Immunosuppression Induction With Rabbit Anti-Thymocyte Globulin With or Without Rituximab in 1000 Liver Transplant Patients With Long-Term Follow-Up

Transcription

1 LIVER TRANSPLANTATION 18: , 2012 ORIGINAL ARTICLE Immunosuppression Induction With Rabbit Anti-Thymocyte Globulin With or Without Rituximab in 1000 Liver Transplant Patients With Long-Term Follow-Up Richard S. Mangus, 1 Jonathan A. Fridell, 1 Rodrigo M. Vianna, 1 Paul Y. Kwo, 2 Jeanne Chen, 1 and A. Joseph Tector 1 1 Transplant Division, Department of Surgery, and 2 Hepatology Division, Department of Medicine, Indiana University School of Medicine, Indianapolis, IN Rabbit anti-thymocyte globulin (ratg) based immunosuppression induction is being increasingly used in liver transplantation (LT) in conjunction with steroid-free protocols to delay the initiation of calcineurin inhibitors. This study reports a single-center comparison of transplant outcomes and complications in 3 immunosuppression eras. Data were obtained retrospectively from a center research database, and the analysis included LT patients from 2001 to The immunosuppression consisted of ratg induction in 3 doses (6 mg/kg in all): (1) the first dose was administered perioperatively [the rabbit anti-thymocyte globulin in the operating room (ratg-or) era]; (2) the first dose was delayed until 48 hours after transplantation [the rabbit anti-thymocyte globulin after a delay (ratg-d) era]; or (3) the first dose was delayed until 48 hours after transplantation, and a single dose of rituximab was added 72 hours after transplantation [the rabbit anti-thymocyte globulin after a delay plus rituximab (ratg-d-ritux) era]. The initial maintenance immunosuppression was tacrolimus monotherapy, which was started on postoperative day 2. There were 166 patients (16%) in the ratg-or era, 259 patients (26%) in the ratg-d era, and 588 patients (58%) in the ratg-d-ritux era (1013 patients in all). Demographically, the latter eras were characterized by higher recipient and donor ages; greater percentages of liver-kidney transplants, hepatocellular carcinoma (HCC), donation after cardiac death (DCD), and imported organs; and shorter graft ischemia times. There were no significant differences between the 3 immunosuppression groups in unadjusted patient survival 3 and 5 years after transplantation (80% and 75% for the ratg-or era, 75% and 67% for the ratg-d era, and 79% and 71% for the ratg-d-ritux era, P ¼ 0.15). The 5-year survival rates for patients with hepatitis C virus (HCV) and HCC were 65% and 68%, respectively. The factors included in the Cox regression model for patient death included the Model for End-Stage Liver Disease score [hazard ratio (HR) ¼ 1.03, P ¼ 0.001], HCV (HR ¼ 1.28, P ¼ 0.04), donor age (HR ¼ 1.01, P ¼ 0.001), recipient age (HR ¼ 1.01, P ¼ 0.05), and DCD (HR ¼ 1.55, P ¼ 0.11). ratg-based induction immunosuppression can be safely used in adult LT recipients with excellent survival and low rejection rates and without increases in immunosuppression-related side effects. Liver Transpl 18: , VC 2012 AASLD. Received November 19, 2011; accepted January 7, Abbreviations: CI, confidence interval; CMV, cytomegalovirus; DCD, donation after cardiac death; DVT, deep venous thrombosis; GFR, glomerular filtration rate; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; HR, hazard ratio; LT, liver transplantation; MELD, Model for End-Stage Liver Disease; PTLD, posttransplant lymphoproliferative disorder; ratg, rabbit antithymocyte globulin; ratg-d, rabbit anti-thymocyte globulin after a delay; ratg-d-ritux, rabbit anti-thymocyte globulin after a delay plus rituximab; ratg-or, rabbit anti-thymocyte globulin in the operating room. The data collection for this study was supported by the Roche Organ Transplant Research Foundation. Address reprint requests to Richard S. Mangus, M.D., M.S., Transplant Division, Department of Surgery, Indiana University School of Medicine, 550 North University Boulevard, Room 4601, Indianapolis, IN Telephone: ; FAX: ; rmangus@iupui.edu DOI /lt View this article online at wileyonlinelibrary.com. LIVER TRANSPLANTATION.DOI /lt. Published on behalf of the American Association for the Study of Liver Diseases VC 2012 American Association for the Study of Liver Diseases.

