretransplant DSA (41.7%) than in those without (1.6%) (p < 0.001). The median cumulative MFI (cMFI) of the group DSA+/AMR+ was 5680 vs 2208 in DSA+/AMR (p = 0.058). With univariate logistic regression a threshold value of 5280 cMFI was predictive for acute AMR. DSA cMFI’s ability to predict AMR was also explored by ROC analysis. AUC was 0.728 and the best threshold was a cMFI of 4340. Importantly pretransplant DSA > 5280 cMFI had a detrimental effect on 5-year graft survival. Conclusions: Preformed DSA cMFI values were clinically-relevant for the prediction of acute AMR and graft survival in kidney transplantation. A threshold of 4300–5300 cMFI was a significant outcome predictor. Ó 2016 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights reserved.
1. Introduction The presence of pretransplant donor-specific antibodies (DSA) is recognized to be a risk factor for antibody-mediated rejection
Abbreviations: AIC, Akaike’s information coefficient; AMR, antibody-mediated rejection; CDC, complement dependent cytotoxicity; cMFI, cumulative mean fluorescence intensity; CsA, cyclosporine; DSA, donor-specific antibodies; FCXM, flow cytometric crossmatch; HLA, human leukocyte antigen; ivIg, intravenous immunoglobulin; MMF, mycophenolate mofetil; PRA, panel reactive antibody; SPA, solid phase assay; Tx, transplantation. ⇑ Corresponding authors at: CTO BH07, CHUV, Rue du Bugnon 46, 1011 Lausanne, Switzerland. E-mail addresses: [email protected] (I. Salvadé), [email protected] (M. Pascual).
(AMR) and worse kidney allograft survival, even with a negative complement-dependent cytotoxicity (CDC) cross-match [1,2]. Since highly sensitive analytical technologies based on solid phase assays (SPA) have been routinely used in the transplant immunology laboratory, anti-HLA antibody detection has become much more sensitive than cell-based crossmatches but their precise clinical relevance (MFI, number, IgG subclass, complement fixing ability) remains to be established [2–8]. Over recent years, studies to identify which characteristics of pretransplant DSA (classes, strength, IgG-subclasses, capacity to activate complement, etc.) were predictive of adverse outcomes have been performed in order to improve our understanding of immunological risk stratification. As previously shown, not all pretransplant DSA, especially those at low levels, have a negative impact on graft function or acute
http://dx.doi.org/10.1016/j.humimm.2016.04.010 0198-8859/Ó 2016 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights reserved.
Please cite this article in press as: I. Salvadé et al., Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation, Hum. Immunol. (2016), http://dx.doi.org/10.1016/j.humimm.2016.04.010
2
I. Salvadé et al. / Human Immunology xxx (2016) xxx–xxx
rejection [9,10]. However, the amount of DSA, as measured by mean fluorescence intensity (MFI) with the Luminex, could play a role [4]. The analysis of IgG subclasses and the ability to activate complement may be important for identifying high-risk pretransplantation DSA in kidney and lung transplantation, although this remains controversial [10–15]. A recent study has suggested that the amount of pre-transplant DSA, as measured by MFI, may predict post-transplant AMR and kidney allograft survival [16]. Furthermore, recent data from our own program have indicated that cumulative MFI values above 5000 are highly predictive of a positive flow cytometric crossmatch (FCXM) [17]. Interestingly, a recent editorial discussed the issue and significance of clinicallyrelevant DSA in organ transplantation [18]. In the current study, we analyzed whether the strength of pre-transplant DSA, as measured by cumulative MFI (cMFI), may correlate with the incidence of acute AMR and allograft survival.
2. Materials and methods 2.1. Patients All 280 consecutive adult patients undergoing kidney transplantation at our Transplantation Center at the CHUV in Lausanne, between January 2008 and March 2014, were analyzed. No patient was excluded from the analysis. Baseline patient characteristics are shown in Table 1. All patients had a negative pretransplant T- and B-cell CDC crossmatch on current and peak sera. Data regarding
Table 1 Patients characteristics. All
DSA negative
DSA positive
Patients, n (%) Male, % Age, mean ± SD
280 66.8 54.6 ± 15.8
256 (91.4) 69.1 54.5 ± 16.0
24 (8.6) 41.7 55.4 ± 14.5
0.011 0.788
Kidney disease, % Diabetes Hypertension Glomerulonephritis Genetic diseases Other
8.6 11.4 31.1 19.6 29.3
8.6 12.5 30.5 18.7 29.7
8.3 0.0 37.5 29.2 25.0
1.00 – 0.494 0.279 0.815
Living donor, %
49.3
51.6
25.0
0.018
HLA mismatches, % 0 1 2 3 4 5 6
4.6 2.1 9.3 18.2 27.8 25.0 13.9
4.7 2.3 9.0 18.0 26.6 25.0 14.4
4.2 0.0 12.5 20.8 29.2 25.0 8.3
1.00 – 0.476 0.782 0.811 1.00 0.548
HLA DQ mismatches, n 0 11 1 22 2 14
10 21 10
1 1 4
1.00 0.705 0.023
Cytotoxic PRA I at tx, % 0% 91.1 <10% 4.6 10–49% 3.6 50–79% 0.7 P80% 0.0
93.0 3.9 2.3 0.8 0.0
70.8 12.5 16.7 0.0 0.0
0.002 0.089 0.006 – –
Ac anti-HLA, % No Class I Class II Class I + II
81.2 5.9 6.2 6.6
0.0 16.7 12.5 70.8
– 0.067 0.215 <0.001
74.3 6.8 6.8 12.1
p-value (DSA+ vs DSA )
DSA. = donor-specific antibody; HLA. = human leucocyte antigen; PRA. = panel reactive antibody; SD. = standard deviation; tx. = transplantation.
pretransplant FCXM were available in 59% of cases (n = 166) and pretransplant FCXM were positive in 10 patients (6.0% of patients). Patient data were obtained by analyzing the clinical records and electronic databases of the CHUV Transplantation Center. All patients were regularly followed in the outpatient clinic from the time of transplantation until death, allograft failure (defined as the need of chronic renal replacement therapy), up to April 2015. Allograft survival was analyzed after censoring for death with a functioning graft. 2.2. Detection of anti-HLA antibodies and definition of DSA All pretransplant sera were tested for the presence of anti-HLA antibodies using the multiplex technology SPA. Anti-HLA class I (i.e., HLA-A/B/Cw) and anti-HLA class II (i.e., HLA-DR/DP/DQ) antibodies were tested using LabScreen LS1A04 Lot 007/008 and LS2A01 Lot 008/009/010 (OneLambda, Canoga Park, CA) commercialized in Switzerland by InGen. Briefly, 20 ll of serum samples were incubated with HLA class I–coated and HLA class II–coated microspheres, respectively, for 30 min in the dark under gentle agitation. The specimens were then washed five times before being incubated with anti-human phycoerythrine-conjugated IgG in the same conditions as in the first incubation. The Labscan 100 flow analyzer (Luminex, Austin, TX) was used for beads and data acquisition. Data were then exported to HLA Visual software (One Lambda) for analysis. The cut-off level was defined as a baseline normalized 500 mean fluorescence intensity units (MFI). The presence of DSA was assigned by comparing the various HLA specificities proposed by the software analysis with the HLA typing of the donor for all transplanted patients. The analysis was carried out on the last available pre-transplant serum. DSA were considered positive if MFI was superior to 500. If several DSAs were present, we calculated the cMFI by adding together the MFI of each single antibody. 2.3. Donor HLA typing and virtual crossmatch Between 2008 and 2011, HLA typing was done without DQ specificity. A retrospective HLA-DQ typing was realized using donor’s DNA in cases of anti-HLA DQ antibody production in the recipient, to look for the presence of DSA. Since 2012, all donors were typed for class I and class II HLA molecules using PCR-SSP technology (Olerup SSP HLA-A-B-DR-DQ Combi-tray, Stockholm, Sweden) distributed in Switzerland by Milan Analytica AG or PCR-SSO technology (LabType SSO, One Lambda Canoga Park, CA, USA distributed in Switzerland by InGen). No detection of DSA against HLA-Cw and HLA-DP could be performed. 2.4. Complement-dependent cytotoxicity crossmatch Prospective donor T- and B-cell crossmatches were performed at the time of transplantation using the classical CDC-technique. Patient sera were incubated with donor’s T and B cells, respectively for 20 and 30 min. The cells were then incubated and stained with rabbit complement sera mixed with acridine orange/ethidium bromide for 35 min (T cells) and 45 min (B cells), and observed for cytotoxicity in a fluorescence microscope. The CDC crossmatch results were considered positive when cell death exceeded 20%. 2.5. Induction protocol and maintenance immunosuppression Induction therapy was used for all patients. For low-risk patients, it consisted of an anti-IL-2 receptor monoclonal antibody (Basiliximab NovartisÒ, 20 mg twice, at day 0 and 4). For high-risk patients (defined by the presence of high levels of lymphocytotoxic antibodies (PRA > 50%), previous transplantations or delayed graft
Please cite this article in press as: I. Salvadé et al., Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation, Hum. Immunol. (2016), http://dx.doi.org/10.1016/j.humimm.2016.04.010
