Is there an optimal therapeutic regimen for antimitochondrial antibody-negative primary biliary cirrhosis (autoimmune cholangitis)?

Authors


Correspondence to: Dr J. P. Gisbert, Servicio de Gastroenterología y Hepatología, Hospital Universitario de la Princesa, Diego de León, 62, 28006 Madrid, Spain. E-mail: gisbert@meditex.es

Summary

Testing for antimitochondrial antibodies is the most useful laboratory procedure in the diagnosis of primary biliary cirrhosis; nevertheless, 5–10% of patients with typical features of primary biliary cirrhosis do not have detectable antimitochondrial antibodies, their condition being referred to as antimitochondrial antibody-negative primary biliary cirrhosis or ‘autoimmune cholangitis’. Uncertainty exists whether antimitochondrial antibody-positive and -negative primary biliary cirrhosis represent distinct entities.

We reviewed studies that compared: (i) the clinical, laboratory and histological characteristics of antimitochondrial antibody-positive and -negative primary biliary cirrhosis; (ii) the response to treatment of both conditions; and (iii) the response of autoimmune cholangitis to ursodeoxycholic acid and immunosuppressive therapy. Antimitochondrial antibody-positive and -negative primary biliary cirrhosis were characterized by similar clinical, laboratory and histological abnormalities, clinical course and survival.

Antimitochondrial antibody status did not seem to affect the response to ursodeoxycholic acid. At present, the efficacy of therapies for autoimmune cholangitis has not been established in controlled trials. Of 52 patients with autoimmune cholangitis treated with ursodeoxycholic acid in 13 uncontrolled studies, 83% had serum biochemical improvement. Also, a favourable effect of immunosuppressive drugs occurred in 57% of 54 patients with autoimmune cholangitis in 17 uncontrolled studies. Each of these trials included very few patients and most evaluated the effects of treatment on surrogate markers of disease only.

No marker that consistently distinguished patients who would respond favourably to ursodeoxycholic acid or immunosuppression was apparent. Consequently, treatment is, at present, empirical. However, ursodeoxycholic acid may be given when histology reveals bile duct lesions, whereas immunosuppressive therapy should probably be reserved for patients exhibiting interface hepatitis.

Introduction

Primary biliary cirrhosis is a chronic, non-suppurative, destructive cholangitis characterized by the presence of antimitochondrial antibodies (AMA) in serum.1 Testing for AMA is the most useful laboratory procedure in the diagnosis of primary biliary cirrhosis, because these autoantibodies are present in most cases.2 Nevertheless, 5–10% of patients with otherwise classical features of primary biliary cirrhosis do not have detectable AMA,3 a condition that traditionally has been designated AMA-negative primary biliary cirrhosis. In the majority of patients with AMA-negative primary biliary cirrhosis, autoantibodies other than AMA, such as antinuclear antibodies (ANA) and/or anti-smooth-muscle antibodies (ASMA), are present.4–9 Recently, the terms ‘autoimmune cholangiopathy’ and ‘autoimmune cholangitis’ have been introduced for the association of AMA-negative primary biliary cirrhosis with ANA and/or ASMA serum positivity.10, 11 Although some authors have not considered ANA positivity as a diagnostic criterion for autoimmune cholangitis,12–16 many others have inferred that the presence of ANA is a constant feature of this syndrome.

In some patients with primary biliary cirrhosis, AMA titres may be low and/or fluctuating.17 In such patients, false negative results of the indirect immunofluorescence technique, the most common method for the detection of AMA, may occur in clinical laboratories.18, 19 Furthermore, approximately 20% of patients who appear to fulfil the current diagnostic criteria of autoimmune cholangitis may, after more detailed testing, be diagnosed as having classical AMA-positive primary biliary cirrhosis.10, 14, 20 In addition, up to 73% of patients with autoimmune cholangitis are AMA-positive using recombinant mitochondrial autoantigens.21 In contrast, the data of case reports suggest the concept of sequential diseases.22, 23 These observations suggest that, in some instances, whether a diagnosis of primary biliary cirrhosis or autoimmune cholangitis is made might depend on the stage of evolution of a single disease process.