2 LIVER TRANSPLANTATION, Vol. 18, No. 7, 2012 MANGUS ET AL. 787 Immunosuppression after liver transplantation (LT) has improved to the point that clinicians are focused on the minimization of infectious, renal, neurological, and cardiovascular complications rather than the prevention of rejection. Disease recurrence is also a more pressing problem than rejection for LT patients. Modern immunosuppressive regimens should be designed with the consideration of steroid elimination, calcineurin minimization, and the potential for entry into immunosuppression withdrawal trials. Lymphocyte-depleting antibodies have been used in successful experimental models of tolerance induction. Anti-lymphocyte preparations were used in early LT series in conjunction with azathioprine and steroids, but they were replaced with calcineurin inhibitors. Anti-lymphocyte preparations were used as a part of quadruple-drug regimens with cyclosporine, azathioprine, and steroids in the late 1980s and early 1990s. These regimens were associated with severe hepatitis C virus (HCV) recurrence, posttransplant neoplasia, and significant infectious morbidity. 1,2 The direct result of these publications was that biological induction therapy with anti-lymphocyte preparations fell out of favor in clinical LT. In the past 10 years, however, rabbit anti-thymocyte globulin (ratg) has been used in LT as a way of delaying the introduction of calcineurin inhibitors or eliminating steroids or as part of an immunosuppression weaning protocol. 3-8 Although the use of ratg has increased in the past 10 years, the role of lymphocyte-depleting antibodies in modern LT remains unclear. In this article, we report the 7-year evolution of our immunosuppressive regimen in 1013 adult LT recipients who underwent ratg-based induction. There were 2 major changes to our regimen. First, we delayed the introduction of all immunosuppression for 48 hours after transplantation to delay the hemodynamic, pulmonary, renal, and infectious complications that accompany immunosuppressive drugs. The second change to the protocol involved the clinical decision to use a single dose of rituximab (a B cell depleting CD20 antibody) to replace steroids. This addition provided immunosuppression induction with 2 agents: one primarily T cell directed and the other primarily B cell directed. The long-term results of our series may be of value to programs considering the use of biological induction for LT and to programs trying to evaluate the timing of immunosuppression in weaning protocols. PATIENTS AND METHODS The records for all adult LT procedures (deceased donor and whole organ) performed between July 1, 2001 and December 31, 2008 were reviewed retrospectively from the transplant center research database. Recipient and donor demographics, including graft and patient survival statistics, were extracted. Additionally, a thorough review was undertaken to assess HCV and hepatocellular carcinoma (HCC) recurrence rates and the incidence of immunosuppression-related complications, including severe or chronic graft rejection, opportunistic infections, and changes in renal function. Our center maintains a high transplant volume through the regional and national importation of organs, and this results in a very short wait time to transplantation (median <40 days) but does not affect overall survival. 9,10 Our center uses a piggyback hepatectomy technique in >95% of cases. All patients in this study routinely underwent a second posttransplant operation: only skin closure was performed at the time of transplantation, and this was followed by re-exploration with liver biopsy and fascial closure on postoperative days 3 to 5. This approach was instituted to minimize the effects of abdominal compartment syndrome on newly transplanted allografts, many of which came from extended criteria donors. The immunosuppression induction consisted of 3 equal doses of ratg (total dose ¼ 6 mg/kg) with standard premedication given immediately before its administration: solumedrol [500 (first dose), 250 (second dose), and 120 mg (third dose)], acetaminophen (650 mg), and diphenhydramine (25 mg). Maintenance immunosuppression was primarily tacrolimus monotherapy, although some early recipients did receive a short steroid taper with steroid withdrawal within 3 months of transplantation. 3 Approximately one-half of therecipientsinthisstudyreceivedasingledoseofrituximab between the first and second doses of ratg as part of the induction protocol. Tacrolimus therapy was initiated on postoperative day 2. All clinicians strictly adhered to goal tacrolimus levels with daily monitoring for any hospitalized patients and twice weekly monitoring for the first 1 to 2 months for patients outside the hospital. The goal trough levels were 7 to 10 ng/ml in thefirst3monthsand6to8ng/mlthereafter.there were 3 eras during the study period: 1. Rabbit anti-thymocyte globulin in the operating room (ratg-or) era ( ). ratg with solumedrol was administered perioperatively on the day of transplantation and was followed by additional ratg/solumedrol administration on postoperative days 2 and 4; a short steroid taper was used thereafter (<3 months) with maintenance tacrolimus. 2. Rabbit anti-thymocyte globulin after a delay (ratg-d) era ( ). No immunosuppression (including steroids) was given at the time of transplantation or for 48 hours thereafter; ratg was then administered with premedication on postoperative days 2, 4, and 6. Maintenance tacrolimus was used. 3. Rabbit anti-thymocyte globulin after a delay plus rituximab (ratg-d-ritux) era ( ). The protocol of the ratg-d era was followed, but a single dose of rituximab (1.5 mg/m 2 ) was administered between the first and second doses of ratg (on postoperative day 3); maintenance tacrolimus was used.

