Acute antibody mediated rejection (AAMR) is the principal immunological risk factor for development of chronic graft dysfunction and predicts chronic allograft nephropathy, the latter being the main cause of graft loss during the first post-transplant year [1-8].
Advances in the definition of AAMR, in the evolution of plasma replacement techniques and knowledge of transplant immunology  have not been accompanied by efficacious therapeutic strategies. Following current treatment guidelines, up to 80% of cases achieve a response, however it is still poor in episodes occurring 3-6 months after transplant (late AAMR) [3,4,9]. On the other hand, the slow reversion of clinical symptoms due to delayed therapeutic effect and the long-term persistence of donor specific antibodies (DSA), are factors that significantly impede the improvement of results in AAMR .
The current standard treatment consists of a combination of plasma replacement therapy (PRT), intravenous immunoglobulin (IVIg) and pulses of high dose corticoids [2,7,9,11]. The benefit of PRT in this context has been widely demonstrated in various randomized controlled clinical trials  and its use has been recommended by various scientific societies and clinical practice guidelines [2,7,11].
New therapeutic possibilities include agents directed against B lymphocytes (rituximab and alemtuzumab), plasma cells (bortezomib) and complement (eculizumab) which achieve satisfactory results [2,9]. However, routine use of these drugs is limited by their high cost and scarcity of prospective studies evaluating their efficacy and safety [5,12,13].
This study aims to analyze the response of episodes of AAMR treated with PRT, in conjunction with other therapies, and evaluate the safety of the technique.
Patients and Methods
A retrospective review of a prospective database, of all cases of AAMR after ABO compatible renal transplant (ABOcRTx), that were treated with PRT combined with IVIg and rituximab or bortezomib, in our center over 5 consecutive years (2008-2012). All kidney transplants achieved at our center have negative complement-dependent microlymphocytotoxicity crossmatch and absence of DSAs prior to transplant. This study was approved by the Local (Aragon) Clinical Research Ethics Committee according to Declaration of Helsinki (Minutes n 04/2017).
Diagnosis of AAMR
All cases were diagnosed according to Banff 2009  nomenclature, which was the last edition published, at that time. Cases without DSA, absence of histological specific findings and/ or low intensity C4d deposits (<2) were designated as suspected. Serological studies for detecting and monitoring circulating donor specific anti-HLA antibodies during the post-transplant period were performed using the Luminex solid-phase technique (One Lambda, Canoga Park, CA, USA). Until 2011, only class I single-antigen technique was performed (7 patients). The cut-off point to consider positive mean intensity of fluorescence (MFI) for DSAs class I was 1500 and 3500 for DSAs class II.
AAMR treatments were carried out according to the protocol established in our center. All patients signed an informed consent form, and received one cycle of PRT (6 sessions) and three (250 mg to 500 mg) boluses of methylprednisolone (MTP) with a subsequent rapid decrease to reach the basal dose of prednisone. Routine immunosuppression regimen was either switched or intensified. In some individual cases, a new PRT cycle was programmed if plasma creatinine clearance was lower than 20- 30% of the basal rate. Two protocols were used according to the date of the diagnosis. Between 2008 and 2010, patients received a dose of 200mg/Kg of IVIg (Flebogamma, Grifols, Barcelona, Spain) after each replacement (total dose of 1.2g/kg). And from 2011 onwards, a dose of 100mg/Kg of IVIg was given at the end of each replacement for the first 5 sessions, and a final dose of 500mg/Kg of IVIg after the last replacement (total dose of 1g/ kg). Since 2011, patients receive one dose (375 mg/m2) of rituximab (RTX) "off label" (Roche, Welwyn Garden City, GB) prior to initiating or after terminating the cycle (6 sessions) of PRT. Rituximab was administered at least 24 hours before a PRT session, and also, as far as possible from the IVIg.
Plasma replacement therapy
When AAMR is suspected, patients receive a cycle of 6 PRT on alternate days which are carried out using continuous flow cell separators (Cobe Spectra or Optia Spectra, Barcelona, Spain). The process involves removing and processing 1 to 1.2 times the patient's plasma volume. The standard replacement solution is 5% human albumin (Grifols, Barcelona, Spain). To avoid hypocalcemia symptoms, intravenous calcium gluconate is administered as prophylaxis at the beginning of each session and is repeated if symptoms appear. In accordance with the local protocol to prevent dilutional coagulopathy, all patients receive intravenous vitamin K at the end of the procedure.