I. Salvadé et al. / Human Immunology xxx (2016) xxx–xxx
function (DGF)), a polyclonal anti-thymocyte globulin induction (Thymoglobulin Sanofi-AventisÒ, 1–1.5 mg/kg daily for 3–4 days) was administered aiming for a CD3 cell-count < 20/ll after 3–4 doses. All enrolled recipients had similar triple maintenance immunosuppression, consisting of oral tacrolimus (PrografÒ) with mycophenolate mofetil (MMF) and methylprednisolone (MP)/prednisone. In the absence of DGF, tacrolimus was generally initiated on day 2, and the dose was adjusted to maintain trough levels in whole blood between 7 and 9 lg/l during the first 3 months postoperatively, then between 6 and 8 lg/l during 3–6 months after transplant and between 5 and 7 lg/l thereafter. MMF was started at a dose of 2000 mg/day, adjusting the dose to 1000–1500 mg/day postoperatively, depending on white blood cells count, gastrointestinal tolerance and BK viral loads between months 3 and 6. Methylprednisolone was administered intravenously at doses of 500 mg on the day of transplantation, 250 mg at day 1, 125 mg at day 2 and 80 mg on day 3. Oral prednisone was started on day 4 after the operation at the dose of 40 mg, 20 mg the day after and was then tapered to 5 mg/day within 3 months after transplantation. No desensitization protocol was carried out in any of these patients, and no ABO-incompatible transplants were performed. 2.6. Definition and management of rejection In cases of renal dysfunction post operatively (defined as creatinine increase of 20% or more from baseline, new proteinuria over 1 g/day or DGF of >1 week), an indication allograft biopsy was performed. Rejection episodes were all biopsy-proven and defined according to the Banff classification [19]. All biopsy samples were evaluated by light microscopy and immunofluorescence staining for C4d. Briefly, acute AMR was defined by the presence of circulating DSA in plasma, C4d positivity on the biopsy with signs of glomerulitis and capillaritis. Treatment for acute cellular rejection consisted of 3 days of high-dose methylprednisolone (500 mg daily boluses), followed by a prednisone taper. In addition, steroidresistant acute cellular rejections were treated with Thymoglobulin for 4–5 days. In cases of acute AMR the treatment consisted of daily plasmapheresis during five consecutive days, with Thymoglobulin (1 mg/kg) administered at the end of the exchange. After the last session of plasmapheresis the patients received a dose (0.4 g/kg) of i.v. polyclonal immunoglobulin (ivIg). 2.7. Statistical analysis Statistical calculations were performed using R-software [20] and graphics with R-package ggplot 2 [21]. The power of the tests was calculated with software G⁄Power [22]. Continuous data were expressed using mean (standard deviation) or median (interquartile range) and categorical data were described as numbers (frequencies). Categorical data including demographic, clinical, immunological features and pre-transplant anti-HLA antibodies detection were compared using Fisher’s exact test for count data. Continuous variables were compared with Student t-test or Mann-Whitney U test, as appropriate. Graft survival curves were visualized using Kaplan–Meier method, with comparison between DSA+ and DSA patients groups being done by log-rank test [23]. Association between AMR occurrence and MFI strength categories was explored by univariate logistic regression. A multivariable analysis with «backward elimination» was performed to identify the relevants variables for both outcomes – all type of acute rejection and acute AMR –. Receiver operator characteristic (ROC) curves and area under the curve (AUC with the trapezoidal rule) were calculated for DSA MFI values, as well as sensitivity, specificity, positive and negative predictive values [24]. A two-sided p-value of <0.05 was considered as statistically significant.
3
3. Results 3.1. Patient characteristics according to DSA detection Baseline characteristics of the 280 patients (all, divided according to DSA) enrolled in the study are shown in Table 1. Of all the patients, 66.8% were male and 33.2% female, with the percentage of female rising to 58.3% in the DSA positive group. The mean age at transplantation was 54.6 years. The most represented cause of kidney disease was glomerulonephritis (31.1%). Kidney transplants were from a living donor in almost half of cases (49.3%) with this proportion dropping to half (25%) in the DSA positive group of patients. No statistical differences between patients with or without DSA were found regarding age and HLA mismatches (except for double HLA DQ mismatches: 3.9% vs 16.7%, p = 0.023). Cytotoxic PRA-I at transplantation showed less sensitization in DSA negative patients (PRA-I 0% in 93% vs 70.8%, p = 0.002, for DSA and DSA+ groups, respectively) with DSA positive group having more frequently a PRA-I >10% (3.1% vs 16.7%, p = 0.006, for DSA and DSA+ groups, respectively). Pretransplant anti-HLA antibodies were present in 72 patients (25.7%), and 24 (8.6%) had DSA. Patients with anti-HLA antibodies against both classes were more frequently found in the DSA positive group (70.8% vs 6.6%, p < 0.001, for DSA+ and DSA groups, respectively). The antibodies identified were anti-class I DSA in 4 patients, anti-class II DSA in 3 patients, and 17 patients had both anti-class I and anti-class II DSA. Overall 16 patients had one single DSA, four patients had two DSAs, three patients had three DSAs and one patient had four DSAs (mean ± SD: 1.54 ± 0.88; median [IQR]: 1 [1,2]). As already mentioned, there were ten patients with positive FCXM: all of them have been transplanted. Of those ten patients, six had DSAs and four did not. The six DSA positive patients had a median cMFI of 5978 (IQR: 1817–11,030). Two patients had class I DSA, three patients had class II DSA and the last one had DSA of both classes. Four patients had only one single DSA, another had two different DSAs and the last patient had four different DSAs. 3.2. Immunosuppression In 71.1% of patients, induction therapy consisted in basiliximab, while 27.2% received Thymoglobulin (alone or with adjuvants) (Table 2). These proportions were similar in the group without DSA (75.4% basiliximab vs 23% Thymoglobulin), but, as expected, they were different in the DSA positive group (25.0% basiliximab vs 70.8% Thymoglobulin). Long-term maintenance immunosuppression consisted in the majority of cases of tacrolimus (93.6%), but 10 patients received cyclosporine (mainly prescribed because of focal segmental glomerulosclerosis as underlying kidney disease). Analyzing more carefully DSA positive patients (n = 24), no statistically significant differences were found in the choice of induction therapy and the development after transplantation of an acute AMR (Table 3). The incidence of AMR in the group who received basiliximab was 50% vs 22.2% in the group treated with Thymoglobulin alone vs 62.5% in the group who received also ivIg (p = 0.251). DSA strength was not statistically different between patients induced with basiliximab, Thymoglobulin or Thymoglobulin + ivIg (median cMFI (IQR) respectively 3527 (5722) vs 2326 (2814) vs 5145 (7560), p = 0.626 with ANOVA Kruskal-Wallis). 3.3. Correlation of DSA with acute antibody mediated rejection The number and characteristics of acute rejection episodes in the first year after transplantation are shown in Table 2. There were a total of 46 (16.4%) acute rejection episodes (all patients experienced a single episode of acute rejection), 32 (11.4%) being
Please cite this article in press as: I. Salvadé et al., Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation, Hum. Immunol. (2016), http://dx.doi.org/10.1016/j.humimm.2016.04.010
4
I. Salvadé et al. / Human Immunology xxx (2016) xxx–xxx
Table 2 Immunosuppression, acute rejection and survival. All
DSA negative
DSA positive
280
256
24
199 (71.1) 37 (13.2) 31 (11.1)
193 (75.4)
6 (25.0)
<0.001
28 (10.9) 31 (12.1)
9 (37.5) 0 (0.0)
0.001 –
8 (2.9) 5 (1.8)
0 (0.0) 4 (1.6)
8 (33.3) 1 (4.2)
– 0.364
262 (93.6) 10 (3.6) 8 (2.8)
238 (93.0)
24 (100.0)
0.381
10 (3.9) 8 (3.1)
0 (0.0) 0 (0.0)
– –
Rejection at 1 year, n (%) All acute rejection Acute cellular rejection Borderline 1A 1B 2A 2B Acute humoral rejection
46 (16.4) 32 (11.4) 9 (3.2) 15 (5.7) 4 (1.4) 3 (1.1) 1 (0.4) 14 (5.0)
34 (13.3) 30 (11.7) 9 (3.5) 14 (5.5) 3 (1.2) 3 (1.2) 1 (0.4) 4 (1.6)
12 (50.0) 2 (8.3) 0 (0.0) 1 (4.2) 1 (4.2) 0 (0.0) 0 (0.0) 10 (41.7)
<0.001 1.00 – 1.00 0.303 – – <0.001
5-y graft survival, % 5-y patient survival, %
90.8 93.9
92.0 95.5
76.2 74.5
0.248 0.013
Patients, n Induction, n (%) Simulect Thymoglobulin Simulect + Thymoglobulin Thymo + ivIg Other Immunosuppression, n (%) Tacrolimus CsA Other
DSA. = donor-specific = cyclosporine.
antibody;
ivIg. = intravenous
pvalue
immunoglobulin;
CsA.
