EASL Clinical Practice Guidelines

Clinical spectrum

Clinical features of AIH

In the early 1950s, a novel type of chronic hepatitis with several particular features, such as a predilection for young women, a progressive and usually fatal outcome accompanied by arthralgia, endocrine dysfunction, cutaneous striae and acne, and very high levels of immunoglobulins in the serum that correlated with an excess of plasma cells in the liver, was reported firstly by the Swedish physician Jan Waldenström [22] and later by Kunkel et al. [23]. In 1955, the lupus erythematosus cell phenomenon was demonstrated in these patients and therefore, the term “lupoid hepatitis” was introduced by the group of Ian Mackay in 1956 [24], but ten years later this term was replaced by ‘Autoimmune hepatitis’ [25], which after a variety of different terms was accepted in the 1990s by the International AIH Group (IAIHG) as the final one [26].

It is now well established that AIH is a clinically distinct syndrome characterized by a large heterogeneity of clinical, laboratory and histological manifestations (Table 2). Therefore, AIH should be considered in any patient with acute or chronic liver disease, particularly if hypergammaglobulinemia is present, and if the patient has features of other autoimmune diseases (Table 3) [[2], [3], [4], [26], [27], [28]]. The disease can also affect males (ca. 25–30% of all AIH patients) and may present at any age and in all ethnic groups [[8], [9], [10], [11], [12], [13], [29], [30], [31], [32], [33]]. In most studies, a bimodal age pattern at presentation has been reported with one peak during childhood/teenage years and another in middle age between the 4th and 6th decade of life [[8], [11], [13], [33], [34]]. Recent studies have shown that an increasing number of patients are diagnosed also at older ages (above 65 years) [[30], [31], [32], [35]]. Recently it has been shown that appropriate attention should also be paid to the health related quality of life (HRQoL) parameters, since a high rate of previously unrecognised mental impairment with depression and anxiety symptoms are present in patients with AIH [36].

Table 2
Clinical spectrum of autoimmune hepatitis.

AIH, autoimmune hepatitis; HCC, hepatocellular carcinoma; HLA, human leukocyte antigens; ANA, antinuclear antibodies; SMA, smooth muscle antibodies; anti-SLA/LP, soluble liver antigen/liver pancreas antibodies; anti-LKM1, liver/kidney microsomal antibody type 1; anti-LKM3, liver/kidney microsomal antibody type 3; anti-LC1, antibodies against liver cytosol type 1 antigen.

Table 3
Differential diagnosis of autoimmune hepatitis.

The spectrum of clinical manifestations is variable, ranging from no obvious signs or symptoms of liver disease to a severe and almost identical form of an acute or even fulminant episode of viral hepatitis (Table 2) [[3], [4], [37]]. Indeed, approximately 25% of patients present with an acute onset of AIH, which is phenotypically similar to acute hepatitis cases of other causes [[33], [38]]. However, acute presentation of AIH actually may contain two different clinical entities. One is the acute exacerbation of chronic AIH (acute exacerbation form of undiagnosed or misdiagnosed AIH cases) and the other is the genuine acute AIH without chronic histological changes (acute form of AIH; Table 2) [[33], [37], [38], [39]]. Of note, in some patients with acute presentation of AIH, immunoglobulin G (IgG) levels may be within the normal range and antinuclear (ANA) and/or smooth muscle antibodies (SMA) as first screening may be negative and thus, the clinician may not consider AIH [[3], [4], [34], [40], [41]]. A more extensive and sensitive autoimmune liver serology testing could be helpful in such cases. Furthermore, in some patients autoantibodies may only become positive some months later in the disease course. Some of these acute cases of AIH may rarely progress to acute liver failure and this should be kept in mind. The identification of AIH as the aetiology of acute hepatitis and/or fulminant hepatic failure is very important because a delay of the diagnosis and thus delay of initiation of therapy results in poorer prognosis of AIH, whereas administration of immunosuppression with steroids might avoid the need for liver transplantation (LT) [[33], [37], [38], [39], [41], [42], [43]].