3 788 MANGUS ET AL. LIVER TRANSPLANTATION, July 2012 Less than 3% of the patients did not complete the full protocol: some patients had severe infusion reactions or developed severe hypoxia, and some died from transplant-related complications. The development of posttransplant fevers was nearly ubiquitous in patients in the delayed immunosuppression eras and was felt to be related to the presence of the transplanted organ with no immunosuppression. In all cases, tacrolimus was not introduced until at least postoperative day 2. Some patients were started on additional immunosuppressant agents after the first 3-month posttransplant period when they returned to their primary hepatologist. Agents included mycophenolate mofetil, azathioprine, and sirolimus, but the use of these was limited to cases of refractory renal insufficiency and accounted for <5% of the total. Pre-established study endpoints of immunosuppression-related complications included opportunistic infections, posttransplant lymphoproliferative disorder (PTLD), graft rejection, and changes in renal function. Infectious complications required both clinical and laboratory confirmation, except for wound infections, which required a clinical diagnosis only. PTLD required a tissue diagnosis. Rejection was documented for patients with elevated liver function enzymes, biopsy findings consistent with rejection, and a requirement for steroid or antibody therapy. The diagnosis and treatment of both rejection and infection were found to be nonuniform for patients receiving care in the posttransplant period outside our center. Out-of-state patients were followed at our center for 3 months after transplantation and then returned to their home center. This accounted for 17% of the patients, and these patients were excluded only from the rejection and infection portions of the analysis. A diagnosis of rejection required endothelialitis, bile duct damage, and a periportal infiltrate as well as the exclusion of all other potential etiologies. A change in renal function was calculated as the difference between the preoperative glomerular filtration rate (GFR) measured immediately before transplantation and the GFR 30 days after transplantation. The GFR was calculated with the 4-variable Modification of Diet in Renal Disease equation, which has been shown to be the most accurate GFR measure in patients with liver disease or post-lt patients. 11 Patients undergoing combined liver-kidney transplantation were excluded from the renal function analysis (n ¼ 70 or 7%). Additionally, we report the rates of postoperative (30-day) myocardial infarction, stroke, and deep venous thrombosis (DVT), which required confirmation from the discharge summary and appropriate confirmatory testing, including changes in electrocardiograms or cardiac laboratory markers for myocardial infarction, computed tomography or magnetic resonance imaging for stroke, and Doppler ultrasound for DVT. The surveillance protocol for patients with HCC at the time of transplantation includes serial measurements of alpha-fetoprotein levels, computed tomography scans, and chest radiographs 4 and 12 months after transplantation, 18 and 24 months after transplantation, and yearly thereafter. At our center, 34% of HCC patients are outside the Milan criteria, and 50% receive some form of liverdirected therapy before transplantation. This has increased in recent years. The posttransplant HCV surveillance protocol at our center includes protocol biopsy 4 and 12 months after transplantation and then yearly protocol biopsy. Fibrosis severity is graded with the METAVIR system, and clinically significant HCV recurrence is defined as the presence of stage 2 fibrosis. HCV viral loads are not routinely measured. Anti-HCV therapy is initiated on a case-by-case basis with the onset of graft dysfunction and/or with fibrosis progression thought to be related to HCV. No pretransplant or preemptive antiviral therapy is administered. Standard statistical testing was conducted with commercially available software (SPSS 17.0 for Windows 2009, SPSS, Inc., Chicago, IL). Categorical variables were compared with chi-square testing, whereas continuous variables were compared with a 1-way analysis of variance. Cox proportional hazards models were constructed with a direct entry method, and covariates with a P value 0.10 were included in the final regression models. Standard tests for model construction were performed. This retrospective analysis of data from the transplant research database at our center was reviewed and approved by the institutional review board of the Indiana University School of Medicine. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki. RESULTS Patient and recipient demographics, stratified by the induction protocol, are presented in Table 1. The demographics of the 3 study groups differed statistically with respect to recipient sex, recipient age, and donor age. The ratg-d-ritux group differed from the other groups with a higher rate of HCC, more liverkidney transplants, and more donation after cardiac death (DCD) donors. The higher rate of HCC resulted from a referral bias due to the short wait-list time at our center, whereas the liver-kidney and DCD disparities are consistent with national trends. We anticipated worsened survival for the ratg-d-ritux group because of the HCC, liver-kidney, and DCD disparities. The study groups also differed with respect to ischemia times, with the ratg-or group having longer cold and warm ischemia times in comparison with the other groups. Although the cold ischemia times differed, the absolute impact on outcomes is unclear because there is likely little clinical impact from the difference between 6 and 8 hours of cold ischemia. The warm ischemia time is more problematic because a warm ischemia time of 25 minutes is clearly superior to a time of 60 minutes and may in fact have an important impact on clinical outcomes that must be considered. The minimum and median follow-up

4 LIVER TRANSPLANTATION, Vol. 18, No. 7, 2012 MANGUS ET AL. 789 TABLE 1. Demographic Data for LT Recipients and Corresponding Donors Stratified by the ratg Induction Protocol ratg-or Immunosuppression Era* ratg-d ratg-d-ritux ( ) ( ) ( ) P Value Recipient data Recipients [n (%)] 166 (16) 259 (26) 588 (58) Male sex (%) Caucasian race (%) Age (years) 50 (19-69) 52 (18-72) 54 (18-76) 0.02 Body mass index (kg/m 2 ) 27.8 ( ) 28.4 ( ) 27.5 ( ) 0.39 MELD score at transplant 16 (7-40) 17 (6-40) 16 (6-40) 0.42 Retransplantation (%) Combined liver-kidney transplantation (%) Diagnosis HCV (%) HCC [n (%)] 28 (17) 47 (18) 143 (24) 0.04 HCC outside the Milan criteria [n/n (%)] 8/28 (29) 16/47 (34) 49/143 (34) 0.84 Donor data Male sex (%) Caucasian race (%) Age (years) 35 (9-73) 43 (6-78) 43 (5-81) <0.01 Body mass index (kg/m 2 ) 25.7 ( ) 25.8 ( ) 25.7 ( ) 0.51 Cause of death (%) 0.27 Stroke Trauma Anoxia/other DCD (%) <0.01 Organ origin (%) <0.001 Local share Regional share National share Cold ischemia time (hours) 8 (3-17) 7 (3-20) 6 (3-18) <0.001 Warm ischemia time (minutes) 60 (21-203) 40 (14-142) 25 (8-85) <0.001 *In the ratg-or era, ratg induction was initiated intraoperatively at the time of reperfusion; in the ratg-d era, the ratg induction protocol was delayed until 48 hours after transplantation; and in the ratg-d-ritux era, the ratg induction protocol was delayed until 48 hours after transplantation, and rituximab was administered 3 days after transplantation. The data are presented as medians and ranges. times for the entire cohort were 18 and 63 months, respectively. There was a lower risk of intraoperative death and early graft loss in the last immunosuppression period (ratg-d-ritux; Table 2). This improvement may be related to the improvements in surgical techniques and recipient choice that come with increasing transplant volume and experience. The unadjusted 1-year graft and patient survival rates did not differ significantly for the groups. HCV and HCC recurrence was clearly higher over time because these events were time-dependent, but 1-year HCV and HCC recurrence rates did not differ statistically. Figures 1 to 3 and Table 3 present the 5-year Cox proportional hazards models for survival stratified by the immunosuppression era for all patients (Fig. 1), HCV-positive patients (Fig. 2), and patients with HCC (Fig. 3). For all 3 models, the 3 induction protocol groups did not differ significantly in survival up to 5 years after transplantation. Patients infected with HCV did have a lower 5-year survival rate (68%) than non-hcv patients (77%, P ¼ 0.01; data not shown). Patients with HCC and patients without HCC had equivalent 5-year patient survival rates when we accounted for several significant covariates, which included HCV infection, the Model for End-Stage Liver Disease (MELD) score at the time of transplantation, and the recipient and donor ages (73% versus 74%, P ¼ 0.75; data not shown). The causes of all first-year graft losses and patient deaths are reported in Table 4. There were no identifiable differences between the study groups. The 1-year rate of acute allograft rejection was less than 5%, and there were 2 patients who lost their grafts to chronic rejection in the first year after transplantation. The majority of patients were found to have a mixed cellular infiltrate at the time of fascial closure biopsy (between days 3 and 5), and this was consistent with early changes of rejection related to their delayed immunosuppression state. Posttransplant renal function is reported in Fig. 4. All groups experienced a decline in renal function in

5 790 MANGUS ET AL. LIVER TRANSPLANTATION, July 2012 TABLE 2. Graft and Patient Survival Stratified by the ratg Induction Protocol ratg-or ratg-d Immunosuppression Era* ratg-d-ritux ( ) ( ) ( ) P Value Intraoperative death (%) Graft loss within 7 days (%) year graft survival (%) year patient survival (%) HCV recurrence with 1 year (%) Any HCV recurrence (%) <0.01 HCC recurrence within 1 year [n/n (%)] 3/28 (11) 0/47 (0) 6/143 (4) 0.08 Any HCC recurrence with a median follow-up of 52 months (%) HCC exceeding the Milan criteria (%) Acute cellular rejection within 1 year [n (%)] 1(<1) 5 (2) 6 (1) 0.45 *In the ratg-or era, ratg induction was initiated intraoperatively at the time of reperfusion; in the ratg-d era, the ratg induction protocol was delayed until 48 hours after transplantation; and in the ratg-d-ritux era, the ratg induction protocol was delayed until 48 hours after transplantation, and rituximab was administered 3 days after transplantation. Defined as progression to F2 fibrosis (protocol biopsy at 4 and 12 months). Reported for 840 patients who were followed locally (83% of the total). Figure 1. Cox proportional hazards model for adjusted posttransplant graft survival with different immunosuppression induction protocols: ratg-or (n ¼ 166), ratg-d (n ¼ 259), and ratg-d-ritux (n ¼ 588). the first year after transplantation (mean decrease ¼ 14 ml/minute/1.73 m 2, median decrease ¼ 14 ml/ minute/1.73 m 2 ). This decline was greatest at all time periods for the ratg-d group versus the other groups (18 versus 14 and 14 ml/minute/1.73 m 2 ). The study groups differed only slightly in their risk of infectious Figure 2. Cox proportional hazards model for adjusted liver allograft survival with different immunosuppression induction protocols among patients infected with HCV at the time of LT (n ¼ 459). complications (Table 5). The overall risk of any fungal infection within 1 year of transplantation was 4%, and this did not differ between the groups. The most common fungal pathogen was Candida, with the most common forms being oral and urinary candidiasis. Cytomegalovirus (CMV) infections were seen in 2% of the patients within 1 year, and the incidence did not

6 LIVER TRANSPLANTATION, Vol. 18, No. 7, 2012 MANGUS ET AL. 791 a positive result (33%). No patient who received the ratg induction protocol a second time was noted to have any adverse sequelae from the administration of the agent. Figure 3. Cox proportional hazards model for adjusted patient survival with different immunosuppression induction protocols among patients with HCC at the time of LT with a median followup of 63 months (n ¼ 218). differ between the groups. The overall risk of any bacterial infection was 29%, with the most common being urinary infections (8%) and wound infections (7%). Pneumonia (6%), Clostridium difficile colitis (6%), and intra-abdominal abscesses (4%) were also important sources of infection. The incidence of PTLD was negligible [n ¼ 4(<1%)] and did not differ between the groups. Additionally, the risks of postoperative myocardial infarction, stroke, and DVT were similar to those reported in the literature. Side effects related to the administration of ratg were common and included shaking chills, tachycardia, fever, and hypoxia. Clinically significant pulmonary edema was documented in 13 patients (1.3%), and these events did rarely result in a need for intubation and ventilator support. Acute, severe pulmonary edema likely contributed to an early death for 1 elderly patient (0.1%). There were no documented episodes of hemodynamic compromise. Most systemic reactions to ratg administration occurred during the first 1 to 2 hours of the first infusion, and reactions thereafter were very rare. Although there were mild infusion-related reactions to rituximab, these side effects were mild, were easily controlled, and had no identified long-term effects. During the study period, 42 patients who had previously undergone LT underwent transplantation. In this group, 39 received the standard induction protocol, but 3 did not because of an early postoperative death (2) or the surgeon s choice (1). Twelve of the 42 re-lt patients were tested for anti-rabbit antibody before the second administration of ratg, and 4 had DISCUSSION The ideal induction immunosuppressive protocol for LT would result in good patient and graft survival with low rates of rejection, viral infection, fungal infection, and PTLD. The protocol would also have minimal renal toxicity and little adverse impact on blood pressure and lipid metabolism. Additionally, the ideal protocol would allow for minimal chronic immunosuppression and promote a situation favoring the promotion of tolerance. Horse anti-lymphocyte serum was used in the earliest LT series with azathioprine and corticosteroids. This regimen was insufficient to reliably prevent the rejection of liver allografts, and anti-lymphocyte preparations ultimately fell out of favor with the advent of the calcineurin inhibitors cyclosporine and tacrolimus. The addition of calcineurin inhibitors to an anti-lymphocyte agent, an antiproliferative agent, and corticosteroids resulted in significant increases in fibrosing cholestatic HCV, PTLD, and viral and fungal infections after transplantation. The role of biological induction therapy in modern LT is not well defined. Our results demonstrate that ratg-based induction immunosuppression is safe and can yield good results. The use of high-dose immunosuppression in the immediate postoperative period in animal models in which the graft is usually accepted without immunosuppression is associated with the inhibition of graft tolerance. 12,13 Immune activation, rather than the prevention of the initial immune response to the graft, is associated with long-term tolerance in rat liver allografts. We hypothesize that the delayed introduction of immunosuppression might allow the necessary immune activation to occur and result in significant reductions in the requirement for chronic long-term immunosuppression after LT. 14 Results from our clinical series provide several important points regarding induction immunosuppression for LT. ratg can be used with tacrolimus monotherapy in LT and yield low rates of rejection, CMV, PTLD, fungal infection, and severe HCV recurrence. Eason et al. 6 successfully used ratg induction in 119 LT patients who were randomized to receive either a steroid-free regimen with 2 doses of ratg induction or a 3-month steroid regimen. All patients received additional immunosuppression with tacrolimus and mycophenolate mofetil. The 1- and 2-year graft and patient survival rates and the rates of acute rejection were equivalent for the ratg and steroid groups. ratg was not associated with increased rates of infections, and notably, the steroid-treated patients experienced significantly more CMV infections (23% versus 5%, P < 0.05). Similarly, we have reported good success with ratg-based immunosuppression induction and low rejection and complication rates, as have other centers. 3,4 Early

7 792 MANGUS ET AL. LIVER TRANSPLANTATION, July 2012 TABLE 3. Cox Proportional Hazards Covariates for Figures 1 to 3 Figure 1. All Patients: Adjusted Graft Survival (n ¼ 1013) Model Covariate P Value HR 95% CI ratg-or era 0.36 Reference ratg-d era ratg-d-ritux era MELD score at transplant HCV infection Donor age (years) Recipient age (years) DCD Figure 2. Patients With HCV: Adjusted Graft Survival (n ¼ 459) Model Covariate P Value HR 95% CI ratg-or era 0.97 Reference ratg-d era ratg-d-ritux era MELD score at transplant Donor age (years) Recipient age (years) DCD Figure 3. Patients With HCC: Adjusted Patient Survival (n ¼ 218) Model Covariate P Value HR 95% CI ratg-or era 0.96 Reference ratg-d era ratg-d-ritux era HCV infection MELD score at transplant Donor age (years) Recipient age (years) DCD Outside the Milan criteria results showed that recurrent HCV was exacerbated with OKT3 induction with triple therapy versus triple immunosuppression with calcineurin inhibition, azathioprine, and steroids. 15 These results suggest that ratg induction should be accompanied by a reduction in calcineurin trough levels and a reduction or even omission of steroids and antiproliferative agents. Much of the use of ratg in LT in this decade has been centered on the minimization of chronic immunosuppression. Recent data from the United Network for Organ Sharing database suggest that graft survival is improved in HCV-negative recipients who receive ratg and is not adversely affected in HCV-positive recipients. 16 We have demonstrated the feasibility of delaying the introduction of immunosuppression, which has a number of applications and patient benefits. The background for delaying the introduction of immunosuppression lies in critically ill recipients with hepatorenal syndrome; for these patients, it would be desirable to allow their renal function to return before the introduction of nephrotoxic calcineurin inhibitors as part of a maintenance immunosuppressive regimen. Numerous reports have demonstrated the safety of this practice. 4,5 It is also desirable to delay the introduction of immunosuppression in LT patients who have infections at the time of transplantation. This would allow the infections to be cleared with a new liver and antibiotics before the introduction of immunosuppression. A delay in the introduction of immunosuppression may also promote graft acceptance. In rat models of LT, in which there is spontaneous graft acceptance with a complete major histocompatibility complex mismatch, the introduction of corticosteroids in the immediate posttransplant period eliminates the immunosuppression-free tolerant state. 13 Rodent models of liver allograft tolerance require the immune activation, apoptosis, and clearance of recipient T cells in the donor liver for the development of tolerance. Immunosuppression in the posttransplant period immediately after graft reperfusion would impair the immune activation required to pursue this type of tolerance. Initially, we considered developing a trial to remove tacrolimus from our patients, but after

8 LIVER TRANSPLANTATION, Vol. 18, No. 7, 2012 MANGUS ET AL. 793 TABLE 4. Patient Deaths and Graft Losses 1 Year After LT Stratified by the Immunosuppression Induction Era ratg-or Immunosuppression Era* ratg-d ratg-d-ritux Overall ( ) ( ) ( ) Patients [n (%) or n] 1013 (100) year patient deaths [n (%)] 136 (13) 21 (13) 42 (16) 73 (12) Transplant-related complications (perioperative) 49 (5) 8 (5) 17 (7) 24 (4) Sepsis/infection 26 (3) 4 (2) 10 (4) 12 (2) Liver disease recurrence/progression 17 (2) 2 (1) 4 (2) 11 (2) Non-HCC cancer 13 (1) 1 (<1) 2 (<1) 10 (2) Cardiac event 14 (1) 2 (1) 5 (2) 7 (1) Recurrent HCC 3 (<1) 0 (0) 0 (0) 3 (<1) Stroke 4 (<1) 2 (1) 1 (<1) 1 (<1) Lung disease 3 (<1) 0 (0) 1 (<1) 2 (<1) Complications of renal failure 3 (<1) 1 (<1) 1 (<1) 1 (<1) Other/unknown 3 (<1) 1 (<1) 1 (<1) 1 (<1) Chronic allograft rejection 1 (<1) 0 (0) 0 (0) 1 (<1) 1-year graft loss only [n (%)] 17 (2) 4 (2) 5 (2) 8 (1) Primary nonfunction 5 (<1) 2 (1) 0 (0) 3 (<1) Vascular complication/thrombosis 5 (<1) 1 (<1) 3 (1) 1 (<1) Chronic biliary complications 5 (<1) 0 (0) 1 (<1) 4 (<1) Liver disease recurrence/progression 1 (<1) 0 (0) 1 (<1) 0 (0) Chronic allograft rejection 1 (<1) 1 (<1) 0 (0) 0 (0) NOTE: P values were not calculated for this table because the small numbers resulted in invalid statistical analyses. *In the ratg-or era, ratg induction was initiated intraoperatively at the time of reperfusion; in the ratg-d era, the ratg induction protocol was delayed until 48 hours after transplantation; and in the ratg-d-ritux era, the ratg induction protocol was delayed until 48 hours after transplantation, and rituximab was administered 3 days after transplantation. Figure 4. Renal function 1 year after LT stratified by the ratg induction protocol and measured with GFR. Combined liverkidney transplant patients were excluded from this analysis. Pittsburgh was unable to routinely withdraw immunosuppression from a similarly treated group of patients, we decided against moving forward. 17 Our results clearly show that anyone wishing to pursue tolerance protocols involving delays in the introduction of immunosuppression to facilitate immunological engagement can do so without risking graft loss or causing harm to the patient. A third important point from our analysis is that ratg can be used in conjunction with rituximab without significant increases in CMV, PTLD, fungal infections, or HCV recurrence. Rituximab is a genetically engineered, chimeric, murine/human monoclonal antibody directed against the CD20 antigen found on the surface of normal and malignant pre-b and mature B cells. CD20 þ B cells have been implicated as important cells in both liver and renal allograft rejection. In renal allograft rejection, CD20 þ graftinfiltrating cells were associated with steroid-resistant rejection and graft failure. 18 B cells are also important antigen-presenting cells in the allograft response, and the goal of our protocol was to reset as completely as possible the alloantigen response after transplantation. The dose that we used (150 mg/m 2 ) was based on pharmacokinetic studies performed by Vieira et al., 19 who showed similar areas under the receiver operating characteristic curve for doses of 150 and 375 mg/m 2. The low rate of rejection seen in our series is the result of a host of factors and has implications for clinical LT. Our results show that in the face of sufficient immunosuppression, the rejection of a liver allograft is an unusual event, and resistant or difficult-to-treat rejection is very rare if it occurs at all. Elevations in serum transaminases, alkaline phosphatase, or bilirubin in the post-lt patient are far more likely to be caused by biliary tract issues, vascular problems, recurrent liver disease, or idiosyncratic reactions to medication. This information is very important to consider when we are caring for a patient