Table 3 Induction therapy association with acute AMR for DSA positive patients. Induction
Simulect Thymoglobulin Thymoglobulin + ivIg Other
DSA positive (n = 24)
p-value
AMR positive
AMR negative
n
Median cMFI (IQR)
n
Median cMFI (IQR)
3 2 5 0
8725 (4712) 4566 (16) 6779 (10,555)
3 7 3 1
2996 (1304) 1752 (2048) 3510 (3590) 2091
AMR. = antibody-mediated rejection; intravenous immunoglobulin.
DSA. = donor
specific
0.70 0.056 0.57
antibody;
ivIg. =
acute cellular rejections and 14 (5.0%) acute AMR. The incidence of acute AMR was significantly higher in patients with DSA (41.7%) than in those without DSA (1.6%) (p < 0.001). Of the 24 patients with pretransplant DSA, 10 had an acute AMR (DSA+/AMR+) and 14 did not (DSA+/AMR ). The median (IQR) cMFI of the DSA +/AMR+ group was 5680 (2289–9641) vs 2208 (1347–3920) in DSA+/AMR (p = 0.058) (Fig. 1). With univariate logistic regression we could define a threshold value of 5280 cMFI (95% CI: 4205–9235) that was predictive for acute AMR (Fig. 2). Pretransplant DSA cMFI’s ability to predict AMR occurrence was also explored by ROC analysis. AUC was 0.728 (95%CI: 0.474– 0.983) and the best threshold was a cMFI of 4340 with a specificity of 0.928 (95%CI: 0.786–1) and a sensitivity of 0.700 (95%CI: 0.4–0.9); positive predictive value of 0.875 and negative predictive value of 0.812 (Fig. 3). We also analyzed the ability of maximum MFI (the MFI of the ‘‘strongest” DSA present) to predict acute AMR: the discrimination between the two groups DSA+/AMR+ vs DSA+/AMR was weaker and non significant (p-value 0.072) and the ROC analysis showed a lower AUC of 0.690 (95%CI: 0.444–0.867). The median time from transplant to acute AMR in the group DSA+/AMR+ (n = 10) was 13 days (IQR 132 days, min 5 days and max 217 days). Surprisingly, patients with low cMFI had a shorter time to acute AMR than patient with high cMFI (adjusted R-squared: 0.6541, p-value 0.003). The odds-ratio and relative risk of developing an acute AMR in the first year according to cMFI values is shown in Table 4. With multivariable logistic regression and backward elimination of all explicative variables analyzed in our study (age at transplantation, sex, type of donation – cadaveric or living –, class of anti-HLA antibodies, current PRA-I, class of DSA, number of DSA, DSA strength assessed by cMFI, number of HLA mismatches and induction therapy), only the age (p = 0.006) and cMFI (p < 0.001, AIC coefficient 225.61) were significantly associated with acute rejection (all types). Regarding acute AMR, only the value of cMFI was found to have a significant role (p = 0.004, AIC coefficient 69.75, OR 1.97 for each cMFI increment of 1000 (95%CI: 1.46–2.67). The number or anti-HLA class of pretransplant DSA did not. We also analyzed the evolution with time of cMFI values after transplant and we found that the cMFI of DSA detected pretransplant showed a significant decrease (median cMFI (IQR): 2209 (2571) pre-transplant vs 0 (967) at 1 year, p-value 0.005) post transplant, in patients DSA+/AMR (n = 14). However, for patients DSA+/AMR+ (n = 10), the cMFI persisted during the first year post transplant (median cMFI
Fig. 1. Box plot of DSA cMFI for patients with an antibody-mediated rejection (AMR+, n = 10) and without (AMR , n = 14).
Please cite this article in press as: I. Salvadé et al., Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation, Hum. Immunol. (2016), http://dx.doi.org/10.1016/j.humimm.2016.04.010
5
I. Salvadé et al. / Human Immunology xxx (2016) xxx–xxx
Fig. 2. Univariate logistic regression. Probability of experiencing an antibody-mediated rejection (AMR) in function of cumulative mean fluorescence intensity (cMFI) of donor-specific anti-HLA antibodies (DSA). Solid line represent logistic regression curve with grey zone being the CI95%. The cMFI value predictive of AMR (probability > 0.5) is 5280.
Fig. 3. Receiver operator curves of the cMFI for prediction of acute AMR. Area under the curve is 0.728 and the best threshold was a cMFI of 4340 with a specificity of 0.93 and a sensitivity of 0.70.
(IQR): 5680 (7351) pre-transplant vs 11,557 (15,905) at 1 year). Two patients in this group (DSA+/AMR+, n = 10) also developed de-novo DSA at the time of acute AMR. 3.4. Kidney and patient survival Patient survival at 5 years was worse for those with pretransplant DSA+ (74.5%) compared to the DSA group (95.5%; p = 0.013) (Fig. 4). The causes of death within the analyzed population were: 1 sudden death after hemodialysis, 1 hypoglycemia, 1 suicide (patient without DSA), four infections (septic chock or abdominal infection) and four neoplasms. There was also a trend for worse 5-year death-censored graft survival in the DSA+ group
Fig. 4. Survival of patients with donor-specific anti-HLA antibodies (DSA+) and without donor-specific anti-HLA antibodies (DSA ).
(76.2% vs 92.0%, for DSA+ and DSA , respectively), although this did not reach statistical significance (p = 0.248). However, by stratifying and analyzing the patients with the threshold of 5280 cMFI (value derived from the univariate logistic regression), there was a significantly lower graft survival in the DSA+ group with >5280 cMFI vs the other patients (53.3% vs 92.1%; p = 0.025) (Fig. 5). For this analysis, the patients without DSA and those with lowerlevels cMFI DSA (<5280 cMFI) were grouped together because their 5 year graft survival was similar (92.0% vs 94.4%; p-value = 0.950). If we analyze the graft survival of patients with DSAs according to development of acute AMR, we can observe a trend to a deleterious effect of experiencing an acute AMR: 5-year death-censored graft survival for patients DSA+/AMR+ was 57.8% vs 100% for those DSA+/AMR (p = 0.089).
Table 4 Odds ratio and relative risk.
cMFI 500–5280 cMFI P 5280
AMR+
AMR
OR
95% CI
RR
95% CI
p-value
5 5
13 1
11.50
0.96–657.53
3.00
1.31–6.86
0.050
AMR. = antibody-mediated rejection; cMFI. = cumulative mean fluorescence intensity; DSA. = donor-specific antibody; OR. = odds ratio; RR. = relative risk.
Please cite this article in press as: I. Salvadé et al., Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation, Hum. Immunol. (2016), http://dx.doi.org/10.1016/j.humimm.2016.04.010
6
I. Salvadé et al. / Human Immunology xxx (2016) xxx–xxx
Fig. 5. Death-censored kidney graft survival of patients with donor-specific antibody more than 5280 cumulative mean fluorescence intensity (cMFI) and less than 5280 cMFI or without.
4. Discussion In the present study, we have investigated the importance of DSA as a predictive biomarker for antibody-mediated rejection in kidney transplantation. We have hypothesized that pre-transplant DSA levels may serve as predictive biomarker for AMR. The investigation was performed on a cohort of 280 patients receiving kidney transplantation at our institution. The results obtained indicated that, in the transplant recipients with a negative CDC crossmatch at the time of transplantation, the presence of preformed DSA was associated with a significantly higher incidence of biopsyproven acute AMR (41.7%) in the first year after transplantation. The proportion of patients with acute AMR in our study is consistent with previous studies showing that the incidence of AMR ranged between 30% and 50%, depending in part on whether biopsies were systematically performed in all patients or not [2,4,5,16]. The detection of DSA was performed using a quantitative Luminex-based assay. Interestingly, the results obtained indicate that the group of patients with preformed DSA experiencing an acute AMR had higher amounts of circulating DSA (as measured by cMFI on pretransplant sera), as compared to those with preformed DSA but without AMR. The rationale of using cMFI rather than maximum (peak) MFI was based on the assumption that cMFI values may better reflect the degree of HLA incompatibility of the recipient versus the graft at the time of transplantation, and because of their possible better correlation with AMR. Multivariate analysis was performed on a panel of parameters including age, sex, type of donation, class of anti-HLA antibodies, current PRA-I, class of DSA, number of DSA, number of mismatches and induction therapy and the cMFI. The cMFI level was the only and single parameter which correlated with the incidence of acute AMR. We then performed logistic regression and ROC curve analysis in order to further evaluate the role of the levels of cMFI DSA as predictive biomarker of AMR. We showed that levels of cMFI being predictive of AMR ranged between 4300 and 5300, with good sensitivity and specificity (sensitivity 0.700 and specificity 0.928, with negative predictive value of 0.812). In support of these findings, a recent study from our group showed that DSA cMFI levels above 5000 were highly predictive of a positive FCXM [17]. Previous studies have already assessed the relationship between DSA strength and AMR with conflicting results [2,4,8,13,16,25]. The reasons for these discrepancies remain unclear, although the differences in the expression of MFI values (i.e. threshold of positivity, cumulative vs maximum) [1] have rendered somewhat difficult to compare the results from different studies.
