Impact of Rituximab Desensitization on Blood-Type-Incompatible Adult Living Donor Liver Transplantation: A Japanese Multicenter Study

Authors


Abstract

We evaluated the effects of rituximab prophylaxis on outcomes of ABO-blood-type-incompatible living donor liver transplantation (ABO-I LDLT) in 381 adult patients in the Japanese registry of ABO-I LDLT. Patients underwent dual or triple immunosuppression with or without B cell desensitization therapies such as plasmapheresis, splenectomy, local infusion, intravenous immunoglobulin and rituximab. Era before 2005, intensive care unit-bound status, high Model for End-Stage Liver Disease score and absence of rituximab prophylaxis were significant risk factors for overall survival and antibody-mediated rejection (AMR) in the univariate analysis. After adjustment for era effects in the multivariate analysis, only absence of rituximab prophylaxis was a significant risk factor for AMR, and there were no significant risk factors for survival. Rituximab prophylaxis significantly decreased the incidence of AMR, especially hepatic necrosis (p < 0.001). In the rituximab group, other B cell desensitization therapies had no add-on effects. Multiple or large rituximab doses significantly increased the incidence of infection, and early administration had no advantage. In conclusion, outcomes in adult ABO-I LDLT have significantly improved in the latest era coincident with the introduction of rituximab.

Abbreviations
ABO-I

ABO-blood-type incompatible

ACR

acute cellular rejection

AIH

autoimmune hepatitis

AMR

antibody-mediated rejection

AUC

area under the curve

CMV

cytomegalovirus

DSA

donor-specific antibody

FHF

fulminant hepatic failure

ICU

intensive care unit

IHBC

intrahepatic biliary complication

IVIG

intravenous immunoglobulin

LDLT

living donor liver transplantation

MELD

Model for End-Stage Liver Disease

RBC

red blood cell

ROC

receiver operating characteristic

Introduction

Advances in ABO-blood-type-incompatible living donor liver transplantation (ABO-I LDLT) through innovations in B cell desensitization aimed at preventing antibody-mediated rejection (AMR) have expanded the donor pool in Japan. Local infusion through the portal vein or hepatic artery to decrease inflammatory reaction at the epithelium was introduced in 2000, and rituximab prophylaxis was introduced widely in 2004 in Japan [1]. Although there have been several single-center reports of rituximab prophylaxis in ABO-I LDLT, all describe small numbers of patients [2-4]. There is no information about how much, how many times or when rituximab should be administered, and there have been no comparisons of patient outcomes with and without rituximab in a large cohort.

Age is an important prognostic factor for AMR and patient and graft survival [5]. Demand for an effective desensitization method is especially strong in adult ABO-I LDLT. This study aimed to assess the effects of rituximab prophylaxis in ABO-I LDLT and to determine an effective and safe rituximab regimen.

Patients and Methods

Data collection

The Japan Study Group for ABO-Blood-Type-Incompatible Transplantation and a national registry for liver transplantation were established in 2001 by transplant centers performing ABO-I LDLT in Japan. The study group meets yearly to report experiences and has established a consensus for AMR diagnosis, treatment strategies and quality control of antibody titer measurements. Questionnaires are updated yearly and were sent in 2012 to registered surgeons and hepatologists in transplant centers, inquiring about patient characteristics, treatments and clinical courses. Information assayed included age, sex, disease, blood types of the recipient and donor, preoperative status, Model for End-Stage Liver Disease (MELD) score, relation of donor to recipient, peak titer of anti-donor-blood-type antibodies before transplantation and anti-donor antibody titer at the time of operation. Each center was classified as a large (≥10 ABO-I cases) or small (<10 ABO-I cases) volume center. Patients who required hospitalization in an intensive care unit (ICU) or a ward before surgery were classified as “in-ICU” or “in-hospital,” respectively. Patients who required medical care other than in an ICU or ward were classified as “at home” at the time of transplantation. Treatment data included graft type, splenectomy, immunosuppression, local infusion, plasmapheresis, intravenous immunoglobulin (IVIG) and rituximab. Data concerning dose, frequency and timing of rituximab treatment and its adverse effects were collected in 2012. Clinical course data included peak titer of anti-donor-blood-type antibodies after transplantation, as well as rejection, bacterial infection, fungal infection, cytomegalovirus (CMV) disease requiring treatments and patient survival. Data on mortality and cause of death were also collected.

Measurement of anti-A/B antibody levels

Titers of anti-donor-blood-type antibodies were measured at each institution and a quality control survey was performed yearly by The Japan Study Group for ABO-Blood-Type-Incompatible Transplantation [6]. The standard protocol for the test tube agglutination test is described briefly below [6, 7]. For both IgM and IgG assays, red blood cells (RBCs) were combined with the patient's serum sample at a ratio of 1:2 and centrifuged for 15 s. For the IgM assay, serum samples were first serially diluted with saline, and then incubated with RBCs at room temperature for 15 min. For the IgG assay using anti-human globulin, serum samples were preincubated with 0.01 M dithiothreitol at 37°C for 30 min, and then serially diluted and incubated with RBCs at 37°C for 30 min. The final dilution at which the agglutination reactivity was positive (1+), not equivocal (+/−), was determined as the antibody titer.

