The term ‘autoimmune hepatitis’ (AIH) refers to a hepatitic liver disorder not directly due to viral infection, inherited or acquired metabolic abnormalities or drug-associated liver injury, and whose laboratory features (circulating autoantibodies and hypergammaglobulinaemia) suggest an immune aetiology. Extrahepatic syndromes attributed to autoimmunity may be present; response to immunosuppressive treatment is usually dramatic. The target of this immune process is the hepatocyte. Autoimmune hepatitis usually causes severe liver disease in children. Especially in children, but also in adults, AIH can present as either acute or chronic disease. The preferred term for this disorder is therefore simply ‘autoimmune hepatitis’ and other terms like ‘lupoid hepatitis’ and ‘autoimmune chronic active hepatitis’ are obsolete and no longer used. Liver disease resembling AIH clinically may occur with chronic hepatitis B or with chronic hepatitis C: the early clinical reports are difficult to interpret because the patient cohorts included patients with undiagnosed chronic viral hepatitis.
Autoimmune hepatitis (AIH) is an important entity within the broad spectrum of autoimmune hepatobiliary disease comprised of AIH, primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC). Since the 1960s, AIH has been investigated with extensive clinical research aimed at effective therapeutic intervention. It was one of the first liver diseases where treatment was demonstrated to prolong survival. AIH occurs in children, as well as in adults. Its clinical manifestations in children may differ from classic adult AIH. These differences have elucidated certain aspects of AIH and hepatobiliary disease in general. There are two major patterns of AIH: type 1, with anti-smooth muscle antibodies and type 2, with anti-liver/kidney microsomal antibodies. The second type of AIH was first identified in children and is more common in younger patients. AIH often presents as acute disease in children and also in adults: the nomenclature has dropped the allusion to chronicity. Some children who have sclerosing cholangitis present with clinical disease closely resembling AIH; this AIH-like PSC, termed autoimmune sclerosing cholangitis (ASC), is also found in adults. Children with AIH may have identifiable monogenic disorders of immune regulation such as autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED). Like adults with AIH, children with AIH usually respond very favourably to immunosuppressive treatment with corticosteroids ± azathioprine. True cures seem to be rare, although many children achieve a stable remission. Nonetheless children with AIH may develop cirrhosis and some require liver transplantation. Early diagnosis and improved treatment strategies may further improve the outlook for children with AIH.
Autoimmune hepatitis in the context of hepatobiliary autoimmune disease
Indeed, although AIH was one of the first liver diseases in which extensive clinical research established effective treatment, it continues to pose conceptual and practical problems. The evolving nomenclature, mentioned previously, has been part of the problem. Difficulties distinguishing bona fide AIH from other viral or metabolic liver diseases looking clinically like AIH, such as with chronic hepatitis B or Wilson disease, have been another source of confusion. Regaining an awareness that AIH is often not a chronic disease has cleared up a further source of confusion. A certain proportion of children, and some adults, who actually have primary sclerosing cholangitis (PSC), present clinically as if they had AIH. This key observation draws attention to the need for a broader perspective from which paediatric AIH must be viewed. This conceptual framework (Fig. 1A) can be depicted as a triangle with three ‘anchoring’ hepatobiliary diseases whose mechanism involves immune mechanisms: AIH (target cell: hepatocyte); primary biliary cirrhosis (PBC; target cell: epithelial cell of portal bile ducts) and PSC (principal target: epithelial cell of medium-to-large bile ducts). As shown by the bidirectional arrows, there are relationships between these diseases. Figure 1B shows the first layer of complexity in that the clinical phenotype of all three diseases can be found as a result of drug-induced liver injury. Drug-induced forms of PBC and PSC are rare, whereas drug-induced AIH is rather common. Most of the drugs associated with drug-induced AIH are not used in children. In the paediatric age-bracket, minocycline is the medication frequently associated with triggering AIH [1-3] and some would call the resulting disorder a drug-induced AIH to differentiate it from apparently spontaneous AIH. The situation becomes complicated when ‘overlap’ syndromes are added to the picture (Fig. 1C). Overlap between PBC and AIH was identified in 1970s, well before the overlap between AIH and PSC was identified in children [4, 5]. Moreover, there is a subtle overlap between PBC and PSC because in the late stages of both diseases intrahepatic duct paucity may be the dominant histological feature. Thus, the highly prevalent ‘overlap’ nomenclature is ambiguous. The paediatric terminology for AIH/PSC overlap, namely ‘autoimmune sclerosing cholangitis’ (ASC), risks confusion with the entity ‘autoimmune cholangitis’, a variant of PBC, but ASC captures the key features, namely, an AIH-component plus large-duct cholangitis. Over time, each of these overlap syndromes plays out like its biliary component, as shown by dominant unidirectional arrows in Figure 1C.