Nevertheless, taken together, these results suggested that pretransplant cMFI DSA levels may serve as a predictive biomarker for acute AMR. We then investigated the impact of cMFI DSA levels on graft survival and overall patient survival. With regard to graft survival, we found that the association of DSA with AMR influenced negatively graft survival (57.8% for DSA+/AMR+), while in the absence of acute AMR graft survival (100%) was similar to that observed in non-immunized DSA negative patients (92.0%). The presence of pretransplant DSA also had a negative impact on patient survival. Indeed the 5-year patient survival was 95.5% in recipients without DSA vs 74.5% in recipients with DSA (p = 0.013). Whether this finding may reflect a higher incidence of comorbidities, or posttransplant infections and/or tumors, possibly related to more aggressive immunosuppressive regimens, as reported in a previous study [13], remains to be determined. The finding of the importance of acute AMR and/or cMFI DSA levels on graft survival led us to analyze whether other features of DSA or induction therapy may have a role on the incidence of AMR. As also reported by others [2,7], we found that neither the class nor the number of DSA played a role. With regard to immunosuppression, the incidence of acute AMR in the group of patients with DSA was not significantly influenced by the type of induction therapy (basiliximab vs Thymoglobulin vs Thymoglobulin + ivIg). Our results support the data from Amico et al. in a cohort of 67 kidney transplant recipients with preformed DSA, where no differences in AMR incidence were found in patients without induction vs patients who received basiliximab or daclizumab [2]. The present study has a number of limitations. It is a singlecenter study, the results on FCXM were not available in all patients and HLA typing did not include Cw and DP specificity. Finally there are known limitations regarding the use of MFI as a quantitative tool for circulating DSA [26,27]. In conclusion, the results from the current study indicate that preformed DSA cMFI values are clinically relevant to predict acute AMR and graft survival after kidney transplantation. A threshold of 4300–5300 cMFI appears to be a key determinant outcome predictor. Altogether our data support the detrimental role of preformed DSA in kidney transplantation.
Disclosure The authors of this manuscript have no conflicts of interest to disclose. This is an original work not submitted for publication elsewhere. This work was presented as an oral presentation at the ESOT meeting in Brussels (September 2015) by Dr. I. Salvadé.
Authorship IS and MP were responsible for concept, analytic design, data analysis, interpretation of results, and manuscript writing. GP helped with the interpretation of results, and critically reviewed the manuscript. VA, JPV, and DG helped with the data collection, gave thoughtful comments, and reviewed the manuscript. ACS, MM and SR helped with the data collection. All authors critically reviewed and approved the final version of the manuscript.
Funding No funding was required for this study.
Please cite this article in press as: I. Salvadé et al., Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation, Hum. Immunol. (2016), http://dx.doi.org/10.1016/j.humimm.2016.04.010
I. Salvadé et al. / Human Immunology xxx (2016) xxx–xxx
Aknowledgements We thank Stéphanie Gregoretti and Rocco Sugamele for the excellent laboratory assistance, and Prof. Giorgio Salvadé for his valuable assistance in the data analysis. We also thank all the clinicians and nurses, datamanagers, who were involved in patient management and data collection. References [1] S. Mohan, A. Palanisamy, D. Tsapepas, et al., Donor specific antibodies adversely affect kidney allograft outcomes, J. Am. Soc. Nephrol. 23 (12) (2012) 2061–2071. [2] P. Amico, G. Honger, M. Mayr, et al., Clinical relevance of pretransplant donorspecific HLA antibodies detected by single-antigen flow-beads, Transplantation 87 (11) (2009) 1681–1688. [3] J.L. Caro-Oleas, M.F. Gonzalez-Escribano, F.M. Gonzalez-Roncero, et al., Clinical relevance of HLA donor-specific antibodies detected by single antigen assay in kidney transplantation, Nephrol. Dial. Transplant. 27 (3) (2012) 1231–1238. [4] C. Lefaucheur, A. Loupy, G.S. Hill, et al., Preexisting donor-specific HLA antibodies predict outcome in kidney transplantation, J. Am. Soc. Nephrol. 21 (8) (2010) 1398–1406. [5] C. Lefaucheur, C. Suberbielle-Boissel, G.S. Hill, et al., Clinical relevance of preformed HLA donor-specific antibodies in kidney transplantation, Am. J. Transplant. 8 (2) (2008) 324–331. [6] H.G. Otten, M.C. Verhaar, H.P. Borst, R.J. Hené, A.D. van Zuilen, Pretransplant donor-specific HLA class-I and -II antibodies are associated with an increased risk for kidney graft failure, Am. J. Transplant. 12 (6) (2012) 1618–1623. [7] D.S. Tsapepas, R. Vasilescu, B. Tanriover, et al., Preformed donor-specific antibodies and risk of antibody-mediated rejection in repeat renal transplantation, Transplantation 97 (6) (2014) 642–647. [8] M. Willicombe, P. Brookes, E. Santos-Nunez, et al., Outcome of patients with preformed donor-specific antibodies following alemtuzumab induction and tacrolimus monotherapy, Am. J. Transplant. 11 (3) (2011) 470–477. [9] V. Aubert, J.P. Venetz, G. Pantaleo, M. Pascual, Low levels of human leukocyte antigen donor-specific antibodies detected by solid phase assay before transplantation are frequently clinically irrelevant, Hum. Immunol. 70 (8) (2009) 580–583. [10] G. Hönger, M. Wahrmann, P. Amico, H. Hopfer, G.A. Böhmig, S. Schaub, C4dfixing capability of low-level donor-specific HLA antibodies is not predictive for early antibody-mediated rejection, Transplantation 89 (2010) 1471–1475. [11] G. Hönger, H. Hopfer, M.L. Arnold, B.M. Spriewald, S. Schaub, P. Amico, Pretransplant IgG subclasses of donor-specific human leukocyte antigen antibodies and development of antibody-mediated rejection, Transplantation 92 (2011) 41–47.
7
[12] A. Loupy, C. Lefaucheur, D. Vernerey, et al., Complement-binding anti-HLA antibodies and kidney-allograft survival, N. Engl. J. Med. 369 (2013) 1215– 1226. [13] J.D. Smith, M.W. Ibrahim, H. Newel, et al., Pre-transplant donor HLAspecific antibodies: characteristics causing detrimental effects on survival after lung transplantation, J. Heart Lung Transplant. 33 (2014) 1074–1082. [14] N. Khovanova, S. Daga, T. Shaikhina, et al., Subclass analysis of donor HLAspecific IgG in antibody-incompatible renal transplantation reveals a significant association of IgG4 with rejection and graft failure, Transpl. Int. 28 (2015) 1405–1415, http://dx.doi.org/10.1111/tri.12684 (Epub 1 Sept). [15] C. Lawrence, M. Willicombre, P.A. Brookes, et al., Preformed complementactivating low-level donor-specific antibody predicts early antibody-mediated rejection in renal allografts, Transplantation 95 (2) (2013) 341–346. [16] J. Malheiro, S. Tafulo, L. Dias, et al., Analysis of preformed donor-specific antiHLA antibodies characteristics for prediction of antibody-mediated rejection in kidney transplantation, Transpl. Immunol. 32 (2) (2015) 66–71. [17] V. Aubert, R. Sugamele, S. Gregoretti, J.-P. Venetz, G. Pantaleo, M. Pascual, Are donor specific anti-HLA antibodies detected by Luminex predictive of positive flow cytometric crossmatch?, Tissue Antigens 85 (2015) 41 [18] R.R. Hackem, N.L. Reinsmoen, What is the definition of a clinically relevant donor HLA-specific antibody (DSA)?, Am J. Transplant. 15 (2015) 299–300. [19] M. Haas, B. Sis, L.C. Racusen, et al., Banff 2013 meeting report: inclusion of C4d-negative antibody-mediated rejection and antibody-associated arterial lesions, Am. J. Transplant. 14 (2014) 272–283. [20] R Core Team, R: A Language and Environment for Statistical Computing, R Foundation for Statistical Computing, Vienna, Austria, 2014. URL:; Terry M. Therneau, Patricia M. Grambsch, Modeling Survival Data: Extending the Cox Model, Springer, New York, 2000. ISBN 0-387-98784-3. [24] X. Robin, N. Turck, A. Hainard, et al., PROC: an open-source package for R and S+ to analyze and compare ROC curves, BMC Bioinformatics 12 (2011) 77, http://dx.doi.org/10.1186/1471-2105-12-77.
