Review article: current management of primary sclerosing cholangitis
Dr S. Cullen, John Radcliffe Hospital, Oxford OX3 9DU, UK.
The management of primary sclerosing cholangitis (PSC) is hindered by incomplete understanding of the pathogenesis of the disease and the lack of good prognostic models. Few large randomized controlled trials of drug therapy have been published. Best practice in the management of PSC is currently based therefore on careful interpretation of the available evidence, close observation of individual patients and clinical experience of the disease. Drug therapy is useful for alleviating symptoms. Ursodeoxycholic acid may slow progression of the disease and reduce the frequency of complications. Consensus is emerging on the issues of screening for the malignant complications of PSC and the indications for liver transplantation are becoming broader and encompassing the earliest stages of cholangiocarcinoma. In view of the rarity of the disease in the general population, large international collaborations to study PSC are necessary to provide clearer answers in areas of uncertainty, and these are now beginning to emerge.
Primary sclerosing cholangitis (PSC) is a chronic, progressive disease with a median survival from diagnosis to death or liver transplantation currently estimated at 18 years.1 Patients suffer with the symptoms of chronic cholestatic liver disease, particularly itch and fatigue, and as the disease progresses, with symptoms of cirrhosis. PSC cholangitis is associated with a high and currently unpredictable risk of the development of cholangiocarcinoma.
Although there is no curative medical therapy for this disease at present, increasingly effective treatments for the symptoms of the disease have become available and preliminary studies have demonstrated that ursodeoxycholic acid (UDCA) in high dose may retard the rate of disease progression.2, 3 This article reviews the current management of PSC focusing on trials of drug therapy, management of complications and the role of liver transplantation.
Pathogenesis of PSC
PSC is a chronic cholestatic disease of the intra- and extra-hepatic bile ducts. It is characterized by stricturing and dilatation of the biliary tree, and concentric obliterative fibrosis of intrahepatic bile ducts. PSC is closely associated with inflammatory bowel disease (IBD), particularly ulcerative colitis (UC) which is found in approximately two-thirds of northern European PSC patients.4, 5
The aetiology and pathogenesis of PSC are not yet well understood. There is now however, a large and growing body of evidence that immune dysregulation plays a key part in the development of the disease. Cellular immune abnormalities have been demonstrated in PSC, including changes in tissue lymphocyte populations, abnormal cytokine repertoires and the aberrant expression of human leucocyte antigen (HLA) class II molecules on the bile duct epithelium.6–12 All these studies provide indirect evidence that PSC is an immune-mediated disease possibly involving an exaggerated cell-mediated immune response with immunological damage targeted at the biliary epithelial cells.
Primary sclerosing cholangitis has some features which are characteristic of an autoimmune disease. These include the presence of a range of autoantibodies in the serum of PSC patients, the coexistence of other autoimmune diseases with PSC, and the strong association with the ‘autoimmune’ HLA haplotype B8-DRB1*0301-DQB1*0201.13–16 The mechanisms underlying the loss of immune tolerance which allow such ‘autoimmune’ disease processes to occur have not as yet been identified and, interestingly, drugs useful in classical autoimmune diseases, for example, steroids and other immunosupressant drugs have, to date, failed to prove useful in the management of PSC. However, as is discussed below, UDCA, which has a range of potential modes of action including effects on the immune system has shown some promise.
The importance of genetic susceptibility to the development of PSC is currently under investigation. It appears unlikely that susceptibility to PSC will prove to be attributable to a single gene locus, but will be acquired by inheriting one of a number of patterns of genetic polymorphisms, which together cause a predisposition to the development of the disease. Four key HLA haplotypes have been shown to be associated with the development of the PSC and some non-HLA loci are being explored including CTLA-4, CCR-5, interleukin (IL)-1, IL-10 and MMP-3.14, 15, 17–23
The intimate connection between PSC and inflammatory colitis has led to a hypothesis which proposes that the initial event in the pathogenesis of PSC is the ingress of bacteria into the portal circulation through an inflamed and permeable gut wall. Bacterial cell products might then be taken up by hepatic macrophages, setting up an immune response leading to peribiliary fibrosis in immunogenetically susceptible hosts. Although this hypothesis is attractive, to date there is little solid evidence to support it.24 More recently, Grant et al. have proposed the existence of an enterohepatic circulation of lymphocytes, whereby some mucosal lymphocytes produced in the gut during active inflammation persist as memory cells capable of recirculation through the liver.25 Under certain circumstances these gut-derived lymphocytes might become activated resulting in hepatic inflammation. This hypothesis is supported by the finding that some lymphocyte homing receptors are shared by the gut and the liver.26, 27 This concept of dual homing lymphocytes may explain the observation that PSC runs a course independent of inflammation in the bowel and indeed can develop even after proctocolectomy.
