Keywordsautoimmune hepatitis, autoantibodies, autoimmune hepatitis- primary biliary cirrhosis overlap syndrome, autoimmune hepatitis-primary sclerosing cholangitis overlap syndrome, autoimmune liver disease
Demographic features 492
Elderly patients 492
Clinical features 494
Grading and staging 500
Histopathological features of remission and relapse 501
Liver transplantation 502
Differential diagnosis 502
Variant forms and cholestatic liver disease 503
Primary biliary cholangitis (‘AIH-PBC overlap’) 503
Autoimmune cholangitis 505
Primary sclerosing cholangitis (‘AIH-PSC overlap’) 505
Diagnostic difficulties 506
Autoantibody-negative patients 506
Antimitochondrial antibody-positive patients 506
Viral hepatitis 507
Drug reaction mimicking or causing autoimmune liver disease 507
Role of liver biopsy 510
Autoimmune hepatitis (AIH) is an unresolving hepatitis usually associated with hypergammaglobulinaemia and tissue-directed autoantibodies and responding in most cases to immunosuppressive therapy. The pathogenesis appears to be an aberrant autoreactivity in genetically susceptible individuals, with molecular mimicry between viral and ‘self’ antigens the likely basis for the autoimmune response. AIH is subclassified on the basis of serological profile into three types, all diagnosed by the same criteria and with similar histological features and treatment strategies. It can usually be readily distinguished from the other two major autoimmune liver diseases, primary biliary cholangitis (PBC) and primary sclerosing cholangitis (PSC) ( Table 8.1 ), although variant forms and overlap syndromes exist.
|Female/male ratio||4 : 1||9 : 1||1 : 2||9 : 1|
|Predominant liver test elevation||AST, ALT||ALP, GGT||ALP, GGT||ALP, GGT|
|Serum Ig elevation||IgG||IgM||IgG, IgM||IgM|
|Autoantibodies||ANA, SMA, LKM1, SLA/LP, pANCA and others||AMA, AMA-PDH-E2, gp210||pANCA||ANA, SMA|
|HLA association||A3, B8, DR3, DR4||DR8 (weak association)||DR52||B8, DR3, DR4|
|Histology||Interface and lobular hepatitis; prominent plasma cells||Florid bile duct lesion||Fibrosis and obliteration of large bile ducts; ductopenia||Florid bile duct lesion|
|Diagnosis||AIH score >15 for diagnosis of definite AIH||AMA, PDH-E2, cholestatic serum enzyme pattern, compatible histology||Biliary strictures and dilation on cholangiography; cholestatic serum enzyme pattern, pANCA||Cholestatic serum enzyme pattern, AMA negative, ANA or SMA positive; histology c/w PBC|
|First-line medical therapy||Immuno-suppression (corticosteroids +/− azathioprine)||Ursodeoxycholic acid (UDCA); obeticholic acid (OCA) in UDCA non-responders||UDCA use is controversial||UDCA + corticosteroids|
Although autoimmune hepatitis has been recognized as a clinical entity for more than 50 years, the worldwide prevalence of AIH remains largely unknown, partly because earlier studies of incidence of idiopathic chronic active hepatitis did not exclude hepatitis C. In addition, before the development of the scoring system for AIH proposed by the International Autoimmune Hepatitis Group, there were no standardized diagnostic criteria in suspected AIH for treatment or for research purposes. Thus, population-based studies performed before widespread implementation of this system and specific testing for hepatitis C and hepatitis B include somewhat heterogeneous patient populations. Prospective studies of the epidemiology of AIH in which hepatitis C was excluded report the incidence of AIH in western European countries as 1.68–1.9 per 100,000 population and the point prevalence 16.9–18.3 per 100,000 population, about that of PBC and twice that of PSC. A similar study in Sweden reported an incidence of 0.85 and point prevalence of 10.7 per 100,000 population.
The incidence and prevalence of AIH vary widely among different populations, but the disease appears to be increasing in frequency in both women and men, with the incidence rate almost doubling between 1994 and 2012 in Denmark. For Caucasian North American populations, the incidence is estimated as about 1 per 100,000. AIH appears to be less common in Japan, but the prevalence appears to be significantly increased in some Aboriginal or First Nations populations of North America, with prevalence of definite or probable AIH estimated as 42.9 per 100,000 in an Alaska Natives population. In Western Europe and North America, where viral hepatitis prevalence rates are relatively low, AIH accounts for about 20% of chronic hepatitis cases in the white populations. In European populations, up to one-third of cases of cryptogenic hepatitis may represent seronegative AIH, with potential benefit from immunosuppressive therapy.
The genetic basis for AIH is complex and partly overlaps with other autoimmune diseases. The most conclusive associations with AIH to date are related to major histocompatibility complex (MHC) class I and II alleles, which appear to confer susceptibility to AIH and influence the disease course. In white European and North American populations, AIH is associated with the human leukocyte antigen (HLA) DRB1*0301 haplotype (DR3) and with DRB1*0401 (DR4); 75–85% of patients from these populations with type 1 AIH carry one or both haplotypes, which confer a six- to sevenfold risk for AIH. DRB1*0301 appears to be the stronger genetic modifier and is associated with higher IgG levels and more severe disease, whereas DRB1*0401 confers a greater chance of clinical remission, lower risk of cirrhosis and lower need for liver transplantation. In Japan, where the DR3 haplotype is rare, AIH appears to be primarily associated with DRB1*0401 ; this form of the disease occurs in elderly patients in both Japanese and other populations who exhibit mild necroinflammatory activity and a good response to immunosuppressive therapy. In contrast to adults, DRB1*0401 does predispose to AIH in children. In South American populations, HLA-A11 confers a high risk of type 1 AIH among white adults. In Argentinian children, DRB1*1301 confers a higher risk for type 1 AIH and is also associated with a protracted course after hepatitis A infection, suggesting that the genetic predisposition to response to viral infection may trigger AIH in this group.
As with most autoimmune disorders, AIH is more common in female patients, with a male/female ratio of about 1 : 3. Some reports suggest that AIH in males presents at a younger age (39 versus 49 years for women) and is more likely to relapse after withdrawal of immunosuppressive therapy. A recent population-based study from Denmark found that male patients with AIH had higher mortality and greater risk for development of hepatocellular carcinoma than women. AIH can present at any age and in any population, but it often presents acutely and may follow a more aggressive course in children and young adults. Different ethnic groups may present with nonclassic phenotypes; for example, male predominance and cholestatic features appear to be more common in nonwhite populations.
Because it has classically been considered a disease of young women, AIH may be underdiagnosed in elderly patients. However, more than 20% of patients with AIH develop the disease after age 60. The median time to diagnosis is more than twice as long in patients over 65, who are more likely to be cirrhotic on presentation, perhaps from more indolent but progressive liver disease. AIH in the elderly patient is typically responsive to corticosteroid therapy. Older adults in the United States with AIH, like those in Japan, tend to have the DRB1*0401 haplotype, whereas patients under age 30 at diagnosis are more likely to have DRB1*0301. Older patients are also more likely to have concomitant thyroid or rheumatic diseases. Histological cirrhosis, without clinical hepatic decompensation, does not appear to be associated with an impaired treatment response.
Autoimmune hepatitis is by definition a disease or a syndrome of unknown cause. A loss of tolerance against the patient’s own liver tissue is regarded as the underlying cause of the self-perpetuating disease process. Environmental agents (e.g. drugs, chemicals) and infectious agents (e.g. viruses) are suspected as triggers of this autoimmune process in genetically susceptible individuals.
The main target cell of the disease process is the hepatocyte. Our knowledge on the pathogenetic mechanisms is otherwise still limited, especially on how nonparenchymal liver cells contribute to the self-perpetuating disease process. Knowledge of an infection-triggered process is much more advanced in other autoimmune diseases such as diabetes mellitus, which only becomes clinically overt once >90% of insulin-producing β cells are destroyed and may take decades after the disease-triggering event. This also means that the agent has left the body once autoimmune hepatitis is diagnosed by the clinician. Hypotheses on the pathogenesis of AIH are therefore derived from human studies that investigate either components of the humoral and cellular immune system or the genetic background in cohorts of patients of different ethnic background. However, these are only static observations that rarely allow insight into the dynamics of pathogenetic mechanisms responsible for the self-perpetuation of the autoimmune disease process.
