SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Conclusions
  6. Authorship
  7. Acknowledgements
  8. References
  9. Supporting Information

Background

Despite considerable advances over the last two decades in the molecular understanding of cholestasis and cholestatic liver disease, little improvement has been made in diagnostic tools and therapeutic strategies.

Aims

To critically review controversial aspects of the scientific basis for common clinical practice in primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC) and to discuss key ongoing challenges to improve patient management.

Methods

We performed a literature search using PubMed and by examining the reference lists of relevant review articles related to the clinical management of PBC and PSC. Articles were considered on the background of the European Association for the Study of the Liver (EASL) and the American Association for the Study of Liver Diseases (AASLD) practice guidelines and clinical experience of the authors.

Results

Ongoing challenges in PBC mainly pertain to the improvement of medical therapy, particularly for patients with a suboptimal response to ursodeoxycholic acid. In PSC, development of medical therapies and sensitive screening protocols for cholangiocarcinoma represent areas of intense research. To rationally improve patient management, a better understanding of pathogenesis, including complications like pruritis and fatigue, is needed and there is a need to identify biomarker end-points for treatment effect and prognosis. Timing of liver transplantation and determining optimal regimens of immunosuppression post-liver transplantation will also benefit from better appreciation of pre-transplant disease mechanisms.

Conclusion

Controversies in the management of PBC and PSC relate to topics where evidence for current practice is weak and further research is needed.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Conclusions
  6. Authorship
  7. Acknowledgements
  8. References
  9. Supporting Information

Cholestatic liver disease may arise due to defects at any level of bile formation. The aetiologies of these conditions range from molecular abnormities caused by genetic variation or drugs to structural changes due to developmental disorders, autoimmune bile duct injury, tumours and gallstones. In everyday clinical practice, the term is most often applied to primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC). Both these conditions pose major management challenges in adult hepatology. In addition to the general features associated with a broader syndrome of ‘cholestatic liver disease’ (e.g. development of cholestatic liver cirrhosis) (Figure 1), disease specific pathologies (e.g. malignancy risk in PSC) require attention. Considerable advances have been made over the last two decades regarding the molecular understanding of cholestasis and cholestatic liver disease.[1-5] These advances have largely derived from the identification of the genes responsible for the progressive familial intrahepatic cholestasis (PFIC) syndromes throughout the 1990s[6-9] and from the recent genome-wide association studies (GWAS) applications in PBC and PSC.[10] Despite the new knowledge, little improvement has been made in diagnostic tools and therapeutic strategies.

image

Figure 1. The term ‘cholestatic liver disease’ jointly denominates the manifestations of a wide variety of liver diseases. For most of the chronic cholestatic liver diseases, therapeutic opportunities are limited, resulting in progression to liver cirrhosis and risk of cancer development. Disease-associated symptoms (e.g. pruritus and fatigue) represent major clinical challenges. The focus of this review article is primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC).

Download figure to PowerPoint

Primary biliary cirrhosis and PSC (Figure 2) are slowly progressive diseases with a course of one to two decades from the manifestations of early stages of disease until end-stage liver disease. This results in difficulties in establishing a robust level of evidence for the benefit of any management option, including surveillance for complications throughout the disease course. Prognostic modelling taking surrogate markers for disease stage [including alkaline phosphatase (ALP), bilirubin, model for end-stage liver disease (MELD) score, Child-Pugh score, histology and radiology] into account has, to some extent, assisted the assessment in treatment trials, but robust markers for disease activity and disease complications are missing. Interpretation of the present basis for patient management requires caution and awareness of such limitations.

image

Figure 2. Primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC) share features of an autoimmune affection targeting different levels of the biliary tree.[14, 135] Co-occurring autoimmune manifestations outside the liver are common in both conditions. In PSC, the increased risk of biliary tract- and colonic cancer poses particular challenges. The present review elaborates on clinical challenges in PBC and PSC, as well as each of the shared hepatic features shown in the figure. AMA, anti-mitochondrial antibodies; ANCA, anti-neutrophil cytoplasmic antibodies.

Download figure to PowerPoint

In this review, we aimed to critically summarise the evidence for clinical practice in PBC and PSC. The controversies that do exist arise on topics where evidence is weak or conflicting and a further and better research is needed. Therefore, we will highlight both controversies and important challenges. We will, to some extent, discuss general features of cholestatic liver disease (cirrhosis development, pruritus and fatigue) and associated molecular and genetic aspects. We will not provide comprehensive discussions on drug-induced cholestasis and cholestatic liver disease in the context of developmental disorders and gallstone disease, for which the interested reader is guided elsewhere.[11-13]

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Conclusions
  6. Authorship
  7. Acknowledgements
  8. References
  9. Supporting Information

A literature search was conducted 1st September 2013 on PubMed using a broad range of search terms including, but not limited to, ‘primary biliary cirrhosis’ and ‘primary sclerosing cholangitis’ along with ‘ursodeoxycholic acid’, ‘treatment’, ‘clinical trial’ and ‘prognostic score’. Articles were selected for discussion on the basis of the authors' prior knowledge on controversial or challenging topics in the management of PBC and PSC. In addition, important review articles, meta-analyses and Cochrane Systematic Reviews were reviewed and examined for relevant references and practice guidelines. Randomised clinical trials involving a treatment group (UDCA for PSC, non-UDCA for PSC or PBC) vs. placebo or no treatment were included for full table presentation irrespective of blinding or language. Observational studies or studies lacking a control group were excluded from full table presentation, but have been included in the main text to appropriately account for controversial aspects where appropriate. Evaluation of predictive models in PSC was limited to those assessing pre-transplant survival. Articles may have been missed by the adopted approach, and the presentation may also be biased as to the opinion of the authors.

Management of PBC

Diagnosis of PBC

The diagnostic criteria for PBC rely on the fact that the targets of liver infiltrating T cells and the antibody production have been identified as lipoylated domains of three members of the 2-oxo-acid dehydrogenase complex family (mainly the E2 component of the pyruvate dehydrogenase complex; PDC-E2).[14] The diagnosis can be made on the basis of two of three criteria: the presence of biochemical cholestasis (ALP elevation), detection of anti-mitochondrial antibodies (AMA) and typical histological findings. A liver biopsy is not essential in patients with ALP elevation and AMA, but may be required for the diagnosis of concurrent features of autoimmune hepatitis and disease stage. There is little controversy regarding the diagnostic criteria in PBC, but some discrepancy in the evaluation of the level of evidence supporting these.