Interesting work on nitrogen oxide species has demonstrated expression of inducible nitric oxide synthase (NOS2) in all biliary structures in advanced PSC. NOS2 appears to be induced by proinflammatory Th1 cytokines, such as tumour necrosis factor (TNF)-α and interferon-γ, and is responsible for the production of NO from l-arginine. High concentrations of NO may result in the formation of reactive nitrogen oxide species (RNOS) that generate a number of pathobiological effects including interference with cAMP-mediated fluid and electrolyte transport causing bile duct and hepatocellular damage by reducing bile hydration and alkalinity.28 Jaiswal et al. have demonstrated that RNOS effects also include protein nitrosylation and an NO-dependent inactivation of DNA repair enzymes in cholangiocytes. Chronic inflammation therefore, via NO generation may be responsible for cholestatic and long-term oncological complications of PSC.29
Diagnosis of PSC
Most patients with PSC are diagnosed at an asymptomatic stage after being found to have abnormalities in their routine liver function tests performed at IBD follow-up appointments. The prevalence of PSC in patients with UC ranges from 2.4 to 4% of patients.30–32 Symptoms of PSC include pruritus, fatigue, right upper quadrant pain, jaundice and, more unusually, cholangitis. Occasionally, PSC patients present with symptoms of advanced liver disease or cholangiocarcinoma.
Laboratory investigation often reveals a cholestatic picture of liver function tests with elevation of the serum alkaline phosphatase. The liver function tests, however, may be normal and can fluctuate during the course of the disease. A range of non-specific autoantibodies is found in the serum of PSC patients. Atypical perinuclear neutrophil cytoplasmic antibody (p-ANCA) has the strongest association with PSC, being present in 33–88% of cases.33–35 This antibody is almost exclusively found in patients with PSC, autoimmune hepatitis (AIH) or IBD.
The diagnosis of PSC is established by visualization of stricturing and dilatation in the intrahepatic or extrahepatic bile ducts. This has traditionally been performed using endoscopic retrograde cholangiopancreatography (ERCP) but more recently magnetic resonance cholangiopancreatography (MRCP) has proved to be as sensitive as ERCP in diagnosing PSC where the best equipment and operator are available.36, 37 Although the diagnosis of PSC can be made on the basis of typical ERCP or MRCP findings alone, liver biopsy is a useful adjunct to determine the stage of the disease and look for evidence of biliary dysplasia.38 It is also, of course, the only way of diagnosing small duct PSC and therefore should always be arranged for colitic patients with cholestatic liver function tests (LFTs) of unknown cause and a normal ERCP (see Small Duct PSC below). The histological findings range from normality to frank biliary cirrhosis with the typical appearances being of portal inflammation, concentric ‘onion skin’ periductal fibrosis and periportal fibrosis developing into septal and bridging necrosis.39 Diagnostic histological changes are seen in only one-third of PSC patients.
Small duct PSC
A subgroup of patients with typical histological changes of PSC but normal ERCP or MRCP examinations has recently been defined and the term ‘small duct’ PSC has been adopted. Three studies have now been published which address the natural history and outcome of patients with small duct PSC and all have shown similar results.40–42 The clinical course of small duct PSC appears to be relatively benign with only 12% of patients progressing to develop classical PSC and no reports of cholangiocarcinoma. Patients with small duct PSC appear to have a similar rate of underlying IBD as those with classical PSC, although the study published from Oxford and Oslo found that a greater proportion of patients with small duct disease had Crohn's colitis rather than UC.42
Disease modifying treatment
The hydrophilic bile acid UDCA is widely used in the treatment of cholestatic liver disease including primary biliary cirrhosis, intrahepatic cholestasis of pregnancy, cystic fibrosis liver disease, progressive familial intrahepatic cholestasis, chronic graft-vs.-host disease and drug and parenteral nutrition-induced cholestasis.43–49 Hydrophobic bile acids are clearly highly toxic to hepatic cells. The mechanism of bile acid-induced cytotoxicity is probably not, as has previously been assumed, because of a detergent-like effect on cellular membranes, but due to apoptosis via ligand-independent death receptor pathways especially those mediated by the Fas receptor.50
The therapeutic actions of UCDA appear to be complex and the relative importance of each in alleviating cholestasis remains unclear.51 First, UDCA appears to protect cholangiocytes from the cytotoxicity of hydrophobic bile acids predominantly by displacing them from the bile acid pool.52 Secondly, UDCA has a choleretic effect, stimulating hepatobiliary secretion and preventing the retention of potentially toxic biliary constituents in the liver. This effect may be mediated by the stimulation of transporter proteins in the hepatocyte and the targeting and insertion of transporter molecules into the canalicular membrane.53–55 Finally, UDCA appears to have an antiapoptotic effect, inducing a survival signal in hepatocytes, via activation of the epidermal growth factor receptor and mitogen-activated protein kinases (MAPK).56 There are a number of other less well characterized potential modes of action of UDCA. It appears to have an effect on reversing the aberrant expression of HLA class I molecules on hepatocytes, for example, although this might be secondary to its anticholestatic effect. In vitro studies have also suggested a possible immunomodulatory role of UDCA on cytokine secretion of peripheral monocytes.57
A number of trials have attempted to address the clinical efficacy of UDCA in the context of PSC. All the trials performed to date have been limited by small number of patients and relatively short follow-up periods. A summary of these trials is shown in Table 1.