In animal models this becomes of increasing importance, in particular when genetic mouse models allow examination of specific questions raised by observations in human AIH. It must be noted that most of these animal models are not self-perpetuating; they rarely target the same molecules or autoantigens as in human disease and therefore can only mimic some aspects of AIH in humans. Nevertheless, advantages of animal models are that they can be used to study the dynamic disease process from its beginning. Usually, specific questions can be addressed with specific answers obtained.
One initial approach for the last five decades was to study the humoral immune response, such as the specificity of autoantibodies in the peripheral blood. The application and availability of molecular biology and the screening of complementary DNA (cDNA) libraries have allowed the molecular identification of the autoantigens. Table 8.2 summarizes the various autoantibody specificities and their target antigens. The main autoantibodies in AIH are antinuclear antibody (ANA) and smooth muscle antibody (SMA), which are neither liver nor liver disease specific. Of more interest are the intracellular enzymes involved in the metabolism of endogenous and exogenous environmental agents. Examples are members of the cytochrome P-450 (CYP) supergene family responsible for phase I drug metabolism but also involved in the metabolism of endogenous compounds including hormones. A second group of autoantigens in liver disease comprise the uridine diphosphate glucuronosyltransferases (UGTs), a supergene family responsible for drug metabolism and that of important endogenous products such as bilirubin. UGTs mediate glucuronidation and thus are important for phase II drug metabolism. Another important group of autoantigens are intracellular enzymes, especially formiminotransferase cyclodeaminase, which is a target of the anti-liver cytosol 1 (LC-1) antibodies and an RNA-binding protein that is a target of anti-soluble liver antigen/liver pancreas antigen (SLA/LP) antibodies. Anti-SLAs are directed against the transfer ribonucleoprotein complex tRNA (ser) sec, renamed as SEPSECS (Sep[O-phosphoserine] tRNA:Sec [selenocysteine] tRNA synthase). These autoantibodies share some characteristics: the autoantigens are highly conserved molecules, and the autoepitopes are conserved regions within these molecules. In contrast to most of the monoclonal antibodies produced against these autoantigens, the naturally occurring autoantibodies are inhibitory. Expression of these drug-metabolizing enzymes is inducible by cytokines. The genetic organization is similar to that of the immune system, allowing defence against numerous chemical agents that were produced in the past or that will be produced in future decades.
|ANA||Indirect IF||Multiple targets, including chromatin, ribonucleoproteins, ribonucleoprotein complexes||AIH type 1; AIC; PSC; AIH-PSC overlap (autoimmune sclerosing cholangitis, ASC); PBC||Specific IF pattern has no clinical significance.|
|SMA||IF and ELISA||Microfilaments (e.g. F-actin) and intermediate filaments (vimentin, desmin)||Same as ANA||Found alone or in conjunction with +ANA in up to 87% of AIH patients|
|SLA/LP||Immunoblot, RIA or commercial ELISA assay||Transfer ribonucleoprotein complex (tRNP [Ser]Sec)||AIH; rarely, reported in PBC and PSC||High specificity for AIH but limited sensitivity|
|LKM1||Indirect IF, ELISA, IB, RIA||Cytochrome P-450 2D6 (CYP2D6)||AIH type 2||Patients with type 2 AIH usually have only LKM antibodies, which may be found in other conditions, notably hepatitis C.|
|LC-1||Indirect IF, ELISA, IB, RIA||Formiminotransaminase cyclodeaminase (FTCD)||AIH type 2||Also found in chronic hepatitis C. In AIH, associated with severe disease and has prognostic implications.|
|LKM3||Indirect IF, ELISA, IB, RIA||Uridine diphosphate glucuronosyltransferase (UGT)||AIH type 2||Also found in chronic hepatitis D|
|AMA||Indirect IF, ELISA, IB, RIA||E2 subunit of 2-oxo-acid dehydrogenase complexes||PBC; present in low titres in rare cases of AIH type 1 and SLA+ AIH|
|pANCA||Indirect IF||Nuclear lamina proteins||AIH type 1; PSC; ASC||May aid in reclassification of cryptogenic chronic hepatitis as type 1 AIH|
