Risk factors for recurrence of primary sclerosing cholangitis after liver transplantation



Liver transplantation (LT) is the only therapeutic option for end-stage primary sclerosing cholangitis (PSC), but PSC can recur (rPSC) in some patients after LT. The aim of our study was to evaluate the risk factors associated with rPSC. Between 1989 and 2004, 69 patients receiving transplantation for PSC (42 male, mean age 41.9 yr). Clinical and laboratory data, activity/extension and treatment of ulcerative colitis (UC), post-LT cytomegalovirus (CMV) infection, and immunosuppression were evaluated. Determination of rPSC was made by radiological and histological findings. Exclusion criteria were ABO blood group incompatibility, hepatic artery stenosis, and biliary strictures occurring in <3 months post-LT. A total of 48 (70%) patients had PSC and UC pre-LT. rPSC occurred in 7 of 53 (13.5%, 2 patients with de novo UC) who were alive 1 yr after LT and/or met inclusion/exclusion criteria: median 60 (4-120) months. No patient without post-LT UC had rPSC: 0 of 20 vs. 7 of 26 with post-LT UC (P = 0.027). The multivariate logistic regression analysis showed that maintenance steroids for UC (>3 months) post-LT was the only risk factor significantly associated with rPSC (P = 0.025). In conclusion, the presence of UC post-LT, and the need for maintenance steroids post-LT, which is an independent factor, are associated with rPSC. These findings could help elucidate a possible mechanism of PSC pathogenesis. Liver Transpl, 2007. © 2007 AASLD.

Primary sclerosing cholangitis (PSC) is a chronic inflammatory disease of the intra- and/or extrahepatic biliary tree, characterized by progressive fibrosis and strictures of the bile ducts.1 PSC is of unknown etiology. However, there is some evidence that an autoimmune process is implicated in its pathogenesis, as there are antineutrophilic cytoplasmic antibodies, and there are associations with other autoimmune diseases, such as inflammatory bowel disease1, 2; it is estimated that 60 to 70% of patients with PSC have inflammatory bowel disease, mainly ulcerative colitis (UC)3

Currently, no specific treatment is known to alter the natural course of PSC, with the possible exception of high doses of ursodeoxycholic acid.1 Thus, at present, liver transplantation (LT) remains the only effective therapeutic option for PSC patients with end-stage liver disease. However, LT is not without drawbacks; apart from specific complications, such as opportunistic infections, the clinical course of UC often worsens after LT, and PSC can recur with an incidence of 8.6 to 47%.4, 5 The diagnosis of recurrent PSC (rPSC) is difficult, and several conditions, such as hepatic artery stenosis or thrombosis and chronic rejection, may mimic PSC in the post-LT setting.6, 7 In addition, a helpful diagnostic serum marker in the pre-LT setting, antineutrophilic cytoplasmic antibodies,6 is not useful in the post-LT follow-up. Indeed, although rPSC is recognized clinically, it is considered to be a diagnosis of exclusion.

Studies assessing recurrence of PSC after LT have recently been reviewed,8, 9 but only few of them have evaluated potential factors related to rPSC.10–19 The latter studies have either identified no factors,15 or each a different factor for rPSC.10, 14, 17–19 The reviewers concluded that this was likely due to the difficulty in defining diagnostic criteria as well as different inclusion and exclusion criteria for PSC.1, 8 In addition, these studies have usually focused separately on pre- or on post-LT risk factors, but not a combination. Although features of UC have been implicated in rPSC, only 5 studies have evaluated whether the presence or absence of UC was related to rPSC,10, 14, 15, 17, 18 only 2 studies have evaluated the impact of colectomy,17, 19 and only 1 study assessed the severity of UC together with the presence of colorectal cancer.19 In the latter study19 UC activity was not associated with rPSC, but specific details were not given. Finally, it is known that PSC patients without UC pre-LT may develop UC post-LT (de novo UC).4 None of the previous studies have evaluated the potential impact of de novo UC on rPSC.

The aim of our study was to evaluate: 1) the pre- and post-LT risk factors that could be implicated in the recurrence of PSC in patients with or without UC using strict inclusion and exclusion criteria for rPSC; and 2) whether pre-LT colectomy or absence of UC is protective for rPSC.


LT, liver transplantation; PSC, primary sclerosing cholangitis; rPSC, recurrent PSC; UC, ulcerative colitis; CMV, cytomegalovirus.


