Is ABO-Incompatible Living Donor Liver Transplantation Really a Good Alternative for Pediatric Recipients?

Background: ABO-incompatible (ABOi) living donor liver transplantation (LDLT) has been proposed to compensate for donor shortage. To date, few studies have reported detailed ABOi LDLT results in large series of pediatric patients. C4d complement deposition in graft capillaries has been reported to be associated with antibody-mediated rejection in solid organ transplantation. Methods: A retrospective case–control study was conducted, comparing clinical outcomes of each of 34 consecutive pediatric ABOi LDLT recipients with those of 2 non-ABOi pairs (n = 68), matched according to pre-transplant diagnostic criteria, age, and date of transplantation. In addition, we studied the C4d immunostaining pattern in 22 ABOi and in 36 non-ABOi recipients whose liver biopsy was performed within the first 4 post-transplant weeks for suspected acute rejection. Results: The incidence of biliary complications was higher in ABOi recipients (p < 0.05), as were the incidence of acute humoral rejection (p < 0.01) and the incidence of retransplantation (p < 0.05). All children who required retransplantation were older than 1 year at the time of ABOi LDLT. Positive C4d immunostaining was observed in 13/22 (59%) ABOi recipients versus 3/36 (8.3%) non-ABOi recipients (p < 0.0001). Conclusions: ABOi LDLT is a feasible option for pediatric end-stage liver disease but carries increased risks for the recipient, especially for children older than 1 year, even with a specific preparation protocol. C4d immunostaining may be a hallmark of acute humoral rejection in ABOi liver transplantation.


Introduction
Orthotopic liver transplantation (LT) nowadays constitutes a validated treatment for acute liver failure, end-stage liver disease, liver tumors, and selected metabolic disorders in children. The high mortality on the waiting list, due to the shortage of size-matched postmortem grafts, has led to the development of alternative techniques. Thomas Starzl described the liver as "a privileged organ" as it resists acute rejection in experimental animal studies [1]. Starzl was therefore the first to describe 11 cases of ABO-incompatible (ABOi) pediatric LT in 1979 because of the difficulty in finding compatible small grafts. He did not describe any case of acute rejection in these patients [2].
In the 1980s and in the early 1990s, several groups subsequently published their results concerning ABOi LT but with poor outcomes. The main encountered complications with ABOi LT were hepatic artery thrombosis, severe rejection, and increased incidence of biliary complications, including diffuse intrahepatic bile duct injuries [3,4]. In ABOi LT, acute humoral rejection (AHR) triggered by antibodies against donor-type iso-agglutinins is the most serious form of rejection and is often associated with graft loss. C4d complement deposition in graft capillaries has been reported to be associated with antibody-mediated rejection in kidney and other solid organ transplantations [5,6].
To overcome complications related to ABOi LT, several protocols were proposed in the 1990s, including high-dose immunosuppression, splenectomy, and plasmapheresis, but they had little effect on poor outcomes and increased the occurrence of specific complications, including infections and sepsis [7,8]. The impact of preformed ABO antibodies against the donor and specific strategies to reduce their titers were then described by several centers in the 1990s as a good strategy to improve ABOi LT outcomes [8][9][10][11]. Subsequently, the introduction of anti-CD20 monoclonal antibody (rituximab) led to a significant improvement in the survival rates of ABOi pediatric and adult living donor liver transplantation (LDLT) [12][13][14][15][16].
The objective of this study was to retrospectively review a single pediatric LT center's 14-year experience with ABOi pediatric LDLT to determine whether outcomes with ABOi donors are truly comparable to ABO-compatible (ABOc) LDLT. Moreover, the impact of recipient age at the time of ABOi LT on post-transplant complications was investigated as it was hypothesized that children younger than 1 year are immunotolerant. It was also hypothesized that positive C4d immunostaining is more frequently observed in ABOi recipients' biopsies taken within the immediate post-transplant period and could be a hallmark of underestimated AHR in ABOi pediatric LDLT.