Definitions

Clinical AMR was diagnosed on the basis of radiological findings and clinical course, as described previously [1, 5]. The clinical manifestations of AMR were hepatic necrosis and intrahepatic biliary complication (IHBC). Hepatic necrosis was diagnosed when hepatic enzyme levels increased markedly in laboratory studies and liver necrosis was observed by computed tomography, usually 1 week after transplantation. IHBC was diagnosed when refractory cholangitis had developed and sclerosing change of the hepatic duct was observed by cholangiography. Diagnosis of acute cellular rejection (ACR) and chronic rejection was based on Banff criteria [8]. Infectious diseases were defined as infections requiring treatment.

Statistical analysis

Survival curves were constructed with the Kaplan–Meier method [1]. In univariate and multivariate analyses, Cox regression and logistic regression were used to evaluate the association between patient characteristics and overall survival and AMR, respectively. In the multivariate analyses, all potential confounders (p < 0.05 in the univariate analysis), including the era of operation, were included, and all patient data, including those for which values were missing, were used to minimize confounding and biases. The incidences of clinical complications were compared by using the chi-squared test.

Receiver operating characteristic (ROC) curves were plotted and areas under the curve were calculated to assess the optimum cut-off values for independent predictors of AMR. In analyses of prognostic factors for AMR and patient survival, the antibody cut-off titers that we calculated previously [1] were used. In the subgroup analysis of patients treated with rituximab, the cut-off titers for antibodies were newly calculated. SAS version 9.3 (SAS Institute, Inc., Cary, NC) was used for statistical analysis, and JMP version 10.0 (SAS Institute, Inc.) was used for the ROC curve analysis.

This study was performed in accordance with the provisions of the Declaration of Helsinki (as revised in Seoul, Korea, October 2008).

Results

Patients

By December 2011, clinical and laboratory data on 663 patients who underwent ABO-I LDLT in 37 institutions were available in the Japanese registry of ABO-I LDLT; of these patients, 381 who were aged 16 years or older were included as adults in the study. All 136 adult patients enrolled in our previous study [1] were included in the current study. The annual number of adults undergoing ABO-I LDLT was higher in 2001 and 2004 than in the previous years (Figure 1).

Figure 1.

Annual numbers of adults undergoing ABO-I LDLT or rituximab prophylaxis at 37 institutions in Japan. ABO-blood-type-incompatible living donor liver transplantation (ABO-I LDLT) without rituximab prophylaxis (black bars); with rituximab prophylaxis (gray bars).

Demographic data on the 381 patients are listed in Table 1. Recipient age ranged from 16 to 70 years (median, 52 years). MELD scores ranged from 17 to 66 (median, 18), and donor age ranged from 18 to 66 (median, 45). Graft type was left-side liver in 146 patients, right-side liver in 231 patients and unknown in 4 patients. The original diseases were hepatocellular carcinoma in 104 patients, hepatitis C cirrhosis in 58 patients, hepatitis B cirrhosis in 22 patients, alcoholic cirrhosis in 14 patients, primary biliary cirrhosis in 57 patients, primary sclerosing cholangitis in 10 patients, cirrhosis secondary to autoimmune hepatitis (AIH) in 5 patients, cirrhosis after Kasai operation for biliary atresia in 24 patients, fulminant hepatic failure (FHF) in 22 patients (including 2 cases of FHF due to AIH), Wilson's disease in 8 patients, cirrhosis secondary to nonalcoholic steatohepatitis in 6 patients, cryptogenic cirrhosis in 5 patients, idiopathic portal hypertension in 5 patients, re-transplantation in 16 patients and other diseases in 25 patients. In an analysis of the impact of the original disease, 7 patients with AIH (5 cases of cirrhosis and 2 of FHF), 57 patients with primary biliary cirrhosis and 10 patients with primary sclerosing cholangitis were classified as having autoimmune disease.