|ASGPR||ELISA, RIFA||Asialoglycoprotein receptor (ASGPR)||AIH, PBC||Also found in drug-induced hepatitis and chronic hepatitis B, C, D|
|LM||IF, IP, ELISA||CYP1A2||APECED hepatitis||Also occurs in dihydralazine-induced hepatitis|
Autoantibodies are regarded as useful diagnostic tools but not as pathogenetic mediators of disease itself. In addition, they may be helpful in identifying targets of disease processes ultimately mediated by T cells. Autoantibodies may also help in the identification of aetiological agents that may trigger the process. The CYP enzymes 2D6 and 2A6 share sequence homology with T-cell epitopes of the hepatitis C virus (HCV) proteins, with the immediate early protein (IE) 175 of the herpes simplex virus (HSV), as well as with Epstein–Barr virus (EBV) and cytomegalovirus (CMV). This suggests that more than one agent may be capable of triggering AIH in humans. Genetic studies have identified risk alleles located mainly in the MHC locus but also non-MHC gene polymorphisms such as tumour necrosis factor alpha (TNFα), CTLA-4, C4A QO or the vitamin D receptor. Within the MHC locus DRB1*0301, DRB1*0401 are genetic risk factors in Europe and North America, within DRB1*1301 are risk factors in South America and within DRB1*0404, DRB1*0405 risk factors in Mexico, Japan and China, as well as DRB1*07, DQB1*0201 for type 2 AIH. The genetic background may be responsible for either risk of disease occurrence or severity of the disease process. The genetic background may vary among different ethnic groups worldwide.
One particular human syndrome of interest for a genetically mediated autoimmune disease process is the autoimmune polyendocrine syndrome type 1 (APS-1), or autoimmune polyendocrinopathy–candidiasis–ectodermal dystrophy (APECED). This syndrome is caused by single point mutations (single nucleotide polymorphisms, SNPs) in a single gene, the autoimmune regulator protein ( AIRE ), which is a transcription factor expressed in cells involved in induction and maintenance of immune tolerance, in particular medullary cells of the thymus and dendritic cells. This protein is significant for the maintenance of self-tolerance. APS-1 patients develop multiorgan autoimmune diseases, including AIH. Once the liver is involved, CYP 1A2 and 2A6 are targets of circulating autoantibodies. If adrenal insufficiency becomes part of the clinical phenotype, patients develop autoantibodies against adrenal CYP enzymes involved in the biosynthesis of corticosteroids. Consequent corticosteroid hormone deficiency is responsible for the clinical disease phenotype of adrenal insufficiency.
AIH does not occur in a mouse model when animals are deficient for natural killer T (NKT) cells, an important component of the cellular immune system. NKT cells are usually enriched within the human liver and therefore may also be crucial for mediating the disease process of AIH in humans. Another recently developed animal model uses mice that express human CYP2D6 as a transgene in their livers, in addition to the mouse equivalent of the human cytochrome P-450. If these mice are infected with an adenovirus expressing human CYP2D6 but not wild-type adenovirus, the liver develops a self-perpetuating autoimmune process with CD4+ T cell-enriched liver infiltrates. The humoral autoantibody profile shows anti-CYP2D6 antibodies that react with the identical autoepitopes to that in human AIH type 2. Thus, for the first time, this model shows an infection-triggered autoimmune response that seems to be self-perpetuating and that targets the same autoantigen as in humans.
These studies from humans to mice, and vice versa, from bedside back to the laboratory, give insights into the complex nature of the pathogenetic mechanisms in AIH. Observations in humans identified the molecular targets, and specifically designed animal models then become useful to answer mechanistic questions on the pathogenesis of AIH. Recently, regulatory T cells (Tregs) have attracted the interest of immunologists, and studies on the role of Tregs in AIH have just started. Understanding the molecular mechanisms relevant to AIH will also reveal new potential molecular targets for specific immune interventions in AIH therapy. Fig. 8.1 summarizes the current views on the molecular pathogenesis of AIH.