Basic Immunosuppressive Protocol
Primary baseline immunosuppression for all LT recipients consisted of a bi-therapy including a calcineurin inhibitor (tacrolimus; Prograft ® ; Astellas, Tokyo, Japan) and a monoclonal anti-CD 25 antibody (basiliximab; Simulect ® ; Novartis, Bâle, Switzerland). We applied a basic immunosuppressive protocol without steroids [22]. Basiliximab was administered for intravenous induction on day 0 and day 4 after LT, the dose being adapted based on the recipient's body weight (<15 kg, 5 mg; 15-25 kg, 10 mg; >25 kg, 20 mg). Patients received oral tacrolimus from the day of transplantation, with an initial dose of 0.1 mg/kg/12 h. Target tacrolimus levels varied according to each patient's clinical condition but were generally 8-10 ng/mL during the first 2 months after LT, 6-8 ng/mL during the third month after transplantation, and 4-6 ng/mL between 3 months and 1 year post-transplantation. When acute cellular rejection (ACR) was suspected on the basis of abnormal liver function tests and/or on clinical parameters (unexplained fever, severe unexplained ascites, etc.), a liver biopsy was performed. Proven ACR on liver biopsy (according to BANFF classification [22]) was treated with high-dose methylprednisolone (Solumedrol ® , Medrol ® ; Pfizer, New York, NY, USA): 5 mg/kg for 3 days, 3.75/kg mg for 1 day, 2.5 mg/kg for 1 day, and 1.25 mg/kg for 1 day to maintain a dose of 1 mg/kg for 1 week. Depending on clinical and biological evolutions, patients were switched to oral steroid therapy, and steroid dosage was subsequently tapered to reach 0.25 mg/kg/day at 3 months post-LT and 0.5 mg/kg/2 days until the systematic post-LT biopsy at 6 months.

Immunosuppressive Protocol for ABOi Pediatric LDLT
We performed a total of 34 ABOi pediatric LDLTs, in which 18 patients (52.9%) were less than 1 year old at the time of transplantation. Anti-donor blood group antibodies were systematically monitored in all candidate ABOi recipients during the pre-transplant workup. Natural anti-ABO titers were measured using a simple serial test tube dilution method. For the immune part of the anti-A and/or anti-B antibodies, a column agglutination technique was used after hydrolysis of the patients' plasma with dithiothreitol. If natural and/or immune iso-agglutinin titers were ≥1/16 at pre-transplant workup, a specific preparation protocol was applied: 375 mg/m 2 rituximab (Rituxan ® ; Mabthera, Roche, Welwyn Garden City, UK) on day 14 before LT. In patients with anti-donor blood group antibodies ≥1/16 on day 7 before transplantation, plasmapheresis was performed as well. If iso-agglutinins remained at ≥1/16 on day 5 before transplantation, patients received both a further plasmapheresis and an immunoglobulin injection (sandoglobuline ® ; CSL Behring, Pennsylvania, PA, USA) (100 mg/kg). Additional plasmaphereses were done on day 1 before LT if iso-agglutinins remained at ≥1/16 despite the previous preparation protocol. All ABOi patients also received the baseline immunosuppressive protocol described above. The target level of tacrolimus was the same as in non-ABOi cases. For ABOi LDLT cases, steroids were added to immunosuppressive therapy and initiated with an injection of 2 mg/kg of methylprednisolone before graft perfusion during surgery. Recipients received an intravenous injection of 1 mg/kg of methylprednisolone during postoperative day (POD) 1-2, 0.75 mg/kg during POD 3-5, 0.5 mg/kg during POD 6-10, 0.4 mg/kg during POD 11-15, 0.3 mg/kg during POD 16-20, and 0.25 mg/kg during POD 25-50. Patients were switched to oral methylprednisolone as soon as possible, and steroid dosage was subsequently tapered (as previously explained) until the result of the routine post-transplant liver biopsy at 6 months (depending on the results of the latter).

Monitoring of Iso-Agglutinins in the Post-Transplant Period for ABOi Recipients and Adjustment of Immunosuppressive Therapy
Natural and immune iso-agglutinins were monitored daily for the first 8 days after ABOi LT and then at POD 10, 15, 20, and 30 (using the technique previously described). Additional dosages of immune and natural iso-agglutinins were performed depending on the clinical and biological course of each patient. Depending on the post-LT iso-agglutinin titers, additional immunosuppressive treatments were applied: plasmapheresis, rituximab, and/or immunoglobulins.

Histopathological Studies on Post-LT Liver Biopsies for Suspected ACR
Pathological diagnosis was made on a routine basis by 3 pathologists. The minimal quantitative requirement for the diagnosis of rejection was a biopsy containing at least 5 portal spaces. Liver biopsies were formalin-fixed (10%), paraffin-embedded, and cut into 4 µm sections. Staining methods for routine histological evaluation included hematoxylin and eosin, Period acid-Shiff-diastase (PAS-D), Masson's trichrome, iron, and immunostaining for cytokeratin 7.
When acute rejection was suspected based on abnormal liver function tests and/or on clinical parameters (unexplained fever, severe unexplained ascites, etc.), a liver biopsy was performed. Twenty-two (64.7%) ABOi recipients and thirty-six (52.9%) ABOc recipients underwent a liver biopsy according to these criteria within the first 4 weeks post-transplant. ACR was diagnosed using Banff criteria [22]. Histological AHR was suspected when peri-portal edema and necrosis, or portal hemorrhagic edema was associated with elevated antidonor A/B antibody titers [23,24].
In each case of ACR and AHR, portal inflammation, bile duct damage, and venular endothelialitis were assessed separately using the rejection activity index of Banff criteria [22]. Immunostaining for C4d (rabbit monoclonal SP91, abcam, Cambridge, UK) was additionally performed, but AHR definition was not based on this C4d immunostaining. We considered the immunostaining for C4d positive when endothelial cells of portal veins and/or portal capillaries were strongly positive [25].

