Does Race Influence Outcomes after Primary Liver Transplantation? A 23-Year Experience with 2,700 Patients
Johnny C. Hong,Kambiz Kosari,Elizabeth Benjamin,John P. Duffy,R. Mark Ghobrial,Douglas G. Farmer,Hasan Yersiz,Junming Xu,Jonathan R. Hiatt,Ronald W. Busuttil
DOI: https://doi.org/10.1016/j.jamcollsurg.2007.12.019
2008-01-01
Journal of the American College of Surgeons
Abstract:Study Design This was a retrospective case series. Among 2,728 adult patients who underwent primary LT from 1984 to 2007, 1,566 (57%) were Caucasian, 761 (28%) were Hispanic, 290 (11%) were Asian, and 111 (4%) were AA. The primary immunosuppressive agent was cyclosporine from 1984 to 1993 (Era I, n = 817) and tacrolimus from 1994 to 2007 (Era II, n = 1922). Results In Era I, the 1-, 5- and 10-year patient and graft survival figures for AA and Asian recipients were considerably lower compared with Caucasian and Hispanic recipients. In Era II, patient and graft survival figures were comparable for all groups. Statistically significant independent predictors of diminished patient survival included LT in Era I; recipient or donor age greater than 55 years; and liver failure secondary to cryptogenic cirrhosis, malignancy, or hepatitis C. Predictors of graft failure included LT in Era I; recipient or donor age greater than 55 years; prolonged cold ischemia time; liver failure secondary to hepatitis C, cryptogenic cirrhosis, or malignancy; and acute rejection. Patient and graft survival were independent of race in both eras. Conclusions This is the first study to demonstrate equivalent longterm results after LT for AA and other races. Modern immunosuppression with tacrolimus substantially lowered rejection rates and improved graft and patient survival after LT. Abbreviations and Acronyms AA African American CyA cyclosporine DGF delayed graft function ESLD end stage liver disease LT liver transplantation PNF primary graft nonfunction TAC tacrolimus For many diseases, there are racial differences in incidence, severity, methods of treatment, and access to care. In 2002, the Institute of Medicine identified racial disparities in treatment of cancer, cardiovascular disease, HIV/AIDS, diabetes, mental health, and other chronic and subacute diseases. 1 End stage liver disease (ESLD), the 12th leading cause of death in the US, 2 is another example of the problem. Among African-American (AA) patients, both incidence and mortality rates for cirrhosis have remained substantially higher than for other ethnic groups despite an overall decline in cirrhosis death rates since 1973. 3 Although liver transplantation (LT) has been established as a durable therapy for all forms of ESLD, 4,5 several studies have identified racial barriers to LT, 6-8 and AA patients appear to be under-represented among LT recipients. In 2005, only 6.8% of all patients on the liver waiting list and 9.4% of LT recipients were AA, compared with the US population, in which AA represented 12.9% of the total. In contrast, the fraction of Hispanics on the waiting list (16%) has nearly tripled in the last decade, and Hispanics comprised 13% of LT recipients and 12.5% of the US population. 9 Success rates for transplantation also are related to race, 10 with studies showing poorer outcomes for renal transplantation in AA patients when compared with outcomes in other races. 11,12 Data concerning the influence of race on survival outcomes after LT are limited and conflicting. 7,8,13,14 These reports suffer from small sample size and may not reflect current management of LT recipients, with refinement in surgical techniques, improvement in perioperative care, and advances in immunosuppression. 5,15 This large single center study was undertaken to evaluate longterm outcomes for LT in AA recipients compared with those in other races and to determine factors responsible for any observed differences. Methods Data collection Using a prospectively collected transplant database, we performed a retrospective review of 2,844 patients 19 years of age and older who underwent primary LT at the Dumont-UCLA Transplant Center from February 1, 1984 through February 28, 2007. Races other than AA, Caucasian, Hispanic, or Asian (n = 116) were excluded, leaving 2,728 patients for analysis. The study was approved by the UCLA Institutional Review Board. The median followup time was 4 years. Primary liver diseases Among the patients included in the analysis, diagnoses included hepatitis C (n = 559), alcoholic liver disease (n = 547), malignancy (n = 480), cholestatic disorders (n = 355), cryptogenic cirrhosis (n = 259), acute liver failure (n = 178), hepatitis B (n = 153), and other disorders (n = 197). The diagnosis of malignancy included hepatocellular carcinoma (n = 461) and other tumors (n = 19). Cholestatic disorders included primary biliary cirrhosis (n = 192) and primary sclerosing cholangitis (n = 163). Patient management All patients with ESLD were evaluated for LT by a multidisciplinary team, as previously described. 