Table 1. Trials of ursodeoxycholic acid (UDCA) in primary sclerosing cholangitis (PSC)
|O'Brien et al.182||1991||12||10||Open-label||30||Y||Y||Not done|
|Beuers et al.58||1992||6||13–15||Double-blind placebo controlled||12||Y||N||Y|
|Lo et al.183||1992||23||10||Double-blind placebo controlled||24||Y||N||N|
|Stiehl et al.59||1994||20||750 mg||Double-blind placebo controlled||12–48||Y||N||Y|
|De Maria et al.184||1996||59||600 mg||Double-blind placebo controlled||24||N||Not done||Not done|
|Lindor185||1997||105||13–15||Double-blind placebo controlled||34||Y||N||N|
|van Hoogstraten et al.186||1998||48||10||Double-blind||24||Y||N||Not done|
|Mitchell et al.2||2001||26||20–25||Double-blind placebo controlled||24||Y||N||Y|
|Harnois et al.3||2001||30||25–30||Open-label||12||Y||Not done||Not done|
|Okolicsanyi et al.60||2003||86||8–13||Double-blind placebo controlled|| ||Y||Y||N|
|Olsson et al.62||2004||110||17–23||Double-blind placebo controlled||60||Y||N||Not done|
Although almost all these trials have documented an improvement in the liver function tests of patients on UDCA, only the small trial of Beuers et al.58, Stiehl et al’s.59 trial in 1994 and the high-dose trial of Mitchell et al.2 showed significant improvement in liver histological appearance. Most trials have used the standard dose of UDCA used for primary biliary cirrhosis (PBC) of 13–15 mg/kg. This dose does not appear to make a significant impact on the progression of PSC despite the improvement in liver biochemistry. In only two of the low-dose trials does UDCA appear to provide symptomatic relief. The largest of these was the Okolicsanyi et al.60 trial using a group of southern European patients. These patients had a particularly low incidence of concurrent IBD (37%), which might suggest a different immunogenetic background to the northern European or American patients used in the other trials.
Recent studies on the biliary bile acid composition of PSC patients taking different doses of UDCA has demonstrated that biliary enrichment of UDCA increases with increasing dose and reaches a plateau at 22–25 mg/kg. If biliary enrichment of UDCA is the most important factor for its clinical effect, it seems likely that UDCA doses of up to 22–25 mg/kg may be more effective than lower doses.61 The Oxford group tried high-dose UDCA in 26 patients on the hypothesis that higher doses might be needed to provide sufficient enrichment of the bile acid pool in the context of cholestasis and that these doses might also have an enhanced immunomodulatory effect. Interestingly, this small trial did appear to show significant improvements both in the cholangiographic appearances of PSC and in the degree of liver fibrosis. No significant side-effects were noted, and in particular, no worsening of underlying colitis. An open study from the Mayo clinic showed a significant improvement in projected survival using the Mayo risk score in patients treated with high dose, but not conventional doses of UDCA. The importance of this study is however, limited by the fact that no direct measurement of progression of disease was used other than liver function tests. The largest and most recent trial to date by Olsson et al. has so far been published in abstract form only. This found a trend towards improved survival in the UDCA-treated group but was insufficiently powered to produce a result of statistical significance.62
Ursodeoxycholic acid also appears to have a beneficial influence on the rates of hepatobiliary and colonic malignancies in PSC and this will be discussed in more detail below.