At present, there is uncertainty as to whether autoimmune cholangitis and AMA-positive primary biliary cirrhosis represent clinically and immunologically distinct entities.10, 24 There appears to be a consensus that both conditions have common histological characteristics.25 Indeed, it has been suggested that autoimmune cholangitis can be defined as AMA-negative primary biliary cirrhosis.26 On the other hand, others have suggested that autoimmune cholangitis is an entity distinct from typical primary biliary cirrhosis and not merely an ‘AMA-negative’ variant.12 Autoimmune cholangitis, but not primary biliary cirrhosis, is characterized by the presence of antibodies to carbonic anhydrase;27 however, testing for these autoantibodies did not discriminate between autoimmune cholangitis and primary biliary cirrhosis in other studies.28–30 Recently, 73% of patients with primary biliary cirrhosis, who were AMA-negative by the immunofluorescence assay and were considered to have autoimmune cholangitis, were found to be AMA-positive by reactivity to recombinant autoantigens.21

The differentiation between primary biliary cirrhosis, autoimmune cholangitis and autoimmune hepatitis is important, because of the potential different responses of these chronic liver diseases to specific therapeutic strategies. Patients with primary biliary cirrhosis may not benefit from immunosuppressive therapy or suffer serious side-effects,1 whereas ursodeoxycholic acid therapy may be associated with a lower rate of progression to cirrhosis, survival benefit and a delay in the timing of liver transplantation.9, 31–33 In contrast, most patients with autoimmune hepatitis exhibit an early beneficial response to immunosuppressive treatment.34 Our aims in this investigation were to review: (i) studies comparing the clinical, laboratory and histological characteristics of AMA-positive and -negative primary biliary cirrhosis; (ii) studies comparing the response to treatment of AMA-positive and -negative primary biliary cirrhosis; and (iii) studies evaluating the response of autoimmune cholangitis to treatment with ursodeoxycholic acid and/or immunosuppressive therapy.

Methods

Search strategy and selection criteria

Bibliographical searches were performed in the PubMed (Internet) database and in the Cochrane Controlled Trials Register, including studies available up until January 2002. The following terms were sought (all fields): (‘autoimmune cholangitis’) OR (‘autoimmune cholangiopathy’) OR (‘immunocholangitis’) OR (‘antinuclear antibody/ies AND primary biliary cirrhosis’) OR (‘antimitochondrial antibody/ies and primary biliary cirrhosis and negative’) OR (‘autoimmune hepatitis and variant syndromes’) OR (‘atypical primary biliary cirrhosis’). Abstracts of the articles found were reviewed to determine if they fulfilled the inclusion criteria relevant to one of the three pre-defined objectives. References cited in reviews on autoimmune cholangitis or AMA-negative primary biliary cirrhosis, and from the articles previously selected, were also examined in search of studies that met the inclusion criteria. Articles published in any language were reviewed. In the present review, ‘AMA-negative primary biliary cirrhosis’ and ‘autoimmune cholangitis’ are considered as a single entity and the two terms are used interchangeably.

Data extraction

In reports of the response of autoimmune cholangitis to treatment, the following variables were extracted in a pre-defined form (see Tables 1 and 2): author, number of patients treated, type of treatment (ursodeoxycholic acid or immunosuppressive therapy), follow-up after treatment (months), end-point of response assessment (biochemical or histological) and response rate (percentage of patients who underwent improvement). Finally, notes were also included when relevant in some cases [e.g. presence of an appreciable component of autoimmune hepatitis, referral for orthotopic liver transplantation, previous therapy with other drugs and details of partial responses (e.g. reduction of inflammation but persistence of cholestasis and bile duct lesions in patients treated with immunosuppressive drugs)].

Table 1.  Studies of patients with autoimmune cholangitis treated with ursodeoxycholic acid (UDCA)
ReferenceNumber of patientsType of treatmentFollow-up (months)End-pointResponse (% with improvement)Notes
  • AIH, autoimmune hepatitis.

  • Weighted mean.