Table 1. Prognostic factors for overall survival and antibody-mediated rejection: univariate analysis (n = 381)
CharacteristicsCategoryNOverall survivalAntibody-mediated rejection
Hazard ratio95% CIp-Valuep-Value (global association without unknown)Odds ratio95% CIp-Valuep-Value (global association without unknown)
Cox regression analysisLogistic regression analysis
  • CMV, cytomegalovirus; IVIG, intravenous immunoglobulin; MELD, Model for End-Stage Liver Disease.
  • *p < 0.05.
Characteristics before transplantation
SexMale1691.0001.000
Female2121.0620.762–1.4790.7231.4550.759–2.7890.259
Center sizeLess than 10 cases491.0001.000
10 cases or more3321.1020.684–1.8450.7051.1710.438–3.1320.749
EraUp to 2000201.0000.002*1.000<0.001*
2001–2004790.6280.335–1.1780.1470.6400.214–1.9150.425
2005 onward2820.3910.217–0.7080.002*0.1880.065–0.5390.002
Autoimmune diseaseNo3041.0001.000
Yes741.0320.685–1.5530.8822.4111.217–4.7770.012*
Unknown32.6120.642–10.620.1800.000N/AN/A
Preoprative statusAt home1431.0000.013*1.0000.022*
In-hospital1781.2220.837–1.7860.2991.4600.692–3.0800.320
In-ICU402.1531.289–3.5960.003*3.6391.438–9.2080.006*
Unknown201.4890.727–3.0480.2770.5750.071–4.6730.605
Recipient's blood typeA911.0000.8601.0000.116
B870.8960.548–1.4640.6601.0500.353–3.1280.930
O2031.0040.671–1.5020.9842.0810.878–4.9320.096
Donor's blood typeA1831.0000.6541.0000.654
B1170.9490.643–1.4000.7930.7570.363–1.5800.458
AB811.1660.772–1.7620.4650.7260.311–1.6930.459
Antigen blood typeA2171.0000.5281.0000.965
B1530.9920.705–1.3960.9621.0240.537–1.9510.943
AB111.5970.696–3.6620.2690.7680.094–6.2560.805
Donor relativeNo1881.0001.000
Yes1850.7770.558–1.0830.1361.0180.543–1.9110.955
Unknown80.3500.049–2.5230.2980.000N/AN/A
IgM (peak before transplantation)Low (<256)2731.0001.000
High (≥256)621.1800.767–1.8170.4510.6830.275–1.6990.413
Unknown460.9080.528–1.5630.7290.1420.019–1.0600.057
IgG (peak before transplantation)Low (<64)1551.0001.000
High (≥64)1821.2290.863–1.7490.2532.3521.159–4.7710.018*
Unknown441.1120.627–1.9730.7170.5680.122–2.6370.470
IgM (at transplantation)Low (<16)2451.0001.000
High (≥16)821.2310.828–1.8280.3041.1830.577–2.4290.646
Unknown541.0070.613–1.6530.9790.1300.017–0.9760.047
IgG (at transplantation)Low (<16)1911.0001.000
High (≥16)1241.1720.809–1.6990.4012.6721.334–5.3540.006*
Unknown661.3360.855–2.0890.2041.1730.436–3.1610.752
MELDLow (<23)2401.0001.000
High (≥23)881.6191.095–2.3930.016*3.1721.565–6.4280.001*
Unknown532.0391.325–3.1380.0012.1930.898–5.3520.085
Desensitization therapies
Local infusionNo651.0001.000
Yes3120.9040.582–1.4050.6550.9290.410–2.1050.861
Unknown41.3680.323–5.7950.6710.000N/AN/A
SplenectomyNo1351.0001.000
Yes2410.8410.599–1.1810.3171.0940.564–2.1220.0790
Unknown50.8740.213–3.5870.8520.000N/AN/A
Rituximab prophylaxisNo1191.0001.000
Yes2590.5010.358–0.702<0.001*0.2140.111–0.414<0.001*
Unknown31.5540.380–6.3580.5400.000N/AN/A
Prophylactic IVIG after transplantationNo3251.0001.000
Yes560.8590.523–1.4090.5470.3920.117–1.3130.129
Anti-lymphocyte antibodiesNo3451.0001.000
Yes361.2320.732–2.0730.4320.9530.320–2.8360.931
PlasmapheresisNo471.0001.000
Yes3200.7230.454–1.1520.1721.1320.422–3.0380.806
Unknown140.9130.368–2.2630.8440.6460.069–6.0410.702
Plasmapheresis (times)0471.0000.2401.0000.247
1680.6390.353–1.1550.1380.8130.233–2.8370.745
2890.8650.505–1.4830.2771.1850.386–3.6370.767
3930.6220.355–1.0910.0980.6840.205–2.2830.537
4281.1590.597–2.2490.6642.8010.793–9.8880.110
≥5280.6590.302–1.4390.2951.0080.222–4.5840.992
Unknown280.6160.282–1.3460.2241.8260.478–6.9730.378
Short-term outcomes
IgM (peak posttransplantation)Low (<64)2511.0001.000
High (≥64)941.6891.180–2.4180.004*7.9353.973–15.85<0.001*
Unknown361.0460.571–1.9160.8840.000N/AN/A
IgG (peak posttransplantation)Low (<64)2051.0001.000
High (≥64)1261.4841.043–2.1100.028*10.4534.467–24.46<0.001*
Unknown501.1420.671–1.9450.6241.8050.450–7.2440.405
Acute rejectionNo2961.0001.000
Yes780.9640.640–1.4530.8621.1330.533–2.4080.745
Unknown72.0230.746–5.4870.1660.000N/AN/A
Chronic rejectionNo3491.0001.000
Yes51.9050.703–5.1580.2051.8270.199–16.740.594
Unknown271.7501.006–3.0440.0480.2810.037–2.1260.219
Bacterial infectionNo2541.0001.000
Yes1244.1602.965–5.835<0.001*1.8430.975–3.4850.060
Unknown33.6500.890–14.970.0720.000N/AN/A
Fungal infectionNo3421.0001.000
Yes345.7183.772–8.667<0.001*3.7761.666–8.5580.002*
Unknown51.3940.344–5.6480.6410.000N/AN/A
CMV diseaseNo1991.0001.000
Yes1800.7840.562–1.0950.1530.9110.485–1.7130.773
Unknown21.2330.171–8.8700.8350.000N/AN/A
Antibody-mediated rejectionNo3371.000
Yes442.4931.654–3.759<0.001*

Immunosuppression

All patients underwent double (calcineurin inhibitor and steroids; n = 36) or triple (calcineurin inhibitor, steroids and antimetabolites; n = 345) immunosuppression. The calcineurin inhibitor tacrolimus was administered in 364 cases, cyclosporine in 13 cases and an unknown drug in 4 cases. Regarding antimetabolites, cyclophosphamide was administered in 137 cases, mycophenolate mofetil in 286 cases, azathioprine in 18 cases, mizoribine in 20 cases and data were missing in 4 cases. Cyclophosphamide was switched to another antimetabolite in 105 cases. Antibody induction was performed by anti-lymphocytic antibody in 36 cases, anti-lymphocyte globulin in 15 cases, anti-IL-2 receptor antibody in 18 cases, muromonab-CD3 (OKT-3) in 2 cases and an unknown antibody in 1 case.

B cell desensitization

Plasmapheresis (n = 320), local infusion (n = 312), rituximab (n = 259), splenectomy (n = 241) and IVIG (n = 56) were performed. Local infusion, IVIG and rituximab were first used in 2000, 2003 and 2004, respectively. The number of times plasmapheresis was used before transplantation ranged from 0 to 11 (median, 2). Prophylactic IVIG was performed in seven institutions as center-specific policy, and it was performed in 6 patients before transplantation and 56 patients after transplantation. Here, we analyzed the effects of only posttransplantation IVIG. The dose ranged from 0.5 to 0.8 g/kg/injection, and the number of doses in regimens ranged from 2 to 5. There was no significant difference in titers between patients treated, or not treated, with IVIG (data not shown).

In the subgroup analysis of the rituximab group, regimens were classified into the following four groups: rituximab only without splenectomy or local infusion (R; n = 10); rituximab with splenectomy but without infusion (RS; n = 30); rituximab with infusion but without splenectomy (RI; n = 80); and rituximab with both infusion and splenectomy (RIS; n = 137).

Rituximab administration

Doses of rituximab were 500 mg/body in 113 cases, 300 mg/body in 60 cases and 375 mg/m2 in 49 cases. The number of doses administered was 1 in 222 cases, 2 in 22 cases and 3 in 12 cases. The timing of initial administration ranged from preoperative days 0 to 66 and was ≤6 days before transplantation in 22 cases (Figure 2).

Figure 2.

The timing of initial administration of rituximab ranged from preoperative days 0 to 66 and was within 6 days before transplantation in 22 cases.

Analysis for prognostic factors

In univariate Cox regression analyses, prognostic factors that were significantly and favorably associated with patient survival were era (2005 onward), preoperative status (at home), low MELD score (<23), rituximab prophylaxis, low peak IgM and IgG donor-specific antibody (DSA) titers posttransplantation (<64), absence of bacterial and fungal infection and absence of AMR (Table 1). There was no significant factor among pretransplant characteristics and types of desensitization therapy in the multivariate analysis after adjustment for the era effect (Table 2).

Table 2. Prognostic factors for overall survival: multivariate analysis (n = 381)
CharacteristicsCategoryN5-Year survival (%)Hazard ratio95% CIp-Value
  1. MELD, Model for End-Stage Liver Disease.
EraUp to 20002040.01.000
2001–20047950.60.7660.378–1.5510.459
2005 onwards28267.50.7420.346–1.5910.443
Preoperative statusAt home14365.81.000
In-hospital17863.61.0870.7351.6060.676
In-ICU4044.31.3550.7652.3980.297
Unknown2060.00.8830.3951.9740.762
MELDLow (<23)24066.91.000
High (≥23)8857.21.3640.894–2.0800.149
Unknown5348.81.4200.8272.4370.203
Rituximab prophylaxisNo11948.41.000
Yes25969.60.6290.3771.0510.077
Unknown333.31.8750.4457.9000.391

In univariate analyses, significant risk factors for AMR were era (up to 2000 or 2001–2004), autoimmune disease, preoperative status (in-ICU), high peak IgG DSA titer before transplantation (≥64), high IgG DSA titer at transplantation (≥16), high MELD score (≥23), absence of rituximab prophylaxis, high peak IgM and IgG DSA titers posttransplantation (both ≥64) and presence of fungal infection (Table 1). Among pretransplant characteristics and types of desensitization therapy, only the absence of rituximab prophylaxis was a significant indicator of risk of AMR in the multivariate analysis after adjustment for the era effect (Table 3).

Table 3. Prognostic factors for antibody-mediated rejection: multivariate analysis (n = 381)
CharacteristicsCategoryNAMR (%)Odds ratio95% CIp-Value
  • AMR, antibody-mediated rejection; MELD, Model for End-Stage Liver Disease.
  • *p < 0.05.
EraUp to 20002030.01.000
2001–20047921.50.6560.170–2.5340.541
2005 onwards2827.50.6250.143–2.7420.534
Autoimmune diseaseNo3049.51.000
Yes7420.32.0230.9404.3560.072
Unknown30.00.000N/AN/A
Preoperative statusAt home1438.41.000
In-hospital17811.80.9290.404–2.1340.862
In-ICU4025.01.4300.4734.3200.526
Unknown205.00.3220.0303.4430.349
IgG (preoperative)Low (<64)1557.71.000
High (≥64)18216.51.8050.7244.5050.205
Unknown444.60.7440.100–5.5550.773
IgG (at operation)Low (<16)1917.91.000
High (≥16)12418.61.9330.790–4.7310.149
Unknown669.11.0660.2694.2340.927
MELDLow (<23)2407.51.000
High (≥23)8820.52.0260.8784.6750.098
Unknown5315.10.9360.278–3.1540.915
Rituximab prophylaxisNo11923.51.000
Yes2596.20.2480.0890.6900.008*
Unknown30.00.000N/AN/A

AMR was a significant risk for overall survival in the univariate analysis (p < 0.001; Figure 3).