Statistical Analysis
Cases and controls were verified to be correctly matched for pre-transplant diagnosis, age, and date of transplantation. Continuous variables were normalized by Box and Cox transformation. Comparisons of relevant variables between cases and donors were assessed by mixed model regressions, treating the average outcomes of each set of matched patients as a random variable. These regressions were performed according to a mixed linear, logistic, or Weibull model, depending on the nature of the dependent variable. Post-transplant liver biopsy (performed in the case of suspected acute rejection) was not available for all patients, and we had to rely on classical tests rather than mixed models because of the lack of match between the two groups. The Banff score being an ordered variable, we used a regression according to Snell [26]. Comparisons of variables between two age groups were made by Student's t-test, Weibull model regression, or the chi-square test (with Cook's correction), depending on the nature of the variable studied. When the numbers studied did not meet the criteria for a chi-square, Fisher's exact test was used. To minimize the bias related to the asymmetry of the independent variables, we used a sandwich variance for the regressions and calculated the degrees of freedom according to Bell and McCaffrey [27]. Statistical significance was set at two-sided * p < 0.05, ** p < 0.01, and *** p < 0.001.

Overall Results
Characteristics of ABOi pediatric liver recipients (n = 34) and their control group of ABOc LDLT (n = 68) are summarized in Table 1.
The 2-year patient and graft survival rates for ABOi LDLT recipients were 100% and 91.2%, respectively. For ABOc LDLT recipients, the 2-year patient and graft survival rates were 95.6% and 95.6%, respectively. The 5-year patient and graft survival rates remained unchanged in both groups. Three ABOc recipients died during the follow-up period: 1 patient because of sepsis and 2 children transplanted for liver malignancies in the setting of post-LT metastatic disease. Three ABOi recipients required retransplantation for diffuse intrahepatic bile duct injuries.
Positive correlations were observed between both natural and immune iso-agglutinins at pre-LT workup and the recipient's age at transplantation (p = 0.001 and p < 0.01, respectively) (Figure 1). Correlation between pre-LT natural iso-agglutinins titers and recipient's age (p = 0.001).

Post-Transplant Iso-Agglutinins and Additional Post-LT Immunosuppressive Treatments in ABOi LDLT Recipients
Correlation between pre-LT immune iso-agglutinins titers and recipient's age (p < 0.01).
Positive correlations were found between the maximum titers of post-LT natural and immune iso-agglutinins and the recipient's age at transplantation (p < 0.01 and p < 0.05; respectively) ( Figure 2).
The incidences of major post-ABOi LT complications according to recipient age (<1 year versus ≥1 year) are summarized in Table 3.  As shown in Table 3, all ABOi LT recipients who required retransplantation were older than 1 year at the time of transplantation. The indication for retransplantation was diffuse intrahepatic bile duct injuries in all cases, an entity that always occurred within the first 4 months after ABOi LT.
A positive correlation was observed between immediate pre-transplant immune isoagglutinin titers (after the immunosuppressive preparation applied for ABOi LDLT) and the incidence of post-LT AHR (p < 0.05). A positive correlation was also observed between the maximum post-transplant immune iso-agglutinin titer and the incidence of several post-LT complications: ACR (p = 0.01), AHR (p < 0.01), and retransplantation (p < 0.01). Similarly, a positive correlation was observed between the maximum post-transplant natural iso-agglutinin titer and AHR (p < 0.05) and retransplantation (p < 0.05).