Lacerda et al.16 5UDCA> 12Biochemical100 
Sánchez-Pobre et al.7 2UDCA> 12Biochemical
 (histological
 in one case)
100Patients with milder
 abnormality on liver test
 results received UDCA
Omagari et al.41 1UDCA< 12Biochemical100Significant component of AIH
Kim et al.20 8UDCA> 12Biochemical100Two patients received
 orthotopic liver transplantation
Invernizzi et al.14 9UDCA> 12Biochemical 89 
Salóet al.37 5UDCA   6Biochemical100Two patients had previously
 not responded to steroids
 One patient received
 orthotopic liver transplantation
Li et al.36 8UDCA  12Biochemical100 
Mohr et al.42 1UDCA> 12Biochemical100No response to
 immunosuppressive therapy
Primo et al.43 1UDCA  12Biochemical100Partial improvement of
 biochemistry previously
 achieved with steroids was
 maintained with UDCA
Romero Gomez et al.44 2UDCA 4–6Biochemical100 
Cho et al.23 1UDCA> 12Biochemical  0 
Czaja et al.12 8UDCA   6Biochemical
 and
 histological
 12 
Devesa et al.45 1UDCA   1Biochemical100Despite cholestatic
 biochemical improvement,
 liver failure finally
 appeared, and hepatic
 transplantation
 was necessary
Total52    83* 
Table 2.  Studies of patients with autoimmune cholangitis treated with immunosuppressive therapy
ReferenceNumber of patientsType of treatmentFollow-up (months)End-pointResponse
(% with improvement)
Notes
  • AIH, autoimmune hepatitis; UDCA, ursodeoxycholic acid.

  • *

     Weighted mean.

Brunner and  Klinge48 3Steroids and  azathioprine< 12Biochemical100Significant component  of AIH
Ben-Ari et al.6 3Steroids> 12Biochemical
 and
 histological
100Significant component  of AIH
 Partial response:
 inflammation was reduced,
 but cholestasis and
 bile duct lesions persisted
Taylor et al.5 7Steroids and  azathioprine> 12Biochemical  0 
Colombato et al.22 1Steroids> 12Biochemical100Significant component  of AIH
Sánchez-Pobre et al.7 4Steroids and/or  azathioprine> 12Biochemical  (histological in  one case)100Patients with the most  markedly  abnormal liver  function test  results received steroids
Salóet al.37 6Steroids ±  azathioprine   6Biochemical 67 
Czaja50 6Steroids   3Biochemical  0 
Mohr et al.42 1Steroids and  azathioprine   4Biochemical  0Response to UDCA   was demonstrated
Sherlock51 5Steroids> 12Biochemical100Partial response:
 inflammation was reduced,
 but cholestasis and
 bile duct lesions persisted
Mitry et al.52 1Steroids> 12Biochemical100 
Masumoto et al.53 3Steroids> 12Biochemical
 and
 histological
100Partial response:
 inflammation was reduced,
 but cholestasis and
 bile duct lesions persisted
Marinho et al.49 2Steroids> 12Biochemical100Significant component  of AIH
Primo et al.43 1Steroids   3Biochemical100Partial response (less
 than 50% improvement)
Cho et al.23 1Steroids> 12Biochemical100 
Ogata et al.54 1Steroids < 1Biochemical100 
Czaja et al.12 8Steroids ±  azathioprine  12Biochemical and   histological 12 
Castro et al.55 1Steroids> 12Biochemical100 
Total54    57* 

Quality assessment

A priori, it was planned to assess the quality of comparative clinical trials using the scoring system proposed by Jadad et al., which is based on three items (randomization, double-blinding and documentation of withdrawals and dropouts). However, the quality score could not be finally applied as no report of a comparative controlled clinical trial was identified (see ‘Results and discussion’ section).

Subanalyses

In studies that evaluated the response of autoimmune cholangitis to treatment, subanalyses of responses to ursodeoxycholic acid or immunosuppressive drugs were planned, a priori, separately depending on the therapy prescribed.