Figure 3.

Comparison of overall survival between patients with and without antibody-mediated rejection. Patients with antibody-mediated rejection (AMR) had a significantly higher overall survival risk than those without AMR, p < 0.001.

Impact of rituximab on clinical outcomes

The AMR incidence was significantly lower in the rituximab group (6%) than in the nonrituximab group (23%) (p < 0.001; Figure 4, top); a significant difference was also observed for the subset of patients with hepatic necrosis-type AMR (p < 0.001; Figure 4, top). There were no significant differences between the incidences of ACR (Figure 4, top), bacterial infection or CMV disease (Figure 4, bottom) between the rituximab and nonrituximab groups. The rate of fungal infection was significantly lower in the rituximab group (4%) than in the nonrituximab group (19%) (p < 0.001; Figure 4, bottom).

Figure 4.

Comparison of incidences of complications between rituximab and nonrituximab groups. The incidences of antibody-mediated rejection (AMR) and acute cellular rejection (ACR) are shown (top); rates of intrahepatic biliary complication (IHBC) and hepatic necrosis (HN) type AMR were lower in the rituximab group than in the nonrituximab group (chi-squared test, p < 0.0001). The incidences of bacterial infection, fungal infection and cytomegalovirus (CMV) disease are shown (bottom); rates of bacterial infection and CMV disease were similar between the two groups (chi-squared test, p = 0.36), but the rate of fungal infection was significantly lower in the rituximab group (chi-squared test, p < 0.0001).

Adverse effects of rituximab (kidney dysfunction, sepsis, neutropenia or lung edema) were observed in four patients, whose ages ranged from 56 to 62 years. Neutropenia occurred after a single dose of 300 mg/body, and the other complications manifested after the second or third dose of 500 mg/body. The patient with renal dysfunction died from a massive thrombus of the superior mesenteric artery on postoperative day 63, and the patient with sepsis died on postoperative day 202 from sepsis with an unknown focus. The other two patients are doing well.

Subgroup analysis of rituximab group

Because most ABO-I LDLT patients are currently administered rituximab, we analyzed the effects of additional desensitization therapies and the manner of rituximab administration to elucidate a better regimen. In a subgroup analysis of the rituximab group, local infusion, splenectomy, anti-lymphocyte antibodies and IVIG had no significant impact on overall survival or AMR incidence (Table 4).

Table 4. Prognostic factors for antibody-mediated rejection and overall postsurgical survival: univariate analysis of 259 patients given rituximab prophylaxis
CharacteristicsCategoryNOverall survivalAntibody-mediated rejection
Hazard ratio95% CIp-Valuep-Value (global association)Odds ratio95% CIp-Valuep-Value (global association)
Cox regression analysisLogistic regression analysis
  1. IVIG, intravenous immunoglobulin; R, only rituximab; regular dose, 500 mg/body or 375 mg/m2; RI, rituximab and infusion; RIS, rituximab and infusion and splenectomy; RS, rituximab and splenectomy; small dose, 300 mg/body or less.
Local infusionNo401.0001.000
Yes2181.3290.635–2.7790.4512.8820.370–22.4500.312
Unknown1
SplenectomyNo901.0001.000
Yes1690.9850.614–1.5790.9480.8810.309–2.5060.812
Anti-lymphocyte antibodiesNo2441.0001.000
Yes150.8380.306–2.2980.7310.4470.023–8.5470.593
Prophylactic IVIG after transplantationNo2141.0001.000
Yes450.9840.529–1.8300.9600.6640.146–3.0310.598
Timing of rituximab administration before transplantation≤6 days221.0001.000
>7 days2361.2410.535–2.8830.6151.4250.179–11.3300.738
Unknown1
Number of doses of rituximab12251.0000.4431.0000.922
2221.5040.747–3.0310.2530.9470.161–5.5600.730
3121.3770.550–3.4480.4940.5430.027–10.770.689
Dose of rituximabRegular1621.0001.000
Small661.2820.745–2.2070.3702.6550.952–7.4040.062
Unknown31
Dose and number of doses of rituximabRegular × 11341.0000.4611.0000.409
Regular × 2161.4080.589–3.3660.4420.4510.023–8.9020.601
Regular × 3121.5060.580–3.9100.4000.5950.029–12.2400.737
Small × 1601.2640.694–2.3100.4442.0860.738–5.8970.165
Small × 262.7550.844–8.9930.0934.0580.512–32.190.185
Unknown31
RegimenRS301.0000.7001.0000.938
R102.0530.490–8.5970.3250.9370.031–28.370.970
RI811.5680.596–4.1280.3621.6930.266–10.7900.577
RIS1371.6910.667–4.2850.2681.4540.242–8.7430.683
Unknown1

Patients who were administered multiple doses of rituximab, or a regular dose of 500 mg/body or 375 mg/m2, tended toward a lower incidence of AMR, but this was not statistically significant (Table 4). In contrast, patients given multiple doses had significantly greater incidences of fungal infection and CMV disease than those given a single dose, and patients given the regular dose had a greater incidence of CMV disease than those given a small dose of 300 mg/body or less (Table 5). Patients subjected to local infusion together with rituximab prophylaxis (RI and RIS) had greater incidences of CMV disease than patients without local infusion or splenectomy (R) (Table 5). Finally, there were no significant differences among rituximab regimens in terms of AMR incidence or patient survival (Table 4; Figure 5).