Depending on the assay employed,[15-17] 5–10% of patients with PBC present without detectable AMA, even by repeated measures at follow-up. There is little knowledge on the specific pathological features of AMA-negative PBC. Other mitochondrial epitopes than M2 may be relevant.[18] There are also slight differences in the cellular composition and severity of histological lesions between AMA-positive and -negative cases.[19] Robust genetic determinants to differentiate between AMA-positive and -negative PBC patients have not been detected;[20] however, the AMA-negative subset of patients is clearly too small to obtain conclusive statistical power in the analyses. Clinical presentation and behaviour of AMA negative PBC is largely similar to that of AMA-positive PBC,[21] and collated anecdotal data suggest that there is also a similar response to ursodeoxycholic acid.[22] Disease recurrence rates of PBC after liver transplantation also do not seem to be affected by AMA status.[23] In clinical practice, the presence of non-M2 immunoreactivity in a patient with PBC means that a liver biopsy is required for diagnosis. Typically, considerations must also be made as to the presence of small-duct PSC and genetic cholangiopathies,[24-26] and precise distinctions between these conditions may not always be feasible.

Current treatment of PBC

Therapeutic applications in PBC and other cholestatic liver diseases can broadly be divided into bile acid therapy and other approaches. Of bile acid preparations, doses of 13–15 mg/kg/day of ursodeoxycholic acid (UDCA) appear beneficial and are commonly used in PBC.[27-30] AASLD advises in favour of UDCA therapy judging evidence unambiguously as class I (conditions for which there is evidence and/or general agreement that a given diagnostic evaluation, procedure or treatment is beneficial, useful and effective), level A (data derived from multiple randomised clinical trials or meta-analyses). The EASL guidelines evaluate the overall support for UDCA treatment at the same level, but suggest that the evidence for long-term treatment is less robust (category II-2; evidence for recommendation is derived from cohort or case–control analytic studies, grade B; further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate, recommendation strong). This slight discrepancy may partly be due to a still ongoing dispute as to the quality of trials performed and the true efficacy of UDCA,[31] largely deriving from two negative Cochrane-based meta-analyses of survival benefit from UDCA.[32, 33] Another meta-analysis,[34] focusing on studies with long-term follow-up, however, claims a significant improvement in transplant-free survival in patients on UDCA and that duration of studies need to be taken into account when evaluating drug efficacy in PBC. UDCA therapy also appears to be associated with reduced costs compared with placebo/conservative treatment[28, 35] and is by many authorities considered standard care.

A favourable biochemical response to UDCA, incorporating specific improvements in ALP (‘Barcelona criteria’),[30, 36] or in ALP, bilirubin and aspartate transaminase (AST) (‘Paris criteria’),[37] seems to associate with improvement in transplant-free survival in PBC. In brief, the sum of these biochemical responses and related assessments (e.g. the Mayo risk score[38] and baseline parameters like ductopenia[39]) strongly suggests that there is heterogeneity of the PBC patient population as to the efficacy of UDCA. The mechanism of action of UDCA is reviewed elsewhere,[40] yet which aspects of the pathogenesis of PBC that are reflected by heterogeneity in the response to UDCA are unknown. For research purposes, the response-indices may serve useful in stratifying patients in studies aiming to determine the pathophysiology of such aspects. PBC patients with little or no improvement in suggested response-indices are also candidates for trials of supplementary or alternative therapies.

Future treatment of PBC

No other bile acids than UDCA are currently in use in clinical practice. Obeticholic acid, a derivative of chenodeoxycholic acid, has (unlike UDCA) strong activating effects on the nuclear receptor farnesoid X receptor (FXR) and is currently entering phase III clinical trials following promising phase II results (http://clinicaltrials.gov/ct2/show/NCT01473524).[41] A derivative of UDCA, nor-UDCA, is currently in phase II clinical trials for PSC (see below), but has not been tested in PBC. Of the nonbile acid therapies, bezafibrate combination therapy with UDCA has shown the most promising results[42, 43] and is also at time of writing at the recruiting stage for phase III trials (http://www.clinicaltrials.gov/ct2/show/NCT01654731). Like obeticholic acid, bezafibrate exerts its likely main mechanism of action via nuclear receptors, targeting the pregnane X receptor (PXR; also called the steroid and xenobiotic receptor, SXR) and peroxisome proliferator-activated receptor alpha (PPARα).[43] The glucocorticoid receptor agonist budesonide also activates PXR.[44] Budesonide has shown some efficacy, which so far has not translated into standard care,[45-48] largely due to concerns on osteoporotic side effects.[49] In our opinion, glucocorticoid receptor-centred adjuvancy to UDCA in PBC should be restricted to patients with features of autoimmune hepatitis (see below).

In many ways, PBC is a prototypical autoimmune disease,[14] with a well-defined autoantigen, a relatively homogenous disease expression and a genetic susceptibility background similar to that of other autoimmune diseases (Figure 2).[50] For this reason, the poor efficacy of immune-targeted therapies that have been tested so far (Table 1) remains somewhat of a paradox. There is little controversy as to this observation, yet there is renewed interest in immune-target therapies on the basis of specific pathways highlighted by GWAS findings. Phase II trials on ustekinumab (a monoclonal anti-p40 antibody, http://www.clinicaltrials.gov/ct2/show/NCT01430429) are ongoing based on the prominent genetic associations with several components of the interleukin 12 (IL12) and IL23 signalling pathway in the study populations of European ancestry.[51, 52] As a note of caution in this regard, the biological implications of these genetic associations are as of yet not understood. Moreover, they seem not to be a prerequisite for PBC development as shown by the absence of similar associations in Asian study populations.[53] The efficacy of ustekinumab in other IL12/23-related diseases varies; e.g. in psoriasis, ustekinumab is part of the established treatment armamentarium,[54] whereas in multiple sclerosis, it has been concluded to be of no benefit.[55] As the relationship between this response heterogeneity and the genetic associations in implicated diseases is unknown, efficacy in PBC of ustekinumab and similar approaches is hard to predict.

Table 1. Overview of clinical studies of non-ursodeoxycholic acid treatment in primary biliary cirrhosis (PBC). References for studies refer to the online supplementary reference list (Data S1 for tables 1–5)
StudyYearTreatmentN (treat/control)Study durationLabHistologyOLT-free survivalOutcome
  1. ALP, alkaline phosphatase; AMA, anti-mitochondrial antibodies; HT, hypertension; ND, not done; UDCA, ursodeoxycholic acid; OLT, orthotopic liver transplantation.

  2. The interested reader is guided to the following references for treatment reports not fulfilling the criteria for Table presentation, i.e. randomized allocation to either a treatment group or a control group receiving placebo, no treatment or UDCA.[16-34]

  3. a

    No controls included, but comparison with 180 patients in a UDCA vs. placebo study conducted during the same period.

  4. b

    No controls, cross-over design with exchange of treatment protocols between the two groups after 6 months.

Heathcote et al.[1]1976Azathioprine45 (22/23)5 yearsNo effect on liver tests, histology or survival
Christensen et al.[2]1985Azathioprine248 (127/121)11 yearsNo effect on liver tests, histology or survival
Minuk et al.[3]1988Ciclosporine vs. placebo12 (6/6)1 year+NDNDImproved liver tests, but high incidence of side effects
Wiesner et al.[4]1990Ciclosporine vs. placebo29 (19/10)2 years++NDImproved liver tests (ALP, bilirubin) and AMA, improved histology compared with placebo
Lombard et al.[5]1993Ciclosporine vs. placebo3496 years+

Prolonged survival in Cox multivariate analysis was not confirmed in the univariate analysis

Side effects (9% HT, 11% renal insufficiency)

Lindor et al.[6]1995Methotrexate/UDCA32a2 yearsND

Seven patients withdrawn because of side effects

No effect on liver tests or histology

Gonzalez-Koch et al.[7]1997Methotrexate/UDCA vs. UDCA25 (13/12)48 weeksNDNo effect on liver tests or histology
Leuschner et al.[8]1999Budesonide/UDCA vs. UDCA39 (20/19)2 years++NDImproved liver tests, IgM, IgG, and histology
Nakai et al.[9]2000Bezafibrate/UDCA vs. UDCA23 (10/13)1 year+NDNDCombination therapy improved liver tests compared with UDCA
Kurihara et al.[10]2000Bezafibrate vs. UDCA24 (12/12)1 year+NDNDImproved liver tests more than UDCA
Itakura et al.[11]2004Bezafibrate/UDCA vs. UDCA16b6 months+NDNDImproved liver tests
Combes et al.[12]2005Methotrexate/UDCA vs. UDCA2657.5 years (4.6–8.8 years)No effect on death, OLT,varices encephalopathy, bilirubin, histological progression by two stages or to cirrhosis
Rautiainen et al.[13]2005Budesonide/UDCA vs. UDCA77 (41/36)3 years+NDBudesonide improved histology, but did not result in additional improvement in liver tests
Iwasaki et al.[14]2008

Bezafibrate vs. UDCA

Bezafibrate/UDCA vs. UDCA

45 (20/25)

 

22 (12/10)

1 year

 

1 year

Equal

 

+

ND

 

ND

ND

 

ND

Bezafibrate and UDCA monotherapy equally improved liver tests. Additional improvement with combination
Mason et al.[15]2008Lamivudine/zidovudine/UDCA vs. UDCA596 months+NDNDImproved ALP, but did not improve primary end-points

A definite controversy in PBC exists as to the possibility of an infectious aetiology.[56, 57] It is beyond the scope of this article to review the scientific basis of this dispute, which has triggered the execution of completed[58] and planned (http://clinicaltrials.gov/ct2/show/NCT01614405) anti-retroviral regimens. As of yet, there is no proven benefit of such approaches. In epidemiological data (reviewed elsewhere),[59] associations have also been suggested between exposure to urinary tract pathogens (Novosphingobium aromaticivorans in particular) and other potential environmental co-factors for the ongoing immune response in PBC. Of particular interest is 2-Octynoic acid,[60] which is present in commonly used cosmetic products and food flavourings and has the potential to modify PDC-E2 in an immunogenic direction. Elimination studies, like for gluten in coeliac disease, have not yet been performed to justify any particular advice on 2-Octynoic acid-related products for patients.[61] Unlike the case in PSC, antibiotic therapy in PBC has not been attempted outside the context of pruritus (rifampicin).[62, 63] Of the reported environmental risk factors in PBC,[64] smoking seems to be the only one to be accounted for in clinical counselling so far, given associations with an increased rate of liver fibrosis.[65, 66]

Liver transplantation in PBC

In general, disease progression is more predictable in PBC than in PSC, and the utility of prognostic models (bilirubin, Mayo risk score and MELD, in particular) in the appropriate timing of liver transplantation is established.[67-69] Pruritus may, in a few cases, serve as the sole indication for liver transplantation. The improvement of fatigue following liver transplantation is less predictable and pronounced than other symptoms,[70, 71] and fatigue should, therefore, not trigger liver transplantation without the presence of other indications. There is little or no controversy as to incorporating symptomatic indications in the transplant assessments, even in areas with MELD-based graft allocation programmes. Hepatocellular carcinoma (HCC) may develop in PBC patients with advanced disease stages,[72, 73] and may also be associated with nonresponse to UDCA.[74] The patients should be followed accordingly.[75-77]

Disease recurrence occurs in up to 30–35% of liver allografts in PBC recipients.[78] The variable frequencies between studies probably reflect differences in practice as to protocol biopsies and diagnostic criteria. There is no consensus as to treatment of UDCA in transplant recipients with recurrent PBC, but hepatic biochemistries improve upon administration like in the native liver disease.[78-80] Whether the choice of calcineurin inhibitor influences recurrence rates is debated.[81] Tacrolimus seems to associate with higher frequencies of recurrence and shorter recurrence-free survival than ciclosporine in many series.[81] Meta-analysis of available data as of January 2006, however, did not support a significant difference in PBC recurrence rates according to immunosuppressive regimen. On this basis, and as short- to mediate-term impact on overall and graft survival from PBC recurrence is neglible,[82] tacrolimus-based regimens remain standard at most centres. As very long-term (>10–20 years of follow-up) impact from PBC recurrence on re-transplantation rates is unknown, further considerations may become relevant as data accumulate.

Management of PSC

Diagnosis of PSC

There is emerging evidence that previous estimates of a frequency of PSC in IBD of 2–4% may be too low and that up to 10% of patients with IBD may show changes compatible with PSC on magnetic resonance cholangiography (MRC).[83, 84] The association between IBD and PSC remains one of the key features of PSC for which the basis is still not established. Hypotheses range from endotoxin leakage,[85] via pathogenic changes to the gut microbiota,[86] to aberrant homing of intestinally activated lymphocytes to engage in biliary inflammatory processes.[87, 88] If the pathological processes involved can be dissected, it may be possible to develop diagnostic tools for early, potentially pre-clinical, detection of PSC in IBD patients. On the other hand, it is currently not known to what extent failure of medical therapy in PSC can be ascribed to an advanced disease stage with manifest strictures already at the time of diagnosis. As of yet, procedures to diagnose PSC are restricted to IBD patients in whom abnormal regular biochemical tests suggest the presence of liver disease. There is now consensus in both AASLD and EASL guidelines that MRC is the method of choice, unless endoscopic interventions or sampling is called for and endoscopic retrograde cholangiography (ERC) is indicated. Likely, the practice is sound, but IBD patients with normal biochemistries in whom an intensified colorectal screening programme could be justified if they proved to have PSC will be missed by this approach.[83]

The topic of elevated levels of immunoglobulin G4 (IgG4) in 10–20% of patients with PSC has received much attention over the last years.[89-93] An unknown fraction of these patients is likely to have cholangitis in the context of autoimmune pancreatitis (AIP) as determined by the HISORt criteria[94] first reported in Asian populations.[95, 96] Modified criteria have been adopted for the application in patients with PSC, i.e. accounting also for patients with two or more main manifestations (elevated serum IgG4, suggestive pancreatic imaging findings, other organ involvement and bile duct or papilla Vateri biopsy with >10 IgG4 positive cells/hpf) in combination with a significant corticosteroid treatment response defined as markedly improved biliary strictures allowing stent removal, liver enzymes <2 × ULN and significant decreases in serum IgG4 and CA19-9 levels.[97] Corticosteroid responsiveness is a central feature of these criteria. Serum IgG4 levels are also typically higher in patients with AIP than in PSC, although precise cut-off levels for serum IgG4 to distinguish the conditions are not firmly established.[98] It may be argued that IgG4 elevations in PSC arising outside the context of AIP is of uncertain importance and associates with pathogenetic processes in PSC per se.[99, 100] On the other hand, it has been suggested that disease behaviour in patients with PSC and even mild elevations of IgG4 shows atypical features, which include a more rapid disease progression.[90, 101] A pragmatic approach is to adopt IgG4 measurements as part of the routine diagnostic work-up of patients with suspected PSC (as already proposed by EASL and AASLD guidelines),[67, 102] and to keep the threshold at a reasonable level when selecting suspected cases for assessment of corticosteroid responsiveness.[97] Until refined criteria are available, sound clinical judgment, also accounting for corticosteroid side effects, is needed to guide decisions on the topic.

Cancer surveillance in PSC

Historically, it is of interest to note the 10 years that passed from the first report on an increased risk of colorectal cancer in PSC[103] until the association was corroborated in meta-analysis of available and contrasting reports.[104] Colorectal cancer surveillance annually or biannually in PSC patients with IBD is now advised in both the EASL and AASLD guidelines. PSC presentation may precede that of IBD, and IBD may even present post-liver transplantation.[105] Furthermore, IBD in PSC typically runs a quiescent course,[106-108] and may thus be missed. Regular re-evaluation by total colonoscopy and histology, for instance every 4–5 years, is therefore recommended by many authorities.

We recently reviewed the present data on diagnosis of cholangiocarcinoma in PSC.[109] The main problem is that of early diagnosis, and so far, there are no consensus guidelines as to robust screening protocols. A recent population-based patient series from the Netherlands questions previous notions that late presentation of cholangiocarcinoma is a rare event,[110] with 37% of cholangiocarcinomas presenting more than 10 years after diagnosis of PSC.[111] Clinical and biochemical deterioration at any point in the disease course should therefore always result in reflections as to a potentially malignant explanation. There is some evidence that imaging routinely should be incorporated in PSC patient follow-up,[112] and MRC or ultrasound is included in patient follow-up at many centres. The EASL guidelines advise for annual ultrasound to detect gall-bladder mass lesions,[67, 113] but neither this practice nor the utility of other imaging modalities in early cholangiocarcinoma detection has so far been validated. The ability of FUT2/FUT3 genotyping to improve CA19-9 test performance also needs prospective evaluation and validation.[114]

With few exceptions, investigations of brush cytology specimens, obtained at ERC by brushing in a region of a suspicious stricture, have a specificity of 90–100% for cholangiocarcinoma in PSC in published series (Table 2). The problem with the method is highlighted by the variable sensitivity. Operator variability as well as differences in the criteria defining a positive test is likely to contribute to the variable sensitivity.[115] Technical aspects of brushing at ERC and intrinsic variability in cholangiocarcinomas may also play a role. Repeated brushings can improve the sensitivity of this method,[116] a practice also employed at our centre on inconclusive findings in a suspicious clinical context. As shown in Table 2, various means of structural chromatin/DNA assessments, by fluorescent probe hybridisation, digitised image analysis or flow cytometry, generally improve sensitivity and are incorporated in clinical routine diagnostics of brush cytology specimens at many referral centres. As reviewed elsewhere,[109] molecular biomarkers for cholangiocarcinoma in PSC are awaited and expected to improve sensitivity even further. However, the path forward to clinical validation of these makers is likely to still take some years, due to scarcity of patients and only recent establishing of systematic prospective patient biobanking in the international PSC research network (www.ipscsg.org).

Table 2. Results of studies on the diagnostic utility of biliary brush cytology with and without DNA analysis [by fluorescence in situ hybridisation (FISH), digital image analysis (DIA) or flow cytometry] in primary sclerosing cholangitis. Differences in classification algorithms, brush cytology specimen collection procedures, as well as operator variability make direct comparisons difficult and are likely to underlie parts of the differences. Implementation of DNA analysis in the diagnostic protocols has a growing amount of scientific basis, but there is controversy as to the management of detected abnormities. References for studies refer to the online supplementary reference list (Data S1 for tables 1–5)
AuthorTime periodCountryPSC patients (n)Brush samples (n)Cytology classificationMethodSensitivity (%)Specificity (%)Comment
  1. CCA, cholangiocarcinoma; HR, hazard ratio; ND, not done.