Outcome variable/end-point

In studies in which the response of autoimmune cholangitis to treatment was evaluated, the outcome variable was ‘response’, defined as the percentage of patients who underwent a detectable improvement. The response variables assessed were either biochemical or histological in all studies. Overall, the mean percentage of patients who underwent a favourable response was calculated, and expressed as the weighted mean, to make due allowance for the number of patients included in each study.

Results and discussion

Comparison of clinical features of AMA-positive and -negative primary biliary cirrhosis

A decisive question is whether primary biliary cirrhosis and autoimmune cholangitis are distinct entities or part of the same disease spectrum.4, 7, 13, 14, 16, 35 Indeed, it has been suggested that autoimmune cholangitis may represent variant forms of diverse conditions, a transition stage or a separate entity with varying clinical, laboratory and histological manifestations.12

The clinical, laboratory and histological characteristics and the clinical course of AMA-positive and -negative primary biliary cirrhosis have been compared in several studies. The results are summarized in chronological order of their publication. In 1994, Michieletti et al. reported a group of 20 patients with autoimmune cholangitis and compared them with another group of 20 patients with AMA-positive primary biliary cirrhosis matched for sex and serum bilirubin.4 An analysis of clinical findings, hepatic histology and serum bilirubin, alkaline phosphatase, alanine aminotransferase and immunoglobulin G disclosed no significant differences between the two patient groups. However, the use of bilirubin as a matching criterion in this study may explain, at least partially, the lack of differences between the courses of the two groups. Goodman et al. reviewed 200 patients in whom hepatic histology was consistent with primary biliary cirrhosis;13 40 cases were AMA-negative and ANA-positive. They observed no significant differences in the clinical characteristics, laboratory results or hepatic histopathology between the two groups; however, there was an appreciably higher proportion of patients with established cirrhosis in the AMA-positive group. Lacerda et al. studied 597 primary biliary cirrhosis patients;16 35 of these were negative for AMA. No differences were found between the AMA-positive and -negative patients with respect to age, gender and serum biochemistry; however, serum immunoglobulin M and gamma-globulin levels were higher in the AMA-positive group. Sánchez-Pobre et al. compared the clinical, laboratory and liver biopsy features of seven patients with AMA-negative primary biliary cirrhosis and seven patients with AMA-positive primary biliary cirrhosis.7 The AMA-negative group had a significantly higher incidence of asthenia and liver failure, and higher serum immunoglobulin G levels. There were no significant differences between the two groups with respect to other laboratory test results, although there was a tendency for serum bilirubin and aminotransferase levels to be higher and for serum alkaline phosphatase and immunoglobulin M levels to be lower in the AMA-negative group. It was noteworthy that the liver histology was similar in both groups. Invernizzi et al. studied 297 patients with clinical and histological features of primary biliary cirrhosis.14 AMA-negative and AMA-positive patients were compared. Both groups had similar abnormal clinical and serum biochemical findings. ANA and ASMA were more frequently positive in the AMA-negative patients; however, this serological profile did not seem to be associated with a different presentation or clinical course of the disease. Most importantly, the incidence of complications of cirrhosis and development of liver failure did not differ significantly between the two groups. Li et al. compared several variables in primary biliary cirrhosis patients who were AMA-positive or -negative.36 There were no significant differences between the two groups with respect to age, sex, clinical manifestations, routine serum biochemical liver tests, hepatic histological findings and survival; multivariate analysis revealed that lower serum immunoglobulin M levels were the only predictor to negative serum AMA in primary biliary cirrhosis patients. Nakanuma et al. examined the clinicopathological features of 30 patients with AMA-negative primary biliary cirrhosis and 38 with AMA-positive primary biliary cirrhosis.15 There were few differences between the clinical and serological features of the two groups, and the hepatic histopathological features were also similar. Salóet al. reviewed a series of 13 patients with features of autoimmune cholangitis and compared them with 13 patients with typical primary biliary cirrhosis.37 No relevant clinical or hepatic histological differences between the two groups of patients were apparent. Kinoshita et al. compared the features of 21 cases of autoimmune cholangitis, 37 cases of AMA-positive primary biliary cirrhosis and 16 cases of autoimmune hepatitis.38 By clinical, biochemical and hepatic histological criteria, autoimmune cholangitis and primary biliary cirrhosis were similar, and both could be readily distinguished from autoimmune hepatitis. Finally, Czaja et al. identified 20 patients with autoimmune cholangitis and compared them with 242 patients with a well-defined form of autoimmune liver disease.12 Patients with autoimmune cholangitis were distinguished from those with primary biliary cirrhosis by the presence of higher serum levels of aspartate aminotransferase and bilirubin and lower serum concentrations of immunoglobulin M, and a higher incidence of non-AMA autoantibodies.

In summary, it appears that patients with AMA-positive and -negative primary biliary cirrhosis do not have substantially different clinical, serum biochemical and hepatic histological features or course of chronic autoimmune liver disease. In particular, long-term follow-up reveals that the incidence of complications of cirrhosis and survival are similar for AMA-positive and -negative patients.

Comparison of response of patients with AMA-positive and -negative primary biliary cirrhosis to treatment

Lacerda et al. compared, retrospectively, the responses of serum biochemistry and the Mayo risk score of five patients with AMA-negative primary biliary cirrhosis treated with ursodeoxycholic acid with the corresponding responses of patients with typical AMA-positive primary biliary cirrhosis receiving ursodeoxycholic acid; the outcomes were similar in both groups.16 Kim et al. reported the results of ursodeoxycholic acid treatment in eight patients with AMA-negative primary biliary cirrhosis.20 Serum biochemical liver tests showed a similar pattern of improvement to that which occurred in similarly treated AMA-positive primary biliary cirrhosis patients. After a median of 36 months of follow-up, the clinical outcome was also similar in AMA-positive and -negative patients. Invernizzi et al. compared the response of nine AMA-negative and 88 AMA-positive patients to treatment with ursodeoxycholic acid for more than 1 year.14 Responses to therapy were similar in the two groups, both in terms of improvement in serum biochemical liver tests and the incidence of treatment failures. Finally, Li et al. compared the responses of AMA-positive and -negative primary biliary cirrhosis patients to 1 year of ursodeoxycholic acid; the responses of the two groups were similar.36 In summary, AMA status does not seem to affect the response of patients with histological features of primary biliary cirrhosis to treatment with ursodeoxycholic acid.

Response of patients with autoimmune cholangitis to treatment with ursodeoxycholic acid

Ursodeoxycholic acid has been shown to improve serum biochemical liver tests, but its effects on disease progression and survival are controversial.39, 40 Some studies (Table 1) have assessed the effect of ursodeoxycholic acid on the clinical, serum biochemical and hepatic histological features of autoimmune cholangitis.7, 12, 14, 16, 20, 23, 36, 37, 41–45 The number of patients included in each of these studies was low. Only studies in which more than one patient was included are reviewed here. Lacerda et al. reported an improvement in both serum biochemical liver test results and the Mayo risk score in five patients with AMA-negative primary biliary cirrhosis treated with ursodeoxycholic acid.16 Sánchez-Pobre et al. treated two patients with AMA-negative primary biliary cirrhosis and mild abnormalities of serum biochemical liver tests with ursodeoxycholic acid; treatment was associated with the disappearance of symptoms and improvement of serum biochemical liver test results in both patients.7 One of these patients, who had histological stage II disease before treatment, underwent a post-treatment liver biopsy which revealed regression to stage I disease. Kim et al. reported their experience with eight patients with AMA-negative primary biliary cirrhosis treated with ursodeoxycholic acid; a favourable response of serum biochemical liver test results occurred after a median of 36 months of follow-up.20 Invernizzi et al. treated nine AMA-negative primary biliary cirrhosis patients with ursodeoxycholic acid and reported a serum biochemical response in most (89%) cases.14 Salóet al. administered ursodeoxycholic acid to five patients with autoimmune cholangitis, and subsequently observed an improvement of serum biochemical liver test results in each case, including two patients who had not responded previously to corticosteroids.37 Nevertheless, the improvement was only slight in one patient, who was subsequently referred for orthotopic liver transplantation because of intense pruritus. Li et al. prescribed ursodeoxycholic acid to eight patients with AMA-negative primary biliary cirrhosis and, after 1 year of treatment, serum levels of bilirubin, alkaline phosphatase and alanine aminotransferase and the Mayo risk scores decreased.36 Romero Gomez et al. treated two women with autoimmune cholangitis with ursodeoxycholic acid, and serum biochemical improvement occurred in both cases.44 Czaja et al. treated eight autoimmune cholangitis patients with ursodeoxycholic acid; only one (12%) satisfied the authors' criteria for remission after 6 months of therapy.12

In summary, an apparently favourable effect of ursodeoxycholic acid on autoimmune cholangitis has been reported in most studies, although none included significant ‘hard’ end-points. Some improvement was associated with ursodeoxycholic acid in 83% (weighted mean) of 52 patients (Table 1). However, only partial serum biochemical responses occurred in some studies. In the majority of studies, evidence of histological response was not sought. Furthermore, it should be stressed that, in several patients, serum biochemical responses associated with ursodeoxycholic acid did not prevent subsequent referral for liver transplantation.

Response of patients with autoimmune cholangitis to treatment with immunosuppressive therapy

Corticosteroids and azathioprine have not been shown to have a favourable risk/benefit ratio in the treatment of primary biliary cirrhosis,1 although, in some cases, corticosteroids may achieve biochemical improvement.46, 47 In addition, drugs of this class accentuate osteoporosis and may precipitate pathological fractures.1, 47 However, some authors have reported encouraging results of immunosuppressive therapy in autoimmune cholangitis. Only studies in which more than one patient was included are reviewed here. Firstly, in 1987, Brunner and Klinge reported a prompt, dramatic decrease in serum aminotransferase levels after prednisone and azathioprine were administered to three patients with autoimmune cholangitis.48 Sánchez-Pobre et al. evaluated the responses to immunosuppressive treatment in AMA-negative primary biliary cirrhosis.7 In four patients with more severe biochemical abnormalities, therapy with azathioprine and/or corticosteroids was associated with clinical and serum biochemical improvements. Salóet al. prescribed corticosteroids to six patients with autoimmune cholangitis; treatment-associated responses were inconsistent.37 In four patients, there were both clinical and serum biochemical improvements, whereas, in the other two, no beneficial response was apparent. Finally, Marinho et al. studied two patients with features of both primary biliary cirrhosis and autoimmune hepatitis.49 Liver biopsies from both patients revealed cirrhosis with appreciable bile duct lesions. After the initiation of prednisolone therapy, elevated serum aminotransferase levels rapidly normalized.

However, in other studies of the treatment of autoimmune cholangitis with immunosuppressive drugs, no trends suggesting improvement of the disease process occurred. Taylor et al. treated seven patients with autoimmune cholangitis with prednisone and azathioprine, but no clinical benefit was apparent.5 Indeed, during follow-up of 1–5 years, there was evidence of disease progression in all patients. Czaja administered corticosteroids to six patients with autoimmune cholangitis; no remissions occurred, in contrast with the favourable responses to corticosteroids observed in patients with autoimmune hepatitis in the same study.50 Finally, Czaja et al. also prescribed corticosteroids to eight patients with autoimmune cholangitis; only one patient (12%) exhibited clinical, serum biochemical and histological evidence of remission after 12 months of therapy.12

In summary, an apparently beneficial effect of immunosuppressive treatment has been reported in approximately one-half of patients with autoimmune cholangitis. Thus, of the 54 patients included in Table 2,5–712, 22, 23, 37, 42, 43, 48–55 some improvement associated with immunosuppressive therapy was observed in 57% of cases (weighted mean). However, it should be pointed out, as in the case of ursodeoxycholic acid treatment, that the responses in some studies were only partial, and that the indices of response were usually only clinical or serum biochemical markers of disease (surrogate markers). In general, hepatic histology was not used as an end-point, and the effects on survival or the incidence of referral for liver transplantation were not adequately assessed.

Patients with characteristics of autoimmune hepatitis and other autoimmune liver diseases are considered to have ‘overlap syndromes’. If they have findings that are inconsistent with a definitive diagnosis of autoimmune hepatitis, the term ‘outlier syndromes’ may be used. Both groups of patients are considered to constitute examples of the so-called ‘variant syndromes’.56 According to these definitions, autoimmune cholangitis would appear to be an ‘outlier syndrome’ and should probably not be considered as an ‘overlap syndrome’.9–11, 56–58 Some patients, however, share features of both autoimmune cholangitis and autoimmune hepatitis; such patients may be classified as having an ‘overlap syndrome’. The inclusion in certain studies of some cases of autoimmune cholangitis with features of autoimmune hepatitis, or ‘overlap syndromes’, may explain some of the reports of encouraging responses to immunosuppressive therapy6, 9, 22, 48, 49, 59–61 (Table 2). It has been suggested that the response to corticosteroids may be useful in the differentiation of primary biliary cirrhosis from autoimmune hepatitis with cholangitic features,46 indicating that a favourable response to corticosteroids may be attributable to the autoimmune hepatitis component of the ‘overlap syndrome’ and not to the cholangitic component. Therefore, it may be concluded that, in cases of autoimmune cholangitis that exhibit appreciable overlap with autoimmune hepatitis, corticosteroids alone or in combination with azathioprine may be particularly effective. Immunosuppressive therapy may decrease hepatic inflammation, but probably does not affect the cholestatic features. Ben-Ari et al. described four patients with features of overlap between AMA-negative primary biliary cirrhosis and autoimmune hepatitis.6 Three underwent rapid clinical and serum biochemical remissions with prednisolone. Serial liver biopsies indicated that treatment was associated with reduced hepatic inflammation, but persistent bile duct lesions. Such patients may constitute a subgroup of autoimmune hepatitis type 1 with significant bile duct damage. Sánchez-Pobre et al. administered immunosuppressive therapy to four AMA-negative primary biliary cirrhosis patients with markedly abnormal serum biochemical liver test results.7 Clinical and biochemical improvements occurred in all four patients. In one, a post-treatment liver biopsy showed reduced inflammatory infiltration but unchanged bile duct lesions. Masumoto et al. treated three autoimmune cholangitis patients with prednisolone, and a marked improvement in the biochemical indices of disease occurred.53 Repeat liver biopsy during treatment showed marked amelioration of hepatocellular damage in all three patients, but persistence of bile duct lesions in two. Sherlock reported a series of five autoimmune cholangitis patients treated with corticosteroids.51 The response to prednisolone therapy was partial; elevated serum transaminase levels fell and there was a reduction in liver inflammation, but serum gamma-glutamyl transpeptidase levels remained elevated and repeat liver biopsies revealed persistent bile duct lesions. Finally, it has been suggested that patients with autoimmune cholangitis who appear to respond to immunosuppressive therapy may belong to a subgroup of autoimmune hepatitis with significant bile duct damage, although this hypothesis has not been particularly addressed or proved by this review. Other patients with autoimmune cholangitis, who appear to have a clinical and serum biochemical response to ursodeoxycholic acid, may represent typical AMA-negative primary biliary cirrhosis; such patients may have a similar clinical course and response to therapy as patients with AMA-positive primary biliary cirrhosis. In summary, in patients with autoimmune cholangitis, treatment with corticosteroids may improve the clinical and cytolytic biochemical indices, and may also reduce hepatic inflammation. However, corticosteroids have little effect on the biochemical indices of cholestasis or histological features of bile duct lesions.

Finally, corticosteroid therapy in patients with chronic cholestatic liver diseases tends to accelerate the development of osteopenic bone disease and increases the incidence of pathological fractures.47 Therefore, the possible benefits should be weighed against the substantial risks associated with the long-term use of corticosteroids. Recently, some authors have suggested that, when the potential risks and benefits of treatment are considered, ursodeoxycholic acid rather than corticosteroids should be the first therapeutic option for patients with autoimmune cholangitis.10

Response of patients with autoimmune cholangitis to treatment with both ursodeoxycholic acid and immunosuppressive therapy

Li et al. reported a patient with AMA-negative but ANA-positive chronic cholestatic liver disease with hepatic histological features of both autoimmune cholangitis and autoimmune hepatitis.59 Combination therapy with immunosuppressive agents (prednisone and azathioprine) and ursodeoxycholic acid was given. On treatment, serum aminotransferases and the biochemical indices of cholestasis improved. A follow-up liver biopsy revealed improvement of both hepatic necro-inflammation and bile duct lesions. This case illustrates that a combination of corticosteroids and ursodeoxycholic acid may be effective in patients with features of both autoimmune cholangitis and autoimmune hepatitis.

Conclusions

It is still uncertain whether autoimmune cholangitis is an entity distinct from primary biliary cirrhosis or whether it should be classified as primary biliary cirrhosis without detectable serum AMA. It seems that AMA-negative primary biliary cirrhosis and autoimmune cholangitis are not associated with any substantial differences in clinical features, course of the disease, incidence of complications of cirrhosis or survival. Furthermore, AMA status does not seem to affect the response of primary biliary cirrhosis patients to treatment with ursodeoxycholic acid. At present, there is no consensus regarding the optimal therapy for autoimmune cholangitis because, due to the low prevalence of this disease, no controlled therapeutic trials have been undertaken. A favourable effect of ursodeoxycholic acid on the serum biochemical markers of patients with autoimmune cholangitis has been demonstrated in most studies (approximately 80% of patients). Additionally, a favourable effect of immunosuppressive treatment has also been demonstrated in approximately one-half of autoimmune cholangitis patients. However, these studies have important limitations; each included few patients and most evaluated the effect of treatment (ursodeoxycholic acid or immunosuppressive drugs) on biochemical markers only. It is uncertain whether serum biochemical variables are valid surrogate markers of clinically important end-points. Relatively ‘hard’ end-points include unequivocal hepatic histological improvement, referral for liver transplantation and transplantation-free survival.

The inclusion in some studies of cases of autoimmune cholangitis with appreciable features of autoimmune hepatitis, or ‘overlap syndromes’, may explain the apparently beneficial effect of corticosteroid treatment in some patients in several studies. In patients with mixed chronic hepatocellular and cholestatic disease, treatment with corticosteroids may improve the clinical and serum biochemical indices of hepatocellular necrosis and histological evidence of periportal necro-inflammatory lesions, but is unlikely to decrease cholestasis or the histological features of bile duct lesions. Finally, publication bias may partly explain the reports of relatively favourable trends associated with ursodeoxycholic acid or immunosuppressive treatment, as negative results in one or a few patients are unlikely to be published.

There is no feature that can consistently distinguish autoimmune cholangitis patients who would respond favourably to ursodeoxycholic acid or immunosuppression. Accordingly, treatment strategies are empirical. When there is marked serum biochemical evidence of cholestasis and histological features of florid bile duct lesions, ursodeoxycholic acid therapy may be initiated, although its efficacy is uncertain. Immunosuppressive therapy should probably be reserved for patients exhibiting substantial evidence of hepatic inflammation and necrosis. However, primary biliary cirrhosis patients with features of autoimmune hepatitis may respond adequately to ursodeoxycholic acid.62 An initial, short therapeutic trial with immunosuppressive therapy may be of diagnostic value.42, 57 Patients whose serum aminotransferase levels respond rapidly to immunosuppressive therapy may have atypical, predominantly cholestatic autoimmune hepatitis and might benefit from long-term immunosuppression. In cases in which immunosuppression does not induce a rapid serum biochemical response, ursodeoxycholic acid may be given as long-term treatment. For patients with features of both autoimmune cholangitis and autoimmune hepatitis, it would be reasonable to treat with a combination of an immunosuppressive agent and ursodeoxycholic acid. Randomized, controlled studies involving larger number of patients will be necessary to assess fully the differences in the response of patients with autoimmune cholangitis to treatment with ursodeoxycholic acid and immunosuppressive drugs. The end-points should be ‘hard’ (histological improvement, referral for liver transplantation, transplantation-free survival).

Ancillary