Commonly (about one third of patients), the clinical presentation is characterized by the presence of one or more of several non-specific symptoms listed in Table 2 [[8], [11], [13], [18], [21], [29], [33], [44], [45]]. Amenorrhea is also common whereas maculopapular skin rash and unexplained low-grade fever are rare features. Physical findings may be normal, but sometimes hepatomegaly, occasionally painful, splenomegaly and, when frank cirrhosis has developed, signs and symptoms of chronic liver disease like palmar erythema and spider naevi may be found. In advanced stages, the clinical picture of portal hypertension dominates including ascites, oesophageal varices and portal hypertensive gastropathy, cytopenias due to hypersplenism as well as hepatic encephalopathy.

Up to a third of patients have an insidious onset and gradual progression without apparent symptoms at diagnosis (asymptomatic) and the final diagnosis is usually established during investigation for unexplained elevation of serum aminotransferases on routine testing or testing performed for other reasons [[8], [11], [13], [29], [31], [32], [44], [45]]. However, approximately one third of adult patients and about half of children at diagnosis have already developed advanced disease with the presence of cirrhosis, which in most studies is associated with lower overall survival irrespective of the presence of symptoms or not [[8], [13], [29], [44], [45], [46], [47]]. The latter finding along with the presence of histological evidence of chronic disease on liver biopsy in a proportion of patients with acute presentation imply that they probably have had subclinical disease for a long time [[37], [38], [42]]. Actually, this is one important diagnostic challenge, because subclinical disease often precedes the onset of the disease symptoms, whereas long periods of subclinical disease may also occur after presentation.

According to the pattern of autoantibodies detected, a sub-classification of the disease into two or three subtypes has been proposed. Initially, two major types, AIH-1 and AIH-2, have been proposed (Table 2). AIH-1 is characterized by the presence of ANA and/or SMA [[3], [4], [27], [28], [34], [40]]. AIH-2 is characterized by the detection of specific anti-liver/kidney microsomal antibody type 1 (anti-LKM1) or infrequently anti-LKM type 3 (anti-LKM3) and/or antibodies against liver cytosol type 1 antigen (anti-LC1) [[3], [4], [27], [28], [34], [40]]. This distinction was initially based on circulating autoantibodies alone but thereafter other differences have been reported (Table 2). Similarly, the discovery of antibodies against soluble liver antigens (anti-SLA), later found to be identical with previously described antibodies against liver pancreas (anti-LP) and therefore called anti-SLA/LP antibodies, lead to the definition of a third subtype, AIH-3 (Table 2) [48]. Differences between AIH-1 and AIH-3 seemed less pronounced than between AIH-1 and AIH-2, but some authors postulated more severe disease and the need for lifelong immunosuppression in most if not all AIH-3 patients [[48], [49], [50]]. The validity of these sub-classifications, however, is questionable and subject of an ongoing debate [3].

Specific clinical features and presentations of AIH

Variant forms of AIH and cholestatic liver disease

Some patients within the spectrum of AIH present either simultaneously or consecutively, with clinical, biochemical, serological, and/or histological characteristics of PBC or PSC [51]. Vice versa, some patients with a diagnosis of PBC or PSC show or develop features of AIH. So far, several terms have been used to describe these phenomena, in particular “overlap syndromes”, but also “the hepatitic forms of PBC”, “secondary autoimmune hepatitis”, “autoimmune cholangitis”, “autoimmune sclerosing cholangitis” or “combined hepatitic/cholestatic syndromes” to describe patients with features of both AIH and PBC or PSC [[51], [52], [53], [54]]. A descriptive terminology of these variant forms (e.g. PBC with features of AIH) is probably the most appropriate terminology in the absence of a clear pathogenetic understanding of these variants.

Internationally agreed criteria defining these variant conditions are lacking, and therefore the characteristics of these entities vary between studies making it difficult to give standardised recommendations. Recently, on behalf of the IAIHG, an international working party critically reviewed “overlap syndromes” and found a low sensitivity of the scoring systems for AIH diagnosis (either revised or simplified) in clinically defined “overlap” patients [51], which is in keeping with results of previous studies (Table 4) [55]. As a consequence, use of these AIH scoring systems is not generally recommended for the distinction of these particular patients. Interface hepatitis is a fundamental component of hepatitis and histology is therefore vital in evaluating patients with overlap presentation. The degree of interface hepatitis can be considered a measure of AIH-like disease activity irrespective of co-existence or underlying cholestatic liver disease [51].

Table 4
Specific characteristics and features of autoimmune hepatitis.

AIH, autoimmune hepatitis; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; AMA, antimitochondrial antibodies; IAIHG, International AIH Group; DILI, drug-induced liver injury; TNF, tumour necrosis factor; HCV, hepatitis C virus; APECED, autoimmune polyendocrinopathy candidiasis ectodermal dystrophy; APS-1, autoimmune polyglandular syndrome type 1.

The pathogenesis of these “variant forms” is debated and it remains unclear whether this syndrome forms a distinct entity or a variant of PBC, PSC or AIH. It has been suggested that features of AIH develop in patients with immune-mediated cholestatic liver disease and a genetic susceptibility for AIH as shown by the high prevalence of the AIH-susceptibility HLA-genes DR3 or DR4 in PBC patients with features of AIH, leading to the term “secondary AIH” in patients with PBC and overlapping features of AIH [56]. In this regard, the name “overlap” that strongly suggests the presence of two distinct diseases could be a misnomer. It should be kept in mind that “variant forms” of AIH should not be over-diagnosed in order not to expose PBC or PSC patients unnecessarily to the risk of steroid side effects.

Features of both AIH and PBC

With no codified diagnostic approach, reported prevalence figures are variable, but it is generally assumed that the prevalence of AIH-PBC variant is approximately 8–10% of adult patients with either PBC or AIH [[57], [58]]. The “Paris criteria” are currently the most commonly used tool for diagnosing AIH-PBC variant form as attested by the presence of at least two of the three accepted key criteria of each disease namely, for PBC: 1) alkaline phosphatase (ALP) ⩾2× upper normal limit (ULN) or γ-glutamyl-transpeptidase (γ-GT) ⩾5 × ULN; 2) presence of antimitochondrial antibodies (AMA); 3) a liver biopsy specimen showing florid bile duct lesions; and for AIH: 1) alanine aminotransferase (ALT) ⩾5 × ULN; 2) serum IgG levels ⩾2 × ULN or presence of SMA, 3) a liver biopsy showing moderate or severe periportal or periseptal lymphocytic piecemeal necrosis [57]. Indeed, a recent study [59], has shown that the criteria strictly defined previously by Chazouilleres et al. [57] could identify patients with a clinical diagnosis of AIH-PBC “variant” with high sensitivity (92%) and specificity (97%). In addition, the 2009 EASL guidelines on the management of cholestatic liver diseases endorsed these diagnostic criteria, but specified that histologic evidence of moderate to severe lymphocytic piecemeal necrosis (interface hepatitis) was mandatory. It was stated that these “variants” should always be considered once PBC has been diagnosed and the patient responds poorly to ursodeoxycholic acid (UDCA) because of potential therapeutic implications (i.e. the need of immunosuppression) [60]. Simultaneous presence of features of both diseases is the usual presentation, but it should be noted that occasionally the onset of AIH and PBC is temporally dissociated, usually with PBC presenting first. Interestingly, in most cases, it is possible to define one primary disorder (“dominant” disease), usually PBC [51].