Definition of variables
Analytical variables were classified following recommendations made by the International Committee for Standardization in Hematology . Response was evaluated 30 days after finishing the PRT cycle and was defined as dialysis independence. Other definitions were, acute "mixed" AAMR: histological findings of AAMR associated to other histological lesions, early rejection: the one established the first 6 months after the ABOcRTx and Late rejection: later than 6 months after ABOcRTx. Follow-up in the long-term was calculated from the date of the start of PRT to the date of last control or death and was analyzed in the 16 end AAMR patients.
The bivariate analysis was performed using Student's t-test, to analyze relationships between quantitative variables and the categorical variables. When the assumptions of normal distributions of t-test were not met, we applied the nonparametric Mann-Whitney U test. Association between two categorical variables were determined by chi-square test, yates' continuity correction or Fisher's exact test were applied, if necessary. Median survival of patients and renal graft was performed using the Kaplan-Meier test. The level of confidence selected was 95%, in all the tests performed, P values less than or equal to 0.05 were considered statistically significant. All statistical analyses were performed using PASW Statistic version 18 software (SPSS Inc., Chicago, Illinois, USA).
During the study period, we have performed 322 kidney transplants. Prevalence of AAMR in our center cohort was 6.5%. Twenty-one patients with AAMR were treated with PRT, but 5 were excluded because of incomplete data. From 16 end patients, 4 were suspected cases. Episodes of AAMR developed 7.4 (0.23-147.8) months after ABOcRTx. Table 1 shows the initial characteristics of all the patients included in this study. The most frequent clinical presentation in the series was oliguria/anuria (68.8%), microscopic hematuria (62.5%) and hypertension (50%). Parameters prior to initiating PRT were: maximum serum creatinine 4.24 [1.05-7.11] mg/dL, simple urine sample proteinuria 0.7 [0.0-21] g/dL, and simple urine protein/ creatinine ratio 7.7 (0-270) mg/g. Twelve patients presented anemia, 2 thrombocytopenia and 6 hypoalbuminemia.
Renal biopsies demonstrated histological signs associated with borderline rejection, acute cellular rejection (ACR), and transplant glomerulopathy, in 1, 3 and 3 patients, respectively. Immunological analyses detected DSA in 6 cases, of these, 2 presented class I DSA, one presented class II DSA, and the remaining 3 presented both, 5 patients developed anti-MICA antibodies. The median DSA MFI for Class I was 4332.8 (2758.2-5509), and 13504.4 for Class II (4513.1-17446.9). The most relevant renal biopsy findings prior to initiating PRT are shown in Figure 1.
Due to earlier acute tubular necrosis with oliguria/anuria, one patient had initiated dialysis three days prior to developing AAMR. All the patients were receiving treatment with at least two immunosuppressants, the most frequent combination being (n=12) mycophenolate, tacrolimus [mean levels in plasma prior to PRT: 6.9 (2.2-12.3) ng/mL] and prednisone. Six patients required dialysis which was performed every-other-day alternating with PRT.
RTX was not administered in 5 patients, 2 of them before 2011, in one case due to clinical fragility, in another case because of relevant comorbidities Charlson Comorbidity index (CCI) of 6.6 and in the last case, due to stabilization of renal function after intensification of immunosuppression, IVIg and PRT. In 2 cases where treatment failed, bortezomib "off label" (Janssen-Cilag, Madrid, Spain) was added, receiving 4 doses of 1.3 mg/m2 per week. Both patients had 2 prior renal transplants, and one of them, achieved a response.
A total of 11 (68.75%) patients achieved better response (3 corresponded to suspected cases). Clinical improvements were observed with decreases in oliguria/anuria and hematuria rates to 56.3% and 50%, respectively. Disappearance of class I DSA was observed in one case, however all other DSA remained unchanged. Although, median DSA MFI descended in both class I (3062.8 [1303-8485]) and class II (11304.5 [9157-12599.5]), DSA levels post-treatment was not associated with response. In twelve patients, maximum serum creatinine levels decreased with a mean reduction of 35.78% (value of this parameter after treatment was 3.68 [0.84-6.61] mg/dL).
Analysis by subgroups
According to the histological findings, patients with better responses (Table 2) were those that developed mixed AAMR episodes (3 ACR, 3 with transplant glomerulopathy and 1 borderline). When ACR and borderline (4/16) patients were analyzed together, the response rate was 100% compared to 58.3% of the remaining patients.