Age and ethnic background reflecting important genetic variations can also influence the clinical features of AIH. Very young children seem to be more likely to have AIH with positive anti-liver/kidney microsomal antibodies and it was initially thought that ASC was an exclusively paediatric disorder. Children may have AIH associated with rare disorders of immune function (Table 1), some of which are genetic, including autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED), periodic fever-aphthous stomatitis-pharyngitis-cervical adenitis (PFAPA)  and 22q13 deletion syndrome . Ethnic background may influence immune reactivity more broadly by determining an individual's immunogenetic (and pharmacogenetic) complexion. Consequently, there can be subtle differences worldwide in the clinical features of AIH and response to therapy.
|Hepatic manifestation of autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED) (autosomal recessive, gene on 21q22.3)|
|Periodic fever-aphthous stomatitis-pharyngitis-cervical adenitis (PFAPA)|
|22q13 deletion syndrome|
|Infantile autoimmune haemolytic anaemia with giant cell hepatitis (AIHA-GCH)|
|Hepatic involvement in systemic lupus erythematosus|
|With syncytial giant cell transformation|
Several different types of AIH can be defined on the basis of which autoantibodies are present. Since healthy children do not ordinarily have circulating autoantibodies, even relatively low titres of these autoantibodies (1:20, or even 1:10 for anti-LKM) may be clinically important. Type 1 AIH (AIH1) is the most familiar form of this disease: anti-smooth muscle antibodies are present and antinuclear antibodies are also usually found. Anti-smooth muscle antibodies are directed against actin, found in the cytoskeleton of hepatocytes. Type 2 AIH (AIH2) is less common. It definitely occurs in North America and accounted for 21% of patients in one extensive North American paediatric series (E. A. Roberts, unpublished observations). In AIH2 the anti-smooth muscle antibodies are negative, but anti-liver/kidney microsomal antibodies (anti-LKM1) are typically present. These antibodies are directed against the apoprotein of a particular form of cytochrome P450, namely, that of P450 2D6 (CYP2D6), a polymorphic cytochrome P450 associated with hydroxylation of debrisoquine . Some AIH2 patients express the anti-LC1 antibody directed at formiminotransferase cyclodeaminase (FTCD) ; expression is variable, but it may be the only detectable antibody in some childhood cases of AIH.
Where immunological diagnosis and disease classification become difficult is with patients who have AIH or liver disease resembling AIH but do not express these more typical autoantibodies. In some patients various P450s may be the target of expressed autoantibodies. In AIH found in children with APECED, due to an autosomal recessive mutation in the gene for the AIRE protein, the anti-LKM antibody is directed against CYP1A2 and/or CYP2A6 and/or CYP2D6 . Anti-LKM antibodies are found with various liver diseases resembling AIH. Some patients with chronic hepatitis C have anti-LKM antibodies; these patients also express the autoantibody known as anti-GOR, which reflects hepatitis C virus-specific autoimmunity . An anti-LKM antibody directed against CYP2C9 is associated with ticrynafen-induced hepatitis. A few patients with AIH2 have an anti-LKM, known as anti-LKM3, directed against a Phase 2 drug-metabolizing enzyme, namely a UDP-glucuronosyltransferase (in the UGT1A family); however, some of these patients have chronic hepatitis D with features resembling AIH. Some patients with classic AIH have anti-soluble liver antigen (SLA) antibodies, usually – although not always – in conjunction with more conventional autoantibodies. In children, anti-SLA has been detected in both AIH1 and AIH2 . The anti-SLA antibody is equivalent to the anti-liver/pancreas antibody (anti-LP); the target, SLA, is a tRNA complex, recently renamed SEPSECS . Anti-SLA antibodies are highly specific for AIH and are associated with the HLA allele DRB1*03. Some patients have no detectable autoantibodies at all and are classified as having ‘cryptogenic’ AIH; in these patients looking for anti-SLA may be informative.