Please cite this article in press as: I. Salvadé et al., Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation, Hum. Immunol. (2016), http://dx.doi.org/10.1016/j.humimm.2016.04.010
Contents lists available at ScienceDirect
www.ashi-hla.org
journal homepage: www.elsevier.com/locate/humimm
Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation Igor Salvadé a,⇑, Vincent Aubert a, Jean-Pierre Venetz a, Dela Golshayan a, Anne-Catherine Saouli a, Maurice Matter b, Samuel Rotman c, Giuseppe Pantaleo a, Manuel Pascual a,⇑ a b c
Transplantation Center and Service of Immunology and Allergy, Lausanne University Hospital, 1011 Lausanne, Switzerland Service of Visceral Surgery, Lausanne University Hospital, 1011 Lausanne, Switzerland Institute of Pathology, Lausanne University Hospital, 1011 Lausanne, Switzerland
a r t i c l e
i n f o
Article history: Received 27 October 2015 Revised 8 April 2016 Accepted 8 April 2016 Available online xxxx Keywords: Kidney transplantation DSA cMFI Acute AMR
a b s t r a c t Background: Pretransplant anti-HLA donor-specific antibodies (DSA) are recognized as a risk factor for acute antibody-mediated rejection (AMR) in kidney transplantation. The predictive value of C4d-fixing capability by DSA or of IgG DSA subclasses for acute AMR in the pretransplant setting has been recently studied. In addition DSA strength assessed by mean fluorescence intensity (MFI) may improve risk stratification. We aimed to analyze the relevance of preformed DSA and of DSA MFI values. Methods: 280 consecutive patients with negative complement-dependent cytotoxicity crossmatches received a kidney transplant between 01/2008 and 03/2014. Sera were screened for the presence of DSA with a solid-phase assays on a Luminex flow analyzer, and the results were correlated with biopsy-proven acute AMR in the first year and survival. Results: Pretransplant anti-HLA antibodies were present in 72 patients (25.7%) and 24 (8.6%) had DSA. There were 46 (16.4%) acute rejection episodes, 32 (11.4%) being cellular and 14 (5.0%) AMR. The incidence of acute AMR was higher in patients with pretransplant DSA (41.7%) than in those without (1.6%) (p < 0.001). The median cumulative MFI (cMFI) of the group DSA+/AMR+ was 5680 vs 2208 in DSA+/AMR (p = 0.058). With univariate logistic regression a threshold value of 5280 cMFI was predictive for acute AMR. DSA cMFI’s ability to predict AMR was also explored by ROC analysis. AUC was 0.728 and the best threshold was a cMFI of 4340. Importantly pretransplant DSA > 5280 cMFI had a detrimental effect on 5-year graft survival. Conclusions: Preformed DSA cMFI values were clinically-relevant for the prediction of acute AMR and graft survival in kidney transplantation. A threshold of 4300–5300 cMFI was a significant outcome predictor. Ó 2016 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights reserved.
1. Introduction The presence of pretransplant donor-specific antibodies (DSA) is recognized to be a risk factor for antibody-mediated rejection
Abbreviations: AIC, Akaike’s information coefficient; AMR, antibody-mediated rejection; CDC, complement dependent cytotoxicity; cMFI, cumulative mean fluorescence intensity; CsA, cyclosporine; DSA, donor-specific antibodies; FCXM, flow cytometric crossmatch; HLA, human leukocyte antigen; ivIg, intravenous immunoglobulin; MMF, mycophenolate mofetil; PRA, panel reactive antibody; SPA, solid phase assay; Tx, transplantation. ⇑ Corresponding authors at: CTO BH07, CHUV, Rue du Bugnon 46, 1011 Lausanne, Switzerland. E-mail addresses: [email protected] (I. Salvadé), [email protected] (M. Pascual).
(AMR) and worse kidney allograft survival, even with a negative complement-dependent cytotoxicity (CDC) cross-match [1,2]. Since highly sensitive analytical technologies based on solid phase assays (SPA) have been routinely used in the transplant immunology laboratory, anti-HLA antibody detection has become much more sensitive than cell-based crossmatches but their precise clinical relevance (MFI, number, IgG subclass, complement fixing ability) remains to be established [2–8]. Over recent years, studies to identify which characteristics of pretransplant DSA (classes, strength, IgG-subclasses, capacity to activate complement, etc.) were predictive of adverse outcomes have been performed in order to improve our understanding of immunological risk stratification. As previously shown, not all pretransplant DSA, especially those at low levels, have a negative impact on graft function or acute
http://dx.doi.org/10.1016/j.humimm.2016.04.010 0198-8859/Ó 2016 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights reserved.
Please cite this article in press as: I. Salvadé et al., Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation, Hum. Immunol. (2016), http://dx.doi.org/10.1016/j.humimm.2016.04.010
2
I. Salvadé et al. / Human Immunology xxx (2016) xxx–xxx
rejection [9,10]. However, the amount of DSA, as measured by mean fluorescence intensity (MFI) with the Luminex, could play a role [4]. The analysis of IgG subclasses and the ability to activate complement may be important for identifying high-risk pretransplantation DSA in kidney and lung transplantation, although this remains controversial [10–15]. A recent study has suggested that the amount of pre-transplant DSA, as measured by MFI, may predict post-transplant AMR and kidney allograft survival [16]. Furthermore, recent data from our own program have indicated that cumulative MFI values above 5000 are highly predictive of a positive flow cytometric crossmatch (FCXM) [17]. Interestingly, a recent editorial discussed the issue and significance of clinicallyrelevant DSA in organ transplantation [18]. In the current study, we analyzed whether the strength of pre-transplant DSA, as measured by cumulative MFI (cMFI), may correlate with the incidence of acute AMR and allograft survival.
2. Materials and methods 2.1. Patients All 280 consecutive adult patients undergoing kidney transplantation at our Transplantation Center at the CHUV in Lausanne, between January 2008 and March 2014, were analyzed. No patient was excluded from the analysis. Baseline patient characteristics are shown in Table 1. All patients had a negative pretransplant T- and B-cell CDC crossmatch on current and peak sera. Data regarding
Table 1 Patients characteristics. All
DSA negative
DSA positive
Patients, n (%) Male, % Age, mean ± SD
280 66.8 54.6 ± 15.8
256 (91.4) 69.1 54.5 ± 16.0
24 (8.6) 41.7 55.4 ± 14.5
0.011 0.788
Kidney disease, % Diabetes Hypertension Glomerulonephritis Genetic diseases Other
8.6 11.4 31.1 19.6 29.3
8.6 12.5 30.5 18.7 29.7
8.3 0.0 37.5 29.2 25.0
1.00 – 0.494 0.279 0.815
Living donor, %
49.3
51.6
25.0
0.018
HLA mismatches, % 0 1 2 3 4 5 6
4.6 2.1 9.3 18.2 27.8 25.0 13.9
4.7 2.3 9.0 18.0 26.6 25.0 14.4
4.2 0.0 12.5 20.8 29.2 25.0 8.3
1.00 – 0.476 0.782 0.811 1.00 0.548
HLA DQ mismatches, n 0 11 1 22 2 14
10 21 10
1 1 4
1.00 0.705 0.023
Cytotoxic PRA I at tx, % 0% 91.1 <10% 4.6 10–49% 3.6 50–79% 0.7 P80% 0.0
93.0 3.9 2.3 0.8 0.0
70.8 12.5 16.7 0.0 0.0
0.002 0.089 0.006 – –
Ac anti-HLA, % No Class I Class II Class I + II
81.2 5.9 6.2 6.6
0.0 16.7 12.5 70.8
– 0.067 0.215 <0.001
74.3 6.8 6.8 12.1
p-value (DSA+ vs DSA )
DSA. = donor-specific antibody; HLA. = human leucocyte antigen; PRA. = panel reactive antibody; SD. = standard deviation; tx. = transplantation.
