Risk factors for recurrence of primary sclerosing cholangitis after liver transplantation


  • See Editorial on Page 130


Orthotopic liver transplantation (OLT) is the only effective treatment for end-stage liver disease due to primary sclerosing cholangitis (PSC). Recurrence of PSC has recently emerged as a leading cause of allograft failure in the long term. There is limited data on risk factors for recurrence of PSC. We performed a retrospective analysis of 69 consecutive patients who underwent a first OLT for PSC over a 14-year period. Baseline characteristics and clinical and laboratory test results post-LT were recorded. Cholangiograms and liver histopathology were reviewed in a blinded manner by an experienced radiologist and hepatopathologist. Recurrent PSC was diagnosed using previously published Mayo Clinic cholangiographic or histologic criteria. Of 69 patients, 7 (10%) developed recurrent PSC at a median of 68 months (range, 24-134 months) post-LT. The following variables were associated with recurrent PSC in our cohort: presence of human leukocyte antigen (HLA)-DRB1*08 (29% versus 2%; P= 0.026; odds ratio [OR], 24.4; 95% confidence interval [CI], 1.8-318.1), acute cellular rejection (ACR) (71% versus 22%; P= 0.015; OR, 8.7; 95% CI, 1.5-49.9), and steroid-resistant ACR (29% versus 0%; P= 0.012). Despite the strong linkage disequilibrium between DRB1*08 and DQB1*04, DRB1*08-positive subjects with recurrence were negative for DQB1*04, whereas the single DRB1*08-positive subject without recurrent PSC was positive for DQB1*04. A history of ACR and presence of HLA-DRB1*08 are associated with increased risk of recurrent PSC, suggesting an immunologic mechanism for this syndrome. Further studies are required to confirm these observations and to understand the underlying mechanisms. Liver Transpl 14:245–251. 2008. © 2008 AASLD.

Primary sclerosing cholangitis (PSC) is a chronic liver disease of uncertain etiology. Up to 45% of patients with PSC are asymptomatic at the time of diagnosis, but the disease may progress to biliary cirrhosis, eventually leading to end-stage liver disease over a period of several years.1 Orthotopic liver transplantation (OLT) is the only therapeutic option for long-term survival once the complications of end-stage liver disease have set in.2 Currently, PSC is the leading indication for OLT in Scandinavian countries, and the fifth most common indication in the United States.3

PSC is among the indications of liver transplantation (LT) with the best posttransplantation graft and patient survival.4 However, the disease recurs in 6%5 to 37%6 of the patients who undergo LT for PSC. Previous studies suggested that recurrent PSC after LT may not be associated with decreased patient or graft survival.7–10 More recent data suggest that recurrence may account for graft loss in approximately 20% of patients over a period of 7.0 ± 1.6 years.2 Recurrent PSC may infrequently lead to graft loss and contribute to death over a period as short as 13 months posttransplantation.11 Even when recurrence does not adversely affect patient or graft survival, it can result in considerable morbidity and additional health costs.

Identification of risk factors for recurrence of PSC is important because it may potentially alter the management of patients posttransplantation,12 and might offer clues to the pathogenesis of PSC.13 A handful of studies have tried to identify the risk factors for recurrence of PSC after LT.6, 8, 9, 12, 14, 15 However, there is little agreement between the different studies on the frequency and risk factors for recurrent PSC. Recurrence has been associated with older recipient age,14 younger recipient age,8 male gender,6 gender mismatch between donor and recipient,9 presence of inflammatory bowel disease (IBD),14 presence of intact colon after OLT,6 clinically significant cytomegalovirus (CMV) infections,8 recurrent acute cellular rejection (ACR),8 steroid-resistant ACR,12 and Orthoclone OKT3 (OKT-3) therapy15 (Table 1). Many previous studies have also been limited by the lack of standardized histological and radiological criteria for definition of recurrent PSC and have not excluded patients with confounding factors.

Table 1. Risk Factors for Recurrent PSC Identified in Previous Studies
 Jeyarajah et al. (1998)8Abu-Elmagd et al. (1999)14Vera et al. (2002)6Khettry et al. (2003)9Kugelmas et al. (2003)15Brandsaeter et al. (2005)12
  1. Abbreviations: ACR, acute cellular rejection; CMV, cytomegalovirus; IBD, inflammatory bowel disease; OLT, orthotopic liver transplantation.

OLT for PSC118380152517149
Recurrence of PSC after OLT (%)18 (15)52 (14)56 (37)6 (12)15 (21)9 (18)
Risk factors for recurrenceClinically significant CMV hepatitisOlder recipient ageMale genderRecipient-donor gender mismatchUse of OKT3Steroid-resistant ACR
Absence of donor HLA DR52Presence of IBDIntact colon after OLT   
Younger recipient age     
More than 1 episode of ACR    

Specific class II major histocompatibility complex (MHC-II) haplotypes encoded in the human leukocyte antigen (HLA)-D locus have been associated with a variety of hepatic or other autoimmune disorders, implying a role for T cells bearing the CD4 molecule, which binds to these MHC-II molecules. Indeed, some MHC-II haplotypes are known to increase the risk of PSC in the native liver, and may influence its natural history after LT.9 For example, liver allografts from HLA DR52-positive donors have been reported to protect patients from recurrence of PSC after OLT.8 The goal of this retrospective study was to review the incidence and risk factors of recurrent PSC in a cohort of patients who underwent their first LT for PSC at a single U.S. center.


ACR, acute cellular rejection; CI, confidence interval; CMV, cytomegalovirus; HLA, human leukocyte antigen; IBD, inflammatory bowel disease; MHC, major histocompatibility complex; OLT, orthotopic liver transplantation; OR, odds ratio; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis.


We reviewed the clinical, biochemical, cholangiographic, and histologic findings of all patients with PSC who underwent a first OLT at the University of Washington Medical Center during the period between 1990 and 2003. The study was approved by the Institutional Review Board of the University of Washington. Inclusion criteria included: a pretransplant diagnosis of PSC based on: (1) the presence of a cholestatic biochemical profile with an elevated serum alkaline phosphatase level for at least 6 months; (2) typical cholangiographic findings of stricturing, beading, or irregularities in the intrahepatic and/or extrahepatic biliary tree; and (3) explant histology consistent with the disease while not suggestive of other potential causes of chronic cholestatic liver disease such as primary biliary cirrhosis (PBC) and sarcoidosis.16

The following data were collected from medical records: age, gender, presence of IBD, history of past biliary surgery, hepatitis B surface antigen and anti-hepatitis C virus status, donor/recipient blood groups, donor/recipient CMV antibody status, donor/recipient HLA-DRB1 and DQB1 haplotypes, cold/warm ischemia times, construction of end-to-side Roux-en-Y choledochojejunostomy for biliary drainage, amount of blood loss during surgery, presence of intact colon after transplantation, immunosuppressive regimen, symptomatic CMV infection, patency of hepatic artery, duration of follow-up, and graft and patient survival.

All available liver histological slides of the study subjects were retrieved from the pathology archives and reviewed by a single hepatic pathologist (M.M.Y.) who was blinded to the clinical diagnosis. Each slide was evaluated for presence of ductal damage, ductopenia, ductular reaction, lymphocytic cholangitis, fibro-obliterative lesions, concentric fibrosis, fibrous scars, and nodules. In addition, histological features suggestive of ACR, chronic rejection, and viral causes of liver dysfunction were documented.

In patients who did not have cholangiographic confirmation of recurrence, recurrent PSC was diagnosed on the basis of serial biopsies showing bile ductular reaction, inflammation, and fibrosis in the initial stages, and ductal damage, ductal atrophy, ductopenia, and lymphocytic cholangitis in the subsequent biopsies, without evidence of chronic rejection (such as centrilobular necrosis or dropout in earlier biopsies or foam cell arteriopathy) or clinical or histologic evidence of an alternate cause for these changes. Diagnosis of ACR was based on the Banff criteria, which require the presence of at least 2 of the following features: (1) mononuclear as well as mixed portal inflammation; (2) bile duct infiltration and damage; and (3) venous endothelitis of the portal tracts.17, 18 Those with histological features such as CMV inclusion bodies, neutrophil clusters, lobular inflammation, spotty necrosis, acidophilic bodies, or perivenular cell loss were considered to have viral causes of liver dysfunction, and were therefore not diagnosed as ACR.18

A single gastrointestinal radiologist (C.C.) who was blinded to the clinical diagnosis reviewed serial cholangiograms and documented the presence, location, and number of biliary strictures. The diagnosis of recurrent PSC was based on the Mayo Clinic criteria proposed by Graziadei et al.,16 which requires: (1) a confirmed diagnosis of PSC before LT; (2) cholangiograms showing nonanastomotic biliary strictures of the intrahepatic and/or extrahepatic biliary tree with beading and irregularity occurring >90 days posttransplantation; or (3) a liver biopsy showing fibrous cholangitis and/or fibro-obliterative lesions with or without ductopenia, biliary fibrosis, or biliary cirrhosis; and exclusion of: (1) hepatic artery thrombosis/stenosis; (2) ductopenic rejections; (3) donor-recipient ABO blood type incompatibility; (4) anastomotic stricturing alone; and (5) nonanastomotic strictures before day 90 post-OLT.

The group of patients with recurrence of PSC was compared with those without recurrence. Fisher's exact test was used for comparing discrete variables, and Student t test for continuous variables. A P value less than 0.05 was considered statistically significant. Variables that were found to have a significant association with recurrence were entered into a logistic regression analysis and the odds of recurrence associated with each variable were calculated.


During the period from January 1990 to December 2003, 69 patients underwent a first OLT for PSC (57 males; median age, 49 years; range, 21-69 years). As this was a retrospective study, data pertaining to some variables were not available in all patients. The relevant baseline demographic and clinical features of the study population are summarized in Table 2.

Table 2. Selected Baseline Demographic and Clinical Features of Subjects Studied
 With RecurrenceWith No RecurrenceP Value
  1. NOTE: Age and ischemia times are shown as mean ± standard deviation. All other parameters are shown as number of patients (%).

  2. Abbreviations: CMV, cytomegalovirus; IBD, inflammatory bowel disease; OLT, orthotopic liver transplantation.

Number of patients (%)7 (10)62 (90) 
Recipient gender, male:female (%)6:1 (86:14)51:11 (82:18)>0.05
Age at transplantation (years)41 ± 12.849 ± 11.3>0.05
Past biliary surgery (%)09 (15)>0.05
Presence of IBD (%)7 (100)52 (95)>0.05
Presence of intact colon after OLT (%)5 (71)44 (82)>0.05
Cold ischemia time (minutes)313 ± 103.1437 ± 149.2>0.05
Warm ischemia time (minutes)46 ± 9.546 ± 11.5>0.05
Total ischemia time (minutes)359 ± 96.9470 ± 160.8>0.05
CMV positive recipient (%)3 (50)30 (65)>0.05
CMV positive donor (%)2 (100)18 (78)>0.05
CMV in donor or recipient (%)3 (100)36 (92)>0.05
CMV positive donor/negative recipient (%)1 (50)6 (27)>0.05
Symptomatic CMV infection (%)1 (14)4 (7)>0.05

All donor livers had been stored at 4°C in University of Wisconsin solution. Roux-en-Y choledochojejunostomy was used for the biliary reconstruction in all except 4 patients (8%) in whom end-to-end anastomosis was done. The immunosuppressant agents used in these patients included glucocorticoids, azathioprine, cyclosporine, tacrolimus, sirolimus, and mycophenolate. Episodes of ACR were treated with high-dose pulsed corticosteroids. Steroid-resistant rejections were treated with lymphocyte antibodies, either OKT3 or thymoglobulin. Patients diagnosed to have recurrent PSC received ursodiol, and stent placement or dilation if appropriate.

Protocol liver biopsies were performed immediately following reperfusion, at 3, 6, and 12 months after transplantation, at 5 years, and every 5 years thereafter. Additional liver biopsies were performed as indicated clinically. None of the protocol biopsies were missed in 52 patients (75.4%); 2 out of 7 patients with recurrence and 15 out of 62 patients without recurrence missed at least 1 protocol biopsy, but the difference was not statistically significant (P = 1). Percutaneous, endoscopic (in patients with a choledochocholedochostomy), or magnetic resonance cholangiographies were done when clinically indicated, or in the presence of histologic features suggestive of recurrent PSC. Doppler ultrasonography was performed to assess patency of hepatic artery immediately postoperatively, at regular intervals, and whenever clinically indicated. Hepatic angiography was used to ascertain vascular patency when Doppler ultrasonography was inconclusive.

The median duration of follow-up was 50 months (range, 1-173 months). A total of 7 patients (10%) developed recurrent PSC as defined by the Mayo Clinic criteria. The Kaplan-Meier curve for cumulative risk of recurrence of PSC is shown in Fig. 1. Cholangiographic features of recurrent PSC were present in 4 (6%) patients and histological features in 5 (7%). The median time to cholangiographic recurrence was 68 months (range, 24-110 months), and that to histologic recurrence was 68 months (range, 48-134 months). Of the 30 patients followed up for at least 5 years, 7 (23%) developed recurrence. Of the 7 patients with recurrence, 1 (14%) subsequently required retransplantation 73 months after the initial transplant. The patient is alive and doing well at this time.

Figure 1.

Kaplan-Meier curves for the recurrence of PSC after liver transplantation.

Histologically-proven ACR was present in 18 patients, half of whom underwent transplantation between 1990 and 1998 and the rest underwent transplantation between 1999 and 2003. Recurrent PSC developed in 2 patients in the initial group of 9 and in 3 patients in the latter group of 9 with ACR. Overall, recurrent PSC showed statistically significant association with any episode of ACR (P = 0.015) and steroid-resistant ACR (P = 0.012), and a trend toward association with 2 or more episodes of ACR (P = 0.086) (Table 3). On logistic regression analysis, the odds ratio for recurrence associated with ACR was 8.7 (95% confidence interval, 1.5-49.9). A multivariate analysis was not performed as the sample size was too small. The Kaplan-Meier curves for cumulative risk of recurrence of PSC in patients with and without ACR are shown in Fig. 2.

Table 3. Risk Factors for Recurrence of Primary Sclerosing Cholangitis After Liver Transplantation
 With Recurrence (%)With No Recurrence (%)P Value
  1. Abbreviation:ACR, acute cellular rejection.

HLA DR8 in donor or recipient2 (29)1 (2)0.026
Acute rejection5 (71)13 (22)0.015
More than 1 episode of ACR2 (29)3 (6)0.086
Steroid-resistant ACR2 (29)00.012
Figure 2.

Kaplan-Meier curves for the recurrence of PSC in patients with and without ACR. Abbreviations: ACR, acute cellular rejection.

Age, gender, previous biliary surgery, presence of inflammatory bowel disease, presence of an intact colon after LT, CMV status of the donor or recipient, symptomatic CMV infection, cold/warm ischemia times, Roux-en-Y choledochojejunostomy, and amount of operative blood loss were not found to be significantly different between patients with or without recurrence. Of note, none of the 4 patients without Roux-en-Y reconstruction developed recurrent PSC.

HLA data was available for all the 69 donors and 65 (94%) recipients. This included HLA data of all the donors and recipients for the 7 subjects with recurrent PSC. Despite the small number of cases of recurrent PSC available for this study, a significantly higher frequency of HLA-DRB1*08 alleles (in either donor or recipient) was detected in cases with recurrent PSC (29%) relative to those without recurrence (1.6%, representing a single subject) (P = 0.026). On logistic regression analysis, the odds ratio for recurrence associated with HLA-DRB1*08 haplotype was 24.4 (95% confidence interval, 1.8-318.1).

The HLA-DRB1*03, DQB1*02 genotype was enriched among our OLT recipients with PSC, being present in 49% of these recipients overall, compared to only 23% of donors (P = 0.002), with DQB1*02 cosegregating with DR3 in nearly all cases. However, the frequency of the DRB1*03, DQB1*02 genotype in donors (29%) or recipients (50%) was not elevated specifically in those cases with recurrent PSC (P > 0.05). Similarly, although DRB1*13 alleles were somewhat enriched in PSC subjects overall (35% of recipients versus 22% of donors, P > 0.05), they were not more common in those OLT recipients with recurrent PSC (29% in those with recurrence versus 36% in those without recurrence; P > 0.05). Although the percentage of donors with HLA-DRB1*13 alleles was higher among donor livers in which PSC recurred (43%) than those in which it did not (19%), this finding did not reach statistical significance (P > 0.05) due to a small sample size.

Additionally, DQB1*04 was present in the single DRB1*08-positive subject without recurrent PSC (both alleles in the donor), whereas no HLA-DQB1*04 was identified in any donor or recipient in the recurrent PSC group, despite enrichment for DRB1*08 alleles in this group. Thus, unlike PBC, recurrent PSC cases with HLA-DRB1*08 were DQB1*04-negative in our study.


Recurrent PSC was found in 7 (10%) patients in this retrospective study of a cohort of 69 patients who underwent a first OLT for PSC. ACR, steroid-resistant ACR, and HLA-DRB1*08 were predictive of an increased risk of recurrence.

Our observation of recurrence of PSC in 10% of the patients is within the range of 6%5 to 37%6 reported in the literature. The wide range of reported incidence is attributed to variation in study designs and diagnostic criteria.14 Posttransplantation biliary strictures caused by ischemia (for example, hepatic arterial thrombosis and prolonged cold preservation), infections (for example, CMV infection and cholangitis consequent to Roux loop), immune mediated injury (for example, ABO blood type incompatibility) and chronic ductopenic/arteriopathic rejection make identification of recurrent PSC difficult.19 In this study, histologic and cholangiographic findings were independently reviewed by a single pathologist and a single radiologist, and the objective Mayo Clinic criteria were used to ensure accuracy of the diagnosis of recurrent PSC.

We found ACR and steroid-resistant ACR to be predictive of an increased risk of recurrent PSC. Other studies have also found factors related to ACR to be associated with recurrence. Jeyarajah et al.,8 in a retrospective study of 118 OLTs, in which 20 allografts developed recurrent PSC, occurrence of 2 or more episodes of ACR was found to be significantly associated with recurrence. In another study on association of different immunosuppressive regimens with recurrence of PSC, Kugelmas et al.15 observed that OKT3 therapy, which is used in treatment of steroid-resistant ACR, is associated with a higher risk of PSC recurrence. Recently, 49 patients who underwent OLT for PSC were evaluated by Brandsaeter et al.12 with magnetic resonance cholangiography and steroid-resistant ACR was found to be predictive of recurrent PSC.

There are several potential explanations for the association between ACR and recurrent PSC. Injury of the biliary epithelium resulting from ACR been postulated to increase autoimmune epitopes, leading to immune mediated ductal damage.8 Alternatively, some other variable could be a common factor predisposing to both ACR and recurrence of PSC.8 Indeed, ACR has been found to occur more commonly in OLTs for PSC than for other liver diseases.12 This association could imply some link between the pathophysiology of PSC and ACR, such as a hyperactive component of the immune system or a defective mechanism for immune autoregulation. Yet another hypothesis is that it could be the immunosuppressive therapy for rejection, and not rejection per se, that predisposes to recurrent PSC.12 Indeed, some immunosuppressants, such as calcineurin inhibitors20, 21 and antithymocyte globulin22 have been associated with reduced numbers of CD4+ CD25+ regulatory T cells. The latter have been strongly implicated in preventing murine models of IBD, and so may likewise have a role in preventing human PSC, such that their depletion with immunosuppressive medications could paradoxically facilitate PSC recurrence after OLT. More studies are required to delineate the interrelationship between ACR, immunosuppression, and recurrence of PSC. A clearer understanding of this interrelationship would help decide whether the strategy to prevent recurrence should be a higher standard dose of immunosuppressive therapy to prevent rejection in the first place, or a different approach to treatment of ACR in OLTs for PSC.3

In the native liver, HLA-DRB1*0323 and HLA-DRB1*1324 alleles have been correlated with PSC development, and, DRB1*03 coexpressed with an HLA-DQB1*02 haplotype has been associated with a more aggressive course of PSC.25 Although we confirmed that the HLA-DRB1*03, DQB1*02 genotype was more common in subjects with PSC than in their donors, neither this genotype nor DRB1*13 correlated significantly with PSC recurrence. Similarly, no significant increase in DRB1*04 allele frequency in donors and/or recipients was noted with recurrent PSC, despite a correlation between DRB1*04 and autoimmune hepatitis, with which PSC occasionally overlaps.

Although allografts from HLA DR52-positive donors have previously been negatively associated with recurrent PSC,8 we found HLA-DRB1*08 alleles in either donor or recipient to be predictive of an increased risk of PSC recurrence. Although DRB1*08 has not previously been associated with de novo PSC occurrence, it has been linked to juvenile rheumatoid arthritis and PBC, usually in association with HLA-DQB1*04.26–28 There is substantial linkage disequilibrium between DRB1*08 and DQB1*04, with both alleles cosegregating in up to 99% of subjects in some countries.27 Although this DRB1*08, DQB1*04 cosegregation was indeed seen in the single DRB1*08-positive subject without recurrent PSC (both alleles being in the donor liver), no DQB1*04 was seen in any DRB1*08-positive subject with recurrent PSC. Given the rarity with which this linkage disequilibrium is disrupted, there may be a unique immunological environment in DRB1*08-positive, DQB1*04-negative individuals that predisposes to PSC recurrence.

Although native PBC and recurrent PSC differ in their association with DQB1*04, the finding that both of these immune-mediated biliary diseases share an association with DRB1*08 alleles suggests that there may be a common immunological mechanism between these two disease states. Although rare, there is precedent for an overlap syndrome between PBC and PSC.29 Additionally, there is precedent for autoimmune liver diseases to present de novo after OLT for an alternative indication, such as autoimmune hepatitis occurring after OLT for hepatitis C, in the setting of interferon therapy.31

We acknowledge that the present study has limitations. First, this study was retrospective in nature, and some data were not available in all patients. Second, though liver biopsies were performed on a protocol basis in this study, cholangiograms were performed only when clinically indicated, or in the presence of histologic features suggestive of recurrent PSC. Finally, the sample size in the present study is smaller than in some previous reports. Studies involving larger numbers of patients are needed to confirm the findings of this study. Future studies could also benefit from the longer follow-up data that would accrue with time, as well as from improved technical quality of magnetic resonance cholangiography, which would enable noninvasive and accurate assessment of the biliary tract to be done more frequently and for longer periods.

In conclusion, ACR, steroid-resistant ACR, and HLA-DRB1*08, particularly in the absence of HLA-DQB1*04, are associated with an increased risk of recurrence of PSC following OLT and suggest an immunologic basis for the etiopathogenesis of PSC. Further studies are required to confirm the associations observed in this study and to explore prospective interventions to reduce the rate of recurrent PSC. If the association between HLA haplotypes and risk of recurrence is validated in further studies, HLA typing would be useful in donor selection as well as provide valuable prognostic information right at the time of transplantation, thus identifying a cohort of transplant recipients in whom more rigorous screening for PSC recurrence and subsequent cirrhosis or cholangiocarcinoma would be appropriate. Although at present the only effective treatment for these conditions would be a repeat liver transplantation, there are a number of potential therapies under investigation,31 which may ultimately prove to have prophylactic value in patients at high risk for recurrent PSC.