Between May 1989 and December 2004, 69 patients (42 men; median age 43 yr; range 17-66 yr) with end-stage liver disease due to PSC underwent LT in our center. The diagnosis of PSC was confirmed by pre-LT cholangiography and histological examination of the liver explant in all cases. Patients with pre-LT diagnosis of PSC who had a follow-up of at least 1 yr were further evaluated. For each patient, demographic and clinical data, characteristics of rejection episodes and concomitant pre- or post-LT UC, as well as donor characteristics, perioperative data, cytomegalovirus (CMV) post-LT infection, human leukocyte antigen profile of donor and recipient, and cause of death were recorded.

The diagnosis of rPSC was assessed according to the diagnostic criteria put forward by the Mayo Clinic.7, 15 Thus, each patient underwent liver biopsy in the presence of cholestatic biochemical profile post-LT (elevated serum alkaline phosphate). In the absence of cellular rejection, we performed magnetic resonance cholangiography and/or endoscopic retrograde cholangiography (percutaneous cholangiogram for those with Roux-en-Y anastomosis). Protocol liver biopsies were also performed at 1, 3, 5, 7, and 10 yr if the patient consented. The diagnosis of rPSC was based on a liver biopsy and radiological assessment (with magnetic resonance cholangiography and/or endoscopic retrograde cholangiography) showing changes consistent with PSC (defined as nonanastomotic focal strictures in the intra- and/or extrahepatic biliary tree more than 3 months after LT) in the absence of ABO blood group-incompatible allograft, dominant anastomotic stricture, or hepatic artery thrombosis identified by Doppler ultrasonography and/or hepatic angiography. Histological criteria of rPSC recurrence were defined by the presence of periductal fibrosis, obliterative ductular lesions, and bile duct loss in the absence of chronic allograft rejection.20

The number of rejection (acute or chronic) episodes post-LT, as well as the type, total dose, and duration of immunosuppression that was given for their management, and the maintenance immunosuppressive regimens post-LT were also evaluated in detail. Protocol liver biopsies for rejection were performed within 10 days of transplantation in all patients and whenever there was an evidence of graft dysfunction. All liver biopsies were assessed by experienced liver pathologists. Acute and chronic ductopenic rejection was diagnosed according to established histologic criteria.21 Acute rejection episodes were treated with 1 g daily methylprednisolone for 3 days, intravenously. Steroid-resistant rejection was treated with lymphocyte antibodies Orthoclone (OKT-3) or antithymocyte globulin. Steroids were tapered within 3 months if possible. In our study, patients who received steroids for more than 3 months after LT were defined as those with long-term maintenance therapy. A diagnosis of PSC recurrence was not an indication to alter immunosuppressive therapy.

The diagnosis of UC was based on combined evaluation of clinical, endoscopic, and histological findings, and exclusion of other causes of colitis. All patients with UC were evaluated in detail4 and the course before or after LT was recorded: type, extent, and duration of UC, maintenance treatment, number and severity of exacerbations, number of hospital admissions, courses and total duration of steroid therapy, utilization and duration of azathioprine therapy, and colectomy. All patients with PSC underwent total colonoscopy and colonic mucosal biopsy specimens were taken within 6 months before LT during the workup evaluation, and yearly surveillance was undertaken post-LT. The definitions used for the clinical course of UC before and after LT have been described elsewhere.4 The course of post-LT UC over the entire follow-up period was compared with that in the period up to 5 yr before LT. Truelove and Witts22 criteria were used for classification of severity of UC exacerbation. The course of UC was considered worse when it was becoming more active post-LT than pre-LT, e.g., from quiescent to active, or to becoming steroid-dependent or requiring colectomy. All patients with known UC were given long term mesalazine 1.5 to 2 g/day within the first 3 months after LT and, in this study period, none received ursodeoxycholic acid.

The type of anastomosis (Roux-en-Y choledochojejunostomy or duct-to-duct) and the presence of intraductal cholangiocarcinoma (detected pre-LT or at explant) were also recorded. CMV viremia (using in-house CMV polymerase chain reaction and CMV polymerase chain reaction TaqMan-based assay) was evaluated 3 times per week during the first 3 postoperative months. The serological typing for human leukocyte antigen class I (A and B) and II (DR and DQ) antigens of donor and recipient was performed using the microcytotoxicity assay technique.

Statistical Analysis

All data were analyzed using the statistical package SPSS (version 13.0; SPSS, Inc., Chicago, IL). Quantitative variables were expressed as mean values ± 1 standard deviation (SD), and/or median values (range) if their distribution was skewed. Chi-squared test was used for categorical variables and the Mann-Whitney U test was used for comparison of quantitative variables. Significance testing was 2-sided and set to less than 0.05. To identify the factors that were independently associated with the recurrence of PSC after LT, variables close to or under the statistical significance in the univariate analysis (P < 0.10) were included in the multivariate analysis using logistic regression models.


Total Population

The median follow-up was 110 (range 12-185) months (Table 1). A total of 9 patients underwent re-LT (3 patients for hepatic artery thrombosis or primary nonfunction, 3 for chronic rejection, and 3 for rPSC) at a median time of 17 (0.3-96) months and follow-up terminated at retransplantation.

Table 1. Characteristics of Patients Who Underwent LT for PSC Included in the Study
Patient characteristicsValue
No. of patients53
Age (median, range), year43 (17–66)
Follow up (median, range), months110 (12–185)
Patients died in <1 year, n (%)13 (19)
Cholangiocarcinoma, n (%)4 (7)
 Before LT2
 After LT2
Patients with UC, n (%)36 (68)
Duration of UC pre-LT (months)120 (2–480)
Pre-LT course, n (%) 
 Quiescent19 (36)
 Active12 (23)
 Colectomy5 (9.5)

The cause of death in the first 12 months after LT was multiple organ failure/sepsis in 7, primary nonfunction in 2, severe rejection in 1, rupture of hepatic hematoma in 1, and recurrence of cholangiocarcinoma in 2 patients. None of the above patients had any evidence of rPSC. Thus, 56 (81%) patients were alive 1 yr after LT (Table 1). The other causes of death (>1 yr) were: post-LT lymphoproliferative disorder in 3, nonhematological malignancies in 4 (2 cholangiocarcinoma and 2 colon cancer), rPSC in 2, intestinal obstruction in 1, and unknown in 1.

In total, 16 patients were excluded from the study, since they died in less than 1 yr after LT (n = 13) and/or met the exclusion and inclusion criteria (hepatic artery thrombosis or primary nonfunction, n = 3). Thus, 53 patients underwent further evaluation. A total of 36 (68%) patients had PSC and UC pre-LT: 33 had total UC and 3 distal UC. Pre-LT UC was quiescent in 19 (36%) and active in 12 (23%). A total of 5 (9.5%) patients had undergone total colectomy pre-LT. A total of 3 patients developed de novo UC after LT. In the 3 patients with chronic rejection, 1 had no UC, 1 had a colectomy pre-LT, and 1 had mild UC post-LT, similar to pre-LT activity.

The diagnosis of rPSC was made in 7 (13.5%, 5 male/2 female) of the 53 patients, at a median time of 60 (4-120) months, initially based on biochemical profile with abnormal cholestatic tests and then by histological and radiological findings in 6 patients and on radiological findings alone in 1 patient. All these 7 patients had UC (either pre-LT or de novo after LT), and none of them had undergone total colectomy before LT. Long-term (more than 1 yr) patient survival was not different in patients with or without rPSC (5 yr: 85% vs. 88%, 10 yr: 76% vs. 77%; P > 0.05), but rPSC patients underwent re-LT more frequently, compared to those without rPSC (3/7 for rPSC vs. 3/46 for chronic rejection; P = 0.02).

Risk Factors for Recurrence of PSC

Total Population

The characteristics of the patients with rPSC, compared to those without rPSC are shown in Table 2. Demographic characteristics, such as age and gender of recipient and donor as well as perioperative variables were not associated with rPSC (Table 2). None of the PSC patients without UC before or after LT (n = 14) had rPSC vs. 7 (18%) of 39 PSC patients with UC pre-LT or de novo UC, but this difference was not significant (P = 0.089). However, patients without post-LT UC (i.e., when we analyzed all patients with intact colon and without UC [pre- or post-LT] and those with pre-LT colectomy together) had significantly less rPSC, compared to those with post-LT UC (i.e., de novo or pre-LT UC without colectomy) (0/20 vs. 7/26; P = 0.027) (Table 2).

Table 2. Predictive Factors of Recurrence of PSC After LT in the 53 Patients Included in the Study (Univariate Analysis)
Patient characteristicsPSC post-LT (n = 53)P value
Recurrence (n = 7)No recurrence(n = 46)
  1. NOTE: Values in bold indicate statistical significance (p < 0.05).

  2. Abbreviations:AOKT-3: lymphocyte antibodies Orthoclone, ATG: antithymocyte globulin, CYC: cyclosporine, AZA: azathioprine, FK: tacrolimus, MMF: mycophenolate mofetil.

Gender (M/F)5/225/210.39
Age (years)35 ± 1542 ± 130.17
Donor gender (M/F)6/127/190.18
Donor age (years)33 ± 1844 ± 140.11
Blood requirements (units)7.5 ± 54.5 ± 30.12
Cold ischemic time (minutes)720 ± 160640 ± 1800.30
Presence of de novo UC (yes/no)2/51/450.005
Presence of UC after LT (no UC-pre-LT colectomy/UC with intact colon-de novo)0/720/260.027
Duration of pre-LT UC, months9 (0–180)40 (0–480)0.75
Admission for UC during the last 3 years1.5 (0–9)0.07 (0–1)0.011
Pre-LT course of UC (quiescent/active)4/115/110.21
Extent of UC before LT (total/distal)5/130/20.43
Pre-LT colectomy (yes/no)0/76/400.31
Post-LT colectomy (yes)160.59
Steroid dependence UC before LT (yes/no)150.74
Azathioprine for pre-LT UC1/54/390.57
Colon histology pre-LT (normal/mild/moderate/severe)3/2/1/112/5/12/20.51
Maintenance immunosuppression (CYC/CYC+AZA/FK/FK+AZA/FK+MMF)2/3/1/-/18/13/9/9/70.68
Patients under CYC vs. FK (in total)5/221/250.29
OKT-3 or ATG therapy2110.79
Azathioprine post-LT in general3220.84
Rejection episodes post-LT (no/mild/moderate/severe)1/1/2/39/12/12/130.84
Duration of steroids for rejection (months)6 (0–13)2 (0–13)0.015
Duration of steroids >3 months (n, %)5 (72)1 (23)0.024
Duration of steroids >6 months (n, %)2 (28)4 (9)0.12
CMV viraemia post-LT (n)4150.22
Type of anastomosis (Roux-en-Y/duct-to-duct)4/315/310.22

In the univariate analysis (Table 2), patients with rPSC (n = 7), compared to those without rPSC (n = 46), received maintenance (>3 months) steroids post-LT more frequently (5 [72%] vs. 11 [23%]; P = 0.024) and steroids for longer periods (medians: 6 [0-13] months vs. 2 [0-13] months; P = 0.015). This was due to the fact that their inflammatory disease was more active. In addition, they had more frequently: 1) admission to hospital for UC exacerbation during the 3 yr prior to LT (1.5 [0-9] vs. 0.07 [0-1]; P = 0.011), 2) de novo UC (2 [28.5%] vs. 1 [2.1%]; P = 0.005), and 3) had UC after LT (de novo or pre-LT UC without colectomy) (7 [100%] vs. 26 [56%]; P = 0.027).

Patients with rPSC did not differ from those without rPSC in other characteristics of UC or immunosuppressive therapy. Thus, rPSC was not associated with UC disease extension (distal or total) nor activity, the performance of colectomy before or after LT (P = not significant) or the course of UC after LT (e.g., severity, number of admission for UC, utilization of azathioprine or steroids for UC exacerbations) (Table 2). In addition, OKT3 therapy for intractable acute rejection episodes and the number of acute rejection episodes were similar between the 2 groups (P = 0.79 and P = 0.84, respectively) (Table 2). Post-LT maintenance immunosuppression was monotherapy with cyclosporine (10 mg/kg/day) (n = 10) or tacrolimus (0.1 mg/kg/day) (n = 10) both in 2 divided doses or combination therapies (azathioprine [1 mg/kg/day] plus cyclosporine [n = 16] or azathioprine plus tacrolimus [n = 9] or mycophenolate [MMF] [1 gm twice daily] plus tacrolimus [n = 8]). Steroids were withdrawn within 6 months after LT in all but 6 patients, who received prednisolone for 7 to 13 months. Patients with rPSC received cyclosporine or tacrolimus maintenance therapy, in a similar fashion compared to those without rPSC (P = 0.68) (Table 2). Finally, rPSC was not associated with CMV infection post-LT (P = 0.22), or the type of surgical anastomosis (P = 0.22) or human leukocyte antigen disparity at any locus between donor and recipient.

The multivariate logistic regression analysis showed that the need for maintenance (>3 months) steroids post-LT (P = 0.025) was the only risk factor significantly associated with rPSC. Moreover, in the univariate, but not in the multivariate analysis, the absence of UC post-LT (due to pre-LT colectomy or not) was protective for rPSC. The results were the same even when we excluded the 3 patients with the diagnosis of chronic rejection from our analysis or when we considered these 3 patients as having diagnosis of rPSC, i.e., we assumed chronic rejection was rPSC (data not shown).

Population With or Without UC Before or After LT

None of the PSC patients without UC before LT and/or colectomy before LT (n = 20) had recurrence of PSC after LT. Therefore, we analyzed further only those patients with pre-LT UC and no pre-LT colectomy and those with de novo post-LT UC (n = 33). In the multivariate analysis, we found that only the need for maintenance steroids post-LT was significantly associated with rPSC (P = 0.029). In total, 7 (21%) of the 33 patients required total colectomy after LT, which was not associated with rPSC (rPSC occurred in 1 [14%] of the 7 post-LT colectomy patients vs. 6 [24%] of the 26 patients without post-LT colectomy, P = 0.61). In the patient with rPSC and colectomy post-LT, the recurrence occurred before colectomy, and in the other 2 patients who had a de novo UC with rPSC, the recurrence was diagnosed after de novo UC.


The diagnosis of PSC is based on clinical, laboratory, histological, and, mainly, cholangiographic findings, but none of these are specific for PSC. Thus, because of the lack of a diagnostic gold standard, the diagnosis of rPSC after LT remains difficult, and well-defined cholangiographic and histological criteria are mandatory.7, 8 The diagnosis is complicated by the fact that other potential causes (e.g. hepatic artery stenosis, chronic rejection) after LT leading to bile duct lesions suggesting PSC have to be excluded.1, 6, 23

Another issue is the differences in exclusion/inclusion criteria and methods of diagnosis, which we feel are the reason for the discrepancies among studies of the incidence and the risk factors associated with rPSC.7, 17–19, 24 In some studies,25 the diagnosis of rPSC was based only on histological findings, but it is known that liver biopsy is inaccurate for the diagnosis of PSC due to sampling problems. In other studies, the diagnosis of rPSC was made by ERC/MRC performed in the early postoperative period.11, 12 In our study, we adopted strict criteria in order to eliminate other known causes of nonanastomotic biliary strictures before 3 months (as before this time they are usually due to ischemic induced injury),7, 18 and histological lesions that could mimic rPSC, such as ABO incompatibility between donor and recipient and hepatic arterial occlusion. Moreover, although there are some difficulties associated with distinguishing rPSC from chronic rejection,20 histopathological lesions in combination with the clinical history and the radiological findings are usually provide sufficient data for diagnosis of rPSC.20 In our study, we established the diagnosis of rPSC with both liver biopsy and ERC/MRC performed more than 3 months post-LT, and all the patients had a follow-up of at least 1 yr. Our findings were the same when we excluded the 3 patients who developed chronic rejection in our series from our analysis or when we considered them as having rPSC. Our recurrence rate was 13.5%, similar to a recent review in which it was 11%, but with twice the median follow-up of our study (110 vs. 58 months).8

The pathogenesis of rPSC is considered multifactorial and different factors before and/or after LT may affect the recurrence of PSC in the allograft,6, 7, 14, 17–19 all of them within the context of a genetic predisposition. In our study, in contrast to several of the previous studies, we evaluated concomitantly the impact of potential risk factors in detail, such as the immunosuppression (regimens, duration, etc), the pre- and post-LT UC (severity, therapy, colectomy, etc), CMV infection after LT, the human leukocyte antigen disparity, and the type of biliary anastomosis.

In the univariate analysis we found that both pre-LT factors (more frequent patient admissions to hospital for UC exacerbation and absence of UC-pre-LT or colectomy), and post-LT factors (de novo UC as well as longer maintenance of steroids) were associated with rPSC. These findings are in support of and go further than previous studies. In the study by Vera et al.,19 in which 152 patients were evaluated, male gender and an intact colon post-LT were the strongest predictors of rPSC. Thus, 56 (37%) had rPSC during the follow-up period (range 1.4-120 months), and only 1 (6%) of 17 with pre-LT colectomy had rPSC 18 months after LT. However, it is not stated if the absence of UC before LT and/or the occurrence of UC after LT (de novo UC) were, either negatively or positively correlated with rPSC. In our study, rPSC occurred in 7 (13.5%) of 53 patients who were alive 1 yr after LT and/or met inclusion/exclusion criteria. In addition, none of the PSC patients without UC before or after LT (n = 14), as well as those with pre-LT total colectomy (n = 6) had rPSC. Thus, adding to the study by Vera et al.,16 it seems that the absence of UC after LT (due to pre-LT colectomy or not) is an important factor preventing rPSC. These findings may give support to the hypotheses that UC and PSC have a common pathogenesis.26 However, Gautam et al.,8 acknowledging the lack of detailed data, found no difference between the presence or absence of inflammatory bowel disease. In addition, we did not find male gender to be associated with rPSC, as it developed in 2 of 23 females vs. 5 of 30 males (P = 0.39), a finding different from Vera et al.16

Although Vera et al.16 did not find a significant correlation between severity of pre- or post-LT UC and rPSC, they suggested that their findings are in support of a hypothesis that a leak of bacterial toxins from the inflamed colon are pathogenetic for PSC. In our study, we found (in the univariate analysis) that severity of pre-LT UC (which may lead to higher leakage from intestinal mucosa) or development of de novo UC after LT are associated with rPSC. Although we were not able to find a direct correlation between severity of post-LT UC and rPSC, it is known that the course of UC after LT remains stable or becomes worse after LT, and rarely does it improve,4, 27 restricting the range of variability of severity of UC that is seen. Thus, usually any patient with severe UC pre-LT has usually severe or steroid-dependent UC after LT.

Similar to our study, Vera et al.19 (in the univariate analysis) found that duration of steroid treatment (in months) was associated with rPSC. We presume that this was related to the activity of the UC, but this was not stated. Brandsaeter et al.17 (in a multivariate analysis) and Kugelmas et al.18 found that steroid-resistant rejection (in which OKT3 or antithymocyte globulin were used) and utilization of OKT3 for refractory rejection episodes, respectively, were significantly associated with rPSC (whether both of these were related to the same patients it is unclear). Moreover, Kugelmas et al.18 found a trend between maintenance steroids and rPSC (P = 0.21). Similarly, in our study, rPSC was independently correlated with maintenance steroids, which were given to suppress activity of UC, both in the total cohort as well as in the subgroup of patients with UC and intact colon or de novo UC post LT. Thus, maintenance steroids mainly reflected UC activity and not graft dysfunction.

In the above studies,17, 18 the only data that were assessed regarding the UC were: its presence or absence,17, 18 the development of colon cancer18, and colectomy if performed,17 but no association was found with rPSC. Only Graziadei et al.15 found an association between the presence of UC and rPSC, but this was not significant (P > 0.05). However, data such as severity of UC, de novo UC, or if colectomy was performed pre or post-LT were not evaluated. Moreover, Vera et al.,19 did not document in detail any relationship between immunosuppression and activity of UC. In our study, in which immunosuppressive therapy, activity of UC, and, for the first time, the impact of de novo UC after LT were evaluated, we found that the presence of severe or de novo UC after LT is a risk factor for rPSC. This may be reflected by the fact that maintenance steroids (>3 months) (which was the only independent factor for rPSC) was given for activity of UC.

Similarly to previous studies,15, 17 we found that post-LT CMV infection, human leukocyte antigen disparity between donor and recipient, and the type of biliary anastomosis were not significantly associated with rPSC. Thus, although many patients with PSC undergo a Roux-en-Y anastomosis at the time of LT, which leaves the bile ducts exposed to potential bacterial infections,16 rPSC could not be explained by this potential pathogenetic mechanism.11, 25

In our study, we did not find a significant difference in the 5- or 10-yr survival after LT in patients with rPSC, compared to those without rPSC. Thus, similar to previous study,7 rPSC did not have a negative impact on patient survival. However, this may be related to the fact that patients with rPSC had to undergo re-LT more frequently secondary to recurrent disease (3 of 7 patients in our study); longer follow-up beyond 10 yr may be required to assess the impact of survival of rPSC.