pretransplant FCXM were available in 59% of cases (n = 166) and pretransplant FCXM were positive in 10 patients (6.0% of patients). Patient data were obtained by analyzing the clinical records and electronic databases of the CHUV Transplantation Center. All patients were regularly followed in the outpatient clinic from the time of transplantation until death, allograft failure (defined as the need of chronic renal replacement therapy), up to April 2015. Allograft survival was analyzed after censoring for death with a functioning graft. 2.2. Detection of anti-HLA antibodies and definition of DSA All pretransplant sera were tested for the presence of anti-HLA antibodies using the multiplex technology SPA. Anti-HLA class I (i.e., HLA-A/B/Cw) and anti-HLA class II (i.e., HLA-DR/DP/DQ) antibodies were tested using LabScreen LS1A04 Lot 007/008 and LS2A01 Lot 008/009/010 (OneLambda, Canoga Park, CA) commercialized in Switzerland by InGen. Briefly, 20 ll of serum samples were incubated with HLA class I–coated and HLA class II–coated microspheres, respectively, for 30 min in the dark under gentle agitation. The specimens were then washed five times before being incubated with anti-human phycoerythrine-conjugated IgG in the same conditions as in the first incubation. The Labscan 100 flow analyzer (Luminex, Austin, TX) was used for beads and data acquisition. Data were then exported to HLA Visual software (One Lambda) for analysis. The cut-off level was defined as a baseline normalized 500 mean fluorescence intensity units (MFI). The presence of DSA was assigned by comparing the various HLA specificities proposed by the software analysis with the HLA typing of the donor for all transplanted patients. The analysis was carried out on the last available pre-transplant serum. DSA were considered positive if MFI was superior to 500. If several DSAs were present, we calculated the cMFI by adding together the MFI of each single antibody. 2.3. Donor HLA typing and virtual crossmatch Between 2008 and 2011, HLA typing was done without DQ specificity. A retrospective HLA-DQ typing was realized using donor’s DNA in cases of anti-HLA DQ antibody production in the recipient, to look for the presence of DSA. Since 2012, all donors were typed for class I and class II HLA molecules using PCR-SSP technology (Olerup SSP HLA-A-B-DR-DQ Combi-tray, Stockholm, Sweden) distributed in Switzerland by Milan Analytica AG or PCR-SSO technology (LabType SSO, One Lambda Canoga Park, CA, USA distributed in Switzerland by InGen). No detection of DSA against HLA-Cw and HLA-DP could be performed. 2.4. Complement-dependent cytotoxicity crossmatch Prospective donor T- and B-cell crossmatches were performed at the time of transplantation using the classical CDC-technique. Patient sera were incubated with donor’s T and B cells, respectively for 20 and 30 min. The cells were then incubated and stained with rabbit complement sera mixed with acridine orange/ethidium bromide for 35 min (T cells) and 45 min (B cells), and observed for cytotoxicity in a fluorescence microscope. The CDC crossmatch results were considered positive when cell death exceeded 20%. 2.5. Induction protocol and maintenance immunosuppression Induction therapy was used for all patients. For low-risk patients, it consisted of an anti-IL-2 receptor monoclonal antibody (Basiliximab NovartisÒ, 20 mg twice, at day 0 and 4). For high-risk patients (defined by the presence of high levels of lymphocytotoxic antibodies (PRA > 50%), previous transplantations or delayed graft
Please cite this article in press as: I. Salvadé et al., Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation, Hum. Immunol. (2016), http://dx.doi.org/10.1016/j.humimm.2016.04.010
I. Salvadé et al. / Human Immunology xxx (2016) xxx–xxx
function (DGF)), a polyclonal anti-thymocyte globulin induction (Thymoglobulin Sanofi-AventisÒ, 1–1.5 mg/kg daily for 3–4 days) was administered aiming for a CD3 cell-count < 20/ll after 3–4 doses. All enrolled recipients had similar triple maintenance immunosuppression, consisting of oral tacrolimus (PrografÒ) with mycophenolate mofetil (MMF) and methylprednisolone (MP)/prednisone. In the absence of DGF, tacrolimus was generally initiated on day 2, and the dose was adjusted to maintain trough levels in whole blood between 7 and 9 lg/l during the first 3 months postoperatively, then between 6 and 8 lg/l during 3–6 months after transplant and between 5 and 7 lg/l thereafter. MMF was started at a dose of 2000 mg/day, adjusting the dose to 1000–1500 mg/day postoperatively, depending on white blood cells count, gastrointestinal tolerance and BK viral loads between months 3 and 6. Methylprednisolone was administered intravenously at doses of 500 mg on the day of transplantation, 250 mg at day 1, 125 mg at day 2 and 80 mg on day 3. Oral prednisone was started on day 4 after the operation at the dose of 40 mg, 20 mg the day after and was then tapered to 5 mg/day within 3 months after transplantation. No desensitization protocol was carried out in any of these patients, and no ABO-incompatible transplants were performed. 2.6. Definition and management of rejection In cases of renal dysfunction post operatively (defined as creatinine increase of 20% or more from baseline, new proteinuria over 1 g/day or DGF of >1 week), an indication allograft biopsy was performed. Rejection episodes were all biopsy-proven and defined according to the Banff classification [19]. All biopsy samples were evaluated by light microscopy and immunofluorescence staining for C4d. Briefly, acute AMR was defined by the presence of circulating DSA in plasma, C4d positivity on the biopsy with signs of glomerulitis and capillaritis. Treatment for acute cellular rejection consisted of 3 days of high-dose methylprednisolone (500 mg daily boluses), followed by a prednisone taper. In addition, steroidresistant acute cellular rejections were treated with Thymoglobulin for 4–5 days. In cases of acute AMR the treatment consisted of daily plasmapheresis during five consecutive days, with Thymoglobulin (1 mg/kg) administered at the end of the exchange. After the last session of plasmapheresis the patients received a dose (0.4 g/kg) of i.v. polyclonal immunoglobulin (ivIg). 2.7. Statistical analysis Statistical calculations were performed using R-software [20] and graphics with R-package ggplot 2 [21]. The power of the tests was calculated with software G⁄Power [22]. Continuous data were expressed using mean (standard deviation) or median (interquartile range) and categorical data were described as numbers (frequencies). Categorical data including demographic, clinical, immunological features and pre-transplant anti-HLA antibodies detection were compared using Fisher’s exact test for count data. Continuous variables were compared with Student t-test or Mann-Whitney U test, as appropriate. Graft survival curves were visualized using Kaplan–Meier method, with comparison between DSA+ and DSA patients groups being done by log-rank test [23]. Association between AMR occurrence and MFI strength categories was explored by univariate logistic regression. A multivariable analysis with «backward elimination» was performed to identify the relevants variables for both outcomes – all type of acute rejection and acute AMR –. Receiver operator characteristic (ROC) curves and area under the curve (AUC with the trapezoidal rule) were calculated for DSA MFI values, as well as sensitivity, specificity, positive and negative predictive values [24]. A two-sided p-value of <0.05 was considered as statistically significant.
3
3. Results 3.1. Patient characteristics according to DSA detection Baseline characteristics of the 280 patients (all, divided according to DSA) enrolled in the study are shown in Table 1. Of all the patients, 66.8% were male and 33.2% female, with the percentage of female rising to 58.3% in the DSA positive group. The mean age at transplantation was 54.6 years. The most represented cause of kidney disease was glomerulonephritis (31.1%). Kidney transplants were from a living donor in almost half of cases (49.3%) with this proportion dropping to half (25%) in the DSA positive group of patients. No statistical differences between patients with or without DSA were found regarding age and HLA mismatches (except for double HLA DQ mismatches: 3.9% vs 16.7%, p = 0.023). Cytotoxic PRA-I at transplantation showed less sensitization in DSA negative patients (PRA-I 0% in 93% vs 70.8%, p = 0.002, for DSA and DSA+ groups, respectively) with DSA positive group having more frequently a PRA-I >10% (3.1% vs 16.7%, p = 0.006, for DSA and DSA+ groups, respectively). Pretransplant anti-HLA antibodies were present in 72 patients (25.7%), and 24 (8.6%) had DSA. Patients with anti-HLA antibodies against both classes were more frequently found in the DSA positive group (70.8% vs 6.6%, p < 0.001, for DSA+ and DSA groups, respectively). The antibodies identified were anti-class I DSA in 4 patients, anti-class II DSA in 3 patients, and 17 patients had both anti-class I and anti-class II DSA. Overall 16 patients had one single DSA, four patients had two DSAs, three patients had three DSAs and one patient had four DSAs (mean ± SD: 1.54 ± 0.88; median [IQR]: 1 [1,2]). As already mentioned, there were ten patients with positive FCXM: all of them have been transplanted. Of those ten patients, six had DSAs and four did not. The six DSA positive patients had a median cMFI of 5978 (IQR: 1817–11,030). Two patients had class I DSA, three patients had class II DSA and the last one had DSA of both classes. Four patients had only one single DSA, another had two different DSAs and the last patient had four different DSAs. 3.2. Immunosuppression In 71.1% of patients, induction therapy consisted in basiliximab, while 27.2% received Thymoglobulin (alone or with adjuvants) (Table 2). These proportions were similar in the group without DSA (75.4% basiliximab vs 23% Thymoglobulin), but, as expected, they were different in the DSA positive group (25.0% basiliximab vs 70.8% Thymoglobulin). Long-term maintenance immunosuppression consisted in the majority of cases of tacrolimus (93.6%), but 10 patients received cyclosporine (mainly prescribed because of focal segmental glomerulosclerosis as underlying kidney disease). Analyzing more carefully DSA positive patients (n = 24), no statistically significant differences were found in the choice of induction therapy and the development after transplantation of an acute AMR (Table 3). The incidence of AMR in the group who received basiliximab was 50% vs 22.2% in the group treated with Thymoglobulin alone vs 62.5% in the group who received also ivIg (p = 0.251). DSA strength was not statistically different between patients induced with basiliximab, Thymoglobulin or Thymoglobulin + ivIg (median cMFI (IQR) respectively 3527 (5722) vs 2326 (2814) vs 5145 (7560), p = 0.626 with ANOVA Kruskal-Wallis). 3.3. Correlation of DSA with acute antibody mediated rejection The number and characteristics of acute rejection episodes in the first year after transplantation are shown in Table 2. There were a total of 46 (16.4%) acute rejection episodes (all patients experienced a single episode of acute rejection), 32 (11.4%) being
Please cite this article in press as: I. Salvadé et al., Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation, Hum. Immunol. (2016), http://dx.doi.org/10.1016/j.humimm.2016.04.010
4
I. Salvadé et al. / Human Immunology xxx (2016) xxx–xxx
Table 2 Immunosuppression, acute rejection and survival. All
DSA negative
DSA positive
280
256
24
199 (71.1) 37 (13.2) 31 (11.1)
193 (75.4)
6 (25.0)
<0.001
28 (10.9) 31 (12.1)
9 (37.5) 0 (0.0)
0.001 –
8 (2.9) 5 (1.8)
0 (0.0) 4 (1.6)
8 (33.3) 1 (4.2)
– 0.364
262 (93.6) 10 (3.6) 8 (2.8)
238 (93.0)
24 (100.0)
0.381
10 (3.9) 8 (3.1)
0 (0.0) 0 (0.0)
– –
Rejection at 1 year, n (%) All acute rejection Acute cellular rejection Borderline 1A 1B 2A 2B Acute humoral rejection
46 (16.4) 32 (11.4) 9 (3.2) 15 (5.7) 4 (1.4) 3 (1.1) 1 (0.4) 14 (5.0)
34 (13.3) 30 (11.7) 9 (3.5) 14 (5.5) 3 (1.2) 3 (1.2) 1 (0.4) 4 (1.6)
12 (50.0) 2 (8.3) 0 (0.0) 1 (4.2) 1 (4.2) 0 (0.0) 0 (0.0) 10 (41.7)
<0.001 1.00 – 1.00 0.303 – – <0.001
5-y graft survival, % 5-y patient survival, %
90.8 93.9
92.0 95.5
76.2 74.5
0.248 0.013
Patients, n Induction, n (%) Simulect Thymoglobulin Simulect + Thymoglobulin Thymo + ivIg Other Immunosuppression, n (%) Tacrolimus CsA Other
DSA. = donor-specific = cyclosporine.
antibody;
ivIg. = intravenous
pvalue
immunoglobulin;
CsA.