Overall, 66.7% patients with confirmed AAMR achieved a response versus 75% of suspected cases. This small difference could be due to the distribution of diverse variables in the two groups, despite the fact that there were no differences between them with respect to the treatment applied (Table 2).
The presence of circulating anti-MICA antibodies (Ab), either alone or in association with other anti-HLA antibodies, worsens the response rates (75% in cases with only class II Ab, 60% in cases with only class I Ab, 40% in cases with only MICA Ab and 0% in patients with MICA Ab in association with class I or class II Ab). With regard to the time up to the start of PRT, better results were observed in patients who initiated PRT during the first 14 days after AAMR diagnosis. The response rates were, 75%, 100% and 50% in patients who initiated sessions during the first 6 days, between 7 and 13 days, and 14 days after the AAMR episode, respectively.
Patients who received combined treatment with RTX and MTP were more likely to achieve a response than those who did not (p=0.034 [RR: 2.25, CI 95 (1.08-4.67)]). But this is not the case of patients treated with only PRT and RXT (Table 3).
Finally, patients with a greater comorbidity (Charlson Index ≥ 4) demonstrated a certain tendency towards worse results (p=0.077).
A median of 6 (5-12) sessions were performed per patient and were initiated 15 (4-36) days after being diagnosed with AAMR. Four patients required more than 6 sessions (11 or 12). PRT sessions were performed through a central venous catheter, Hickman type in 7 cases, arteriovenous fistula (AVF) in 6, and Shaldon type in 3. Median exchanged volume was 3075.5 (2098-3683) mL. Out of the 118 sessions carried out, 3 (2.5%) presented catheter related complications (low access pressure, superficial thrombosis and phlebitis) and 5 (4.2%) clinical complications (2 patients with hypotension and dizziness, 1 with perioral paresthesia, 1 with a ruptured vein and another bleeding through the AVF).
Patient and graft survival
Altogether, 81.3% of patients were still alive after a median follow-up of 4.7 (0.3-8.06) years. Two patients died due to respiratory infection and there was one sudden death in a patient with advanced age and a high comorbidity index. The median time between AAMR and death was 24.3 (4.37-53.1) months. At the end of follow-up period, the renal graft survival rate was 50%, and censored graft survival was 56% (one deceased patient with functioning renal graft). Seven patients recommenced dialysis, of which 2 underwent a second transplant 27 (22.6 and 31.5) months after terminating PRT. And neither of the latter patients had suffered graft rejection. Despite the fact that accumulated survival rate, at the end of the follow-up period was 60% in patients who failed to achieve response versus 90.9% in patients who did respond, this difference was not found to be statistically significant by Kaplan-Meier test (Log Rank: 0.172) (Figure 2).
Taking into account the inherent limitations of a retrospective, single center (a reference center for apheresis techniques and renal transplant) study, with a low number of subjects, this paper includes characteristics that have rarely appeared in the literature, such as analytic parameters, associated comorbidities and stratified clinical variables which may become factors that predict response . Graft survivals curves show marked variations in the various published articles. In the short-term, there are reports of between 62% and 80% graft survival 12 months after an episode of AAMR [1,4,7,8,16]. However, the long-term results are clearly inferior, with graft survival between 55% and 70% [17,18], similar to that observed in our series.
Eight patients initiated PRT more than 14 days after diagnosis and, with respect to the rest of the series, presented lower creatinine levels (3.4 vs 4.6 mg/dL), greater age (45.5 vs 38) and higher CCI (63% with ICC>4); these factors could probably influence a later start of PRT.
Due to their lack of response during the first PRT cycle, 4 patients received a second cycle of treatment. Despite the fact that all these cases demonstrated a favorable response one month after terminating the second cycle of PRT, in the medium-term (4 to 8 months) 50% of the patients suffered graft loss and had to reinitiate dialysis . The majority of patients who achieved a response were AAMR confirmed cases (8/11). Nevertheless, when comparing the results with suspected cases, the response rate of the latter group was higher. Suspected cases had some factors in their favor, 75% were female, 3 corresponded to mixed AAMR, 100% were early AAMR, and the group was younger (37.9 [28.5-48.5] vs 45.2 [33.2-75.5] years).