Table 5. Prognostic factors for infectious complications: univariate analysis of 259 patients given rituximab prophylaxis
CharacteristicsCategoryNBacterial infectionFungal infectionCMV disease
Odds ratio95% CIp-Valuep-Value (global association)Odds ratio95% CIp-Valuep-Value (global association)Odds ratio95% CIp-Valuep-Value (global association)
Logistic regression analysisLogistic regression analysisLogistic regression analysis
  • IVIG, intravenous immunoglobulin; R, only rituximab; regular dose, 500 mg/body or 375 mg/m2; RI, rituximab and infusion; RIS, rituximab and infusion and splenectomy; RS, rituximab and splenectomy; small dose, 300 mg/body or less.
  • *p < 0.05.
Local infusionNo401.0001.0001.000
Yes2181.4490.671–3.1280.3450.8300.173–3.9930.8162.9451.373–6.3190.006*
Unknown1
SplenectomyNo901.0001.0001.000
Yes1690.5880.342–1.0110.0550.9130.260–3.2080.8871.0710.641–1.7910.793
Anti-lymphocyte antibodiesNo2441.0001.0001.000
Yes152.0100.703–5.7470.1931.6500.197–13.820.6441.0490.369–2.9820.929
Prophylactic IVIG after transplantationNo2141.0001.0001.000
Yes451.7920.925–3.4710.0841.9220.489–7.5590.3501.6260.851–3.1060.141
Timing of rituximab administration before transplantation≤ 6 days221.0000.383–2.5010.9641.0001.000
>7 days2360.9790.4020.081–1.9880.2641.0120.421–2.4350.978
Unknown1
Number of doses of rituximab12251.0000.5131.0000.010*1.0000.004*
2220.6380.227–1.7980.3961.5430.181–13.170.6923.0381.256–7.9800.019*
3121.5490.475–5.0500.46810.2882.278–46.470.002*36.7424.737–999.90.017*
Dose of rituximabRegular1621.0001.0001.000
Small661.7420.948–3.2030.0740.1220.000–0.9840.1520.4550.249–0.8320.011*
Unknown31
Dose and number of doses of rituximabRegular × 11341.0000.2831.0000.040*1.0000.001*
Regular × 2160.6790.182–2.5260.5632.2430.220–12.320.41214.8023.517–137.30.003*
Regular × 3122.1010.625–7.0580.2308.5421.756–37.860.006*35.8054.548–999.90.018*
Small ×  1601.8280.955–3.5010.0690.1920.001–1.7340.2700.7800.412–1.4510.440
Small × 261.4710.258–8.3900.6642.1080.015–23.080.6570.1100.000–0.9640.167
Unknown31
RegimenRS301.0000.2661.0000.6851.0000.034*
R102.6110.574–11.710.2213.1050.232–41.870.3662.6090.574–11.710.221
RI812.3510.929–6.6700.0890.9000.141–9.5670.9173.1761.264–8.9820.021*
RIS1371.5660.642–4.3180.3570.9800.195–9.6540.9834.0531.688–11.070.004*
Unknown1
Figure 5.

One-year survival of patients in the rituximab group. R, rituximab without splenectomy or local infusion (n = 10); RI, rituximab with infusion but without splenectomy (n = 81); RIS, rituximab with both infusion and splenectomy (n = 137); RS, rituximab with splenectomy but without infusion (n = 30). There were no significant differences among regimens with additional desensitization in patients with rituximab prophylaxis.

Early administration of rituximab had no significant impact on AMR incidence or patient survival (Table 4). Twenty-two FHF patients underwent LDLT, and six of them were given rituximab immediately before or during transplantation (three treated with RIS, two with RI and one with RS). All 6 patients survived transplantation without AMR, whereas AMR occurred in 7 patients and 1-year survival was 44% in the other 16 patients who were not given rituximab.

Peak IgG DSA titer before transplantation, IgG DSA titer at transplantation and peak IgG and IgM DSA titers posttransplantation showed a significant positive association with AMR incidence in the total cohort of adult ABO-I LDLT patients in the univariate analysis (Table 1). In the rituximab group, peak IgG and IgM DSA titers posttransplantation were significantly greater in patients with AMR than in those without AMR (Table 6). When the AMR incidence in the rituximab group was compared between high and low titers according to optimum cut-off values calculated from ROC curves, there were significant differences in peak IgG titers before transplantation (10% [10/104] vs. 3% [4/125] titer ≥128 vs. <128, p = 0.042), peak IgM titers posttransplantation (22% [10/45] vs. 3% [6/194], titer ≥64 vs. <64, p < 0.001) and peak IgG titers posttransplantation (19% [10/54] vs. 2% [3/171], titer ≥128 vs. <128, p < 0.001).