Table 3 Induction therapy association with acute AMR for DSA positive patients. Induction
Simulect Thymoglobulin Thymoglobulin + ivIg Other
DSA positive (n = 24)
p-value
AMR positive
AMR negative
n
Median cMFI (IQR)
n
Median cMFI (IQR)
3 2 5 0
8725 (4712) 4566 (16) 6779 (10,555)
3 7 3 1
2996 (1304) 1752 (2048) 3510 (3590) 2091
AMR. = antibody-mediated rejection; intravenous immunoglobulin.
DSA. = donor
specific
0.70 0.056 0.57
antibody;
ivIg. =
acute cellular rejections and 14 (5.0%) acute AMR. The incidence of acute AMR was significantly higher in patients with DSA (41.7%) than in those without DSA (1.6%) (p < 0.001). Of the 24 patients with pretransplant DSA, 10 had an acute AMR (DSA+/AMR+) and 14 did not (DSA+/AMR ). The median (IQR) cMFI of the DSA +/AMR+ group was 5680 (2289–9641) vs 2208 (1347–3920) in DSA+/AMR (p = 0.058) (Fig. 1). With univariate logistic regression we could define a threshold value of 5280 cMFI (95% CI: 4205–9235) that was predictive for acute AMR (Fig. 2). Pretransplant DSA cMFI’s ability to predict AMR occurrence was also explored by ROC analysis. AUC was 0.728 (95%CI: 0.474– 0.983) and the best threshold was a cMFI of 4340 with a specificity of 0.928 (95%CI: 0.786–1) and a sensitivity of 0.700 (95%CI: 0.4–0.9); positive predictive value of 0.875 and negative predictive value of 0.812 (Fig. 3). We also analyzed the ability of maximum MFI (the MFI of the ‘‘strongest” DSA present) to predict acute AMR: the discrimination between the two groups DSA+/AMR+ vs DSA+/AMR was weaker and non significant (p-value 0.072) and the ROC analysis showed a lower AUC of 0.690 (95%CI: 0.444–0.867). The median time from transplant to acute AMR in the group DSA+/AMR+ (n = 10) was 13 days (IQR 132 days, min 5 days and max 217 days). Surprisingly, patients with low cMFI had a shorter time to acute AMR than patient with high cMFI (adjusted R-squared: 0.6541, p-value 0.003). The odds-ratio and relative risk of developing an acute AMR in the first year according to cMFI values is shown in Table 4. With multivariable logistic regression and backward elimination of all explicative variables analyzed in our study (age at transplantation, sex, type of donation – cadaveric or living –, class of anti-HLA antibodies, current PRA-I, class of DSA, number of DSA, DSA strength assessed by cMFI, number of HLA mismatches and induction therapy), only the age (p = 0.006) and cMFI (p < 0.001, AIC coefficient 225.61) were significantly associated with acute rejection (all types). Regarding acute AMR, only the value of cMFI was found to have a significant role (p = 0.004, AIC coefficient 69.75, OR 1.97 for each cMFI increment of 1000 (95%CI: 1.46–2.67). The number or anti-HLA class of pretransplant DSA did not. We also analyzed the evolution with time of cMFI values after transplant and we found that the cMFI of DSA detected pretransplant showed a significant decrease (median cMFI (IQR): 2209 (2571) pre-transplant vs 0 (967) at 1 year, p-value 0.005) post transplant, in patients DSA+/AMR (n = 14). However, for patients DSA+/AMR+ (n = 10), the cMFI persisted during the first year post transplant (median cMFI
Fig. 1. Box plot of DSA cMFI for patients with an antibody-mediated rejection (AMR+, n = 10) and without (AMR , n = 14).
Please cite this article in press as: I. Salvadé et al., Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation, Hum. Immunol. (2016), http://dx.doi.org/10.1016/j.humimm.2016.04.010
5
I. Salvadé et al. / Human Immunology xxx (2016) xxx–xxx
Fig. 2. Univariate logistic regression. Probability of experiencing an antibody-mediated rejection (AMR) in function of cumulative mean fluorescence intensity (cMFI) of donor-specific anti-HLA antibodies (DSA). Solid line represent logistic regression curve with grey zone being the CI95%. The cMFI value predictive of AMR (probability > 0.5) is 5280.
Fig. 3. Receiver operator curves of the cMFI for prediction of acute AMR. Area under the curve is 0.728 and the best threshold was a cMFI of 4340 with a specificity of 0.93 and a sensitivity of 0.70.
(IQR): 5680 (7351) pre-transplant vs 11,557 (15,905) at 1 year). Two patients in this group (DSA+/AMR+, n = 10) also developed de-novo DSA at the time of acute AMR. 3.4. Kidney and patient survival Patient survival at 5 years was worse for those with pretransplant DSA+ (74.5%) compared to the DSA group (95.5%; p = 0.013) (Fig. 4). The causes of death within the analyzed population were: 1 sudden death after hemodialysis, 1 hypoglycemia, 1 suicide (patient without DSA), four infections (septic chock or abdominal infection) and four neoplasms. There was also a trend for worse 5-year death-censored graft survival in the DSA+ group
Fig. 4. Survival of patients with donor-specific anti-HLA antibodies (DSA+) and without donor-specific anti-HLA antibodies (DSA ).
(76.2% vs 92.0%, for DSA+ and DSA , respectively), although this did not reach statistical significance (p = 0.248). However, by stratifying and analyzing the patients with the threshold of 5280 cMFI (value derived from the univariate logistic regression), there was a significantly lower graft survival in the DSA+ group with >5280 cMFI vs the other patients (53.3% vs 92.1%; p = 0.025) (Fig. 5). For this analysis, the patients without DSA and those with lowerlevels cMFI DSA (<5280 cMFI) were grouped together because their 5 year graft survival was similar (92.0% vs 94.4%; p-value = 0.950). If we analyze the graft survival of patients with DSAs according to development of acute AMR, we can observe a trend to a deleterious effect of experiencing an acute AMR: 5-year death-censored graft survival for patients DSA+/AMR+ was 57.8% vs 100% for those DSA+/AMR (p = 0.089).
Table 4 Odds ratio and relative risk.
cMFI 500–5280 cMFI P 5280
AMR+
AMR
OR
95% CI
RR
95% CI
p-value
5 5
13 1
11.50
0.96–657.53
3.00
1.31–6.86
0.050
AMR. = antibody-mediated rejection; cMFI. = cumulative mean fluorescence intensity; DSA. = donor-specific antibody; OR. = odds ratio; RR. = relative risk.
Please cite this article in press as: I. Salvadé et al., Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation, Hum. Immunol. (2016), http://dx.doi.org/10.1016/j.humimm.2016.04.010
6
I. Salvadé et al. / Human Immunology xxx (2016) xxx–xxx
Fig. 5. Death-censored kidney graft survival of patients with donor-specific antibody more than 5280 cumulative mean fluorescence intensity (cMFI) and less than 5280 cMFI or without.
4. Discussion In the present study, we have investigated the importance of DSA as a predictive biomarker for antibody-mediated rejection in kidney transplantation. We have hypothesized that pre-transplant DSA levels may serve as predictive biomarker for AMR. The investigation was performed on a cohort of 280 patients receiving kidney transplantation at our institution. The results obtained indicated that, in the transplant recipients with a negative CDC crossmatch at the time of transplantation, the presence of preformed DSA was associated with a significantly higher incidence of biopsyproven acute AMR (41.7%) in the first year after transplantation. The proportion of patients with acute AMR in our study is consistent with previous studies showing that the incidence of AMR ranged between 30% and 50%, depending in part on whether biopsies were systematically performed in all patients or not [2,4,5,16]. The detection of DSA was performed using a quantitative Luminex-based assay. Interestingly, the results obtained indicate that the group of patients with preformed DSA experiencing an acute AMR had higher amounts of circulating DSA (as measured by cMFI on pretransplant sera), as compared to those with preformed DSA but without AMR. The rationale of using cMFI rather than maximum (peak) MFI was based on the assumption that cMFI values may better reflect the degree of HLA incompatibility of the recipient versus the graft at the time of transplantation, and because of their possible better correlation with AMR. Multivariate analysis was performed on a panel of parameters including age, sex, type of donation, class of anti-HLA antibodies, current PRA-I, class of DSA, number of DSA, number of mismatches and induction therapy and the cMFI. The cMFI level was the only and single parameter which correlated with the incidence of acute AMR. We then performed logistic regression and ROC curve analysis in order to further evaluate the role of the levels of cMFI DSA as predictive biomarker of AMR. We showed that levels of cMFI being predictive of AMR ranged between 4300 and 5300, with good sensitivity and specificity (sensitivity 0.700 and specificity 0.928, with negative predictive value of 0.812). In support of these findings, a recent study from our group showed that DSA cMFI levels above 5000 were highly predictive of a positive FCXM [17]. Previous studies have already assessed the relationship between DSA strength and AMR with conflicting results [2,4,8,13,16,25]. The reasons for these discrepancies remain unclear, although the differences in the expression of MFI values (i.e. threshold of positivity, cumulative vs maximum) [1] have rendered somewhat difficult to compare the results from different studies.