Despite the evidence of immune dysregulation in PSC, there have been few studies exploring the effect of immunosuppressive therapy on the progression of PSC. These patients are however, regularly exposed to courses of corticosteroids when they are prescribed for their coexistant UC and these courses have not been observed to have any impact on their liver disease. Systemic and topical corticosteroid therapy has been evaluated in a number of small and often uncontrolled trials. Two studies of 10 patients with PSC treated with corticosteroids in 1981 and 1984 came to differing conclusions regarding efficacy and a controlled but non-randomized trial of 12 patients treated with a combination of low-dose prednisolone and colchicine failed to find any benefit in terms of slowing of disease progression or improved survival.63–65 There has been a suggestion that there is a small subgroup of patients with overlap syndromes between PSC and AIH who might benefit from steroid treatment, but this has not yet been clearly established.66, 67 Tjandra et al. have demonstrated a reduction in glucocorticoid receptors on hepatic T lymphocytes in a rat model of cholangitis and it is possible that a similar mechanism might explain the ineffectiveness of steroids in human PSC.68
The anxiety regarding the side-effects of long-term use of corticosteroids, particularly in terms of osteoporosis have lead to some trials being undertaken on second generation corticosteroids such as budesonide which has a high first pass metabolism and minimal systemic availability. One study, published in abstract only to date, demonstrated a biochemical improvement at the expense of a degree of bone loss.69 A trial examining the effects of budesonide in combination with UDCA found no additional beneficial effect.70
Topical corticosteroids administered through a nasobiliary drain following ERCP have been reported to be beneficial in three small studies, but a controlled trial from the Royal Free Hospital found no benefit with a high incidence of bacterial colonization of the bile, leading to episodes of bacterial cholangitis.71–74
There have been three trials to date of methotrexate in PSC. An initial open study of 10 patients looked promising but a small, follow-up double-blind placebo-controlled trial comparing oral pulse methotrexate at a dose of 15 mg/week failed to demonstrate any improvement in liver biopsy or ERCP appearances despite a fall in the serum alkaline phosphatase of 31% in patients receiving methotrexate.75, 76 The negative results might have been due to the small number of patients involved, too many patients with advanced disease or too short a period of follow-up. A further trial of a combination of methotrexate and UDCA in 19 PSC patients however, did not find any additional therapeutic effect over UDCA alone.77 These studies, together with the problematic side-effect profile of methotrexate in PBC patients including interstitial pneumonitis and hepatotoxicity, makes it unlikely that further trials will be undertaken.
No control trials of azathioprine in PSC have been published and there have been only a few, conflicting case reports of the efficacy of this drug in the context of PSC.78, 79 Ciclosporin has been shown in one randomized-controlled trial of 34 PSC patients to prevent progression of liver histological change over a 2-year follow-up period, as well as controlling the associated UC and without serious side-effects.80 There was no improvement in biochemical indices on this treatment. A combination of ciclosporin and prednisolone has been reported to have been beneficial in a 65-year-old man with PSC and associated pancreatic duct abnormalities.80
Tacrolimus (FK506), an immunosuppressive macrolide antibiotic, has also been used in a twice daily oral regimen in a small open study of 10 patients with PSC. Substantial improvements in liver function tests occurred after 1 year of treatment but liver histology was not assessed. No significant adverse events were recorded in this preliminary study and a formal controlled trial is awaited.81
Other drugs with actions on the immune system which have been tested for efficacy in PSC include cladribine, pentoxifylline and entanercept. Cladribine is a nucleoside analogue with antilymphoctye properties and was found to reduce peripheral and hepatic lymphocyte counts when given to six patients with early disease for 6 months. Disappointingly however, no significant effect was seen on symptoms, liver function tests or cholangiograms after a 2-year follow-up.82
The hepatobiliary injury which occurs in rats with experimental bacterial overgrowth appears to result from the peptidoglycan polysaccharide-mediated activation of Kupffer cells which in turn release cytokines such as TNF-α. This liver injury can be prevented in rats by pentoxifylline. In an open pilot study of 20 patients with PSC, no effect on symptoms or liver function tests was demonstrated after pentoxifylline treatment at a dose of 400 mg q.d.s. for 1 year.83 Etanercept, an inhibitor of TNF, has been given to 10 PSC patients in a pilot study for 6 months with no evidence of clinical efficacy.84
The potential for using a combination of treatments with theoretical benefits in PSC was explored in an important pilot study by Schramm et al85 who treated 15 PSC patients with low-dose UDCA (500–750 mg daily), prednisolone 1 mg/kg daily and azathioprine 1–1.5 mg/kg daily for a median period of 41 months. All patients had improvements in their liver biochemistry, including seven previously treated with UDCA alone who had no biochemical response until immunosuppression was added. Of particular interest was the finding that six of 10 patients who had follow-up liver biopsies were found to have undergone histological improvement and only one of 10 had significant radiological deterioration at ERCP. The combination therapy was well tolerated in 13 of the 15 patients. These results are encouraging and suggest that formal controlled trials of combinations of URSO and immunosuppressant drugs should be undertaken.85
Other drugs used to modify disease
d-Penecillamine is used in Wilson's disease to aid the elimination of copper ions. Increased hepatic copper levels are found in cholestasis of all causes including PSC, and this was the rationale for testing this drug in a group of 39 PSC patients in a placebo-controlled trial from the Mayo clinic.86 Over a follow-up period of 36 months, there was no improvement in liver histology, symptoms, liver biochemistry or overall survival in the treated group and 21% of patients had to discontinue the drug because of major side-effects including pancytopenia and proteinuria.
Similarly disappointing results were produced by a placebo-controlled trial of 1 mg daily of colchicine in 44 PSC patients over a 3-year period.87 This drug originally appeared to have an antifibrogenic effect in PBC but more recent longer term studies have failed to confirm these results.
In view of the strong inverse relationship between both UC and PSC and cigarette smoking, two trials have been undertaken to determine if nicotine has a beneficial effect in PSC.88–90 Angulo et al. ran a small trial of nicotine 6 mg q.d.s. for 1 year in eight PSC patients.91 No beneficial effects were produced and three patients terminated the study early due to side-effects from the medication. A transdermal nicotine patch trial in 11 patients also failed to demonstrate any improvement in symptoms or liver biochemistry.92Table 2 summarizes the drugs used in trials for treatment of PSC to date.
Table 2. Drugs used in trials for treatment of primary sclerosing cholangitis
Biliary strictures and cholangiocarcinoma
Primary sclerosing cholangitis is characterized by multiple short strictures in the biliary tree. Particularly tight strictures, especially in the extrahepatic biliary tree can cause acute deterioration of liver function and more rapid progression to biliary cirrhosis. Such lesions are known as ‘dominant biliary strictures’, occur in 15–20% of PSC patients and can be very difficult to differentiate from cholangiocarcinoma at ERCP. Dominant strictures can be treated endoscopically by balloon dilatation or by placing a stent across the affected area. Alternatively, some non-cirrhotic PSC patients may be best managed by a bilioenteric bypass.93
The interpretation of results of endoscopic therapy trials for PSC is difficult due to the small numbers of cases tested and the variety of endoscopic techniques used. Short-term stenting has been shown to provide long-term benefits in some patients. A study from Amsterdam described 32 patients with dominant biliary strictures, treated with endoscopic placed stents for a mean of 11 days. Improvements in symptoms and cholestasis were seen in all patients and these improvements were maintained for several years. Only seven transient procedure-related complications occurred in 45 procedures.94
The results of long-term stenting was explored in a retrospective trial of 25 patients presenting with worsening biochemical tests of liver function and major bile duct strictures.95 Endoscopic therapy was successful in 21 (84%), and in all these cases a significant improvement in liver function tests was noted within 6 months of the insertion of a prosthesis. Stents were changed electively at 2–3 months or earlier if necessary. During the median follow-up of 29 months, 12 (57%) patients remained asymptomatic with stable liver function tests. Four patients had clinical and biochemical deterioration which responded to further endoscopic treatment.
A prospective trial published by Stiehl et al.96 considered the survival of 106 PSC patients treated with 750 mg of UDCA per day for up to 13 years along with balloon dilatation of major dominant strictures or placement of biliary stents whenever necessary. Ten patients had dominant strictures at the beginning of the trial and a further 43 patients (45%) developed dominant strictures during the follow-up period despite the UDCA treatment. This combined approach appeared to significantly improve overall survival rates compared with predicted values but as the study was uncontrolled it was not possible to ascertain whether the UDCA, the endoscopic therapy or the combination of the two was the important factor.96, 97
Cholangiocarcinoma develops in 6–20% of patients with long standing PSC at a rate of 1–5% per year.98 Prediction of hepatobiliary malignancy is extremely difficult. In a recent retrospective study of all Nordic PSC patients listed for liver transplantation over a 12-year period, the only independent risk factors for subsequent diagnosis with hepatobiliary malignancy were; clinical suspicion of cancer, recent diagnosis of PSC, no previous UDCA treatment and previous colon cancer.99 High alcohol consumption and smoking have also been implicated as risk factors for cholangiocarcinoma in the context of PSC.100, 101
Early disease is asymptomatic but patients may present with symptoms of biliary obstruction including jaundice, pale stools, dark urine and new onset, or worsening, pruritus. Pain, malaise and weight loss may indicate more advanced disease. Most patients will be investigated initially with ultrasound scanning. The diagnosis should be suspected when the intrahepatic, but not extrahepatic, ducts are dilated. Colour Doppler can detect tumour-induced compression or thrombosis of the portal vein or hepatic artery.102 Most patients will then go on to cross-sectional imaging. Magnetic resonance imaging (MRI) is currently the recommended technique for diagnosis and staging of cholangicarcinoma as it provides information on the local extent of the tumour, extent of duct involvement, presence of liver metastases and hilar vascular involvement.102
The ERCP is still useful in the management of some cases of cholangiocarcinoma. Tumour extent is sometimes best defined by formal cholangiography, and this procedure also allows tissue sampling by brush cytology or biopsy.103 Endoscopic ultrasonography can also be used both to visualize biliary strictures and to provide a fine-needle aspirate to diagnose malignancy. Intraductal ultrasonography (IDUS) may allow even better images of the proximal bile ducts and surrounding structures but, despite some interesting preliminary studies, its role in diagnosis and staging of cholangiocarcinoma is not yet entirely clear.102, 104–106 Positron emission tomography (PET) shows promise in the differentiation of benign and malignant biliary disease. The technique is based on the excessive accumulation of the glucose analogue [18F]fluoro-2-deoxy-d-glucose (FDG) in cholangiocarcinoma cells relative to hepatocytes which then shows up as ‘hot spots’ on scanning. This technique appears to have good specificity for malignant cells but in some studies its sensitivity has been inadequate.107–109
The accurate differentiation between benign and malignant tumours remains a major challenge in the management of PSC. So far there is no one simple diagnostic test that can perform this differentiation accurately. Exfoliative brush cytology has a high specificity and positive predictive value for the diagnosis of cholangiocarcinoma but low sensitivity and negative predictive value.110 Neoplasia is particularly difficult to differentiate from the cellular atypia seen in chronic bile duct inflammation in PSC. This technique can however, be very valuable in diagnosing malignancy and high grade dysplasia, sometimes at an early enough stage to treat with liver transplantation.38, 111, 112 Regular brushing of dominant strictures to detect early neoplastic transformation is recommended by most centres. Open or percutaneous biopsy of potentially resectable cholangiocarcinoma is not generally recommended due to the risk of seeding of the tumour.102
The level of accuracy achieved in diagnosing malignancy in cytological specimens may be improved by the use of molecular methods. Cholangiocarcinoma is often associated with the inactivation of tumour suppressor genes, for example, p53, APC and Smad-4.113 Dysregulation of apoptosis by overexpression of bcl-2, and p16 gene mutations have also been observed in cholangiocarcinomas in PSC.114 Mutations in oncogenes such as K-ras have been described in a variable proportion of cholangiocarcinomas, and K-ras analysis has been proposed as a diagnostic adjunct to conventional light microscopy.115, 116 Other mutated oncogenes detected in bile duct cancer include c-myc, c-erbB-2, c-met and c-neu.117 Telomerase activity is present in 85–90% of all human cancers and in situ hybridization has been used to detect telomerase RNA in endoscopic brushings as a marker for malignancy.118 Although molecular profiling, perhaps using a combination of markers, offers great promise in the future, currently there is no established clinical role for these techniques.
There are no tumour markers which are specific for cholangiocarcinoma and no evidence that measurement of tumour marker levels is useful in monitoring the progression of this cancer. In the context of PSC however, a serum CA19-9 level of >100 U/mL has been reported to have a sensitivity of 75% and a specificity of 80% for presence of cholangiocarcinoma.119, 120 The combination of CA19-9 and carcinoembryonic antigen (CEA) have been shown to be useful in identifying PSC patients with occult tumours, and more recently results of brush cytology, DNA analysis, serum CA19-9 and serum CEA have been combined in an attempt to improve further on this diagnostic sensitivity.119, 121
Endoscopic stenting is usually the best palliative option once the diagnosis of established cholangiocarcinoma has been made. In patients who might be suitable for orthoptic liver transplantation (OLT), biliary manipulation is best avoided as it increases the risk of stricturing and bacterial cholangitis and may jeopardize the chance of a successful OLT.
Cirrhosis and liver failure
All the usual complications of end-stage liver disease, including oesophageal and gastric varices, ascites and porto-systemic encephalopathy are seen in the advanced stages of PSC and the management of these is the same as for any cause of cirrhosis. The prevalence of gastro-oesophageal varices in the PSC population remains uncertain. It has recently been suggested that a platelet count of <150 × 103/dL may be a useful indicator of the presence of underlying varices.122 PSC patients who have undergone a colectomy and formation of an ileostomy for concurrent IBD may develop peristomal varices which can be severe and difficult to treat. Local treatment with injection of sclerosant, venous ligation and ileostomy revision are often unsuccessful. Porto-systemic shunts, i.e. transjugular intrahepatic porto-systemic shunt (TIPSS) can be useful in this situation or, alternatively this may be an indication for hepatic transplantation.
Liver transplantation in the context of PSC
Liver transplantation is the only long-term treatment for PSC and, with the exception of patients who have developed established cholangiocarcinoma prior to the operation, the results of transplantation in this group are comparable to those achieved in patients with primary biliary cirrhosis and autoimmune liver disease (see Table 3). In view of the variable clinical course of the disease and the unpredictable risk of cholangiocarcinoma development however, determining the optimal timing of transplantation remains controversial. Transplantation earlier in the course of the disease reduces the operative risk and reduces the risk of the development of hepatobiliary malignancy.123 However, PSC recurs in at least 30% of grafts, and colon cancer is a major cause of death in PSC patients after OLT.124, 125 Although prognostic models using age, bilirubin, albumin and histological stage have proved useful in studying large populations, they are not applicable to individuals so these judgements remain extremely difficult.126–129
Table 3. Survival of patients with primary sclerosing cholangitis (PSC) at 1, 2, 3, 5 and 10 years post-liver transplantation
|Mahesehwari et al. 2004187||3309||88||86||83||80||67|
|Solano et al.151||111||84.5|| ||84.5||83.4||68.9|
|Liden et al.188||61||82|| || ||73||64|
|Grazaidei et al.125||150||93.7||92.2|| ||86.4||69.8|
|Saldeen et al.189||41|| ||70|| || || |
|Goss et al.130||127||90||86|| ||85|| |
|Farges et al.123||51|| || || ||89|| |
|Narumi et al.136||37||97||92|| ||88|| |
The indications for transplantation are summarized in Table 4. They include those accepted for liver disease of all causes as well as symptoms seriously affecting quality of life (e.g. itch and fatigue) in the absence of cirrhosis. High grade biliary dysplasia and cholangiocarcinoma in situ are becoming less controversial indications and the incidental discovery of cholangiocarcinoma in the explanted liver without spread to regional lymph nodes does not appear to influence long-term survival post-transplant in most studies.99, 130 Recent work from the Nordic PSC population, however, found a poor prognosis in the small group of PSC patients found during or after transplantation to have incidental cholangiocarcinoma (n = 4, 2 year survival: 0%).99 Results of transplantation for patients with established cholangiocarcinoma prior to the procedure remain very poor, with even the most encouraging results to date still showing survival rates of 60%, 36% and 36% at 1, 3 and 5 years respectively.131 Most centres currently consider a definite finding of cholangiocarcinoma to be a contraindication to liver transplantation, however, the discovery of cholangiocarcinoma without clear evidence of extrahepatic spread during transplant surgery should not lead to cancellation of the procedure.99 Two studies have suggested improved survival in patients transplanted for cholangiocarcinoma who were given adjuvant chemoirradiation.132, 133 This approach has not been tried in the context of PSC to date.
Table 4. Indications for liver transplantation in primary sclerosing cholangitis
|(1) Cirrhosis complicated by|
|• intractable ascites|
|• variceal haemorrhage uncontrolled by endoscopic methods|
|• muscle wasting|
|• recurrent bacterial peritonitis|
|(2) Intractable itch or fatigue|
|(3) Recurrent cholangitis|
|(4) Jaundice which cannot be treated endoscopically or medically (e.g. steroids in ‘overlap’ syndromes)|
|(5) Hepatocellular carcinoma (if no extrahepatic growth and tumour is within accepted size)|
|(6) Biliary dysplasia or cholangiocarcinoma in situ|
There has recently been a suggestion that previous colorectal cancer is a predictor of hepatobiliary malignancy and might even be a contraindication to liver transplantation for PSC.99 All patients referred for consideration of OLT should undergo colonoscopy to diagnose and assess the severity of underlying IBD. Patients with colitis or any evidence of dysplasia should be considered for colectomy prior to, or even during liver transplantation in view of the risks of deterioration in control of IBD post-transplant and colorectal malignancy.
Post-transplantation survival has been shown to be affected by a number of pretransplantation factors, including previous hepatobiliary surgery, elevated creatinine, poor nutritional status and the presence of hepatobiliary malignancy.99, 130, 134, 135 Recurrent PSC in the liver graft occurs in as many as 20–40% of transplant recipients can be difficult to distinguish from biliary strictures associated with the transplant operation itself.125, 130, 136–146 Biliary strictures can occur in patients transplanted for any liver disease and are caused by biliary ischaemia, hepatic artery thrombosis, extended cold ischaemia time, ABO incompatibility, chronic ductopenic rejection and possibly secondary to ascending infections from the gastrointestinal tract. An interesting paper by Vera et al. has suggested that pretransplantation colectomy is associated with a lower risk of recurrent PSC in the graft.137 Other proposed risk factors include cytomegalovirus (CMV) infection and donor–recipient gender mismatch.138, 142 The prognosis of recurrent PSC in the allograft has traditionally been thought to be favourable, with reports from the Mayo clinic suggesting no effect on patient or graft survival with follow-up periods of up to 5 years.125, 147 There has, however, been a growing appreciation of the requirement for re-transplantation as a result of recurrent PSC over longer time periods, and this must now be considered a significant cause of late graft loss.125, 130, 148–151 Longer term studies are however, required to characterize the nature of recurrent PSC and its impact on graft survival in more detail.
Management of inflammatory bowel disease associated with PSC
The IBD associated with PSC is typically a total colitis which runs a relatively mild clinical course.4, 152–154 The risk of colorectal neoplasia however, is very significantly higher in UC associated with PSC with 31% of patients in one series developing colonic neoplasia at 20 years after the diagnosis of UC compared with 5% of patients with UC alone.155 The increased risk of colonic neoplasia in patients with PSC has been thoroughly studied with a total of 14 studies addressing this issue since Broome et al.'s original publication in 1992.155–169 Some of this work has been difficult to interpret due to the tendency of the UC associated with IBD to be extensive, which is an independent risk factor for colorectal neoplasia, and mild, which might lead to an underestimation of the duration of the disease prior to diagnosis. Nevertheless there is a general consensus that patients with both colitis and PSC are at increased risk of colonic dyplasia and should undergo annual screening colonoscopy with multiple biopsies.170 Recent work has suggested that UDCA may have a role in preventing colonic neoplasia and this would have major significance for the use of this drug in the context of PSC.171, 172
High faecal and serum levels of DCA have been demonstrated in UC patients with colorectal dysplasia or cancer and in non-UC patients with adenomas and cancers, implicating DCA as a potential carcinogen.173–177 Whilst it is established that the hydrophilic bile acid UDCA inhibits injury by hydrophobic bile acids, hepatocyte cell death from bile acid-induced toxicity occurs more frequently as a result of apoptosis than necrosis. It has been demonstrated that UDCA inhibits DCA-induced apoptosis by modulating mitochondrial transmembrane potential and reactive oxygen species production.178 UDCA moreover inhibited in vitro DCA stimulated growth in several tumour cell lines including colon cancer, leading to the suggestion that UDCA may reduce the risk of dysplasia and cancer in the biliary tree and the colon.179 A small study of 59 PSC patients undergoing 3 year colonoscopic surveillance found a significantly reduced risk in the prevalence of colonic dysplasia in patients taking UDCA compared with those not on this treatment.172 There are potential problems with the interpretation of this study particularly in view of the exceptionally high rate of dysplasia in the control group. The largest study to date, of 120 patients with UC and PSC demonstrated a small reduction of the risk of dysplasia and cancer in the UDCA-treated group which failed to reach significance.180 The most recent work however from Pardi et al. of 59 patients with PSC and UC is more promising, finding the use of UDCA to be associated with a significant risk reduction in the odds for developing dysplasia (odds ratio 0.14, 95% confidence interval: 0.03–0.64).171 To date, there is no evidence that the high rate of bile duct cancer seen in PSC is reduced by UDCA therapy.
The risk of colorectal neoplasia is clearly an important factor when deciding whether PSC patients should undergo colectomy for refractory UC. Recent reports, which suggest that the course of UC tends to become more severe after liver transplantation despite immunosuppressive drug regimens, would also encourage a more favourable view of a surgical approach to the management of colitis. However, the problems of peristomal varices addressed above and the high risk of pouchitis in patients with pouch-anal anastamosis should also be taken into consideration. A small study of 16 PSC patients with pouch-anal anastamosis after proctocolectomy for UC also found a higher prevalence of moderate or severe mucosal atrophy in the pouchs of PSC patients compared with UC controls and a suggestion that this might be associated with a higher risk of DNA aneuploidy and dysplasia.181 Further work is needed to determine if there is a genuinely increased risk of malignancy in the pouchs of PSC patients.
Primary sclerosing cholangitis remains a highly challenging disease to manage. Despite some advances in the management of the complications and symptoms of PSC, there is still no established, effective medical treatment. UDCA continues to be the most interesting therapeutic option in terms of potential modes of action, slowing disease progression and reducing the risk of colonic malignancy. High-dose UDCA therapy looks promising but large, randomized-controlled trials of this drug, possibly in combination with immunosuppressive agents or antibiotics are now required. Cholangiocarcinoma remains a potential and unpredictable risk in PSC patients and there is, as yet, no reliable screening process to manage this problem. Liver transplantation is the only option for patients with advanced disease, but the timing of the procedure in relatively young patients, with a disease which tends to have a variable course, can be difficult. More information is now emerging on the effects of PSC on the clinical characteristics of concurrent IBD and this is helpful in managing this aspect of the disease, and particularly the high risks of colonic malignancy in this group. Fascinating work into the pathogenesis of PSC is ongoing and provides the best hope for providing targets for drug therapy in a disease where little effective intervention is currently available.