5 From 1984 to 2001, patients were listed according to the United Network for Organ Sharing (UNOS) status categories; from 2002 to the present, the current Model for End Stage Liver Disease (MELD) system was used. 16 Results were analyzed by transplantation era according to primary immunosuppressive agent. Era I (1984 to 1993) The primary immunosuppressive agent was cyclosporine (CyA, Sandimmune, or Neoral, Novartis Pharmaceuticals). A double regimen with prednisone was used from 1984 to 1987, and a triple regimen with addition of azathioprine (Imuran, Prometheus Laboratories Inc) from 1987 to 1993. Era II (1994 to 2007) The primary immmunosuppressive agent was tacrolimus (Prograf, Astellas Pharmaceuticals Inc). Most patients received triple immunotherapy with steroids and either azathioprine or mycophenolate mofetil (CellCept, Roche Pharmaceuticals). In select instances, patients received induction therapy with OKT3 (Orthoclone, Orth Biotech Products), or daclizumab, an interleukin-2 receptor antibody (Zenapax, Roche Pharmaceuticals). Beginning in 1995, steroids were discontinued 3 to 6 months after LT in hepatitis C patients who did not exhibit rejection episodes. All patients with evidence of hepatitis B virus infection before LT received a combination of longterm immune prophylaxis with hepatitis B immunoglobulin (HBIg, North American Biologicals) and antiviral therapy with lamividine (Epivir, GlaxoSmithKline) as previously described. 17 Statistical analysis Patient and graft survival curves were computed using Kaplan-Meier methods and compared using log rank tests. Medians were compared using the Wilcoxon test and proportions using the chi-square test. Both univariate and multivariate analyses were conducted using Cox's proportional hazard model. Backward stepwise procedure was used for variables selection, with retention criteria at a p value of ≤ 0.25 level of significance. In the multivariate analysis, a p value of < 0.05 was considered significant. Statistical analysis was performed using SAS software, version 9.1 (SAS Institute). Results Recipient characteristics Among the 2,728 adult patients who underwent primary LT over the 23-year study period, 1,566 (57%) were Caucasian, 761 (28%) were Hispanic, 290 (11%) were Asian, and 111 (4%) were AA. Patient characteristics and distribution of diagnoses for both eras are shown in Table 1 . In Era I, AA LT recipients were younger than those of other races and more often required urgent transplantation for acute liver failure. Cryptogenic cirrhosis also was more common, but malignancy, hepatitis B, and alcohol-induced cirrhosis were less common in AA recipients. In Era II, AA recipients were more often women than in other races. Urgent transplantation for acute liver failure remained more common among AA recipients, although frequency of cryptogenic cirrhosis declined. Malignancy was a more frequent indication for liver transplantation in Asian recipients in both eras because of the prevalence of vertically transmitted hepatitis B virus infection. Comparing AA recipients in both eras, causes of liver failure that have been shown to portend a worse prognosis were similar. These included hepatitis C (p = 0.100), acute liver failure (p = 0.845), and malignancy (p = 0.424). Donor and operative characteristics Organ donors for AA recipients were younger than those for other races in Era I (p = 0.026). Other donor and operative characteristics were similar for all groups both within and between eras, including use of extended criteria donors, 18 donor type (deceased or living), and graft type (whole or partial). Only deceased donors and whole grafts were used in Era I; similar numbers of living donors and partial grafts were used among the racial groups in Era II. We have identified four extended criteria (donor age greater than 55 years, donor hospital stay greater than 5 days, cold ischemia time longer than 10 hours, and warm ischemia time greater than 40 minutes) to predict outcomes after LT. 18 There was no statistical difference among racial groups for any of these factors in either era. Complications In both eras, rates of hepatic artery thrombosis (1.9% versus 2.2%; p = 0.631) and portal vein thrombosis (0.8% versus 0.6%; p = 0.478) were comparable among all races. There were more biliary complications in Era II versus Era I (1.9% versus 6.3%, p < 0.0001) because of the use of partial grafts either from living or deceased donors. In Era I, biliary complications occurred in 11% of AA recipients, compared with 3% for Caucasian, Hispanic, and Asian recipients (p < 0.006), although the rates were similar among all races in Era II. Primary graft nonfunction (PNF) and delayed graft function (DGF) occurred more frequently in Era I compared with Era II (9.1% versus 6.5%; p = 0.015). Although there were no differences for PNF and DGF by racial group in Era I, both occurred more frequently in Asian recipients than in AA, Caucasian, and Hispanic recipients (10% versus 6% versus 5% versus 8%; p = 0.013) in Era II. More patients required retransplantation in Era I compared with Era II (16% versus 10%; p < 0.0001), but the rates were similar among racial groups (Era I, p = 0.81; Era II, p = 0.34). Retransplantation occurred with similar frequency among all racial groups for PNF and DGF in both Era I (p = 0.16) and Era II (p = 0.15). Rejection Table 2 compares ratios of incidence rates for acute and chronic rejection. There were racial differences in both eras. Acute rejection In Era I, both AA and Hispanic recipients were at markedly increased risk for acute rejection, compared with Caucasian recipients. In Era II, the rate ratio for acute rejection was higher for AA recipients, but the difference was not notable. Most of the acute rejection episodes were reversed with pulse steroid treatment. There was no difference in the incidence of steroid-resistant acute rejection by era (Era I, 3.8% versus Era II, 4.9%; p = 0.642) or the distribution among racial groups (p = 0.81). Chronic rejection In Era I, Asian recipients were at increased risk for chronic rejection, compared with Caucasian recipients, although the rate declined in Era II. Graft failure secondary to chronic rejection was more frequent in AA than Caucasian recipients both in Era I (29% versus 10.2%) and Era II (14% versus 9%), but the differences were not statistically significant. Patient survival Patient survival curves for both eras are compared in Figure 1 . In Era I, the 1-, 5- and 10-year survivals for AA (76%, 54%, and 50%) and Asian recipients (53%, 40%, and 36%) were considerably lower compared with those for Caucasian (83%, 70%, and 61%) and Hispanic recipients (84%, 74%, and 68%). In Era II, survival figures were comparable for all groups: AA (78%, 64%, and 64%), Asian (83%, 73%, and 68%), Caucasian (87%, 73%, and 65%), and Hispanic (86%, 75%, and 70%). Multivariate analysis of overall patient survival is shown in Table 3 . Statistically significant independent predictors of diminished survival included LT in Era I; recipient or donor age greater than 55 years; and liver failure secondary to cryptogenic cirrhosis, malignancy, or hepatitis C. Race was not a predictor of patient survival. Graft survival Graft survival curves for both eras are compared in Figure 2 . In Era I, the 1-, 5- and 10-year graft failure-free survival rates for AA (67%, 44%, and 41%) and Asian recipients (46%, 34%, and 30%) were considerably lower compared with Caucasian (70%, 57%, and 49%) and Hispanic (71%, 59%, and 54%) recipients. In Era II, graft survival figures were comparable for all groups: AA (69%, 53%, and 48%), Asian (72%, 61%, and 56%), Caucasian (79%, 64%, and 57%), and Hispanic (76%, 66%, and 60%). A multivariate analysis of overall graft survival is shown in Table 4 . Predictors of graft failure included LT in Era I; recipient and donor age greater than 55 years; prolonged cold ischemia time; ESLD secondary to hepatitis C, cryptogenic cirrhosis, or malignancy; and acute rejection. Race was not a predictor of graft survival. Survival by era Table 5 compares patient and graft survival by era. In Era I, AA recipients and patients with chronic rejection had diminished patient survival, and graft survival was diminished in patients with acute rejection. In Era II, neither race nor chronic rejection affected patient survival. Graft survival was independent of race, but was diminished with acute rejection. Discussion This series is the largest single-center analysis investigating the role of race as a determinant of outcomes after primary liver transplantation. Earlier data on the subject have been conflicting. Single-center studies by Eckhoff and colleagues, 6 Dawson and associates, 13 and Gonwa and coworkers 8 demonstrated no short-term survival disparities between AA and Caucasian LT recipients; Nair and coauthors 14 used pooled data from the United Network of Organ Sharing (UNOS) transplant registry and reported inferior outcomes in AA and Asians compared with Caucasians and Hispanics. Our study showed equivalent longterm results after LT in AA and other races with tacrolimus-based immunosuppression. Acute rejection, transplantation era, and liver failure from hepatitis C, cryptogenic cirrhosis, and malignancy affected survival negatively; race did not. Our finding of an increased frequency of LT for acute liver failure in AA recipients is consistent with results from a previous report. 