DIAGNOSTIC CRITERIA

An international panel codified the diagnostic criteria of AIH in 1992, and an expanded panel updated them in 1999.³ The propensity for an acute, rarely fulminant, presentation has been recognized,² and the requirement for six months of disease activity to establish chronicity has been waived.³ Cholestatic histologic changes, including bile duct injury and ductopenia, are incompatible findings, but trivial biliary changes within the background of classic histologic features do not preclude the diagnosis.^4,5

The serologic tests essential for diagnosis are assays for antinuclear antibodies (ANA), smooth muscle antibodies (SMA), and antibodies to liver-kidney microsome type 1 (anti-LKM1).^1,6,7 These assays are based on the indirect immunofluorescence of rodent tissues or Hep-2 cell lines or on enzyme immunoassays using microtiter plates with adsorbed recombinant or highly purified antigens. Atypical perinuclear anti-neutrophil cytoplasmic antibodies (atypical pANCA) are common in type 1 AIH, PSC, and chronic ulcerative colitis.^6,7 They are directed against antigens within the nucleus, rather than the cytoplasm, of granulocytes, and the reactivities localize to the proteins within the lamina of the nucleus. Because these antibodies are directed against nuclear envelope antigens rather than cytoplasmic antigens, a better term for them is anti-neutrophil nuclear antigens or “ANNA.”⁸ Atypical perinuclear anti-neutrophil cytoplasmic antibodies have been useful in evaluating patients who lack the conventional autoantibodies.^6,7 Celiac disease can be associated with liver disease that resembles AIH and should be excluded in patients with cryptogenic chronic hepatitis by screening for immunoglobulin (Ig) A antibodies to tissue transglutaminase or endomysium.^6,7,9 IgA endomysial antibodies are more predictive of celiac disease in patients with AIH than are IgA antibodies to tissue transglutaminase, which can be stimulated by hepatic inflammation and fibrogenesis.^6,7

New autoantibodies continue to be characterized in the hope of improving diagnostic specificity and prognostic value, but none has been incorporated into conventional diagnostic algorithms.^6,7,10 Antibodies to soluble liver antigen (anti-SLA), actin (anti-actin), chromatin (anti-chromatin), asialoglycoprotein receptor (ASGPR), and liver cytosol type 1 (anti-LC1) have been associated with severe AIH, poor treatment response, and relapse after drug withdrawal.^10–13 Their major clinical limitation has been their infrequent individual occurrence in AIH.

CLINICAL CRITERIA

The definite diagnosis of AIH requires exclusion of other similar diseases; laboratory findings that indicate substantial immunoreactivity; and histologic features of interface hepatitis (Fig. 88-3).³ A probable diagnosis is justified when findings are compatible with AIH but insufficient for a definite diagnosis (see Fig. 88-3).³ Patients who lack conventional autoantibodies but who are seropositive for investigational markers, such as antibodies to ASGPR, SLA, actin, or LC1, are classified as having probable disease.³

Figure 88-3. Diagnostic algorithm for autoimmune hepatitis. Diagnosis requires predominant elevation of the serum aminotransferase levels, exclusion of other similar disorders (especially Wilson disease, drug-induced hepatitis, and viral hepatitis), interface hepatitis on histologic examination, and manifestations of immunoreactivity, including serum gamma globulin elevation and seropositivity for antinuclear antibodies (ANA), smooth muscle antibodies (SMA), or antibodies to liver-kidney microsome type 1 (anti-LKM1). The degree of immunoreactivity and the presence of confounding etiologic factors, such as alcohol or drug exposure, distinguish definite from probable autoimmune hepatitis. Classification into one of the descriptive categories of type 1 and type 2 autoimmune hepatitis is based on the nature of the autoantibodies. AMA, antimitochondrial antibodies; AST, aspartate aminotransferase; anti-HAV, antibody to hepatitis A virus; anti-HCV, antibody to hepatitis C virus; HBsAg, hepatitis B surface antigen; IgM, immunoglobulin M.

SCORING CRITERIA

The original scoring system proposed by the International Autoimmune Hepatitis Group accommodates the diverse manifestations of AIH and renders an aggregate score that reflects the net strength of the diagnosis before and after glucocorticoid treatment (Table 88-1).³ By weighing each component of the syndrome, discrepant features can be accommodated and biases associated with isolated inconsistencies can be avoided. The original scoring system ensures the comparability of study populations in clinical trials, provides a comprehensive template for systematically assessing all features of the disease, and can render a diagnosis of AIH in patients with few or atypical features.³ The original scoring system is not a discriminative diagnostic index, and it should not be used to distinguish AIH from other liver diseases.