Nevertheless, taken together, these results suggested that pretransplant cMFI DSA levels may serve as a predictive biomarker for acute AMR. We then investigated the impact of cMFI DSA levels on graft survival and overall patient survival. With regard to graft survival, we found that the association of DSA with AMR influenced negatively graft survival (57.8% for DSA+/AMR+), while in the absence of acute AMR graft survival (100%) was similar to that observed in non-immunized DSA negative patients (92.0%). The presence of pretransplant DSA also had a negative impact on patient survival. Indeed the 5-year patient survival was 95.5% in recipients without DSA vs 74.5% in recipients with DSA (p = 0.013). Whether this finding may reflect a higher incidence of comorbidities, or posttransplant infections and/or tumors, possibly related to more aggressive immunosuppressive regimens, as reported in a previous study [13], remains to be determined. The finding of the importance of acute AMR and/or cMFI DSA levels on graft survival led us to analyze whether other features of DSA or induction therapy may have a role on the incidence of AMR. As also reported by others [2,7], we found that neither the class nor the number of DSA played a role. With regard to immunosuppression, the incidence of acute AMR in the group of patients with DSA was not significantly influenced by the type of induction therapy (basiliximab vs Thymoglobulin vs Thymoglobulin + ivIg). Our results support the data from Amico et al. in a cohort of 67 kidney transplant recipients with preformed DSA, where no differences in AMR incidence were found in patients without induction vs patients who received basiliximab or daclizumab [2]. The present study has a number of limitations. It is a singlecenter study, the results on FCXM were not available in all patients and HLA typing did not include Cw and DP specificity. Finally there are known limitations regarding the use of MFI as a quantitative tool for circulating DSA [26,27]. In conclusion, the results from the current study indicate that preformed DSA cMFI values are clinically relevant to predict acute AMR and graft survival after kidney transplantation. A threshold of 4300–5300 cMFI appears to be a key determinant outcome predictor. Altogether our data support the detrimental role of preformed DSA in kidney transplantation.
Disclosure The authors of this manuscript have no conflicts of interest to disclose. This is an original work not submitted for publication elsewhere. This work was presented as an oral presentation at the ESOT meeting in Brussels (September 2015) by Dr. I. Salvadé.
Authorship IS and MP were responsible for concept, analytic design, data analysis, interpretation of results, and manuscript writing. GP helped with the interpretation of results, and critically reviewed the manuscript. VA, JPV, and DG helped with the data collection, gave thoughtful comments, and reviewed the manuscript. ACS, MM and SR helped with the data collection. All authors critically reviewed and approved the final version of the manuscript.
Funding No funding was required for this study.
Please cite this article in press as: I. Salvadé et al., Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation, Hum. Immunol. (2016), http://dx.doi.org/10.1016/j.humimm.2016.04.010
I. Salvadé et al. / Human Immunology xxx (2016) xxx–xxx
Aknowledgements We thank Stéphanie Gregoretti and Rocco Sugamele for the excellent laboratory assistance, and Prof. Giorgio Salvadé for his valuable assistance in the data analysis. We also thank all the clinicians and nurses, datamanagers, who were involved in patient management and data collection. References [1] S. Mohan, A. Palanisamy, D. Tsapepas, et al., Donor specific antibodies adversely affect kidney allograft outcomes, J. Am. Soc. Nephrol. 23 (12) (2012) 2061–2071. [2] P. Amico, G. Honger, M. Mayr, et al., Clinical relevance of pretransplant donorspecific HLA antibodies detected by single-antigen flow-beads, Transplantation 87 (11) (2009) 1681–1688. [3] J.L. Caro-Oleas, M.F. Gonzalez-Escribano, F.M. Gonzalez-Roncero, et al., Clinical relevance of HLA donor-specific antibodies detected by single antigen assay in kidney transplantation, Nephrol. Dial. Transplant. 27 (3) (2012) 1231–1238. [4] C. Lefaucheur, A. Loupy, G.S. Hill, et al., Preexisting donor-specific HLA antibodies predict outcome in kidney transplantation, J. Am. Soc. Nephrol. 21 (8) (2010) 1398–1406. [5] C. Lefaucheur, C. Suberbielle-Boissel, G.S. Hill, et al., Clinical relevance of preformed HLA donor-specific antibodies in kidney transplantation, Am. J. Transplant. 8 (2) (2008) 324–331. [6] H.G. Otten, M.C. Verhaar, H.P. Borst, R.J. Hené, A.D. van Zuilen, Pretransplant donor-specific HLA class-I and -II antibodies are associated with an increased risk for kidney graft failure, Am. J. Transplant. 12 (6) (2012) 1618–1623. [7] D.S. Tsapepas, R. Vasilescu, B. Tanriover, et al., Preformed donor-specific antibodies and risk of antibody-mediated rejection in repeat renal transplantation, Transplantation 97 (6) (2014) 642–647. [8] M. Willicombe, P. Brookes, E. Santos-Nunez, et al., Outcome of patients with preformed donor-specific antibodies following alemtuzumab induction and tacrolimus monotherapy, Am. J. Transplant. 11 (3) (2011) 470–477. [9] V. Aubert, J.P. Venetz, G. Pantaleo, M. Pascual, Low levels of human leukocyte antigen donor-specific antibodies detected by solid phase assay before transplantation are frequently clinically irrelevant, Hum. Immunol. 70 (8) (2009) 580–583. [10] G. Hönger, M. Wahrmann, P. Amico, H. Hopfer, G.A. Böhmig, S. Schaub, C4dfixing capability of low-level donor-specific HLA antibodies is not predictive for early antibody-mediated rejection, Transplantation 89 (2010) 1471–1475. [11] G. Hönger, H. Hopfer, M.L. Arnold, B.M. Spriewald, S. Schaub, P. Amico, Pretransplant IgG subclasses of donor-specific human leukocyte antigen antibodies and development of antibody-mediated rejection, Transplantation 92 (2011) 41–47.
7
[12] A. Loupy, C. Lefaucheur, D. Vernerey, et al., Complement-binding anti-HLA antibodies and kidney-allograft survival, N. Engl. J. Med. 369 (2013) 1215– 1226. [13] J.D. Smith, M.W. Ibrahim, H. Newel, et al., Pre-transplant donor HLAspecific antibodies: characteristics causing detrimental effects on survival after lung transplantation, J. Heart Lung Transplant. 33 (2014) 1074–1082. [14] N. Khovanova, S. Daga, T. Shaikhina, et al., Subclass analysis of donor HLAspecific IgG in antibody-incompatible renal transplantation reveals a significant association of IgG4 with rejection and graft failure, Transpl. Int. 28 (2015) 1405–1415, http://dx.doi.org/10.1111/tri.12684 (Epub 1 Sept). [15] C. Lawrence, M. Willicombre, P.A. Brookes, et al., Preformed complementactivating low-level donor-specific antibody predicts early antibody-mediated rejection in renal allografts, Transplantation 95 (2) (2013) 341–346. [16] J. Malheiro, S. Tafulo, L. Dias, et al., Analysis of preformed donor-specific antiHLA antibodies characteristics for prediction of antibody-mediated rejection in kidney transplantation, Transpl. Immunol. 32 (2) (2015) 66–71. [17] V. Aubert, R. Sugamele, S. Gregoretti, J.-P. Venetz, G. Pantaleo, M. Pascual, Are donor specific anti-HLA antibodies detected by Luminex predictive of positive flow cytometric crossmatch?, Tissue Antigens 85 (2015) 41 [18] R.R. Hackem, N.L. Reinsmoen, What is the definition of a clinically relevant donor HLA-specific antibody (DSA)?, Am J. Transplant. 15 (2015) 299–300. [19] M. Haas, B. Sis, L.C. Racusen, et al., Banff 2013 meeting report: inclusion of C4d-negative antibody-mediated rejection and antibody-associated arterial lesions, Am. J. Transplant. 14 (2014) 272–283. [20] R Core Team, R: A Language and Environment for Statistical Computing, R Foundation for Statistical Computing, Vienna, Austria, 2014. URL:
Please cite this article in press as: I. Salvadé et al., Clinically-relevant threshold of preformed donor-specific anti-HLA antibodies in kidney transplantation, Hum. Immunol. (2016), http://dx.doi.org/10.1016/j.humimm.2016.04.010