Features of both AIH and PSC

The co-existence of features of AIH and features of PSC variant has been described both in children and adults and is assumed to exist in a considerable part of mainly young patients with autoimmune liver disease [[51], [52], [53], [54], [61]]. Unfortunately, diagnostic criteria for these conditions are even less well-defined than in AIH-PBC variant cases. As a result, reported prevalence figures vary greatly but an approximate prevalence of 7–14% is generally assumed [51]. The diagnosis of large duct PSC should always be established on typical cholangiographic findings, keeping in mind that an intrahepatic biliary tree which simulates a sclerosing pattern can also be observed in any liver disease with extensive fibrosis and nodular regeneration or in cirrhosis [62]. In addition, magnetic resonance cholangiopancreatography (MRCP) may lead to false positive diagnosis due to its limited specificity. Some cases of small duct PSC (normal cholangiogram)-AIH variant forms have also been reported, but it can be argued that approximately 10% of patients with typical AIH, with or without ulcerative colitis, may have histological features of bile duct injury, thus making this diagnosis particularly uncertain [63]. In clinical practice, the diagnosis of AIH-PSC “variants” is made in a patient with overt cholangiographic or histological features of PSC, alongside robust biochemical, serological and histological features of AIH. It appears that patients with features of AIH and PSC also require immunosuppression [[64], [65]].

It should be noted that in children with AIH a specific entity has been described in almost half of patients characterized by lesions of both AIH and sclerosing cholangitis. Thus, the term “autoimmune sclerosing cholangitis (AISC)” was introduced by Mieli-Vergani’s group [52] suggesting also the need of an investigation of the biliary tree at least with MRCP in all children with a diagnosis of AIH (Table 4) [[52], [54]]. At present, this variant seems unique for children, as a prospective study in adults with AIH was negative and thus, in the absence of cholestatic indices, MRCP screening does not seem justified in adult-onset AIH [62]. However, particularly in young adults with AIH and cholestatic features, and in AIH patients with remaining cholestasis despite adequate immunosuppression, MRCP for the detection of possible underlying or co-existent PSC is recommended.

IgG4-related AIH

In the emerging era of IgG4-related diseases, the role of IgG4 response has been investigated in AIH patients [[66], [67]]. Typically IgG4 disease in the liver manifests as a differential diagnosis of PSC with features of cholangiopathy and jaundice. Despite anecdotal reports from Japan, confirmation is lacking. Therefore it is difficult to judge, if an entity of AIH-like IgG4 disease exists and presents a separate disease entity.

In summary, based on the current, very limited knowledge about the aetiopathogenesis of AIH, PBC, and PSC, definition of diagnostic criteria for these “variant forms” of AIH are very difficult to be established and can only be arbitrary. Consequently, patients with autoimmune liver diseases should rather be categorized according to the primary clinical and histological manifestation of the disease as AIH, PBC, or PSC, and additional features of the respective other immune-mediated liver disease should be listed as such (e.g. PBC with features of AIH). In addition, the low prevalence of these variants has made it impractical to perform randomised controlled trials for their management. However, as these variant conditions do occur quite frequently, specific therapeutic considerations may be required in patients with PBC or PSC with features of AIH [68]. In general, features of AIH should be managed like AIH, as untreated AIH has a poor prognosis, but response to therapy is generally very good.

DILI and AIH

The relationship between DILI and AIH is complex and not fully understood. In principle, three scenarios are possible [[69], [70]]:

  1. DILI with a strong immunoallergic component mimicking AIH
  2. AIH mimicking as DILI due to drug exposure in recent weeks and spontaneous improvement after cessation of drug exposure
  3. AIH triggered by an offending drug (DILI-induced AIH)

It appears that all three scenarios occur. As both immunoallergic DILI and AIH are presumably mediated by specific immune reactions to antigens in hepatocytes, clinical and histological overlap between these conditions is not surprising. Nonetheless, the differential diagnosis between these conditions and the implications for the pathogenesis of AIH are both important.

Drug-induced AIH has been particularly well described for drugs no longer in use such as tienilic acid and dihydralazine [[71], [72]]. Reactive metabolites created through hepatic metabolism of drugs have been shown to bind to cellular proteins such as components of CYP450, i.e. CYP2C9 in the case of tienilic acid and CYP1A2 in the case of dihydralazine. These can then be recognized by the immune system as neoantigens [[71], [72]]. Among drugs still widely used, drug-induced AIH has been well documented for nitrofurantoin and minocycline [73]. When comparing patients with drug-induced AIH to those with genuine AIH, the two groups were found to have quite similar clinical and histological patterns, although the former has lower histological activity and does not seem to require long-term immunosuppression [74].

Histologically distinguishing DILI from AIH remains a challenge. A recent study has suggested that sufficient differences exist so that pathologists can use the pattern of injury to suggest a correct diagnosis in many cases [75]. Nevertheless, the differentiation is often very difficult, because DILI lacks a reliable diagnostic test and, like AIH, the diagnosis is mainly based on clinical and serological grounds [76]. Although the frequency of drug-induced AIH-like syndrome is difficult to be assessed, it can account for approximately 9–12% of cases with classical features of AIH [[74], [77]]. An important element in the identification of this syndrome is the patient’s history that should focus on recent exposure to drugs that can induce AIH-DILI [74]. In 30% of cases it can be associated with features of hypersensitivity such as fever, rash and eosinophilia [78]. The absence of cirrhosis at presentation can also be an element in favour of AIH-DILI [78]. Severe AIH-DILI usually responds to high doses of steroids in the same way as severe AIH, if treatment is started without delay. Sometimes only the follow-up can differentiate between AIH and DILI: steroid treatment can be discontinued without relapse in DILI, whereas in genuine AIH relapse will occur universally, if immunosuppression is stopped within a few months. A trial of steroid treatment and close observation upon steroid tapering and possible withdrawal is therefore recommended for uncertain cases (see treatment algorithm Fig. 1).


Fig. 1
Suggested diagnostic algorithm for autoimmune hepatitis using routine autoantibody testing by indirect immunofluorescence (IFL) and enzyme linked immunosorbent assay (ELISA) testing with a set of four autoantibodies. A liver biopsy is always required for the demonstration of inflammatory hepatitis, as well as for staging and grading of the liver disease. ∗Long-term follow-up is advised in order not to miss a late relapse of AIH (e.g. 6 monthly for 3 years).

View Large Image | View Hi-Res Image | Download PowerPoint Slide
AIH and pregnancy

The disease is very rarely diagnosed during pregnancy, but, like other autoimmune diseases, may notably manifest in the post-partum period. In patients with known AIH, improvement or even spontaneous remissions during pregnancy can be observed, while flares after delivery are frequently observed [[79], [80], [81], [82], [83], [84]]. This is presumably due to immune reconstitution following delivery. Therefore, the possibility of AIH should be strongly considered in the differential diagnosis of liver dysfunction, particularly accompanied by hypergammaglobulinemia with selective IgG elevation, in the post-partum period, but even during pregnancy, as flares can also occur anytime during pregnancy. Effective immunosuppression has enabled the occurrence of pregnancy in young females with AIH presenting initially with amenorrhea, and immunosuppression should almost always be continued during pregnancy with generally good pregnancy outcome.

Viral hepatitis and AIH

It has been suggested, that in susceptible individuals, AIH may be induced by viral infections, and a number of possible cases have been reported [[3], [4], [85], [86]]. Molecular mimicry between viral epitopes and epitopes of autoantigens have supported the concept of virally induced AIH. On the other hand, the few cases reported might also represent a diagnostic bias in two forms: firstly, patients with subclinical AIH previously overlooked may become diagnosed when suffering from an acute incidental viral hepatitis; secondly, patients with acute AIH and marked hypergammaglobulinaemia might display false positive results on serology for viral markers. On the other hand, the development of AIH, or of features of AIH, has also been reported in some patients with hepatitis C virus (HCV) after treatment with interferon-alpha [[87], [88]] and rarely in acute HCV infection even after viral clearance [89]. The differentiation between AIH and chronic HCV was a challenge in the past, particularly because of the immunostimulatory side effects of interferon-alpha, but due to the advent of interferon-free treatment regimens, this represents no longer a difficult clinical problem: HCV infection should be treated primarily, and if inflammatory liver disease persists, the diagnosis of AIH should be considered.

De novo AIH in liver transplant recipients

AIH, or an AIH-like syndrome, can develop after LT undertaken for other liver diseases, both in adults and children. This situation has been called “de novo AIH” [[90], [91]], although it has been suggested that alternative nomenclature such as “post-transplant immune hepatitis” or “graft dysfunction mimicking AIH” or “post-transplant plasma cell hepatitis” may be more appropriate as the transplanted hepatocytes are not strictly “self” and thus the conditions not strictly “autoimmune” [[5], [92]]. Nevertheless, the timely recognition of this entity appears to be helpful for avoiding graft rejection, and the need for another LT and for improving long-term survival, as these patients benefit from increased immunosuppression including steroids and azathioprine like in genuine AIH [90].

Associated autoimmune conditions

AIH is associated with the presence of a wide variety of other autoimmune or immune-mediated diseases (Table 4) [[8], [13], [29], [45], [93], [94], [95], [96]]. Actually, concurrent autoimmune diseases are common in patients with AIH and mirror the full range of known autoimmune diseases. Therefore, an extended diagnostic screening for other autoimmune diseases, especially autoimmune thyroiditis, seems reasonable in patients with AIH, both at diagnosis and at regular intervals during follow-up [95]. In addition to the patient being affected by immune-mediated diseases, their occurrence is also more frequent in first-degree relatives of AIH patients, and therefore a careful family history should be undertaken. A careful personal and family history may also help in identifying rare variants of AIH due to autosomal recessive genetic aberrations such as the autoimmune polyendocrinopathy-cadidiasis ectodermal dystrophy syndrome (APECED) also known as autoimmune polyendocrinopathy syndrome type 1 (APS-1) which is caused by mutations in the autoimmune regulator gene (AIRE) and characterized by chronic mucocutaneous candidiasis, ectodermal dystrophy and autoimmune destruction of several endocrine organs, leading mainly to hypoparathyroidism, adrenocortical failure and gonadal failure in females (Table 4) [[8], [13], [29], [45], [93], [94], [95], [96]].

Complications of AIH

In principle the complications of AIH are the same as in any other acute or chronic progressive liver disease. In acute presentations the risk of liver failure and infectious complications are predominant and may be aggravated by immunosuppressive treatment. In chronic disease, especially in patients undiagnosed or insufficiently treated, complications of cirrhosis occur. In particular, hepatocellular carcinoma (HCC) is a known consequence of AIH-related cirrhosis although its occurrence in association with AIH is significantly less frequent compared to most other causes of liver cirrhosis (Table 2) [[3], [11], [13], [97], [98]]. A recent population-based study showed that the risk of hepatic and extra-hepatic malignancy was significantly increased in AIH patients [99]. Studies from Denmark, Germany, Netherlands, UK, USA and Japan identified male gender as a particular risk factor, and the presence of cirrhosis was a universal prerequisite for HCC development, which was observed in the at risk cirrhotic population at a rate of 1–2% per year [[11], [13], [97], [98], [100], [101], [102]]. Surveillance recommendations have not been validated in AIH and cirrhosis, but as the HCC risk appears to be significant, liver ultrasonography every six months in patients with cirrhosis appears reasonable.

In addition to complications of the liver disease, complications of long-term immunosuppression need to be considered, and the two risks may associate. Of note, extra-hepatic malignancies of diverse cell types occur in 5% of patients in an unpredictable fashion with non-melanoma skin cancers being the most common [[99], [103]]. It is likely that this risk is primarily due to the long-term immunosuppression required in most patients. To what extent the risk for extra-hepatic malignancy is different from the normal population is poorly studied. Nonetheless, it appears sensible to apply routine health screening measures for other malignancies in AIH patients.