Ponsioen et al.[35]1987–1996Amsterdam, The Netherlands4347

No abnormalities

Reactive changes

Some atypia

Suspicious

Adenocarcinoma

Cytology

 Suspicious or adenoca

6089 
Siqueira et al.[36]1984–1987Pittsburgh, PA, US151318

Negative

Suspicious

Highly atypical

Positive

Cytology

 Positive

46.4100 
Lindberg et al.[37]1997–2001Stockholm, Sweden20 

Insufficient

Benign

Abnormal, but probably benign

Suspicious, but inconclusive

Probably malignant

Malignant

Cytology

Flow cytometric DNA analysis

71

29

100

92

 
Lal et al.[38]1995–2000Chicago, Illinois, US2164

Benign

Atypical

Malignant

Cytologyn.d.94 
Moff et al.[39]2001–2004Baltimore, MD, US47101

Unsatisfactory

Benign/reactive

Atypical

Malignant

Cytologyn.d.n.d. 
Furmanczyk et al.[40]1995–2004Seattle, WA, US51 (37 with follow-up)107

Negative

Atypical

Suspicious

Positive

Cytology

 Positive

 Positive or suspicious

62.5

87.5

100

86.2

 
Moreno Luna et al.[41]2003–2004Mayo Clinic, MN, US86 

Benign

Atypical

Suspicious

Malignant

Cytology

 Positive

 Positive or suspicious

DIA

 Aneuploid or tetraploid

FISH

 Polysomy or trisomy 7 or 3

 Polysomy

Cytology neither positive nor suspicious:

DIA

 Aneuploid or tetraploid

FISH

 Polysomy or trisomy 7 or 3

 Polysomy

DIA + FISH

 Aneuploidy/tetraploid + polysomy/trisomy

18

41

43

70

47

14

60

20

14

100

97

87

86

100

88

87

100

98

 
Boberg et al.[42]2000–2004Oslo, Norway61121

Insufficient

Normal

Indefinite for dysplasia

Low-grade dysplasia

High-grade dysplasia –adenocarcinoma

Cytology

 High-grade dysplasia

 Low-grade + high-grade dysplasia

73

100

95

84

 
Levy et al.[43] Mayo Clinic, MN, US34 

Inadequate

Negative

Atypical

Suspicious

Positive

Cytology

 Positive

 Positive or suspicious

DIA

FISH

 Trisomy 7 benign

 Trisomy 7 malignant

DIA + FISH

 Trisomy 7 benign

 Trisomy 7 malignant

Cytology neither positive nor suspicious:

DIA

FISH

 Trisomy 7 benign

 Trisomy 7 malignant

DIA + FISH

 Trisomy 7 benign

 Trisomy 7 malignant

7

29

21

29

64

36

64

14

14

57

29

71

100

100

95

100

70

100

70

95

100

71

95

67

 
Charatcharoenwitthaya et al.[44]2000–2006Mayo Clinic, MN, US117216

Benign

Atypical

Suspicious

Malignant

Cytology

 Positive

 Positive or suspicious

DIA

 Aneuploid or tetraploid

FISH

 Polysomy or trisomy

 Polysomy

DIA + FISH

 Aneuploidy/tetraploid + polysomy/trisomy

8

46

63

88

38

63

100

97

80

73

98

90

 
Bangarulingam et al.[45]2003–2008Mayo Clinic, MN, US235 

Negative

Atypical

Suspicious

Positive

FISH

 Polysomy

 Trisomy/tetrasomy

46

25

88

67

Patients with FISH polysomy had outcome similar to patients with CCA

Patients with FISH trisomy/tetrasomy had outcome similar to patients with negative FISH

Conclusion:

FISH should be interpreted with caution

Halme et al.[46]2004–2007Helsinki, Finland102186

Benign

Suspicious

Malignant

Cytology

 Suspicious

DNA content

DNA + cytology for patients with end-point

DNA + cytology for whole cohort

48

48

72

50

88

94

82

77

5 aneuploid cases with benign cytology had dysplasia or cancer on histology
Barr Fritcher et al.[47]2003–2009Mayo Clinic, MN, US30 Selected patients (n = 30) with FISH polysomy, but no definite CCA for follow-up   Patients with serial FISH polysomy are more likely to have CCA on follow-up than those with subsequent non-polysomy FISH
Barr Fritcher et al.[48]2003–2010Mayo Clinic, MN, US102 

Selected only equivocal cytology:

 Atypical

 Suspicious

Univariate Cox model:

 Suspicious cytology: HR 1.90 (0.95–3.81)(P = 0.066)

 FISH polysomy: HR 8.70 (3.79–19.99)(P < 0.001)

Multivariate Cox model:

 FISH polysomy: HR 6.96 (2.97–16.34)(P < 0.001)

   
Current medical treatment of PSC

The AASLD guidelines advise against the use of UDCA as medical therapy in PSC, concluding at a class I, level A recommendation (see above for criteria definitions).[102] The EASL guidelines have a more compound conclusion, suggesting that UDCA improves serum liver tests and surrogate markers of prognosis, but does not reveal a proven benefit on survival. Thus, they conclude that present data do not yet allow a specific recommendation for the general use of UDCA in PSC.[67] The controversy surrounding UDCA use has several reasons. These partly derive from the lack of effect on clinical end-points in partly underpowered clinical trials (Table 3) and partly from a study using a high dose of UDCA (28–30 mg/kg/day), which culminated in an increase in clinical end-points including colorectal dysplasia.[117, 118] There is an increasing enrichment of biliary UDCA in patients with PSC, from 43–47% at normal doses (10–17 mg/kg/day) to 56–59% at higher doses (18–32 mg/kg/day),[119] along with increased delivery of unabsorbed UDCA to the colon for bacterial metabolisation. The mechanisms explaining the detrimental effects of high-dose UDCA have not been fully elucidated and conflicting pilot data on high-dose UDCA in PSC do exist.[120, 121] In effect, many centres still offer UDCA to patients on the basis of clinical experience and tradition and argue that only the high-dose UDCA regimen should be abandoned.[122] To what extent UDCA-responsive subsets of PSC patients can be defined (e.g. on the basis of ALP changes like in PBC[123]) is not yet clear.[124, 125]

Table 3. Overview of clinical studies of treatment with ursodeoxycholic acid in primary sclerosing cholangitis (PSC). References for studies refer to the online supplementary reference list (Data S1 for tables 1–5)
StudyYearDose (mg/kg bw/day)N (treat/control)Study durationLabHistologyCCAOLT-free survivalOutcome
  1. UDCA, ursodeoxycholic acid; ND, not determined; CCA, cholangiocarcinoma; OLT, orthotopic liver transplantation.

  2. The interested reader is guided to the following references for treatment reports not fulfilling the criteria for Table presentation, i.e. randomised allocation to either a treatment group receiving UDCA or a control group receiving placebo or no treatment.[58-74]

  3. a

    Cross-over design: 3 months pre-treatment, 6 months treatment, 3 months withdrawal, 18 months re-treatment.

  4. b

    Uniform dose not adjusted to body weight.

  5. c

    Controls were 52 placebo/52 low-dose UDCA using data from Lindor et al.[57]

O'Brien et al.[49]19911012a2.5 years+NDNDNDImprovement of liver tests in treatment periods and worsening in nontreatment periods
Beuers et al.[50]199213–1514 (6/8)1 year++NDNDSignificant improvement in liver biochemistry
Stiehl et al.[51]1994750/dayb20 (10/10)3 months+NDNDNDSignificant improvement in liver tests
De Maria et al.[52]a1996300 b.d.b40 (20/20)2 years    No effect on liver tests or cholangiography
Lindor et al.[53]199713–15102 (51/51)2.2 years+NDNo significant effect on primary end-points (death, OLT, histology, lab)
Mitchell et al.[54]20012026 (13/13)2 years++NDNDUDCA group had improved liver test results, histology and cholangiography
Harnois et al.[55]200125–3030c1 year+ NDND

Improved Mayo Risk Score for UDCA vs. placebo and for high-dose vs. low-dose UDCA

Improved liver test results

Olsson et al.[56]200517–23198 (97/101)5 years(+)NDNo effect on death, OLT, CCA or liver tests
Lindor et al.[57]200928–30149 (76/73)6 years+ ND

Terminated at 6 years as worse outcome in treatment group for death or OLT

Improved liver tests in UDCA group

Future medical treatment of PSC

Like in PBC, there is no other bile acid than UDCA in clinical use for the treatment of PSC. Based on promising phase I and animal data,[126-128] nor-UDCA is currently in phase II clinical trials (http://clinicaltrials.gov/ct2/show/NCT01755507). A variety of immunosuppressives, antibiotics and other compounds have been tested over the years, often in the context of poor study design and in trials of limited duration. However, some trials are of comparable size and quality as the trials evaluating the use of UDCA in PSC (Table 4). More recently, there has been a renewed interest in antibiotic treatment,[86, 129] largely deriving from rapidly emerging data regarding the interrelationship among bile acid metabolism, systemic and mucosal inflammation and the gut microbiota.[130-133] However, specific roles for the gut microbiota in PSC pathogenesis, albeit an intriguing possibility, are still devoid of robust experimental support. Furthermore, knowledge may emerge on specific mechanisms for which broad antibacterial interventions like antibiotics may not represent the appropriate treatment approach. Therapies targeting fibrogenesis are also of interest in PSC, e.g. lysyl oxidase-like 2[134] (http://clinicaltrials.gov/ct2/show/NCT01672853). There is little controversy as to the application of widely variable treatment approaches in PSC, given uncertainties that still exist regarding pathogenesis.[135] Clinical trials are hampered by inherent limitations such as a slowly progressing disease and lack of applicable prognostic indices (Table 5). This includes a lack of or consensus protocols for assessing disease progression by available imaging modalities (MRC, transient elastography).[136-140] The outcomes of ongoing clinical trials are therefore likely to have flaws intrinsic to study design and disease behaviour as much as choice of treatment targets.

Table 4. Overview of clinical studies of non-ursodeoxycholic acid treatment in primary sclerosing cholangitis. References for studies refer to the online supplementary reference list (Data S1 for tables 1–5)
StudyYearTreatmentN (treat/placebo)Study durationLabHistologyOLT-free survivalOutcome
  1. ALP, alkaline phosphatase; UDCA, ursodeoxycholic acid; ND, not done; OLT, orthotopic liver transplantation.

  2. The interested reader is guided to the following references for treatment reports not fulfilling the criteria for table presentation, i.e. randomised allocation to either a treatment group or a control group receiving placebo or no treatment with the aim of evaluating effect on pre-transplant PSC.[81-98]

La Russo et al.[75]1988Penicillamine70 (39/31)3 yearsNo effect on liver tests, histology or survival
Knox et al.[76]1994Methotrexate24 (12/12)2 years+ (ALP only)Improved ALP. No effect on histology, cholangiography or outcome
Olsson et al.[77]1995Colchicine84 (44/40)3 yearsNo effect on liver tests, histology or survival
Sterling et al.[78]2004Mycophenolate mofetil/UDCA vs. UDCA25 (12/13)2 yearsNo effect on liver tests, histology, cholangiography or Mayo Risk Score
Farkkila et al.[79]2004Metronidazole/UDCA80 (39/41)36 months+(+) Improved liver tests and Mayo Risk Score, but no improvement in histology or cholangiography
Hommes et al.[80]2008Infliximab10 (6/4)12 monthsNDPatient enrolment was prematurely stopped when interim analysis showed no treatment benefit. No effect on liver tests, histology
Table 5. Predictive models for primary sclerosing cholangitis (PSC) progression. References for studies refer to the online supplementary reference list (Data S1 for tables 1–5)
  1. AST, aspartate aminotransferase; IBD, inflammatory bowel disease; INR, international normalised ratio; MELD, model of end-stage liver disease.

Study Year Variables in predictive model
Pugh et al.[99]1973Bilirubin, albumin, INR, encephalopathy, ascites (Child-Pugh score)
Wiesner et al.[100]1989Age, bilirubin, histological stage, haemoglobin, IBD (Mayo risk score)
Farrant et al.[101]1991Age, histological stage, hepatomegaly, splenomegaly, alkaline phosphatase (King's College Hospital)
Dickson et al.[102]1992Age, bilirubin, histological stage, splenomegaly
Broome et al.[103]1996Age, bilirubin, histological stage
Kim et al.[104]2000Age, bilirubin, albumin, AST, variceal bleeding (revised Mayo risk score)
Kamath et al.[105]2001Bilirubin, creatinine, INR (MELD score)
Boberg et al.[106]2002Age at diagnosis, bilirubin, albumin
Ponsioen et al.[107]2002Age, cholangiographic score
Tischendorf et al.[108]2007Age, albumin, bilirubin elevation >3 months, hepatomegaly, splenomegaly, dominant bile duct stenosis, intra- and extrahepatic bile duct changes (‘PSC score’)
Endoscopic treatment in PSC

There are also differences in the practice preferences related to the main treatment options for PSC, comprising endoscopic interventions and liver transplantation. For endoscopy, an ongoing trial aims to clarify whether balloon dilatation or short-term (2–3 weeks) stenting is the preferable management for bile duct strictures (http://clinicaltrials.gov/ct2/show/NCT01398917). Both strategies are in current practice,[141, 142] yet the additional value of stenting is not established and has been claimed to associate with increased frequencies of cholangitis.[143] As to either approach, the threshold for intervention is often left at the discretion of the endoscopist, a decision likely to be biased by spontaneous fluctuations in bile flow and cholestatic parameters in PSC.[144]

Liver transplantation in PSC

Pruritus, more rarely, is the sole indication for liver transplantation in PSC compared with PBC due to correlation of pruritus with biliary stricture severity to a larger degree in PSC than in PBC. The appropriate timing of liver transplantation for PSC vs. the risk of biliary dysplasia and cancer is more complex.[145] The EASL guidelines consider cholangiocyte dysplasia as well as severe recurrent bacterial cholangitis as potential indications for liver transplantation. The AASLD guidelines have a notion of liver transplantation with neoadjuvant chemotherapy and brachytherapy (‘the Mayo protocol’),[102] but do not specifically consider biliary dysplasia. In the Nordic countries, which do not have MELD-based organ allocation, biliary dysplasia detected by brush cytology may serve as the sole indication for liver transplantation. The practice has been criticised on the basis that 1/3 of PSC patients may exhibit cholangiocellular dysplasia in the absence of cholangiocarcinoma, and further validation is needed.[146-148] Clearly, upon improvements in biomarkers for early cholangiocarcinoma detection, an evidence-based shift toward earlier transplantation in these patients will be feasible. As of yet, decision-making is largely based on the experience and preference of each individual transplant programme.

Immunosuppressive regimens in liver-transplanted PSC patients are now mainly tacrolimus-based.[149, 150] There is no firm evidence that the choice of immunosuppressive regimen directly affects risk of PSC recurrence.[81, 149] However, there is an, as of yet not fully understood, interrelationship between post-transplant IBD activity (requiring intensified corticosteroid treatment) and increased PSC recurrence rates.[149, 151] A cause–effect relationship between IBD activity and PSC occurrence (or, more appropriately, re-occurrence) is adopted by most authors as the most likely explanation, but alternative mechanisms (e.g. intrinsically high disease activity of both bile duct and colonic inflammation) may also prove correct. There is also a relationship between acute cellular rejection (against which corticosteroids would protect) and PSC recurrence.[152] The topic is of considerable interest as long-term follow-up indicates that PSC recurrence does increase re-transplantation rates.[153] Interpreting the relationship between disease recurrence and immunosuppression is further complicated by associations between tacrolimus-based regimens and IBD flares following liver transplantation in PSC.[154-156] As of yet, evidence does not allow for firm recommendations either on timing of colectomy (pre-transplant in patients with high-activity IBD pre-transplant?) or choice of immunosuppression (ciclosporine-based regimens in patients with high-activity IBD post-transplant?). Likely, on a more comprehensive appreciation of disease mechanisms in PSC and IBD in PSC, immunosuppression regimens post-liver transplantation in PSC that accounts not only for allograft rejection issues but also for PSC immunopathology can be designed.

Common features in PBC and PSC

Autoimmune hepatitis

A consensus document from the International Autoimmune Hepatitis Group (IAIHG) deals thoroughly with the controversy as to whether ‘overlap syndrome’ (Figure 2) should be considered a separate disease entity.[157] If the patients fulfil a diagnosis of PBC or PSC, this should, according to this consensus document, be the primary diagnosis and co-occurring features of autoimmune hepatitis should be determined using general considerations (ALT at least 5× ULN, IgG at least 2× ULN, histological features of autoimmune hepatitis) rather than the IAIHG scoring systems.[158] Identification of patients for corticosteroid therapy is not straightforward. Treatment response to corticosteroids in PSC and PBC patients with concomitant features of autoimmune hepatitis is generally less pronounced than in patients with autoimmune hepatitis as the primary diagnosis,[159, 160] and more so in PSC than in PBC.[161, 162] It is thus important to be aware of the risk of side effects (osteoporosis in particular) and to assess treatment response regularly, as well as to consider steroid sparing agents for patients requiring long-term treatments.[67]

Cholestatic liver cirrhosis

To what extent defects of bile acid homoeostasis involve in the primary insult in PBC and PSC is still debated.[163] Regardless of the aetiological basis of bile duct injury in PBC and PSC, development of liver fibrosis throughout the disease course most likely involves bile acid toxicity at some level.[1] The hepatocyte initiates adaptive responses during cholestasis (Figure 3), including the (i) downregulation of bile acid synthesis and hepatocellular bile acid uptake, (ii) increased hydroxylation and conjugation to make bile acids more water soluble and (iii) induction of bile acid efflux pumps on the sinusoidal membrane leading to export of metabolised bile acids to the systemic circulation for renal elimination. Together with FXR and PXR, the constitutive androstane receptor (CAR) is a major determinant of the expression of the genes involved in this adaption.[164, 165] The concept of directing or augmenting these adaptive mechanisms during cholestasis forms the basis of several ongoing treatment trials (e.g. obeticholic acid), but so far remains to be proven. At the level of the cholangiocyte, several lines of evidence converge on the importance of apical bicarbonate secretion (the bicarbonate ‘umbrella’ hypothesis) in preventing the protonation of luminal bile acids and thus reducing cellular injury.[163, 166] The bile acid receptor TGR5 is involved in these mechanisms, but so far therapeutic efficacy for TGR5 agonism in cholestatic liver disease seems to occur in conjunction with FXR effects.[167] Importantly, the risk of accelerating cholangiocarcinoma development argues against human applications of TGR5 agonists in PSC.[167] The role of the gut microbiota in modifying the cholestatic bile acid pool within the enterohepatic circulation and vice versa also needs to be defined (Figure 3).[130-133]

image

Figure 3. Traditions of bile acid research and the concept of ‘bile acid toxicity’ serve as the main basis for present therapies in primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC), as well as for most of the ongoing treatment trials. At the level of the liver, protective mechanisms in a context of cholestatic liver disease involve regulation of bile acid transport and metabolism (red, downregulation generally beneficial; blue, upregulation generally beneficial). At the level of the bile ducts, protective mechanisms involve apical bicarbonate secretion by the cholangiocytes (blue). There is emerging evidence that the gut microbiota, during the intestinal phase of the enterohepatic circulation, influences these mechanisms and serves additional regulatory mechanisms. There is an ongoing controversy in PBC and PSC as to the importance of these mechanisms as compared with primarily immune-mediated liver and bile duct injury. Cyp (cytochrome P450), UGT (uridine diphosphate glucuronosyltransferase), SULT (sulfotransferase), GST (glutathione S-transferase), BSEP (bile salt export pump), MRP2, 3 and 4 (multidrug resistance-associated protein 2, 3 and 4), MDR3 (multidrug resistance protein 3), OST (organic solute transporter, OSTα-OSTβ heterodimer), ASBT (apical sodium-dependent bile acid transporter), IBABP (ileal bile acid-binding protein), NTCP (Na+-taurocholate cotransporting polypeptide), OATP (several organic anion transporting polypeptides), CFTR (cystic fibrosis transmembrane conductance regulator), AE2 (anionic exchanger 2), TGR5 (Takeda G-protein coupled receptor 5).

Download figure to PowerPoint

Pruritus

The large number of candidate pruritogens in cholestatic liver disease launched over the years is reviewed elsewhere.[168-170] More recently, two groups have reported on slightly differing aspects of cholestatic pruritus.[171, 172] An Amsterdam group identified lysophosphatidic acid (LPA) as a pruritogen produced by autotaxin[171] and associating with itching behaviour in mice. In humans, the serum activity of autotaxin correlated with itch intensity and was reduced upon treatment with nasobiliary drainage. Similar correlations were not observed for bile salts, in contrast to findings recently reported from a San Francisco group, which detected correlations with itching behaviour and TGR5 activation status.[172] Intradermal injection of bile acids and a TGR5-agonist induced itch behaviour in the experimental setup, which also included detection of reduced itch behaviour in tgr5-/- mice and spontaneous itching behaviour in TGR5 overexpressing mice. The main criticism against the TGR5 data pertains to the doses of deoxycholic acid used in the experiments,[170] which range far above physiological concentrations in cholestasis.[173] Furthermore, the concentrations used to elicit itching in vivo may also induce mast cell degranulation and thus cause itching via non-TGR5 mechanisms.[174] Further studies are likely to elaborate on both TGR5 and autotaxin-associated aspects of pruritus and thus potentially open up for novel treatment options.

Fatigue

The pathogenetic basis for fatigue in cholestatic liver disease, sometimes pronounced and debilitating, is not known. The most pronounced affections occur in PBC, where more than half of the patients have been reported to suffer from fatigue.[175, 176] Many of these patients report fatigue to represent the predominant cause of reduced quality of life.[177-179] The range of causes proposed to cause fatigue in PBC is large and reviewed elsewhere.[180] Like in pruritus, a consensus biological explanation of fatigue is as of yet lacking. In neither PBC nor PSC is there a correlation of fatigue with severity or stage of liver disease.[176, 181] The cause vs. effect relationship between fatigue and social and psychological complications to the disease burden is hard to dissect.[178, 181] In lack of therapeutic means, appropriate recognition and handling of such complications are likely to be the most efficient management option.

Conclusions

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Conclusions
  6. Authorship
  7. Acknowledgements
  8. References
  9. Supporting Information

There are several scientific traditions running in parallel within the topic of cholestatic liver diseases, predominantly represented by research involving bile acids, liver immunology and luminal gastroenterology. The strategic priorities for research over the years are coloured by these traditions, sometimes emerging as controversies on topics where traditions and research outcomes point in different directions. Often, however, discussions are also flavoured by personal preferences and experience that may contrast that of frequently limited scientific evidence. The topics in lack of a consensus should serve as focus areas for further research. The discussions should translate into better study designs and renewed efforts to dissect these topics. As to clinical practice in areas of uncertainty, and even controversy, pragmatic considerations are often required. Rational guidance has now also been provided by the EASL and AASLD practice guidelines.[67, 68, 102] Major challenges for future research relate to the accurate assessment of disease activity and progression to allow for evaluation of the treatment effect in clinical trials. Still, PBC and PSC patients account for almost 10% of the European liver transplant programme, highlighting the need for intense research into disease pathophysiology as a basis for novel treatment options. Cholangiocarcinoma represents the cause of death in up to half of the PSC patients, and biomarkers for early detection are sorely needed in this population.

Multiple initiatives in the international community are presently acting upon these topics, and there is also increased appreciation from the broader community (EASL and AASLD) as to the importance of improving patient management in cholestatic liver disease. We anticipate that some of the hurdles will soon be overcome, but as of yet, we have to care for the patients on the basis of sometimes conflicting or even lacking scientific evidence.

Authorship

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Conclusions
  6. Authorship
  7. Acknowledgements
  8. References
  9. Supporting Information

Guarantor of the article: Kirsten Muri Boberg.

Author contributions: T. H. Karlsen wrote the first draft of the manuscript and prepared the figures. M. Vesterhus and K. M. Boberg revised the manuscript for critical content and prepared the tables. All authors approved the final version of the manuscript.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Conclusions
  6. Authorship
  7. Acknowledgements
  8. References
  9. Supporting Information

Declaration of personal interests: We thank Tor Halland for assistance in preparing the figures.

Declaration of funding interests: Prof. Boberg is involved in the nor-UDCA trial from the FALK pharmaceutical company.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Conclusions
  6. Authorship
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Conclusions
  6. Authorship
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
apt12581-sup-0001-DataS1.docxWord document27KData S1. References to Tables 1–5.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.