Histopathologic Studies on Post-Transplant Liver Biopsies
Among the 22 ABOi and 36 ABOc recipients who underwent a liver biopsy during the first 4 weeks post-LT for suspected ACR, the Banff score was not significantly different between the two groups of patients (median Banff for ABOi patients = 5; median BANFF for ABOc recipients = 6; p = 0.8 NS).
On these liver biopsies taken for suspected ACR, positive C4d immunostaining was observed in 13/22 (59%) ABOi recipients versus in 3/36 (8.3%) non-ABOi recipients (p < 0.0001) (Figure 3). As shown in Table 3, all ABOi LT recipients who required retransplantation were older than 1 year at the time of transplantation. The indication for retransplantation was diffuse intrahepatic bile duct injuries in all cases, an entity that always occurred within the first 4 months after ABOi LT.
A positive correlation was observed between immediate pre-transplant immune isoagglutinin titers (after the immunosuppressive preparation applied for ABOi LDLT) and the incidence of post-LT AHR (p < 0.05). A positive correlation was also observed between the maximum post-transplant immune iso-agglutinin titer and the incidence of several post-LT complications: ACR (p = 0.01), AHR (p < 0.01), and retransplantation (p < 0.01). Similarly, a positive correlation was observed between the maximum post-transplant natural iso-agglutinin titer and AHR (p < 0.05) and retransplantation (p < 0.05).

Histopathologic Studies on Post-Transplant Liver Biopsies
Among the 22 ABOi and 36 ABOc recipients who underwent a liver biopsy during the first 4 weeks post-LT for suspected ACR, the Banff score was not significantly different between the two groups of patients (median Banff for ABOi patients = 5; median BANFF for ABOc recipients = 6; p = 0.8 NS).
On these liver biopsies taken for suspected ACR, positive C4d immunostaining was observed in 13/22 (59%) ABOi recipients versus in 3/36 (8.3%) non-ABOi recipients (p < 0.0001) (Figure 3). All ABOi recipients who required retransplantation had positive C4d immunostaining on their liver graft biopsy. We also observed that ABOi recipients with positive C4d All ABOi recipients who required retransplantation had positive C4d immunostaining on their liver graft biopsy. We also observed that ABOi recipients with positive C4d immunostaining had a significantly higher maximum post-LT immune iso-agglutinin titer (p < 0.05) and required plasmapheresis more frequently in the post-transplant period compared with ABOi C4d-negative patients (p < 0.05).