Some patients look as if they have AIH but they actually have sclerosing cholangitis: this disorder is now called ‘autoimmune sclerosing cholangitis’. These patients present a clinical picture very much like AIH, including elevated serum IgG and detectable autoantibodies, usually antinuclear and/or anti-smooth muscle antibodies. They have little clinical evidence of cholestatic liver disease such as jaundice or elevated serum alkaline phosphatase; nevertheless, cholangiography reveals typical changes of sclerosing cholangitis and liver biopsy may demonstrate features of bile duct damage classic for PSC, in addition to interface and lobular hepatitis. Children with ASC often have duct damage limited to the medium and large intrahepatic bile ducts . Lack of extrahepatic duct involvement may make diagnosis extremely difficult. Inflammatory bowel disease, if it occurs at all, may not be evident at the time of hepatic presentation in children .
Disease mechanisms in autoimmune hepatitis
The mechanisms of hepatocellular damage in AIH continue to be elucidated . The best models for this disease are multifactorial and encompass multiple immune processes. Pathogenetic complexity may account for some of the clinical diversity of AIH. It is currently thought that something in the environment perturbs the equilibrium of the immune system, probably through some sort of molecular mimicry, and sets up a T-cell-mediated attack on hepatocytes. Among the genetic predispositions to such a process, the major histocompatibility complex (MHC; in humans: human leucocyte antigen, HLA) is possibly most important, as it regulates antigen presentation causing immune activation and is highly polymorphic . Various DRB alleles (DRB1*0301, DRB1*0401, DRB1*0404, DRB1*0405) are associated with susceptibility to AIH1, and DRB1*0701 and DQB1*0201 appear to be associated with susceptibility to AIH2 . Other alleles may also be important. Of particular interest is the HLA 1301 allele which relates to susceptibility to hepatitis A virus (HAV)-infection . Hepatitis A is common in South America and elsewhere: thus it is possible that anti-HAV immunization may decrease incidence of AIH in children in HAV-endemic areas. Immunogenetic variability relating to production of other proteins which regulate immune function may also be aetiologically important. Overall there is insufficiency of cellular mechanisms which suppress immune reactivity. Regulatory T cells (T-reg) and natural killer (NKT) cells are quantitatively and/or functionally deficient in patients with AIH. T-reg dysfunction, possibly the key feature of AIH , may be partly on a genetic basis. The common pathway is that CD4-positive and CD8-positive T lymphocytes recognize an antigenic component on the hepatocellular plasma membrane and initiate an ‘autoimmune’ attack. Cytokines and chemokines are generated and initiate or enhance other immune mechanisms. Of interest are Th17 T cells, which have a complex process of differentiation that may include activation by endogenous chemical or xenobiotics , and may function reciprocally with T-reg cells . They likely play an important role in AIH.
Clinical features of autoimmune hepatitis in children
Clinically, all types of AIH have similar features. Paediatric AIH can present at any age. In prepubertal children, the incidence appears evenly divided between boys and girls, but females tend to predominate overall (accounting for 75–80% of patients), as with adult patients. Classically, patients present with vague symptoms of malaise and lethargy, possibly joint pain or rash. Secondary amenorrhoea may occur in teen-aged girls. On physical examination, hepatosplenomegaly is usually found. Cutaneous stigmata of chronic liver disease may be present. Since AIH can progress very insidiously, it may be difficult to date the onset of what appears to be a chronic process. Occasionally a child presents with ascites. Jaundice may not be present, but the serum bilirubin is usually elevated. Serum aminotransferases are elevated. In addition to the presence of non-specific tissue autoantibodies, the total serum immunoglobulin G (IgG) is increased polyclonally, usually to at least 30 g/L (3.0 g/dl) or more. Serum complement C3 is normal, but complement C4 may be very low, in which case it likely reflects a congenital defect in complement C4 production [20, 21]. Coagulation studies are normal unless liver damage is extensive.
Approximately one-half of children with AIH present with an apparently acute disease. There is no indication of ongoing hepatic disease until the onset of malaise or jaundice, along with greatly elevated serum aminotransferases. Studies for acute viral infection in these patients are uniformly negative. Occasional patients appear to have a recent significant Epstein–Barr virus infection. Acute hepatitis A can trigger AIH , as can HHV-6 infection . Acute AIH should be suspected in any child with apparent acute hepatitis of no clear aetiology which does not show any sign of resolving by 4–6 weeks of clinical illness. The IgG level and autoantibody profile in these patients are not necessarily distinctive from acute hepatitis at the time of initial presentation. These patients should be followed closely and a timely liver biopsy should be performed to facilitate diagnosis. In some patients this apparently acute illness resolves spontaneously, but signs of liver disease will reappear later, usually with cirrhosis. Irrespective of their presentation (and often despite treatment), patients with AIH of any type tend to develop cirrhosis [24-26].
AIH2 is more common in children than in adults. In one series patients presented between 18 months- and 16 years-old . While sharing most clinical features of AIH1, it tends to be more severe. Presentation as an acute disease is common and rapid progression to cirrhosis may occur. Other diseases, considered to have an autoimmune aetiology, are frequently present. These include vitiligo, autoimmune thyroid disease, insulin-dependent diabetes mellitus, autoimmune anaemia or thrombocytopenia and rheumatoid arthritis  [Concurrence of any kind of AIH and coeliac disease has recently been reported [29, 30]]. In addition to the increased serum IgG, the serum concentration of IgA is typically lower than normal [26, 28].
Diagnosis may be expedited by using a set of standard criteria . Those proposed by the International Autoimmune Hepatitis Group were initially intended for research purposes to ensure uniformity of patient cohorts in research studies. A simplified version has been developed for clinical use , although it has not yet been extensively validated. The problem with these scoring systems is that they are geared to adults. A modification suitable for children has been proposed . The new simplified criteria may not be adequate for diagnosing AIH in children: specific validation is required. Particularly in children, Wilson disease and sclerosing cholangitis need to be excluded by comprehensive biochemical testing and imaging: a scoring system may not attribute enough importance to these diseases in the differential diagnosis. Recognizing different patterns of AIH typical of different ethnic groups is also important and not easily standardized in a scoring system.
Presentation as acute liver failure appears more common with AIH2 than with AIH1 . Even very young children with AIH2 can present with acute liver failure [34, 35]. Alternatively, children with AIH can present with decompensated chronic liver disease resembling acute liver failure, including severe jaundice and coagulopathy, especially if the diagnosis has been missed at an earlier stage. A difficult clinical problem is to make the diagnosis of AIH presenting as acute liver failure when all standard autoantibodies are negative . Experience in adults with AIH presenting as acute liver failure suggests that histological findings on liver biopsy may point to the diagnosis . Percutaneous liver biopsy is generally impossible in children with acute liver failure and transjugular liver biopsy, if available, may be hazardous in these paediatric patients. Thus reliance on histological findings is problematic; however, identifying serum positivity for autoantibodies which are not in routine clinical use may be informative in such patients.
Differential diagnosis of autoimmune hepatitis in children
The differential diagnosis of AIH in children and young adults is fairly broad and requires meticulous attention. As already mentioned, chronic viral hepatitis, principally hepatitis B and C, can produce a similar clinical picture. Appropriate serological tests for hepatitis B and C must be obtained. Wilson disease must be specifically excluded since treatment is entirely different . ASC needs to be identified [4, 5, 39]. Because such patients may account for a sizable proportion of children with autoimmune liver disease, cholangiography is indicated in all paediatric patients thought to have AIH. Drug-induced hepatitis associated with such drugs as α-methyldopa, oxyphenisatin, isoniazid, nitrofurantoin and sulpha antibiotics may resemble AIH. These drugs, except for the sulpha antibiotics, are uncommonly used in children. Although as yet unreported, drug hepatotoxicity from herbal or complementary medications also deserves consideration: black cohosh is a relevant herbal hepatotoxin. Alcoholic hepatitis occasionally shows autoimmune features, although alcoholic liver disease is exceedingly rare in children. Liver disease associated with α1-antitrypsin deficiency may resemble AIH in the paediatric age-bracket, but reports in adults suggest that AIH is merely coincident with partial deficiency such an the PI MZ phenotype . Thus is appears entirely possible that α1-antitrypsin deficiency causes childhood AIH only very rarely. Children with non-alcoholic fatty liver disease may elaborate low-titre autoantibodies: clinical differentiation is usually straightforward, although liver biopsy may sometimes be indicated to distinguish these two disorders.
Histological features on liver biopsy can aid in identifying various entities in this differential diagnosis, as well as in establishing the diagnosis of AIH according to classic histological features . An important finding is a portal inflammatory infiltrate consisting of lymphocytes and plasma cells which extends over the limiting plate and around adjacent periportal hepatocytes (Fig. 2). This is called ‘piece-meal necrosis’ or ‘interface hepatitis’. Spotty lobular inflammation also occurs. Severe inflammation in centrilobular areas is found in some patients, mainly adults. Necrosis of hepatocytes may lead to broad areas of parenchymal collapse. Damage to zone 1 of the hepatic acinus may interfere with normal regeneration. A variable degree of fibrosis is present. Diagnostic histopathological features of other specific liver diseases are absent.
Treatment strategies for children with autoimmune hepatitis
The mainstay of treatment for AIH is immunosuppressive treatment with corticosteroids and azathioprine. The most important consideration is not to delay treatment unnecessarily. At one time it was customary to define chronic disease in adults as being of 6 months’ duration. Waiting for a child to achieve this definition of chronicity will likely compromise the response to treatment for AIH . Even in children with the acute presentation, corticosteroids should be commenced as soon as possible after diagnosis, usually within 6–10 weeks of the onset of clinical disease. Tuberculosis or other tropical diseases, such as strongyloidiasis, which may worsen on prednisone treatment, must be excluded prior to starting treatment in patients with definable risk for such diseases. Recent reports of visceral leishmaniasis resembling AIH are noteworthy .
Outcome of treatment with corticosteroids and azathioprine has been favourable. Despite their evident inadequacies, the early studies in adults established the efficacy of prednisolone or prednisone in low maintenance doses  or combined with azathioprine  whereas azathioprine as sole initial treatment was ineffective . An extended study from the Royal Free Hospital  confirmed the efficacy of corticosteroids to diminish morbidity and prolong life in adults with AIH. Although true cures may rarely occur, late results suggest that many patients stabilize [47, 48]. Resolution of fibrosis or regression of cirrhosis may occur with successful treatment . Since not everyone responds to conventional therapy favourably, development of effective alternate treatment strategies remains a priority.
Treatment regimens in children with AIH vary from one report to another [26, 50-53]. Some begin with large daily doses of prednisone and taper the dose over the first 2–3 months as symptoms and laboratory findings improve; extended treatment on a low-ish maintenance dose continues for approximately 2 years, or longer if relapse occurs with cessation of treatment. Corticosteroids may be used alone initially; azathioprine is then added only if the response to steroids is incomplete. The alternative is to use a combination of prednisone and azathioprine from the beginning [26, 53]. Since many patients require both drugs, the latter regimen may have advantages, although the counterargument (to spare children unnecessary azathioprine and avoid risk of its adverse effects) has merit. Younger children generally need higher doses of prednisone (that is, 2 mg/kg/day initially) than older ones, probably because of more rapid drug metabolism in younger children. The efficacy of alternate-day dosage of corticosteroids in childhood AIH is disputed, but this dosing strategy can be used to reduce side effects of prednisone, especially growth retardation.
Those who respond favourably to corticosteroid treatment, adults and children alike, typically have a prompt and dramatic improvement in serum aminotransferases, often within 2 weeks of commencing treatment. The majority, approximately 80%, achieve remission within 18 months of starting on treatment. Those patients whose response to corticosteroid treatment is sluggish may have very severe disease or they may fail to take the medication because of disinclination or intolerable side effects. Importantly, they may not actually have AIH as such. It is also worth questioning whether normalization of serum aminotransferases really is an adequate criterion for judging clinical response: perhaps normalization of serum IgG, resolution of fatigue and (in post-menarchal girls/premenopausal women) resumption of menses are better endpoints. These criteria might mandate less rapid tapering of corticosteroid dosage than indicated simply by serum aminotransferase levels. In general, avoiding flares of hepatitis occasioned by too rapid tapering of corticosteroids appears to be important for minimizing parenchymal liver damage.
A possible difference between AIH1 in children and adults may be that children can sustain a disease remission, when drug treatment can be stopped. On balance, the reported remission rates are highly variable from one paediatric study to another. Moreover, the tendency to cirrhosis suggests that complete cures are unlikely. The decision to stop immunosuppressive treatment in a child with AIH must be made on an individualized basis. Based on observations in adults , those patients with a fairly rapid response to treatment and sustained normalization of serum aminotransferases, hepatic synthetic function and serum IgG level appear to be best candidates for stopping treatment. However, patients with AIH2 appear to need continuing treatment with very low-dose prednisone over an indefinite period of time to keep disease activity to a minimum. Prolonged remissions are attained in some children, mainly in those with AIH1 or where a clear instigating factor (usually a viral infection or drug exposure) was identified. Early institution of treatment seems to be associated with a better outcome. Children with AIH who relapse after an attempt to stop drug treatment require restarting treatment with corticosteroids and azathioprine. Attention to maintaining bone health, primarily through nutritional and life-style interventions, is important for any child with AIH being treated with corticosteroids over the long term.
Although azathioprine is ineffective as sole initial treatment for AIH, it may be effective as prolonged monotherapy in adults who achieve remission . Continuing treatment with azathioprine after stopping prednisone has been reported in children [53, 55]. The safety of prolonged use of azathioprine in children remains somewhat uncertain. However, it is possible that discontinuing corticosteroids can be achieved more successfully if azathioprine is continued at least until after the prednisone has been fully tapered.
More recently, cyclosporine has been tried in AIH. Its use was initially limited to patients who show no clinical response to prednisone and azathioprine . Its best use may be as a kind of induction therapy [57, 58]. Tacrolimus has been used successfully in some patients with resistant disease, as has mycophenolate mofetil (MMF) . Budesonide shows some promise in adults [60, 61]. Although not yet studied adequately in children, it may be an effective treatment for those with obesity or hypertension or predisposition to diabetes mellitus, as well as for AIH unresponsive to more conventional therapies. However, it is contraindicated in patients with cirrhosis because the decreased hepatic drug metabolism results in adverse side effects . Various other immunomodulators have been used, mainly in adults .
In summary, the outlook for children with AIH is guarded. AIH2 is a somewhat more severe disease than AIH1. The diagnosis of AIH requires a thorough and methodical approach to exclude other diseases which can appear similar clinically. Once the diagnosis of AIH is established, early institution of immunosuppressive treatment offers the best chance for a favourable response to therapy. Some patients definitely show resolution of fibrosis or regression of cirrhosis. Nevertheless, some children require low-dose prednisone treatment indefinitely and a few others later require liver transplantation. In general, the rate of liver transplantation is substantially lower with AIH than with PSC. In AIH patients who are transplanted, however, there is a fairly high rate of disease recurrence . In those who never need liver transplantation, some risk of developing hepatocellular carcinoma exists [65-67] and strategies for surveillance need to be developed. Since the earliest reports , important insights into the clinical features, possible mechanisms and optimal treatment of AIH have been gained from studying AIH in children.