Table 6. Comparison of antibody titers between patients with and without AMR under rituximab prophylaxis
  AMR+AMR−p-Value
  NMedianMean ± SDNMedianMean ± SD
  • AMR, antibody-mediated rejection. p-values are derived from Wilcoxon sum-rank test.
  • *p < 0.05 for AMR+ versus AMR−.
IgMPeak before transplantation1564158 ± 25521164147 ± 1990.881
At transplantation1647 ± 8213416 ± 480.700
Peak posttransplantation1664593 ± 1091223849 ± 181<0.001*
IgGPeak before transplantation14128408 ± 58421564319 ± 7710.221
At transplantation131627 ± 35210834 ± 960.265
Peak posttransplantation132561002 ± 21962121668 ± 187<0.001*

Discussion

Worldwide, the first case report of rituximab prophylaxis in kidney transplantation was published in Japan in 2002 [9]; many rituximab protocols for kidney transplantation have been reported since. Monteiro et al [10] reported the first case of ABO-I liver transplantation using rituximab in 2003, and Usuda et al [3] reported the first case of rituximab prophylaxis in ABO-I LDLT in 2005. In the Japanese registry, the first adult case of rituximab prophylaxis was reported in November 2003. In our previous multicenter study [1] of 291 patients who underwent ABO-I LDLT up to and including March 2006, 44 adult patients were administered rituximab. The current study includes 259 adult patients who underwent rituximab prophylaxis up to and including December 2011.

After 2000, the evolution of innovation in the treatment of small-for-size syndrome in adult LDLT and desensitization for DSA was achieved [11-13]. The era effect on overall survival is significant. In the total cohort of 381 adult patients, after adjustment for era effects in the multivariate analysis, only rituximab prophylaxis was a significant prognostic factor for AMR, but it was not a prognostic factor for overall survival. A prospective study is required to elucidate the effect of rituximab on patient survival; however, it would be difficult to remove rituximab prophylaxis when the current results are so much improved in the most recent era and when this may be attributable to rituximab.

To find the best regimen for rituximab, the impact of additional desensitization therapies and times and doses of rituximab were addressed. Splenectomy used to be considered an essential component of a successful ABO-I desensitization regimen for renal transplantation [14]; however, it has been reported that rituximab can be used in place of splenectomy with similar outcomes [15, 16]. The Kyoto group suggested that splenectomy should be avoided in 2007 [2, 17]. In LDLT, however, splenectomy is performed not only for desensitization but also for portal flow adjustment in patients with small-for-size syndrome and for future anti-viral treatment using interferon in hepatitis C patients. An assessment of the effects of preserving the spleen is required in patients without small-for-size syndrome or hepatitis C infection in future.

Plasma exchange is a standard procedure to reduce DSA titers, but the titer required to prevent AMR is not defined. If titers increase again after plasmapheresis, another plasmapheresis is often performed. When peak titer before transplantation is very low, plasmapheresis is not performed. In other words, the more times the plasmapheresis is performed, the greater the potential for an increase in DSA titer. However, we observed no significant relationship between the number of plasmapheresis procedures and clinical outcomes (Table 1).

IVIG is also a standard procedure, especially for human leukocyte antigen-related DSA in kidney transplantation, and the IVIG dose often ranged from 0.1 to 2 g/kg [18, 19]. In liver transplantation, Ikegami et al [4] reported a small series with desensitization by rituximab and IVIG (0.8 g/kg), and their cases were included here. We found no significant effect of IVIG on overall survival or AMR in the entire adult cohort (Table 1) and no additional effects in the rituximab group (Table 5). We analyzed the AMR incidence in each regimen with IVIG versus without IVIG (Figure 6). The AMR incidence was reduced from 26% to 9% in the local infusion and splenectomy (IS; no rituximab) regimen when IVIG was added, but this difference was not significant (p = 0.19). Among regimens with rituximab (R, RI, RIS and RS), the incidences were similar between with IVIG and without IVIG. IVIG is not approved in Japan and is not covered by insurance. IVIG costs 1.5–2.0 million yen per injection, whereas 500 mg of rituximab costs 0.3 million yen. A prospective study is required to elucidate the effects of IVIG in patients after rituximab prophylaxis.

Figure 6.

Comparison of the incidences of antibody-mediated rejection (AMR) with and without intravenous immunoglobulin (IVIG) in each regimen. IS, local infusion with splenectomy without rituximab; R, rituximab without splenectomy or local infusion; RI, rituximab with infusion but without splenectomy; RIS, rituximab with both infusion and splenectomy; RS, rituximab with splenectomy but without infusion. There were no significant differences in the incidence of AMR.

The incidence of adverse effects of rituximab was 1.6% (4/258), and all patients recovered and underwent LDLT. Rituximab prophylaxis could be tolerated by patients with end-stage liver diseases. The incidences of bacterial infections and CMV disease after transplantation were similar between the nonrituximab and rituximab groups, but the incidence of fungal infection was significantly lower in the rituximab group. Although data for the amount of steroid and trough levels of calcineurine inhibitors were not collected here, the total amount of conventional immunosuppressant might be reduced in light of the expected beneficial effects of rituximab. Lower amounts of conventional immunosuppressants might be a reason for the lower fungal infections.