Discussion
According to our results, (1) despite a specific immunosuppressive protocol, the incidence of biliary complications was higher in ABOi pediatric LT recipients, as was the incidence of AHR, and the incidence of retransplantation; (2) all children who required retransplantation (for diffuse bile duct injuries related to ABO incompatibility) were older than 1 year of age at the time of ABOi LDLT; and (3) liver biopsies performed early after transplantation revealed positive C4d immunostaining mainly in ABOi recipients.
Despite Starzl's description of the liver as "a privileged organ", outcomes of the first studies including ABOi LT were poor. For this reason, ABOi liver grafts were initially limited to emergency situations. Indeed, Gordon et al. reported decreased patient (45%) and graft (26%) survival rates for ABOi LT at 1 year post-transplantation [28]. Similarly, in 1990, Gugenheim et al. published a series of 234 LT cases, in which the 2-year graft survival rates were 76% for ABOc and 30% for ABOi LT [7]. They noted an increased incidence of biliary complications, vascular thrombosis, cellular rejection, and histological findings consistent with antibody-mediated rejection. ABO antigens are expressed on the surface of vascular endothelial cells, bile duct cells, and sinusoidal endothelial cells of the transplanted liver. During ABOi LT, anti-blood-type antibodies bind to corresponding antigens on the surface of vascular endothelial cells and form antigen-antibody complexes that activate the complement system, resulting in complications, such as acute rejection and biliary complications, eventually leading to graft failure [4]. More recently, several groups have described new immunosuppressive strategies to improve graft survival during blood group cross-over: plasmapheresis, splenectomy, intravenous immunoglobulins, and arterial infusion therapy; but these strategies have little effect on poor outcomes and have increased the occurrence of other complications, including infections and sepsis [7,8,29,30].
A key element in ABOi LT was the description of the first case of ABOi LT using rituximab (a chimeric human-murine anti-CD20 monoclonal antibody) in 2003 and the first case of rituximab prophylaxis in ABOi LDLT in 2005 [30,31]. Indeed, the outcomes of ABOi LDLT have been radically improved since the introduction of this therapy. Egawa et al. published a large series including 291 ABOi LDLT cases in Japan, with a 3-year survival rate that increased from 30% to 80% after the introduction of rituximab [32,33]. Pediatric recipients tend to have better postoperative outcomes than adults after ABOi LT, because of the low levels of anti-ABO antibody titers in children, which stem from their immature immune and complement systems. However, outcomes of ABOi LT in pediatric populations remain uncertain and controversial.
According to our results, the incidences of biliary complications, AHR, and retransplantation were significantly higher in ABOi LDLT recipients despite a specific immunosuppressive pre-and post-transplant protocol. In 1994, Tanaka et al. reported 13 cases of ABOi pediatric LDLT, with 77% 1-year patient and graft survival rates [34]. Another study published in 1999 by Varela-Fascinetto et al. observed no significant difference in 10-year patient or graft survival rates in children who received 28 ABOi grafts compared with those who received 72 ABOc grafts [35]. In contrast, Szymczak et al. reported a 57% decrease in 1-year patient and graft survival rates in children, associated with increased acute and chronic rejection of ABOi grafts [36].
It has been suggested that more acceptable ABOi LT outcomes may be achieved in younger children [37,38]. Yandza et al. demonstrated that children under 2 years of age have lower anti-ABO antibody titers and lower morbidity than adults [10]. Gelas et al. found that among infants less than 1 year old, outcomes of ABOi LT are similar to those from ABOc LT, whereas such comparability was not observed in older children [39]. Similarly, Egawa et al. showed that the 5-year patient survival rates in patients younger than 1 year and older than 16 years are 77% and 22%, respectively. They observed that the incidence of intrahepatic biliary complications and hepatic necrosis in ABOi LDLT increases significantly with recipient age. Predictive risk factors for increased morbidity and mortality in this study were age ≥ 1 year and high anti-ABO antibody titers before transplantation [40]. These observations are consistent with our results as we described: (1) positive correlations between pre-and post-transplant immune and natural iso-agglutinins and the age of the recipient at the time of transplantation and (2) positive correlations between pre-and post-transplant immune and natural iso-agglutinin titers and the incidence of several post-LT complications: ACR, AHR, and retransplantation. Indeed, according to our results, children older than 1 year have higher titers of anti-donor blood group antibodies and require heavier immunosuppression protocols before and after ABOi LDLT. Moreover, an interesting finding is that all children who required retransplantation (due to diffuse intrahepatic bile duct injuries) in our series were older than 1 year. We can understand these results if we analyze anti-ABO antibody kinetics during childhood: maternal antibodies disappear from the neonate after 2 weeks, the newborn starts to produce its own IgM and IgG between 8 and 12 weeks after birth, and adult levels are reached around 5 to 10 years of age [41]. Other authors have also studied anti-ABO antibody titers before and after ABOi LDLT, describing (1) high pre-transplant anti-ABO antibody titers in patients ≥1 year of age; (2) a potential to increase anti-ABO antibody titers after LT, even if preoperative titers are low; and (3) a correlation between high antibody titers and severe post-LT complications (i.e., hepatic necrosis and intrahepatic biliary complications) [40,42]. In view of our results, as well as those of other authors [40], it may be questionable whether children older than 1 year of age are good candidates for ABOi LDLT.
In ABOi LT, AHR triggered by antibodies against donor-type iso-agglutinins is the most serious form of rejection and is often associated with graft loss. C4d is a molecule that binds to tissues after complement system activation, and immunostaining has been widely used to demonstrate humoral immunoreactivity or antibody-mediated rejection in kidney and heart transplantation [43,44]. Other authors have also demonstrated that C4d deposits in the portal stroma may be a feature of AHR in ABOi LT [45]. However, few authors have studied C4d status on liver biopsies taken for suspected acute rejection in both ABOi and ABOc pediatric LDLT recipients. When we analyzed C4d immunostaining on liver biopsies taken for suspected ACR within the first 4 weeks after LDLT, C4d-positive immunostaining was observed in 59% of ABOi recipients compared with 8.3% of non-ABOi recipients (p < 0.0001). We also observed that all ABOi recipients who required retransplantation for diffuse intrahepatic bile duct injuries had C4d-positive immunostaining on their posttransplant liver graft biopsy. Our results may suggest that limits between histologic ACR and AHR may be blurred and that C4d immunostaining may be routinely used in cases of suspected ACR in ABOi LT to guide the diagnosis toward underestimated AHR.
Interesting future work could consist of C4d immunostaining studies on systematic liver biopsies performed 6 months and 1 year after transplantation and their correlation with the medium-term clinical and biological course in ABOi versus ABOc pediatric LDLT recipients.

Conclusions
ABO-incompatible LDLT is a feasible option for end-stage pediatric liver disease but entails increased risks for the recipient, with a higher incidence of biliary complications, AHR, and retransplantation despite a specific pre-and post-transplant immunosuppressive protocol. We need to be particularly careful with children older than 1 year as they have higher anti-donor blood group antibody titers, require heavier immunosuppressive protocols before and after ABOi LDLT, and have a higher rate of retransplantation. C4d immunostaining may be used routinely for suspected ACR in ABOi LT and may be a hallmark of underestimated AHR.