14 We and others 4,5,19-21 have reported that outcomes after LT for acute liver failure in adults were inferior compared with outcomes for chronic liver diseases. The principal causes of graft failure were primary nonfunction or delayed function because of a combination of pretransplant morbidity and the likelihood of receiving grafts from extended criteria donors. 22,23 We found no differences in the frequency of PNF or DGF between AA and Caucasian recipients. Racial disparities in outcomes after transplantation have been studied more extensively for renal replacement. Both pre- and posttransplant factors have been implicated to explain the observations, although their relevance to LT remains largely uninvestigated. Regarding pretransplant phenomena, AA renal transplant recipients have traditionally been thought to have higher immunologic risk for rejection and graft loss when compared with Caucasian recipients. 10,24-29 AA recipients were likely to receive deceased-donor allografts bearing a greater number of mismatched donor-recipient HLA or Lewis blood group antigens because of the predominance of Caucasian organ donors. 30 They also tended to exhibit heightened immune responsiveness because of the high degree of antigen polymorphism compared with that in Caucasian recipients. 25,28 In addition, AA recipients demonstrated a higher frequency of presensitization, explained by transfusion with Caucasian blood products. 28 After renal transplantation, AA recipients had inferior graft survival partly because of distinct pharmacokinetic and pharmacodynamic risk factors. AA recipients metabolize CyA and tacrolimus more rapidly than Caucasian recipients do, increasing the number of rejection episodes. 31,32 In addition, Vasquez and colleagues 33 reported higher failure rates with corticosteroid treatment of acute rejection in AA recipients compared with other races. The superiority of tacrolimus over CyA-based immunosuppression in reduction of frequency and severity of acute rejection episodes after LT has been demonstrated in multiple randomized clinical trials. 15,34 In addition, McAlister and associates 35 reported a survival benefit after LT with tacrolimus compared with CyA. Our study showed that tacrolimus substantially reduced the risk of acute rejection, graft loss, and patient death compared with CyA and resulted in a marked improvement of patient and graft survival outcomes in AA recipients after LT. Although steroid-resistant acute rejection remains a difficult problem after renal transplantation, 33,36,37 we found that most of the acute rejection episodes after LT could be reversed with pulse steroid treatment, and there was no difference in the distribution of acute rejection among racial groups. Despite considerable reduction in acute rejection episodes, chronic allograft rejection has been reported to be a major cause for late graft failure after LT. 38,39 Chronic rejection is attributed to multiple risk factors. Immunologic factors include autoimmune liver diseases, HLA mismatch, acute cellular rejection, low CyA trough level, and others, 40-43 and nonimmunologic factors include recipient age, female recipient with hepatic graft from male donor, and donor-recipient pairs of different ethnic origins. 44,45 Inadequate immunosuppression has also been identified as a factor for development of chronic graft rejection. Low trough levels of CyA (mean < 175 ng/mL) during the first month post-transplantation and immunosuppression without azathioprine have been implicated in numerous studies, 42,43,46 Although Nair and associates 14 reported a higher rate of graft failure from chronic rejection among AA patients compared with patients of other races, we found no difference between all racial groups in both eras. Several studies have reported that AA recipients are disproportionately under-represented in the LT population compared with other ethnic groups. 7-9 Analysis of the United Network for Organ Sharing (UNOS) database demonstrated that 7% of the population on the liver transplant waiting list were AA, although the rate of transplantation once on the list was similar for AA and Caucasians. 47 Eckhoff and coworkers 6 reported that referral bias accounted for disproportionately fewer AA patients undergoing evaluation for LT. This was not the case among our patients, where the number of AA recipients (4%) reflected the AA distribution in our population (6.7%). In summary, this is the first study to demonstrate equivalent longterm outcomes after liver transplantation for AA and other patients with comparable liver diseases. An overall reduction in acute rejection episodes and improvement in patient and graft survival were observed with use of tacrolimus as the primary immunosuppressive agent. Finally, we did not find that AA had diminished access to liver transplant evaluation or transplantation, but access to care clearly remains a major goal for all transplant centers. Outreach programs to improve identification of liver disease and early referral for transplantation should be pursued as part of a general effort to improve access to care for underserved populations. Author Contributions Study conception and design: Hong, Ghobrial, Farmer, Yersiz, Hiatt, Busuttil Acquisition of data: Hong, Kosari, Benjamin, Duffy, Xu Analysis and interpretation of data: Hong, Kosari, Duffy Drafting of manuscript: Hong, Kosari, Benjamin, Duffy, Ghobrial, Farmer, Yersiz, Xu, Hiatt, Busuttil Critical revision: Hong, Hiatt, Busuttil Acknowledgment We gratefully acknowledge Daniela Markovic, MS, for her expert statistical analysis of the data. References 1 A. Nelson Unequal treatment: confronting racial and ethnic disparities in health care J Natl Med Assoc 94 2002 666 668 2 A.M. Minino M.P. Heron B.L. Smith Deaths: preliminary data for 2004 Natl Vital Stat Rep 54 2006 1 49 3 Y.H. Yoon Surveillance Report #75 Liver Mortality in the United States 2006 NIAAA Washington, DC 4 M.S. Roberts D.C. Angus C.L. Bryce Survival after liver transplantation in the United States: a disease-specific analysis of the UNOS database Liver Transpl 10 2004 886 897 5 R.W. Busuttil D.G. Farmer H. Yersiz Analysis of long-term outcomes of 3200 liver transplantations over two decades: a single-center experience Ann Surg 241 2005 905 916 6 D.E. Eckhoff B.M. McGuire C.J. Young Race: a critical factor in organ donation, patient referral and selection, and orthotopic liver transplantation? Liver Transpl Surg 4 1998 499 505 7 A.E. Reid M. Resnick Y. Chang Disparity in use of orthotopic liver transplantation among blacks and whites Liver Transpl 10 2004 834 841 8 T.A. Gonwa C.A. Morris M.L. Mai Race and liver transplantation Arch Surg 126 1991 1141 1143 9 E.A. Pomfret J.P. Fryer C.S. Sima Liver and intestine transplantation in the United States, 1996–2005 Am J Transplant 7 2007 1376 1389 10 G. Opelz M.R. Mickey P.I. Terasaki Influence of race on kidney transplant survival Transplant Proc 9 1977 137 142 11 B.L. Kasiske J.F. Neylan 3rd R.R. Riggio The effect of race on access and outcome in transplantation N Engl J Med 324 1991 302 307 12 S. Katznelson D.W. Gjertson J.M. Cecka The effect of race and ethnicity on kidney allograft outcome Clin Transpl 1995 379 394 13 S. Dawson 3rd D.K. Imagawa C. Johnson UCLA liver transplantation: analysis of immunological factors affecting outcome Artif Organs 20 1996 1063 1072 14 S. Nair J. Eustace P.J. Thuluvath Effect of race on outcome of orthotopic liver transplantation: a cohort study Lancet 359 2002 287 293 15 A comparison of tacrolimus (FK 506) and cyclosporine for immunosuppression in liver transplantation The U.S. Multicenter FK506 Liver Study Group N Engl J Med 331 1994 1110 1115 16 R.H. Wiesner S.V. McDiarmid P.S. Kamath MELD and PELD: application of survival models to liver allocation Liver Transpl 7 2001 567 580 17 D.M. Anselmo R.M. Ghobrial L.C. Jung New era of liver transplantation for hepatitis B: a 17-year single-center experience Ann Surg 235 2002 611 619 18 A.M. Cameron R.M. Ghobrial H. Yersiz Optimal utilization of donor grafts with extended criteria: a single-center experience in over 1000 liver transplants Ann Surg 243 2006 748 753 19 H. Bismuth D. Samuel D. Castaing Orthotopic liver transplantation in fulminant and subfulminant hepatitis The Paul Brousse experience Ann Surg 222 1995 109 119 20 J.H. Hoofnagle R.L. Carithers Jr C. Shapiro N. Ascher Fulminant hepatic failure: summary of a workshop Hepatology 21 1995 240 252 21 J. Devlin J. Wendon N. Heaton Pretransplantation clinical status and outcome of emergency transplantation for acute liver failure Hepatology 21 1995 1018 1024 22 D.G. Farmer D.M. Anselmo R.M. Ghobrial Liver transplantation for fulminant hepatic failure: experience with more than 200 patients over a 17-year period Ann Surg 237 2003 666 675 23 K. Bjoro B.G. Ericzon P. Kirkegaard Highly urgent liver transplantation: possible impact of donor-recipient ABO matching on the outcome after transplantation Transplantation 75 2003 347 353 24 D.E. Eckhoff C.J. Young R.S. Gaston Racial disparities in renal allograft survival: a public health issue? J Am Coll Surg 204 2007 894 902 25 D.E. Butkus E.F. Meydrech S.S. Raju Racial differences in the survival of cadaveric renal allografts Overriding effects of HLA matching and socioeconomic factors N Engl J Med 327 1992 840 845 26 R.B. Isaacs S.L. Nock C.E. Spencer Racial disparities in renal transplant outcomes Am J Kidney Dis 34 1999 706 712 27 E. Akalin J.F. Neylan The influence of Duffy blood group on renal allograft outcome in African Americans Transplantation 75 2003 1496 1500 28 R.H. Kerman P.M. Kimball C.T. Van Buren Possible contribution of pretransplant immune responder status to renal allograft survival differences of black versus white recipients Transplantation 51 1991 338 342 29 A.O. Ojo F.K. Port P.J. Held Inferior outcome of two-haplotype matched renal transplants in blacks: role of early rejection Kidney Int 48 1995 1592 1599 30 J.D. Punch D.H. Hayes F.B. LaPorte Organ donation and utilization in the United States, 1996–2005 Am J Transplant 7 2007 1327 1338 31 A. Lindholm M. Welsh C. Alton B.D. Kahan Demographic factors influencing cyclosporine pharmacokinetic parameters in patients with uremia: racial differences in bioavailability Clin Pharmacol Ther 52 1992 359 371 32 J.F. Neylan Racial differences in renal transplantation after immunosuppression with tacrolimus versus cyclosporine FK506 Kidney Transplant Study Group Transplantation 65 1998 515 523 33 E.M. Vasquez E. Benedetti R. Pollak Ethnic differences in clinical response to corticosteroid treatment of acute renal allograft rejection Transplantation 71 2001 229 233 34 Randomised trial comparing tacrolimus (FK506) and cyclosporin in prevention of liver allograft rejection European FK506 Multicentre Liver Study Group Lancet 344 1994 423 428 35 V.C. McAlister E. Haddad E. Renouf Cyclosporin versus tacrolimus as primary immunosuppressant after liver transplantation: a meta-analysis Am J Transplant 6 2006 1578 1585 36 J.C. Hong B.D. Kahan Sirolimus rescue therapy for refractory rejection in renal transplantation Transplantation 71 2001 1579 1584 37 E.S. Woodle J.R. Thistlethwaite J.H. Gordon A multicenter trial of FK506 (tacrolimus) therapy in refractory acute renal allograft rejection A report of the Tacrolimus Kidney Transplantation Rescue Study Group Transplantation 62 1996 594 599 38 J.F. Markmann J.S. Markowitz H. Yersiz Long-term survival after retransplantation of the liver Ann Surg 226 1997 408 418 39 H.R. Doyle F. Morelli J. McMichael Hepatic retransplantation–an analysis of risk factors associated with outcome Transplantation 61 1996 1499 1505 40 G.B. Klintmalm J.R. Nery B.S. Husberg Rejection in liver transplantation Hepatology 10 1989 978 985 41 J.G. O'Grady G.J. Alexander S. Sutherland Cytomegalovirus infection and donor/recipient HLA antigens: interdependent co-factors in pathogenesis of vanishing bile-duct syndrome after liver transplantation Lancet 2 1988 302 305 42 R.H. Wiesner R.M. Goldstein J.P. Donovan The impact of cyclosporine dose and level on acute rejection and patient and graft survival in liver transplant recipients Liver Transpl Surg 4 1998 34 41 43 D.H. Adams J.M. Neuberger Patterns of graft rejection following liver transplantation J Hepatol 10 1990 113 119 44 D. Candinas B.K. Gunson P. Nightingale Sex mismatch as a risk factor for chronic rejection of liver allografts Lancet 346 1995 1117 1121 45 J.J. Devlin J.G. O'Grady K.C. Tan Ethnic variations in patient and graft survival after liver transplantation Identification of a new risk factor for chronic allograft rejection Transplantation 56 1993 1381 1384 46 B. van Hoek R.H. Wiesner J. Ludwig Combination immunosuppression with azathioprine reduces the incidence of ductopenic rejection and vanishing bile duct syndrome after liver transplantation Transplant Proc 23 1991 1403 1405 47 E.C. Seaberg S.H. Belle K.C. Beringer Liver transplantation in the United States from 1987–1998: updated results from the Pitt-UNOS Liver Transplant Registry Clinical Transpl 1998 17 37