In this study, half the patients were given 500 mg/body, a quarter were given 300 mg/body and a quarter were given 375 mg/m2 (corresponding to 430–762 mg/body; median, 600 mg/body). One reason for dose reduction could be concern about potential adverse effects in patients with end-stage liver diseases. In kidney transplantation, Shirakawa et al [20] reported a successful trial to reduce rituximab from 500 to 200 mg/body. Here, there was a tendency toward a higher incidence of AMR in patients treated with ≤300 mg/body compared with 500 mg/body or 375 mg/m2; however, three patients treated with 130 mg/body or 200 mg/body belonged to the same center, and one of them died from severe AMR. More evidence is needed before we can recommend reducing the rituximab dose below 300 mg/body in liver transplantation.

Multiple administrations of rituximab are standard in the treatment of B cell lymphoma. However, because the amount of targeted B cells is expected to be much smaller in transplant patients, a single dose is usually applied. A single dose is standard in kidney transplantation. Here, there were patients with two administrations in six centers and with three administrations in three centers, but the majority of these patients underwent transplantations in 2010 or earlier. All three centers changed their policy to one dose in 2012 on the basis of our data. The current study clearly demonstrates that multiple doses provide no significant benefit in terms of AMR incidence or survival, whereas they increase the incidences of fungal and CMV infections.

The Kyoto group recommended early administration of rituximab to deplete B cells, although the incidence of clinical AMR did not increase significantly in patients with late administration [2]. Here, the timing of rituximab administration had no significant effect on AMR incidence on patient survival. Furthermore, 6 of 22 patients with FHF were given rituximab within 6 days before transplantation and survived without AMR. Hence, administration of rituximab immediately before transplantation is a promising therapeutic strategy.

The titers decrease after desensitization before transplantation and increase or do not change immediately after transplantation, and they usually decrease thereafter when patients survive [1]. Hence, the optimum cut-off values vary among time points, between IgM and IgG. In rituximab-treated patients, peak IgG and IgM DSA titers posttransplantation were significantly greater in those with AMR, and the AMR incidence was significantly higher in patients with peak titers posttransplantation above optimum cut-off values calculated from ROC curves (i.e. IgM, ≥64; IgG, ≥128). Theoretically, it is an option to treat patients preemptively by using other desensitization methods such as IVIG and plasmapheresis when antibody titers are above the cut-off values; however, the decision is still difficult.

This study had limitations. It was an uncontrolled retrospective observational study with many confounders, some of which may have been nonrandom and unaccounted for, and thus despite the use of appropriate multivariate statistics unknown bias was possible. Because of the extent of co-linearity between rituximab and era, estimates of regression coefficients still might be unstable, although we tried to adjust era effects as much as possible. Prospective studies are required to examine the causality of the relationships found.

In conclusion, outcomes in adult ABO-I LDLT have significantly improved in the latest era coincident with the introduction of rituximab.

Acknowledgments

We thank the institutions and members of The Japan Study Group for ABO-Blood-Type-Incompatible Transplantation: Asahikawa Medical University (Dr. Furukawa), Dokkyo University (Dr. Kubota), Ehime University (Dr. Takada), Fujijta Health University (Dr. Hibi), Hiroshima University (Dr. Ohdan), Iwate University (Drs. Takahara and Wakabayashi), Kanazawa University (Dr. Takamura), Hokkaido University (Drs. Shimamura and Taketomi), Jichi Medical University (Dr. Mizuta), Kanagawa Children's Hospital (Dr. Shinkai), Kanazawa Medical University (Dr. Ikawa), Kansai Medical University (Dr. Kaibori), Keio University (Dr. Tanabe), Kobe Municipal Hospital (Drs. Uryuhara and Kaihara), Kumamoto University (Drs. Yamamoto and Inomata), Kyoto University (Drs. Kaido and Uemoto), Kyoto Prefectural University (Drs. Okajima and Yoshimura), Kyushu University (Drs. Shirabe and Maehara), Mie University (Drs. Mizuno and Isaji), Nagasaki University (Drs. Soyama, Takatsuki and Eguchi), Nagoya City University (Dr. Suzuki), Nagoya University (Drs. Kiuchi and Ogura), Nara Medical University (Dr. Nakajima), National Center for Child Health and Development (Drs. Fukuda, Sakamoto and Kasahara), Niigata University (Dr. Sato), Okayama University (Dr. Yagi), Osaka University (Drs. Umeshita, Maruhashi and Nagano), Osaka City University (Dr. Kubo), Osaka Medical College (Dr. Hayashi), Shinshyu University (Dr. Miyagawa), Tokushima University (Dr. Shimada), Tsukuba University (Drs. Hori and Tanaka), Tokyo University (Dr. Sugawara), Tokyo Medical University (Dr. Shimazu), Tokyo Women's Medical University (Drs. Nakajima, Fuchinoue and Yamamoto), Tohoku University (Dr. Kawagishi), Yamaguchi University (Dr. Tamesa) and Yokohama City University (Drs. Takeda and Endo).

Disclosure

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

Ancillary