Presentation of AIH varies widely, ranging from asymptomatic elevations of serum liver enzymes, to massive hepatic necrosis resulting in acute liver failure. Most patients present with nonspecific symptoms such as fatigue, lethargy, anorexia and abdominal pain. Some present with a prolonged prodromal phase characterized by a flu-like illness; about half of patients presenting in this manner will be icteric at some point. About one-third of patients are cirrhotic at presentation, confirming the impression that AIH often has a prolonged subclinical course. Approximately 30% of patients, usually in the younger age-group, will have an acute presentation mimicking acute viral hepatitis; the response to therapy and prevalence of cirrhosis at presentation are the same as in patients who present with more insidious onset. Approximately 20% of patients will be asymptomatic with respect to liver disease at diagnosis, and frequency of cirrhosis in this group at presentation is similar to that for symptomatic patients. In children, acute presentation is common, and the patients are more likely to be cirrhotic (60–80%).
About 50% of patients with AIH will have concurrent autoimmune disorders, most frequently thyroid disease or rheumatoid arthritis. Other associated disorders include type 1 diabetes mellitus, coeliac disease, Sjögren syndrome, vitiligo and Addison disease. The presence of ulcerative colitis, however, should raise questions about the diagnosis of AIH; such patients are more likely to have PSC.
Serum transaminase (ALT, AST) levels are usually elevated in AIH but fluctuate because of the waxing and waning nature of the hepatic necroinflammatory activity and may even be within the normal range at times. Serum bilirubin levels are also quite variable, but alkaline phosphatase (ALP) is usually only mildly raised. Notably, γ-globulin and serum IgG levels are characteristically elevated in the 1.2–3 times normal range, with lesser elevations in serum IgM and normal serum IgA.
About 20–25% of patients with autoimmune hepatitis are asymptomatic at diagnosis, with the disease discovered because of routine liver biochemical testing. Such patients tend to be older (mean age 48) compared with patients symptomatic at diagnosis (mean age 41). Lower serum liver enzymes and IgG levels in asymptomatic patients are reflected in lesser degrees of necroinflammatory activity in liver biopsies, but lymphoplasmacytic inflammation, interface hepatitis and lobular necroinflammatory activity are seen in both symptomatic and asymptomatic patients. About 25% of the asymptomatic patients are expected to develop symptoms during follow-up and are somewhat less likely to respond well to treatment. However, overall prognosis is good, with approximately a 80% 10-year survival, comparable with patients who present with symptoms.
Asymptomatic patients may be divided into two distinct groups: those who present with ‘burned out’ cirrhosis and near-normal liver tests (~25%), and those who present with relatively mild hepatitis but without cirrhosis. The percentage presenting with cirrhosis is similar to that seen in symptomatic patients (36% in a single-centre study). Patients with cirrhosis at presentation have a less favourable outcome (62% versus 94% 10-year survival).
AIH should be considered in all patients presenting with acute liver failure and massive necrosis on liver biopsy. Histological features are unfortunately nonspecific, but prominent zone 3 necroinflammatory activity may be seen (65%), as well as lymphoid follicles (32%) and a plasma cell-rich infiltrate (63%). Diagnosis is facilitated by use of the revised international AIH scoring system ; the simplified criteria may not perform as well in the setting of acute liver failure. In some patients a rapid response to immunosuppressive therapy can support a diagnosis of AIH, although clinical improvement with corticosteroid administration is not seen in all those in whom AIH is suspected. In children, distinguishing AIH from Wilson disease may be particularly problematic. Wilson disease should be considered if haemolysis is evident and ALT elevations are mild, despite significant clotting abnormalities. Other differential diagnoses include severe drug injury and fulminant viral hepatitis; however, a substantial number of massive hepatic necrosis cases remain of indeterminate cause.
The diagnosis of AIH is based on a combination of clinical and laboratory features, with exclusion of other causes of liver disease, such as viral hepatitis, Wilson disease, α1-antitrypsin deficiency, PBC, PSC, alcohol abuse and drug reaction. Diagnosis is straightforward in 50% of patients and is aided by serum studies and the revised scoring system developed by the International Autoimmune Hepatitis Working Group or the simplified system. .
Characteristic biochemical abnormalities include elevated transaminase levels, often accompanied by hyperbilirubinaemia, hypergammaglobulinaemia with a disproportionate increase in serum IgG, and normal or only slightly elevated ALP. As previously noted, transaminases (ALT, AST) and bilirubin levels may vary widely and may even normalize spontaneously for a time, and thus there is no diagnostic threshold level for transaminases. Bilirubin and transaminase levels do not reliably reflect disease severity. Total globulins may be in the normal range, and IgM levels are not as elevated as in PBC.
The three most commonly reported autoantibodies in AIH are antinuclear (ANA), anti-smooth muscle (SMA) and anti-liver–kidney microsomal (LKM) antibodies (see Table 8.2 ). ANA is the most common, showing moderate sensitivity and specificity in the appropriate clinical setting, while SMA shows high specificity and moderate sensitivity. In Europe, indirect immunofluorescence on rodent frozen tissue sections for detection of autoantibodies most significant for autoimmune liver diseases is still the recommended method. In the United States, however, enzyme-linked immunosorbent assay (ELISA) technologies based on recombinant autoantigens are of increasing importance and are less technically challenging. In general, while useful for diagnosis, the autoantibody titre does not reliably reflect disease severity or outcome, although persistence of SMA autoantibodies at titres >1 : 80 has been reported to correlate with inflammatory activity on post-treatment liver biopsy. About 70–80% of patients with AIH have ANA or SMA antibodies or both (≥1 : 40 titre). Because the ANA react mainly with histones and DNA, the pattern is homogeneous, similar to that seen in lupus, but other patterns also occur, with no apparent clinical significance. The SMA reacts with several cytoskeletal components, including F-actin.
Anti-LKM1 antibodies, the hallmark of type 2 AIH, are less frequently seen and are mainly detected in children, without ANA or SMA. This serological marker is more common in southern Europe, but its true prevalence is not known; <4% of patients in the United States have LKM1 antibodies, compared with 14–38% in the United Kingdom. The antibody is directed against CYP2D6 and is associated with DRB1*0701 allele. Antibodies directed against LC-1 also occur in type 2 AIH and bind the folate-metabolizing enzyme formimino transaminase cyclodeaminase. LKM3 autoantibodies that bind to UGTs also occur in AIH type 2 and may be the only serological marker; however, they also occur in some patients with chronic hepatitis D (delta).
Antibodies against the soluble liver antigen/liver pancreas antigen (SLA/LP) are found in 10–30% of patients with AIH. Although testing for this autoantibody is not routine, it appears to have a global distribution. Perinuclear antineutrophil cytoplasmic antibody (pANCA) is not specific but is present in 60–90% of AIH patients.
A working classification devised originally for research purposes but also used in clinical practice has been developed by the International Autoimmune Hepatitis Group and modified in 1999. This system classifies cases as definite or probable AIH, based on weighted parameters ( Table 8.3 ). In multiple reports this system has a high degree of sensitivity (97–100%) for diagnosis of AIH and also effectively excludes AIH in patients with PSC or biliary disorders (96–100% accuracy for exclusion of definite AIH). However, if cases considered probable AIH are included, the overall specificity is reduced. The overall diagnostic accuracy of the scoring system is about 90%.
|ALP/AST (or ALT) ratio|
|Serum globulins or IgG above normal|
|Autoantibodies (ANA, SMA or LKM-1)||Lower titres are considered significant in children and should be scored at least +1.|
|>1 : 80||+3|
|1 : 80||+2|
|1 : 40||+1|
|<1 : 40||0|
|Hepatitis viral markers||The patient should be tested for markers for hepatitis A, B and C infection; tests for other viruses such as EBV and CMV may be considered.|
|Drug history||Recent use of known or suspected hepatotoxic drugs|
|Average alcohol consumption|
|Low (<25 g/day)||+2|
|High (>60 g/day)||−2|
|Liver histology||‘Biliary changes’ refers to bile duct patterns of injury typical of PBC or PSC with ductopenia in an adequate biopsy. |
‘Other features’ are any suggesting an alternative aetiology, e.g. nonalcoholic fatty liver disease.
|Hepatocyte rosette pattern of regeneration||+1|
|None of the above||−5|
|Other autoimmune disorders, in patient or first-degree relatives||+2|
|Optional parameters in patients who are seronegative for ANA, SMA and LKM1:||‘Other defined antibodies’ are those with published evidence of relevance to AIH and include pANCA, anti-LC-1, anti-SLA and anti-ASGPR|
|Seropositivity for other defined autoantibodies||+2|
|HLA DR3 or DR4||+1|
|Response to therapy|
|Interpretation of aggregated scores|
A simplified scoring system was proposed in 2008 by the same group, with the goal of producing a system that is more widely applicable in routine clinical practice. The simplified system incorporates four parameters: associated autoantibodies, IgG level, histological features and exclusion of viral hepatitis, all found to be independent predictors of AIH ( Table 8.4 ). To streamline the system, features of the 1999 system that reflect disease severity rather than the nature of the disease are excluded. Of interest to the pathologist is the inclusion of the following three categories for grading histology in the simplified system:
Typical histology for AIH. Each of the following three features was required to be present:
Interface hepatitis, with lymphocytic/lymphoplasmacytic inflammation in portal tracts and extending into the lobule
Hepatocyte rosette formation
Histology compatible with AIH. Chronic hepatitis pattern of injury with lymphocytic infiltration but lacking some of the features considered ‘typical’.
Atypical histology for AIH. Features suggestive of other diagnoses (e.g., steatohepatitis) are present.
|ANA or SMA +||≥1 : 40||1|
|ANA or SMA +||≥1 : 80||2 *|
|or LKM||≥1 : 40|
|IgG||>Upper limit of normal||1|
|>1.10 times upper limit of normal||2|
|Liver histology||Compatible with AIH||1|
|Absence of viral hepatitis||Yes||2|
|Interpretation of aggregated scores|
This simplified scoring system has superior specificity (90% versus 73%) and accuracy (92% versus 82%) compared with the 1999 system and is better at excluding the diagnosis of AIH in patients with other immune disorders. However, it may underestimate AIH in patients with few or atypical features, and the 1999 system has been found to perform better in this regard. Sensitivity of the simplified system is 90–95% for a diagnosis of probable or definite AIH, compared with 100% for the 1999 system, but only 70% for definite AIH in an independent cohort. Also, the scoring system is more useful and was designed for the definition and characterization of a homogeneous group of patients for interventional studies, rather than helping to establish the diagnosis in an individual patient.
AIH may be subclassified based on the autoantibody profile ( Table 8.5 ). Such subclassification may be useful for research purposes and may define pathogenetically distinct groups, but it has little application in clinical practice at present. Type I AIH, the most common subtype, is defined by ANA and/or SMA positivity and has a bimodal age distribution with peaks at 10–20 and at 45–70 years. Titres of at least 1 : 80 are considered positive in adults, although lower titres in children are accepted as positive, since ANA positivity in the absence of autoimmune disease is rare in childhood.
|Type||Autoantibodies||Age at onset (years)||Comments|
|1||ANA and/or SMA||Bimodal age distribution, with peaks at 10–25 and 45–70||Most common type|
|2||LKM1, LKM3, LC-1; ANA and SMA negative||<15||Severe necroinflammatory activity but responds well to treatment; frequent extrahepatic autoimmune syndromes|
|3||SLA/LP||Mean 37–43||Clinically similar to type 1; identifies patients with a more severe form of AIH; may be more likely to relapse after corticosteroids|
Type 2 AIH is characterized by the presence of anti-LKM1 antibodies; antibodies directed against ANA and SMA are usually not present. Type 2 AIH patients are more likely to present with acute, severely active disease and are younger (usually 2–14 years) at disease onset than patients with type 1 AIH. An association with IgA deficiency has been reported. This subtype of AIH is rare in adults, occurring in only 4% of adult AIH cases in the United States, but may constitute up to 30% of cases of paediatric AIH. It occurs globally but displays regional variation in prevalence. It is more frequently found in Southern than in Northern Europe and the United States ; up to 20% of AIH cases in Germany and France are type 2. These regional variations in disease prevalence may reflect differences in genetic risk factors, and this subtype of AIH has been associated with HLA DRB1*07 in Brazil and Germany. Early diagnosis of type 2 AIH is important because the disease follows an aggressive course if untreated but responds well to immunosuppression. However, because type 1 AIH is much more common than type 2, most young patients with severe disease will actually represent type 1 AIH. Phenotypic differences in adults may be less pronounced, and the clinical significance of subclassification of AIH has been questioned.