9 794 MANGUS ET AL. LIVER TRANSPLANTATION, July 2012 TABLE 5. Post-LT Complications Stratified by the Immunosuppression Induction Protocol ratg-or Immunosuppression Era* ratg-d ratg-d-ritux Overall ( ) ( ) ( ) P Value Patients [n (%) or n] 1013 (100) year infection risk [n (%) or %] Bacterial infection (any) 294 (29) Urinary infection 80 (8) <0.01 Wound infection 76 (8) Bacterial pneumonia (sputum) 61 (6) C. difficile (stool) 60 (6) <0.01 Intra-abdominal abscess 38 (4) <0.01 Cellulitis (any) 21 (2) Sinusitis/bronchitis (any) 28 (3) Fungal infection 39 (4) CMV infection 25 (2) Non-CMV virus 12 (1) PTLD 4 (<1) <1 <1 <1 Perioperative complications (%) Stroke Postoperative cardiac event <0.01 DVT *In the ratg-or era, ratg induction was initiated intraoperatively at the time of reperfusion; in the ratg-d era, the ratg induction protocol was delayed until 48 hours after transplantation; and in the ratg-d-ritux era, the ratg induction protocol was delayed until 48 hours after transplantation, and rituximab was administered 3 days after transplantation. Complications within the first 30 days after transplantation. with persistently elevated liver enzymes in the face of sufficient immunosuppression. In this situation, the problem is rarely rejection, and treating it as such results in overimmunosuppression and its attendant problems and fails to address the real cause of the injury to the allograft. Our approach is based on Demetris et al. s series, 20 which showed that the risk of cellular rejection was overestimated in the immunosuppression weaning protocol for recipients receiving liver allografts for HCV-induced liver disease. Our approach uses a carefully monitored reduction in the goal tacrolimus levels in HCV-positive patients with elevated liver chemistries. Aggressive HCV recurrence usually is the result of large perturbations in immunosuppression (usually too much) after the first 2 to 3 weeks after transplantation. Idiosyncratic drug reactions can cause marked elevations in liver chemistries and on liver biopsy yield few clues about why the numbers are elevated. The decreased rates of recurrent HCV, CMV, PTLD, and fungal infection occurred because of our approach to immunosuppression rather than the intrinsic properties of any single immunosuppressive agent. Our overall 12-month CMV infection rate of 2% (6% in the highest risk group, donor positive and recipient negative) was lower than the rates reported in the literature (4%-27%) and could be attributed to 2 factors. First, all our patients received CMV prophylaxis with valganciclovir for 3 months after transplantation, and second, we were able to eliminate chronic steroids and maintain modest tacrolimus trough levels in the immediate postoperative period (<10 ng/ml) without an antiproliferative immunosuppressive agent (mycophenolate or azathioprine). The lower dose of tacrolimus and the absence of a second immunosuppressive agent could also explain our low fungal infection rate. Our PTLD rate (<1%) was also not adversely affected by the use of ratg: the published rates for PTLD in LT patients are approximately 1% to 3%. 24 In conclusion, we have reviewed more than 1000 consecutive LT recipients (deceased donor and whole organ) during a recent period ( ) with a minimum follow-up of 18 months and a good median follow-up time of 63 months. This is the largest reported series of transplant patients who received ratg-based induction immunosuppression. The reported data suggest that the administration of ratg therapy, either immediately at the time of transplantation or in a delayed fashion, can result in excellent clinical outcomes, including a low incidence of rejection and an acceptable side-effect profile. These results, however, should be viewed in light of their epidemiological limitations: this is a retrospective review of data with no control groups and no randomization. The gold standard remains a randomized controlled trial for comparing induction to no induction and for comparing different induction agents. The addition of rituximab may provide additional benefit; improved graft and patient survival was seen in the ratg-d-ritux group, although this survival advantage did not reach statistical significance. The interpretation of this improved

10 LIVER TRANSPLANTATION, Vol. 18, No. 7, 2012 MANGUS ET AL. 795 survival is complex because of center changes over time, but the year of transplant as a variable was not significant, did not affect the final model, and was excluded. The lack of a demonstrable benefit from the routine administration of rituximab as part of an induction protocol leaves this point open to discussion. The use of biological induction has also been limited in LT because of concerns about overimmunosuppression and opportunistic infections. Our results clearly show that ratg and rituximab can be used without excessive risk of CMV, PTLD, or serious fungal infections. Our series is also interesting with respect to the timing of immunosuppression. Results from our center demonstrate no negative impact on patient or graft survival from the initial withholding of all immunosuppression for up to 48 hours after LT. This may carry important implications both for those centers pursuing clinical tolerance studies and for clinicians desiring to withhold immunosuppression for clinical reasons, including infection or poor renal function. We were unable to use ratg without a steroid pretreatment because of pulmonary complications (hypoxia from pulmonary edema). Eason et al. 6 were able to routinely give ratg without steroids with very good results. This discrepancy in the clinical experience may have resulted from a more vigorous lymphocyte reaction in our patients, who had not received any initial immunosuppression. REFERENCES 1. Rosen HR, Shackleton CR, Higa L, Gralnek IM, Farmer DA, McDiarmid SV, et al. Use of OKT3 is associated with early and severe recurrence of hepatitis C after liver transplantation. Am J Gastroenterol 1997;92: Aucejo F, Rofaiel G, Miller C. Who is at risk for posttransplant lymphoproliferative disorders (PTLD) after liver transplantation? J Hepatol 2006;44: Tector AJ, Fridell JA, Mangus RS, Shah A, Milgrom M, Kwo P, et al. Promising early results with immunosuppression using rabbit anti-thymocyte globulin and steroids with delayed introduction of tacrolimus in adult liver transplant recipients. Liver Transpl 2004;10: Tchervenkov JI, Tzimas GN, Cantarovich M, Barkun JS, Metrakos P. The impact of thymoglobulin on renal function and calcineurin inhibitor initiation in recipients of orthotopic liver transplant: a retrospective analysis of 298 consecutive patients. Transplant Proc 2004;36: Bajjoka I, Hsaiky L, Brown K, Abouljoud M. Preserving renal function in liver transplant recipients with rabbit anti-thymocyte globulin and delayed initiation of calcineurin inhibitors. Liver Transpl 2008;14: Eason JD, Nair S, Cohen AJ, Blazek JL, Loss GE Jr. Steroid-free liver transplantation using rabbit antithymocyte globulin and early tacrolimus monotherapy. Transplantation 2003;75: Soliman T, Hetz H, Burghuber C, Gy ori G, Silberhumer G, Steininger R, et al. Short-term induction therapy with anti-thymocyte globulin and delayed use of calcineurin inhibitors in orthotopic liver transplantation. Liver Transpl 2007;13: De Ruvo N, Cucchetti A, Lauro A, Masetti M, Di Benedetto F, Cautero N, et al. Minimization of immunosuppression with thymoglobuline pre-treatment and HCV recurrence in liver transplantation. Clin Transplant 2005;19: Mangus RS, Fridell JA, Vianna RM, Kwo PY, Chestovich P, Milgrom ML, et al. No difference in clinical transplant outcomes for local and imported liver allografts. Liver Transpl 2009;15: Tector AJ, Mangus RS, Chestovich P, Vianna R, Fridell JA, Milgrom ML, et al. Use of extended criteria livers decreases wait time for liver transplantation without adversely impacting posttransplant survival. Ann Surg 2006;244: Gonwa TA, Jennings L, Mai ML, Stark PC, Levey AS, Klintmalm GB. Estimation of glomerular filtration rates before and after orthotopic liver transplantation: evaluation of current equations. Liver Transpl 2004;10: Sharland A, Yan Y, Wang C, Bowen DG, Sun J, Sheil AG, et al. Evidence that apoptosis of activated T cells occurs in spontaneous tolerance of liver allografts and is blocked by manipulations which break tolerance. Transplantation 1999;68: Wang C, Sun J, Sheil AG, McCaughan GW, Bishop GA. A short course of methylprednisolone immunosuppression inhibits both rejection and spontaneous acceptance of rat liver allografts. Transplantation 2001;72: Bishop GA, McCaughan GW. Immune activation is required for the induction of liver allograft tolerance: implications for immunosuppressive therapy. Liver Transpl 2001;7: Sheiner PA, Schwartz ME, Mor E, Schluger LK, Theise N, Kishikawa K, et al. Severe or multiple rejection episodes are associated with early recurrence of hepatitis C after orthotopic liver transplantation. Hepatology 1995;21: Moonka DK, Kim D, Kapke A, Brown KA, Yoshida A. The influence of induction therapy on graft and patient survival in patients with and without hepatitis C after liver transplantation. Am J Transplant 2010;10: Eghtesad B, Fung JJ, Demetris AJ, Murase N, Ness R, Bass DC, et al. Immunosuppression for liver transplantation in HCV-infected patients: mechanism-based principles. Liver Transpl 2005;11: Martins HL, Silva C, Martini D, Noronha IL. Detection of B lymphocytes (CD20þ) in renal allograft biopsy specimens. Transplant Proc 2007;39: Vieira CA, Agarwal A, Book BK, Sidner RA, Bearden CM, Gebel HM, et al. Rituximab for reduction of anti-hla antibodies in patients awaiting renal transplantation: 1. Safety, pharmacodynamics, and pharmacokinetics. Transplantation 2004;77: Demetris AJ, Eghtesad B, Marcos A, Ruppert K, Nalesnik MA, Randhawa P, et al. Recurrent hepatitis C in liver allografts: prospective assessment of diagnostic accuracy, identification of pitfalls, and observations about pathogenesis. Am J Surg Pathol 2004;28: Paya C, Humar A, Dominguez E, Washburn K, Blumberg E, Alexander B, et al.; for Valganciclovir Solid Organ Transplant Study Group. Efficacy and safety of valganciclovir vs. oral ganciclovir for prevention of cytomegalovirus disease in solid organ transplant recipients. Am J Transplant 2004;4: Arthurs SK, Eid AJ, Pedersen RA, Dierkhising RA, Kremers WK, Patel R, Razonable RR. Delayed-onset primary cytomegalovirus disease after liver transplantation. Liver Transpl 2007;13: Park JM, Lake KD, Arenas JD, Fontana RJ. Efficacy and safety of low-dose valganciclovir in the prevention of cytomegalovirus disease in adult liver transplant recipients. Liver Transpl 2006;12: Jain A, Nalesnik M, Reyes J, Pokharna R, Mazariegos G, Green M, et al. Posttransplant lymphoproliferative disorders in liver transplantation: a 20-year experience. Ann Surg 2002;236: