Telephone: +44 121 627 2414; FAX: +44 121 627 2449
A re-evaluation of the risk factors for the recurrence of primary sclerosing cholangitis in liver allografts
Article first published online: 25 FEB 2009
Copyright © 2009 American Association for the Study of Liver Diseases
Volume 15, Issue 3, pages 330–340, March 2009
How to Cite
Alabraba, E., Nightingale, P., Gunson, B., Hubscher, S., Olliff, S., Mirza, D. and Neuberger, J. (2009), A re-evaluation of the risk factors for the recurrence of primary sclerosing cholangitis in liver allografts. Liver Transpl, 15: 330–340. doi: 10.1002/lt.21679
- Issue published online: 25 FEB 2009
- Article first published online: 25 FEB 2009
- Manuscript Accepted: 23 SEP 2008
- Manuscript Received: 9 JUL 2008
Previously, we have found that the absence of the colon after liver transplantation (LT) protects the patient from recurrent primary sclerosing cholangitis (rPSC). As our previous observation has not been confirmed in other series, we have reviewed our cohort of patients grafted for primary sclerosing cholangitis (PSC) with greater numbers and longer follow-up to reassess the rate, consequences, and risk factors for rPSC. We collected data on patients who underwent LT for PSC between January 1986 and April 2006. Data were collected for cytomegalovirus status, inflammatory bowel disease status, time of colectomy, type of colectomy, donor-recipient gender mismatch, recipient sex, extended donor criteria (EDC), and donor risk index. Accepted criteria were used to diagnose rPSC. Of a total of 230 consecutive adult patients, 61 (27%) underwent colectomy pre-/peri-LT, and 54 (23.5%) developed rPSC at a median of 4.6 (range, 0.5–12.9) years post-LT. A total of 263 deceased donor grafts were used, and 73 were EDC grafts. A diagnosis of rPSC was made in 61 of the 263 grafts (23%). The recurrence-free patient survival was significantly better (P < 0.05) in patients who underwent pre-/peri-LT colectomy and in those with non-EDC grafts. In conclusion, in this larger cohort of 230 patients and with longer follow-up of 82.5 (range, 0.0–238.6) months [in comparison with the previous report of 152 recipients with a follow-up of 52.8 (range, 1–146) months], we have shown that colectomy remains a significant risk factor for rPSC and that colectomy before and during initial LT for PSC confers a protective effect against rPSC in subsequent graft(s). Moreover, we have shown that EDC grafts are also a significant risk factor for rPSC. Liver Transpl 15:330–340, 2009. © 2009 AASLD.
Five years ago, we published a retrospective multivariable analysis of the risk factors associated with recurrent primary sclerosing cholangitis (rPSC) in liver allografts after transplantation and reported that the presence of an intact colon before transplantation was significantly associated with recurrence.1 Other groups have found different risk factors to be significantly associated with rPSC [recipient-donor gender mismatch,2 use of OKT3,3 corticosteroid-resistant rejections,4 maintenance steroids for ulcerative colitis (UC; >3 months) post–liver transplantation (LT),5 acute rejection,6 human leukocyte antigen DRB1*08 (HLA-DRB1*08),6 and the presence of cholangiocarcinoma prior to LT7] or no risk factors8 (Table 1).
|Transplant Center||Year||Method of Diagnosis of rPSC||Risk Factor(s)||Cohort Size||Median Follow-Up (Range) in Months||Recurrence Rate (Median Time)||Reference|
|Queen Elizabeth Hospital, Birmingham, UK||2002||Histological or radiographic evidence||Male gender, intact colon pre-LT||152||52.8 (1–146)||37% (36 months)||Vera et al.1|
|Lahey Clinic Medical Center, Burlington, MA||2003||Histological and/or radiographic evidence||Recipient-donor gender mismatch||51||Not stated (24–168)||11.8% (not stated)||Khettry et al.2|
|University of Colorado Health Sciences Center, Denver, CO||2003||Histological or radiographic evidence||Use of OKT3||71||Not stated (14–91)||21.1% (53 months)||Kugelmas et al.3|
|Rikshospitalet University Hospital, Oslo, Norway||2005||Radiographic evidence||Corticosteroid-resistant rejection||49||76.8 (16.8–182.4)||18% (not stated)||Brandsaeter et al.4|
|Royal Free Hospital, London, UK||2008||Histological and radiographic evidence||Steroids for UC (>3 months) post-LT||69||110 (12–185)||13.5% (60 months)||Cholongitas et al.5|
|University of Washington, Seattle, WA||2008||Histological and/or radiographic evidence||Acute rejection, donor or recipient HLA-DRB1*08||69||50 (1–173)||10% (68 months)||Alexander et al.6|
|University of Colorado Health Sciences Center, Denver, CO||2008||Radiographic evidence||Pre-LT cholangiocarcinoma||130||66 (not stated)||16.9% (not stated)||Campsen et al.7|
Although there is good patient and graft survival after LT for primary sclerosing cholangitis (PSC), with reported median survival rates of 10 to 12 years,9, 10 rPSC has a major impact on graft survival.11, 12 The diagnosis of rPSC is made on the basis of strict diagnostic criteria and the exclusion of confounding entities.8 There is no gold standard to distinguish rPSC from secondary sclerosing cholangitis, but the Banff working group has recently published histopathological criteria that can be used in conjunction with clinical, serological, and radiographic findings to reach the final diagnosis of rPSC.11
Although information on the timing of rPSC was missing from most of the 30 studies in Gautam et al.'s systematic review of recurrent autoimmune liver diseases after LT,13 the mean time to histological diagnosis has been reported as almost 4 years post-LT.8
Our earlier report of the association of colectomy with rPSC has not been confirmed in a recent study.5 To address this, we have re-examined our series with a much larger cohort of patients and with longer follow-up. We have again found that the presence of an intact colon before LT is the strongest predictor of rPSC.
PATIENTS AND METHODS
We prospectively collected risk factor data for rPSC in 230 consecutive adult recipients of deceased donor liver allografts as treatment for PSC between January 1986 and April 2006. All liver allografts were obtained from heart-beating donors, and there were no donations after cardiac death in our series. The diagnosis of PSC was made by pre-LT radiology and/or endoscopic cholangiography and confirmed by histological examination of the liver explant histology in all cases. At transplant, biliary drainage was achieved with a Roux-en-Y choledochojejunostomy in all but 4 transplants, for which duct-to-duct anastomosis was used.
We recorded the following data for each patient: cytomegalovirus status, inflammatory bowel disease (IBD) status, time of colectomy, type of colectomy, donor-recipient gender mismatch, recipient sex, whether or not the graft fulfilled extended donor criteria (EDC),14 donor risk index (DRI),15, 16 corticosteroid-resistant rejection, immunosuppression regime, ursodeoxycholic acid usage, pre-LT albumin, coexisting cholangiocarcinoma, type of graft (whole or split), and OKT3 usage. EDC and DRI are quantitative tools for assessing the risk of donor liver graft failure. EDC offers a scoring system in which there is a greater likelihood of developing early liver allograft dysfunction with increasing donor age and degree of hepatic steatosis,14 and DRI considers donor age, race, and height, the type of graft (whole or split), and factors related to donor death.15 In assessing the effect of the type of colectomy, colectomy operations were grouped as panproctocolectomy, ileoanal pouch surgery, and subtotal colectomy (subtotal colectomy includes hemicolectomy, sigmoid colectomy, and conventional subtotal colectomy). IBD diagnosis was based on a combined evaluation of clinical, endoscopic, and histological findings. All PSC patients underwent colonoscopy or sigmoidoscopy (plus colonic mucosal biopsy) in pre-LT evaluation, and yearly surveillance was undertaken post-LT. All patients with known UC were on long-term mesalazine (1.5–2 g/day) within the first 3 months after LT, and during the study period, ursodeoxycholic acid therapy (10–15 mg/kg/day) was commenced in 2002.
Patients received a calcineurin inhibitor (cyclosporine or tacrolimus), azathioprine, and corticosteroids. Acute rejection episodes were treated with high-dose prednisolone for 3 days. Steroid-resistant rejections were treated with lymphocyte antibodies [Orthoclone (OKT-3)] or sirolimus. Steroids were tapered within 3 months if possible. A diagnosis of PSC recurrence was not an indication to alter immunosuppressive therapy. Immunosuppression management regimens are described in detail elsewhere.17
Diagnosis of rPSC
The diagnosis of rPSC was guided by the Mayo Clinic diagnostic criteria. The diagnosis of rPSC was based on histological and radiological features consistent with PSC in the absence of defined causes of secondary sclerosing cholangitis (eg, blood-supply abnormalities or preservation or reperfusion injury) and nonanastomotic biliary strictures.8 Post-LT, all patients underwent consented liver biopsies at 1 year, 3 years, or as clinically indicated by the post-LT care protocol in our center. If indicated by abnormal liver tests, liver allografts were initially examined with ultrasonography. If the subsequent biopsy result was abnormal, further investigation was carried out with magnetic resonance cholangiography or percutaneous transhepatic cholangiography. Radiological and histological examinations were carried out by experienced liver radiologists and pathologists. Histological recurrence was diagnosed by the presence of fibrosing duct lesions or by compatible features of bile duct loss accompanied by signs of chronic cholestasis (including ductular reaction, periportal fibrosis, and copper-associated protein deposition), as previously described.18 Radiological features of recurrence were defined as intrahepatic or extrahepatic nonanastomotic strictures in the presence of a normal blood supply (any abnormality of flow was excluded by a Doppler scan of the hepatic artery, portal vein, and suprahepatic vein in all patients). Patients were excluded from the analysis if sclerosing cholangitis occurred in the presence of graft ischemia or hepatic artery thrombosis, ABO-incompatible grafts, histological evidence of chronic allograft rejection, or recurrent bacterial or viral cholangitis.
All data were analyzed with the statistical package SPSS (version 16.0, SPSS, Inc., Chicago, IL). Quantitative variables were summarized as mean values ± 1 standard deviation and/or median values (range) if their distribution was skewed. Significance testing was 2-sided with the level set at 0.05. Kaplan-Meier analysis and the Tarone-Ware test were used to identify the factors that were associated with the recurrence of PSC after LT in the whole data set and also in the subset of patients with an intact colon at LT. Hazard ratios were estimated with Cox regression analysis.
A total of 230 adult patients underwent initial transplantation for PSC at a median age of 47.5 (range, 16.4–72.1) years. The majority (95.2%) of patients were Eurocaucasoid, and the rest were Asian-Caucasoid or Black/Negroid. There were 171 males and 59 females (male/female ratio of 2.9:1). The median follow-up was 6.9 (range, 0.0–19.9) years. Of the 230 patients, 162 (70%) had IBD, predominantly UC (146), and 16 had Crohn's colitis (Table 2).
|Cohort size||230 adults|
|Median age at initial LT for PSC||47.5 (range, 16.4–72.1) years|
|Overall survival at 1, 5, and 10 years (%)||80%, 68%, and 57%|
|Overall PSC recurrence, n (%)||54 (23.5%)|
|Time to diagnosis of PSC recurrence post-OLT||4.6 (range, 0.5–12.9) years|
|Sex distribution||171 males, 59 females (2.9:1)|
|Inflammatory bowel disease||162 patients (UC = 146,Crohn's = 16)|
|Median follow-up||82.5 (range, 0.036–238.6) months|
|Presence of an intact colon before OLT, n (%)||201 (87.4%)|
|Tacrolimus primary immunosuppression, n (%)||120 (45.6%)|
|Cyclosporine primary immunosuppression, n(%)||139(52.9%)|
|Cholangiocarcinoma in primary PSC explant, n (%)||18 (7.8%)|
|Corticosteroid-resistant rejection, n (%)||8 (3.0%)|
There were a total of 263 liver transplants in the series, and 27.8% of these were known to be carried out with EDC grafts.14 Primary immunosuppression was tacrolimus in 45.6% of cases and cyclosporine in 52.9%. A diagnosis of rPSC was made in 61 grafts (23%) at a median time of 4.6 (range, 0.5–12.9) years (Fig. 1).
Patient survival at 1, 5, and 10 years was 80%, 68%, and 57%, respectively. Ninety patients died during the study period at a median time of 3.1 (range, 0.003–15.8) years after initial LT. The causes of death were bacterial sepsis (21%), neoplasia (18%), hemorrhage (12%), multiorgan failure (13%), rPSC (8%), chronic rejection (8%), pulmonary complications (4%), fungal sepsis (4%), renal failure (2%), cerebral dysfunction (3%), and non–transplant-related causes (7%).
Graft survival at 1, 5, and 10 years was 75%, 60%, and 50%, respectively. Of the 53 graft failures, 30 failed for reasons other than recurrent disease (rPSC-related causes discussed below). In this group, there were 23 regrafts in 21 patients: indications were chronic rejection (4), graft vascular insufficiency (17), massive hemorrhagic necrosis (1), and primary nonfunction (1). Seven grafts failed as a result of chronic rejection but were not regrafted.
rPSC and Outcome
Of the 61 grafts (in 54 patients) in which rPSC was diagnosed, 23 grafts (in 20 patients) failed as a result of rPSC at a median time of 5.2 (range, 1.4–12.1) years: of these, 13 retransplant operations were carried out in 11 patients at a median time of 5.2 (range, 1.4–11.7) years (Fig. 2A,B). Of the 11 patients regrafted because of rPSC, 2 developed further rPSC in subsequent grafts. Patient survival was worse in the presence of a diagnosis of rPSC versus survival in the absence of rPSC, but this failed to reach statistical significance after the exclusion of all patients who died within the first 6 months (P = 0.28; Fig. 3A). However, on exclusion of all patients who died before 6 months post–liver transplant, restriction to single-transplant patients, and adjustment for age, there was significantly (P = 0.05) better survival in patients without a diagnosis of rPSC (Fig. 3B).
Analysis of Risk Factors for Recurrence
The univariable analysis revealed that the significant factors for rPSC were time of colectomy, type of colectomy, and EDC graft (Table 3).
|Risk Factor||rPSC||P Value|
|Time of colectomy||0.028|
|Type of colectomy||0.024|
|Pre-LT ilcoanal pouch||1||10|
|Pre-LT subtotal colectomy||0||10|
|Post-LT ileoanal pouch||2||0|
|Post-LT subtotal colectomy||5||11|
|Extended donor criteria graft||0.046|
|Use of OKT3||0.727|
|Inflammatory bowel disease||0.065|
|Donor recipient gender mismatch||0.251|
|Donor risk index||0.333|
Time of Colectomy.
The risk of rPSC was significantly lower (P = 0.028) in those who underwent colectomy pre-/peri-LT (1 of 29) for PSC compared to those who underwent colectomy post-LT (14 of 32) or those that did not undergo colectomy at all (39 of 169; Fig. 4).
The recurrence-free patient survival at 1, 5, and 10 years was 100%, 95%, and 95%, respectively, in patients who underwent colectomy pre-/peri-LT; 97%, 70%, and 51%, respectively, in patients who underwent colectomy post-LT; and 97%, 76%, and 60% in patients that did not undergo colectomy. The hazard ratios for post-LT colectomy and no colectomy (with respect to pre-/peri-LT colectomy) were 11.8 (95% confidence interval, 1.55–89.6) and 8.85 (95% confidence interval, 1.22–64.5), respectively.
Type of Colectomy.
There was significantly less risk of rPSC in all the pre-/peri-LT colectomy groups compared to the post-LT colectomy groups. Of those who underwent colectomy pre-/peri-LT, only 1 developed rPSC, and this was diagnosed in a patient with an ileoanal pouch. There were no recurrences in the panproctocolectomy and subtotal colectomy pre-/peri-LT groups. In the post-LT colectomy groups, there was a significantly higher risk of rPSC in patients with ileoanal pouches compared with those who had a panproctocolectomy (P = 0.002) or subtotal colectomy (P = 0.010). Overall assessment of all pre-/peri-LT and post-LT colectomies showed no significant difference in the risk of rPSC between patients that underwent panproctocolectomy and other forms of colectomy in which residual colorectal tissue is left behind (ileoanal pouch and subtotal colectomies).
Extended Donor Criteria (EDC) Grafts.
EDC grafts14 were significantly associated with rPSC (P = 0.046). We also analyzed the effect of DRI high-risk (>2) and low-risk (<2)15, 16 donor grafts but found no significant difference in the risk of rPSC (P = 0.333).
For EDC grafts, recurrence-free patient survival at 1 and 5 years (insufficient 10-year data were available for estimates to be reliable) was 98% and 71%, and the respective figures for non-EDC grafts were 98% and 91%. The hazard ratio for EDC grafts was 5.03 (95% confidence interval, 1.66–15.3).
Lack of data on steatosis for the earlier grafts in our series meant that the EDC status of these grafts could not be calculated (data were available for 147 of 263 grafts). Considering just those patients for whom the graft EDC status was known, we found that none of those with pre-/peri-LT colectomy developed rPSC. For this reason, we looked for risk factors for rPSC in just those who did not have pre-LT colectomy. EDC grafts were significantly associated with rPSC (P = 0.009) when those with pre-/peri-LT colectomy were excluded from the analysis (Fig. 5) with a hazard ratio of 5.01 (95% confidence interval, 1.64–15.3). Type of colectomy was not a significant risk factor for rPSC in those who had a pre-LT colectomy (P = 0.44) or in those with an intact colon at LT (P = 0.25).
Our current study has the largest patient cohort and the longest maximum follow-up time of all the published rPSC studies to date. Our cohort of patients compares favorably with Gautam et al.'s13 pooled cohort but has a slightly higher mean age at initial LT for PSC (47.5 and 43.5 years, respectively), a longer follow-up (medians of 6.87 and 4.83 years, respectively), a higher male/female ratio (2.9:1 and 1.8:1, respectively), and a higher percentage incidence of rPSC (23.5% and 17%, respectively). In our present cohort, the first case of rPSC was diagnosed after 6 months' follow-up time, and the overall recurrence rate on exclusion of all 40 of the 230 patients with less than 6 months' follow-up was higher at 28.3%. The outcome of an analysis of risk factors for all 190 of the 230 patients that survived ≥6 months was unchanged as the analysis of the entire cohort was performed by Kaplan-Meier analysis. After exclusion of all patients with less than 6 months' follow-up, colectomy and EDC grafts remained the only significant risk factors for rPSC. When the data set was adjusted for age and restricted to the 203 single-graft-only transplant recipients, time of colectomy (P = 0.023) and type of colectomy (P = 0.011) remained significant. Also, for all 230 patients, pretransplant colectomy was significant (P = 0.033) when adjusted for age.
Our extended series confirms the previous observation that an intact colon remains a significant risk factor for rPSC and that colectomy pre-/peri-LT for PSC confers a protective effect against rPSC. Despite the higher initial mortality in those without recurrence, recurrence-free survival following pre-/peri-LT colectomy was significantly better than that following post-LT colectomy or without colectomy. Using the criteria of Tekin et al.,14 we also found that an EDC graft is a significant risk factor for rPSC.
The beneficial effect of pre-LT colectomy may shed light on the pathogenesis of rPSC. Inflammation in the gut in the form of IBD (whether UC or Crohn's disease) is closely linked to PSC; such inflammation might also fuel liver inflammation pre-LT and post-LT. Between 2.4% and 7.5% of patients with IBD will develop PSC, whereas 70% and 85% patients with PSC will suffer from IBD during their lifetime.10 The strongest association is with UC (70%–80%), with Crohn's colitis accounting for only 2% and 7.5%.19 The nature and distribution of IBD associated with PSC are, however, significantly different in comparison with patients with UC alone. UC in PSC is often a right-sided colitis with involvement of the terminal ileum, a feature termed backwash ileitis.19 PSC is rare in patients with isolated small bowel Crohn's disease and is almost always associated with large bowel involvement.20, 21
It is recognized that patients may develop PSC for the first time many years after total colectomy for colitis, and colonic inflammation can occur for the first time after patients have undergone LT for PSC despite the presence of immunosuppressive drugs.22 The gut-liver link may be explained by the destructive lymphocytic infiltrate that is common to both PSC and IBD. Evidence in support of an enterohepatic lymphocyte pathway comes from the observation that the molecules that normally only mediate the recruitment of lymphocytes to the gut are expressed in the PSC liver. Mucosal vascular addressin cell adhesion molecule 1 (MAdCAM-1) and the gut-associated chemokine, chemokine (C-C motif) ligand 25 (CCL25), are implicated in the gut-to-liver trafficking seen in PSC.23, 24 Small-bowel mucosal lymphocyte-expressed chemokine (C-C motif) receptor 9 is bound by CCL25 triggering activation of α4β7-integrin–mediated binding to MAdCAM-1 on hepatic vessels, thus permitting recruitment of mucosal lymphocytes to the liver. Also implicated in the liver-to-gut pathway is vascular adhesion protein 1, which supports lymphocyte adhesion to normal human liver vessels25 and is markedly up-regulated in mucosal vessels in IBD.26 On the basis of current evidence, the small bowel is almost certainly the nidus of lymphocytes trafficked via the enterohepatic circulation to the liver. The link with colonic inflammation is explained by the distinct IBD phenotype in PSC patients that predisposes them to backwash ileitis.19 The suggestion that any residual colon is still sufficient to drive mucosal T cells to the liver in the setting of inflammation and backwash ileitis raises important questions about the management of IBD in PSC patients.
Although rPSC has a major impact on patient and graft survival and colectomy at the time of transplant does not add to the morbidity and mortality of the procedure,1 we are not advocating routine colectomy in these patients. The type and timing of colonic surgery can be controversial, so the pros and cons of the available options must be considered for the individual, and the risks and benefits must be considered. If colectomy is indicated for reasons such as malignancy, severe dysplasia, or uncontrolled IBD, panproctocolectomy and end-ileostomy may be preferable in patients with a native diseased liver as there is an increased risk of pouchitis following ileoanal pouch formation.27 A further concern is that those with PSC and UC with ileoanal pouch have a higher risk of neoplasia and dysplasia in the pouch mucosa than UC patients without PSC.28 It is recognized that backwash ileitis in PSC patients with IBD is a risk factor for dysplasia, neoplasia,29 and chronic pouchitis.30 Indeed, 1 recipient with PSC and UC and pre-LT colectomy with ileoanal pouch developed adenocarcinoma in his pouch several years after his LT for PSC.31 In light of the risk of neoplastic change, panproctocolectomy may be more attractive over ileoanal pouch formation. In our post-LT colectomy cohort, there were significantly more cases of rPSC in the ileoanal pouch group (100%) compared to either the panproctocolectomy or subtotal colectomy groups (P≤0.01). In the pre-LT colectomy cohort, there was only 1 case of rPSC, and this occurred in a patient with an ileoanal pouch. The higher incidence of rPSC in the ileal pouch–anal anastomosis group may reflect a greater incidence of ileitis in the context of pouchitis. It was not possible to make any comparisons between the different colectomy groups in the pre-/peri-LT cohort as there was only 1 recurrence overall. The overall multivariable insignificance of type of colectomy suggests that the observed effect is largely due to the significance of time of colectomy.
In a systematic review of rPSC, Gautam et al.13 did not find any significant association between rPSC and IBD even though they included only studies in which raw data were available in order to overcome the effect of publication bias. A single cohort study (71 patients total; rPSC rate of 21.1%) found that although the length of time before rPSC was significantly shorter in the presence of IBD, the incidence of rPSC was unaffected.3 Our data correlate with the findings of Gautam et al., showing that the presence of IBD is not a significant univariable for rPSC (P = 0.065) and does not (P = 0.736) affect the time to histological diagnosis of rPSC. A recent analysis (53 patients; median follow-up, 9.17 years)5 concluded by multivariable logistic regression analysis that maintenance steroids for UC (>3 months) post-LT were significantly associated with rPSC (P = 0.025). The same study also showed that the presence of severe or de novo UC after LT was a risk factor for rPSC in the univariable analysis but not in the multivariable analysis. In support of our earlier work, the study by Cholongitas et al.5 also takes a colocentric view of rPSC risk factors, but it does raise uncertainties. First, subgrouping UC into pre-existing or de novo cases may not account for the contribution of subclinical disease. Second, there were very few cases of de novo UC (n = 3), and this may compromise the validity of the conclusions drawn. Most importantly, as UC does not show significance in the multivariable analysis, it is possible that the prolonged steroid therapy effect is not related to UC. Prolonged steroid therapy may affect rPSC indirectly by altering the immune response. It may exert an effect similar to that of calcineurin inhibitors, which are known to be associated with reduced numbers of CD4+, CD25+ regulatory T cells.32, 33 Therefore, although it is given to suppress activity of UC, prolonged steroid therapy must be interpreted with caution as it may compromise rPSC-free survival of grafts. Just as in our previous study, our current data show that there is no significant effect of UC on rPSC (P = 0.065).
As previously reported,3, 18 we found that primary immunosuppression with either cyclosporine or tacrolimus is not a significant risk factor (P = 0.222). There were significantly more recurrences in the patients treated with cyclosporine than in those treated with tacrolimus, but this is negated when longer follow-up of patients on tacrolimus with respect to those on cyclosporine is taken into account. Steroid-resistant acute rejection, treated with OKT3 (n = 2) or sirolimus (n = 6), was not a significant risk factor, although numbers are very small, so a type 2 error cannot be excluded.
Long-term treatment of PSC patients with high-dose ursodeoxycholic acid is associated with improvement in liver biochemistry and histology, although an effect on survival remains unproven,34–37 and it may be chemopreventive against carcinoma of the colon.38 Our current data do not suggest a protective effect against rPSC (P = 0.614), but longer follow-up may be needed for this to be realized.
In this series, we found that the time to recurrence is greater, and this may reflect the observation that new cases are recognized over time. There is also an impression that with subsequent grafts, the interval to recurrence is reduced, although the numbers of these grafts are small. The median time to recurrence was 4.75 years (n = 50) in the first graft, 4.87 years (n = 8) in the second graft, 2.85 years (n = 2) in the third graft, and 0.56 years (n = 1) in the fourth graft. Distinct HLAB8 expression has been shown in patients who have PSC,39, 40 and HLA-DRB1*08 is associated with increased risk of rPSC.6 This may explain why some patients have a more aggressive course of disease with recurrences in sequential grafts. The delineation of how rPSC differs in the sequential grafts of the same patient will be crucial to the understanding of the complex disease process.
In comparison with our earlier report,1 we found in this larger series a lower rate of and a longer time to developing rPSC (23.5% at a median of 4.1 years compared with 37% at a median of 3 years in the earlier study). The obvious problem of false negatives due to the limited diagnostic indices used is reduced by the fact that our unit routinely follows up all liver transplants with protocol biopsies that are carried out at more frequent intervals if demanded by clinical parameters.
It is unclear why EDC grafts are a significant risk factor for rPSC. There is little evidence available on the effect of EDC grafts on recurrence of liver diseases. Studies looking at the effect of EDC grafts on recurrence of HCV suggest that prolonged cold ischemia41, 42 increases the likelihood of HCV recurrence. Indices of EDC grafts have shown significant associations with recurrent PBC,43 but the effect of EDC grafts has never been specifically analyzed. We speculate that the effect may be related to a greater graft predisposition to autoimmune insult as a result of ischemia-reperfusion injury or donor-specific factors ultimately leading to the development of rPSC. The lack of an association between rPSC and the DRI may reflect the different components of the 2 models: the major factors in the Tekin model are the degree of steatosis and the cold ischemic time, whereas the DRI includes only donor factors. The effect of the cold ischemic time on the outcome of the graft is greatest in steatotic grafts: thus, a degree of reperfusion injury may allow for immunological engagement and so lead to rPSC in much the same way that reperfusion injury may lead to acute rejection. There is, however, the limitation of an incomplete EDC data set, which makes it impossible for us to exclude the possibility of patient selection bias, and it must be borne in mind when we interpret the results of our analyses.
The course and long-term outcome of rPSC are still uncertain. It is likely that the host factors that lead to the development of PSC in the native liver will overlap greatly with those that lead to the development of rPSC in the allograft. There are distinct differences between the conclusions of rPSC risk factor studies published by different liver transplant centers, but by continual updating of the cumulative data in all these series, common risk factors may ultimately begin to emerge.
The authors thank their clinical colleagues and other medical and nursing staff for their clinical care of the patients in this study.
- 9Characterization, outcome, and prognosis in 273 patients with primary sclerosing cholangitis: a single center study. Am J Gastroenterol 2007; 102: 107–114., , , , .Direct Link:
- 34High-dose ursodeoxycholic acid as a therapy for patients with primary sclerosing cholangitis. Am J Gastroenterol 2001; 96: 1558–1562., , , , .Direct Link: