EASL Clinical Practice Guidelines

Primary biliary cirrhosis (PBC)

Diagnosis of PBC

Patients with PBC may present with symptoms as fatigue, pruritus and/or jaundice, but the majority of them are asymptomatic at diagnosis. At first presentation, very few patients present in advanced stage of disease and with complications of portal hypertension (ascites, hepatic encephalopathy or esophageal variceal bleeding). Currently, a diagnosis of PBC is made with confidence on a combination of abnormal serum liver tests (elevation of AP of liver origin for at least 6 months) and presence of AMA (⩾1:40) in serum [6]. The diagnosis is confirmed by disclosing characteristic histological features of florid bile duct lesions. AMA-positive individuals with normal AP carry a high risk to develop PBC during follow-up [7].

Laboratory tests

Biochemical markers: Serum AP and γGT are raised in PBC; serum aminotransferases (ALT, AST) and conjugated bilirubin can also be elevated, but are not diagnostic. Patients with normal AP and γGT, but with serological stigmata of PBC, should be reassessed clinically and biochemically at annual intervals. Patients with PBC typically present elevated levels of immunoglobulin M. Serum cholesterol is commonly elevated like in other cholestatic conditions. Alterations in prothrombin time, serum albumin and conjugated bilirubin are observed only in advanced disease.

Immunological markers: The diagnostic hallmark of PBC is the presence of AMA, which are detected in serum of more than 90% of affected individuals; the specificity of AMA in PBC is greater than 95% [8]. AMA reactivity is classically studied by immunofluorescence and considered positive at a titre ⩾1/40 [9]. The identification of the molecular mitochondrial target antigens has allowed the setting up of immunoenzymatic assays with recombinant proteins that raise the sensitivity and specificity of the test. If available, anti-AMA-M2 (anti-PDC-E2) may be a useful alternative. Non-specific antinuclear antibodies (ANA) are found in at least 30% of PBC sera. However, ANA directed against nuclear body or envelope proteins such as anti-Sp100 and anti-gp210 which present as multiple [[6], [7], [8], [9], [10], [11], [12]] nuclear dots and perinuclear rims, respectively, at indirect immunofluorescence staining show a high specificity for PBC (>95%) and can be used as markers of PBC when AMA are absent. Their sensitivity, however, is low.


A liver biopsy is no longer regarded as mandatory to make a diagnosis of PBC in patients with a cholestatic serum enzyme pattern and serum AMA. It may, however, be useful for assessment of the activity and staging of the disease. Histological staging of PBC (stages 1–4) has been proposed by Ludwig et al.[10] and Scheuer [11] according to the degree of bile duct damage, inflammation and fibrosis. Focal duct obliteration with granuloma formation has been termed the florid duct lesion, and is judged almost pathognomonic for PBC when present. The liver is not uniformly involved, and features of all four stages of PBC can co-exist simultaneously in a single biopsy. The most advanced histological features should be used for histological staging.


Abdominal ultrasound examination is indicated in all patients with elevation of serum AP and γGT to disclose intrahepatic or extrahepatic bile duct dilatation (see above) or focal liver lesions. There are no specific features of PBC on ultrasound; in particular the biliary tree appears normal. Ultrasound findings in advanced PBC resemble those seen in other forms of cirrhosis.


  1. A diagnosis of PBC can be made with confidence in adult patients with otherwise unexplained elevation of AP and presence of AMA (⩾1:40) and/or AMA type M2. A liver biopsy is not essential for the diagnosis of PBC in these patients, but allows activity and stage of the disease to be assessed (III/A1).
  2. A liver biopsy is needed for the diagnosis of PBC in the absence of PBC specific antibodies. A liver biopsy may also be helpful in the presence of disproportionally elevated serum transaminases and/or serum IgG levels to identify additional or alternative processes (III/C1).
  3. AMA-positive individuals with normal serum liver tests should be followed with annual reassessment of biochemical markers of cholestasis (III/C2).

Treatment of PBC

Ursodeoxycholic acid (UDCA)

Over the past two decades, increasing evidence has accumulated indicating that ursodeoxycholic acid (UDCA; 13–15 mg/kg/d) is the treatment of choice for patients with PBC based on placebo-controlled trials and more recent long-term case-control studies. UDCA has been demonstrated to exert anticholestatic effects in various cholestatic disorders. Several potential mechanisms and sites of action of UDCA have been unraveled in clinical and experimental studies which might explain its beneficial effects. Their relative contribution to the anticholestatic action of UDCA might depend on the type of the cholestatic injury. In early-stage PBC, protection of injured cholangiocytes against the toxic effects of bile acids might prevail, and stimulation of impaired hepatocellular secretion by mainly posttranscriptional mechanisms including stimulation of synthesis, targeting and apical membrane insertion of key transporters might be relevant in more advanced cholestasis [12]. In addition, stimulation of ductular alkaline choleresis and inhibition of bile acid-induced hepatocyte and cholangiocyte apoptosis can have a certain role for the beneficial effect of UDCA in PBC [12].

UDCA has been demonstrated to markedly decrease serum bilirubin, AP, γGT, cholesterol and immunoglobulin M levels, and to ameliorate histological features in patients with PBC in comparison to placebo treatment [[13], [14], [15], [16], [17]] although no significant effects on fatigue or pruritus were observed in these large trials. Moreover, long-term treatment with UDCA delayed the histological progression of the disease in patients in whom treatment was started at an early stage [[17], [18]]. Still, a clear-cut beneficial effect of UDCA on survival has not been shown in any of the studies mentioned above, probably due to the limited number of patients and the limited observation periods too short for a slowly progressing disease. A beneficial effect of UDCA on survival has only been demonstrated in a combined analysis of the raw data from the French, Canadian and Mayo cohorts followed up for 4 years [19]. In this analysis, UDCA treatment was associated with a significant reduction in the likelihood of liver transplantation or death. This benefit was seen in patients with moderate and severe disease but not in those with mild disease (serum bilirubin concentration <1.4 mg/dL (24 μmol/L), stage I or II histologic change) in whom progression to end-stage disease did not occur during the 4-year period of observation [19].

The affirmative results on survival have been challenged by meta-analyses which included a majority of studies of up to two years' duration and trials using UDCA doses which are today known to be ineffective [[20], [21]]. Inclusion of trials which have a duration of three months to two years for a disease with an estimated duration of one to two decades without intervention may be suited to analyze biochemical effects of medical treatment, but certainly carries a risk to dilute the information needed for a well-based survival analysis. Therefore, it was not surprising that meta-analyses which excluded studies of short duration (less than 24 months) and those that used an ineffective dose of UDCA (less than 10 mg/kg/d) concluded that long-term UDCA significantly improved transplant-free survival and delayed histologic progression in early-stage patients [[22], [23]].

Recent reports have demonstrated the favorable effects of UDCA on long-term survival in patients with PBC receiving standard doses (13–15 mg/kg/d) [24] over 10–20 years. Treatment with UDCA led to a transplant-free survival similar to that of a healthy control population matched for age and gender in patients with early-stage disease [[25], [26]] and to improved survival in comparison to the estimated survival at the start of treatment as calculated by the Mayo risk score for PBC [[25], [26], [27]]. Interestingly, a “good biochemical response” to UDCA defined as a decrease in AP >40% of pretreatment levels or normalization at one year (“Barcelona criteria”) was associated with an excellent 95% transplant-free survival at 14 years of follow-up, similar to that predicted for the standardized population [27]. The prognostic impact of the “Barcelona criteria” was confirmed in a large independent cohort of PBC patients for which a serum bilirubin ⩽1 mg/dL (17 μmol/L), AP ⩽3× ULN, and AST ⩽2× ULN (“Paris criteria”) after one year of treatment even better identified those with a good long-term prognosis of a 90% (vs. 51%) ten year transplant-free survival [28]. Thus, additional therapeutic options for those patients failing to reach a “good biochemical response” under UDCA are warranted.

Corticosteroids and other immunosuppressive agents

Prednisolone improved serum liver tests and histological features, but markedly worsened bone mineral density in patients with PBC [29] prohibiting its long-term use in PBC. In combination with UDCA (10 mg/kg/d), prednisolone (10 mg/d, 9 months) exerted beneficial effects on various features of liver histology in early-stage PBC in comparison to UDCA alone [30].

Budesonide in combination with UDCA showed favorable results on biochemical and histological parameters in early-stage disease [[31], [32]], but not late-stage disease [[33], [34]]. Studies with a longer follow-up using the combination of budesonide and UDCA in patients with early-stage disease not adequately responding to UDCA alone are warranted to confirm its safety and its effect on postponing or preventing the need for liver transplantation. Development of portal vein thrombosis probably related to short-term budesonide administration was reported in stage 4 patients with portal hypertension [34]. Thus, budesonide should not be administered to cirrhotic patients.

Other immunosuppressive agents like azathioprine [35], cyclosporine A [36], methotrexate [[37], [38], [39]], chlorambucil [40] and mycophenolate mofetil [41] proved to be marginally effective, ineffective or potentially harmful during long-term administration and cannot be recommended for standard treatment in PBC.

Anti-fibrotic agents

Colchicine was inferior to UDCA in the treatment of PBC [42] and did not, when combined with UDCA in comparison to UDCA alone [43], significantly improve symptoms, serum liver tests, serum markers of fibrosis, or histological features. Thus, addition of colchicine to UDCA currently cannot be recommended in the treatment of PBC.

d-Penicillamine is not effective in PBC and can be associated with severe side effects [[44], [45]].

Other drugs

Malotilate [46], thalidomide [47], silymarin [48] and atorvastatin [49] were not effective in the treatment of PBC. Sulindac [50] and the peroxisome proliferator-activated receptor α (PPARα) agonist, bezafibrate [51] improved some serum liver tests in limited groups of patients with an incomplete response to UDCA, and bezafibrate deserves further studies. Tamoxifen [52] decreased AP levels in two women who were taking it after surgery for breast cancer.

Antiretroviral strategies have also been tested in PBC: Lamivudine alone or in combination with zidovudine (Combivir) was associated with minor clinical and biochemical effects. Combivir was also associated with improvement of some histological features, but this finding needs confirmation in randomized studies [53].

Liver transplantation

Liver transplantation has dramatically improved survival in patients with late-stage PBC. Indications are not different from those of patients with other etiologies of liver failure [54]: decompensated cirrhosis with an unacceptable quality of life or anticipated death within a year due to treatment-resistant ascites and spontaneous bacterial peritonitis, recurrent variceal bleeding, encephalopathy or hepatocellular carcinoma. Severe, treatment-resistant pruritus may merit consideration for transplantation. Patients should be referred to a liver transplant center for assessment when their bilirubin approaches 6 mg/dL (103 μmol/L), the Mayo risk score is ⩾7,8, and the MELD score is >12 at the latest.

Survival rates above 90% and 80–85% at one and five years, respectively, have been reported by many centers [55]. Most patients have no signs of liver disease after orthotopic liver transplantation, but their antimitochondrial antibody status does not change. The disease recurs with a calculated weighted disease recurrence of 18% [56], but rarely is associated with graft failure [54].


  1. Patients with PBC, including those with asymptomatic disease, should be treated with UDCA (13–15 mg/kg/d) (I/A1) on a long-term basis (II-2/B1).
  2. Favorable long-term effects of UDCA are observed in patients with early disease and in those with good biochemical response (II-2/B1), which should be assessed after one year. A good biochemical response after one year of UDCA treatment is currently defined by a serum bilirubin ⩽1 mg/dL (17 μmol/L), AP ⩽3× ULN and AST ⩽2× ULN (“Paris criteria”) or by a decrease of 40% or normalization of serum AP (“Barcelona criteria”) (II-2/B1).
  3. There is currently no consensus on how to treat patients with a suboptimal biochemical response to UDCA. One suggested approach is the combination of UDCA and budesonide (6–9 mg/d) in non-cirrhotic patients (stages 1–3) (III/C2). Further studies of this and other combination regimes should be a priority.
  4. Liver transplantation should be strongly considered in patients with advanced disease as reflected by serum bilirubin exceeding 6 mg/dL (103 μmol/L) or decompensated cirrhosis with an unacceptable quality of life or anticipated death within a year due to treatment-resistant ascites and spontaneous bacterial peritonitis, recurrent variceal bleeding, encephalopathy or hepatocellular carcinoma (II-2/A1).