The superiority of the results in patients with early versus those with late episodes (Table 2) coincides with other
published studies [4,8,9,20]. In relation to cases of mixed AAMR, they presented lower creatinine levels (3 vs 5 mg/dL), less comorbidities (CCI >4: 28.6% vs 55.6%) and a greater proportion were treated with RTX and MTP (57.1% vs 33.3%) than the rest of the series. These factors could have contributed to obtaining better results.
Patients with great comorbidity, could had worse outcomes because only one of these 7 cases was treated with RTX and MTP, they were older (52.1[28.5-75.5] vs. 38.7[33.2-48.5] years), and had greater creatinine values (4.43[2.7-7.1] vs. 3.1[1.1-7.1] mg/dL). Rationale for the use of IVIg after each PRT sesion, was based on clinical trials (24) and clinical practice guidelines [2,11] that have demonstrated improvement of graft survival with PRT plus IVIg versus PRT alone or IVIg alone.
Considering the patients were heterogeneously treated, we separately analyzed general results for patients treated with RTX and RTX plus MTP vs patients that had not receive them. The only variable that was clearly related with improved responses was combined treatment with RTX and MTP (Table 3), a fact which confirmed the already recognized synergy of these drugs [2,9]. With respect to the failure to detect DSA at the time of the AAMR episode, this could be justified due to the lack of anti-class II antibody studies in 7 of the patients in the series. Nevertheless, this could also be due to the fact that, in some cases, a certain delay occurred in the appearance of DSA which are detected some months after the AAMR episode. This phenomenon could possibly be due to initial "absorption" of the DSA by the graft itself or a post-transfusional origin.
We can speculate that anti-MICA antibodies worsens the response rates. However, the low number of patients included in our series, prevent us from making definitive conclusions, nevertheless similar findings have previously been published [21-23].
The rate of adverse events (technical and clinical) related to PRT were low in agreement with data published by other authors [1,18] and none of the cases were severe. Regarding to low access pressure, sometimes it solves rechanneling the needle or washing the venous access device, and occasionally with posture changes. If the problem persists, catheter thrombosis should be ruled out. To treat superficial thrombosis and phlebitis we use nonsteroidal anti-inflammatory drugs, warm compresses and compressive bandage. Also, catheter is removed.
In case of hypotension and dizziness, the patient is put into the Trendelenburg position, and the flow speed through the system is reduced. It could be due to the hemodynamic changes of the procedure, or the release of adrenaline or other catecholamines related to illness or procedure. If hypotension persists, hypovolemia must be excluded. Perioral paresthesias are most likely due to citrate anticoagulation calcium chelation. In such cases, we reduce rate of infusion and administer oral or intravenous calcium supplementation, but if symptoms are severe, unless there is a contraindication, citrate can be changed by heparin . It is worth mentioning that patients in this series presented a lower rate of technique-related adverse events compared to a previous series of neurological pathology patients  also studied in our center (technical: 3.4%, clinical: 8.5%). This could possibly be due to better tolerance of blood volume changes in the patients having undergone dialysis. But more catheterrelated complications, compared to patients treated in our center with acute renal failure secondary to vasculitis (1.04%) . With the aim of improving our results, and taking into account the kinetics of eliminating any intravascular substance from plasma, as well as the decreasing efficacy of replacements; we could theorize that insufficient plasma was removed during replacements and it is necessary to evaluate the viability of performing larger plasma replacements (1.5 to 2 volumes) if the patient's condition allows.
Finally, we stress a need to create multidisciplinary standardized treatment protocols to define the role of each drug (and their doses), as well as their pharmacokinetics with concomitant PRT (volumes, frequency, and the moment when immunosuppressants should be administered) in cases of AAMR.
Based on our results, we can conclude that a possible predictor of good response to treatment, is the combination therapy with PRT, IVIg, RTX and MTP. Taking into account the low rate and mildness of the complications associated to the apheresis technique, we can conclude that, in our center, PRT is a safe therapeutic procedure probably regarding our prophylactic measures (intravenous calcium gluconate administration at the beginning of the each session and intravenous vitamin K administration at the end of the procedure). It is probable that the lack of statistically significant differences between responding and non-responding patients was due to the low number of cases in our study.
We would like to express our gratitude to all the personnel in the transfusion and apheresis units of the Hematology and Hemotherapy service of Miguel Servet University Hospital, as well as Nephrology service. We would also like to thank the Spanish Blood Transfusion Society - SETS and Martin Hadley-Adams for translating the manuscript.
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of Interests
The authors declare that there is not conflict of interest.