AIH with autoantibodies directed against LKM epitopes on CYP1A2 and CYP2A6 occur in the setting of autoimmune polyendocrinopathy–candidiasis–ectodermal dystrophy (APECED), a distinctive autosomal recessive syndrome characterized by multiple endocrine organ failure and autoimmune manifestations. As mentioned previously, APECED (another term for the same disease is autoimmune polyendocrine syndrome type 1 or APS-1) is caused by mutation in the AIRE (autoimmune regulator) gene, whose product is expressed in cells such as medullary thymic epithelium and dendritic cells involved in induction and maintenance of immune tolerance. The disease is clinically heterogeneous, perhaps due to the large number of genetic mutations which have been identified, but up to 20% of patients will have an autoimmune-type hepatitis of variable severity and which is sometimes fatal.
Type 3 AIH is characterized by positivity for SLA/LP antibodies, with or without antibodies directed against ANA or SMA, and may be clinically severe. Anti-SLA/LP is often associated with anti-Ro52 antibodies. The existence of this subtype of AIH remains somewhat controversial. Type 3 AIH is rare in adults and is clinically indistinguishable from type 1. However, SLA/LP antibodies seem to characterize a group of AIH patients with a more severe course and an increased risk to relapse after termination of immunosuppressive therapy.
The liver biopsy in AIH varies with the stage of the disease but always reflects a hepatitic pattern of injury. In cases with acute onset, there may be minimal fibrosis, and the overall histopathology is that of an active hepatitis with portal and periportal lymphoplasmacytic infiltrates, interface hepatitis and variable lobular necroinflammatory activity ( Fig. 8.2 ). In mildly active cases, a mononuclear portal inflammatory infiltrate with minimal interface hepatitis may be the only histological abnormality ( Fig. 8.3 ). During more active phases of the disease, interface hepatitis (formerly termed ‘piecemeal necrosis’) is prominent, characterized by extension of lymphocytes and in some cases plasma cells from the portal tract into the periportal parenchyma of the liver ( Fig. 8.4 ). Hepatocytes at the limiting plate show evidence of damage, such as ballooning degeneration ( Fig. 8.5 ), and apoptotic bodies (acidophilic bodies) may also be seen. Engulfment of lymphocytes by hepatocytes, termed emperipolesis, is often found in areas of interface activity ( Fig. 8.6 ) and is classified as a typical feature of AIH in the simplified scoring system.
Plasma cells are often present in considerable numbers in the portal inflammatory infiltrate (see Fig. 8.4 ) and are sometimes seen singly and in clusters in the lobule ( Fig. 8.7 ). However, about one-third of biopsies from patients with well-documented AIH will have few or no plasma cells. Notably, the presence of plasma cells in clusters at the limiting plate and in areas of interface hepatitis, although not specific, favours AIH over viral hepatitis. The plasma cells in AIH are predominantly IgG positive, with few IgM-expressing cells, whereas in PBC, an appreciable number of the plasma cells express IgM. This differential expression has been suggested as a useful technique to distinguish AIH from PBC, but IgG predominance is not specific for AIH, and thus this technique is not useful for cases of overlap syndrome. Portal lymphocytes in AIH are predominantly CD4+, whereas CD8+ lymphocytes are more prominent in areas of interface hepatitis. IgG4-bearing plasma cells have been described in liver disease associated with autoimmune pancreatitis (see Chapter 9 ) but are not present in significant numbers in typical AIH.
The severity of lobular necroinflammatory activity in AIH ranges from mildly active hepatitis, through bridging necrosis, to massive hepatic necrosis ( Fig. 8.8 ). Hepatocyte regeneration may be prominent, with regenerating rosette-like structures ( Fig. 8.9 ) seen in cases with prominent necroinflammatory activity. Lobular changes such as hepatocellular ballooning degeneration, spotty hepatocyte necrosis and apoptotic bodies are common but not specific. Small collections of lymphocytes in the lobule, marking areas of loss of single hepatocytes, are usually found even in mildly active disease ( Fig. 8.10 ), whereas aggregates of Kupffer cells are more typically found in more severely active cases. In some cases, Kupffer cells contain hyaline droplets identical to Russell bodies in plasma cells. Granulomas are not a feature of AIH, and extensive macrovesicular steatosis is suggestive of a concomitant disease such as steatohepatitis. Prominent bile accumulation is not typical, but canalicular and hepatocellular cholestasis may be seen in severely active cases with marked lobular inflammation.