Table 88-1 Revised Original Scoring System for the Diagnosis of Autoimmune Hepatitis

CATEGORY	VARIABLE	SCORE
Gender	Female	+2
AP/AST	>3	−2
	<1.5	+2
Gamma globulin or IgG level above normal	>2.0*	+3
	1.5-2.0*	+2
	1.0-1.5*	+1
	<1.0	0
ANA, SMA, or anti-LKM1 titer	>1 : 80	+3
	1 : 80	+2
	1 : 40	+1
	<1 : 40	0
AMA	Positive	−4
Viral markers	Positive	−3
	Negative	+3
Drug history	Yes	−4
	No	+1
Alcohol	<25 g/day	+2
	>60 g/day	−2
HLA	DR3 or DR4	+1
Immune disease	Thyroiditis, ulcerative colitis, synovitis, others	+2
Other liver-defined autoantibodies	Anti-SLA, anti-actin, anti-LC1, pANCA	+2
Histologic features	Interface hepatitis	+3
	Plasmacytic infiltrate	+1
	Rosettes	+1
	None of above	−5
	Biliary changes	−3
	Other features	−3
Treatment response	Complete	+2
	Relapse	+3
Pretreatment Score
Definite diagnosis	>15
Probable diagnosis	10-15
Post-treatment Score
Definite diagnosis	>17
Probable diagnosis	12-17

AMA, antimitochondrial antibodies; ANA, antinuclear antibodies; anti-LC1, antibodies to liver cytosol type 1; anti-LKM1, antibodies to liver-kidney microsome type 1; anti-SLA, antibodies to soluble liver antigen; AP/AST (or AP/ALT), ratio of serum alkaline phosphatase level to serum aspartate aminotransferase (or serum alanine aminotransferase) level; HLA, human leukocyte antigen; IgG, immunoglobulin G; pANCA, perinuclear anti-neutrophil cytoplasmic antibodies; SMA, smooth muscle antibodies.

* Times upper limit of normal.

Adapted from Alvarez F, Berg PA, Bianchi FB, et al. International Autoimmune Hepatitis Group report: Review of criteria for diagnosis of autoimmune hepatitis. J Hepatol 1999; 31:929-38.

A simplified scoring system has been developed to ease clinical application and is based on four clinical components that include the presence and level of autoantibody expression by indirect immunofluorescence, serum IgG concentration, histologic features, and viral markers (Table 88-2).¹⁴ The original scoring system has greater sensitivity for the diagnosis than the simplified system (100% vs. 95%), but the simplified system has greater specificity (90% vs. 73%) and predictability (92% vs. 82%).¹⁵ Whereas the original scoring system is useful for evaluating patients in whom every component must be assessed because of few or atypical findings, the simplified scoring system is useful for excluding AIH in patients with other conditions and concurrent immune features.

Table 88-2 Simplified Scoring System for Diagnosis of Autoimmune Hepatitis*

CATEGORY	VARIABLE	SCORE
Autoantibodies†
Antinuclear antibodies or smooth muscle antibodies	1 : 40	+1
	1 : 80	+2
Antibodies to liver-kidney microsome type 1	1 : 40	+2
Antibodies to soluble liver antigen	Positive	+2
Immunoglobulin Level
Immunoglobulin G	>Upper limit of normal	+1
	>1.1 times upper limit of normal	+2
Histologic Findings
Morphologic features	Compatible with autoimmune hepatitis	+1
	Typical of autoimmune hepatitis	+2
Viral Disease
Absence of viral hepatitis	No viral markers	+2
Pretreatment Aggregate Score
Definite diagnosis		≥7
Probable diagnosis		6

* Adapted from Hennes EM, Zeniya M, Czaja AJ, et al. Simplified diagnostic criteria for autoimmune hepatitis. Hepatology 2008; 48:169-76.

† Autoantibody titers as determined by indirect immunofluorescence.

PATHOGENESIS

The pathogenic mechanisms of AIH are unknown. The most popular hypotheses invoke a constellation of interactive factors that include a triggering agent, genetic predisposition, and various determinants of autoantigen display, immunocyte activation, and effector cell expansion (Fig. 88-4).^16–18 Proposed triggering factors include infectious agents, drugs, and toxins. The lag time between exposure to the trigger and onset of the disease can be long, and the triggering factor may not be needed for perpetuation of the disorder. The CD4⁺ helper T cell is the principal effector cell, and its activation is the initial step in the pathogenic pathway.

Figure 88-4. Interactive mechanisms that contribute to the development of autoimmune hepatitis in white North American and northern European adults. The initial stimulus for immune activity is an antigenic peptide (upper left corner) that has a negatively charged aspartic acid or glutamic acid at a position within its structure (P4) that can form a salt bridge with a positively charged lysine or arginine residue at position 71 within the DR beta polypeptide chain (DRβ71) of the antigen binding groove of the class II DR molecule of the major histocompatibility complex (top center). The DR molecule-antigen complex of the antigen-presenting cell (APC) then interacts with the antigen receptor of a CD4⁺ T-helper cell (interaction not shown), and the first co-stimulatory signal is completed (1st signal). The CD28 molecule on the surface of the CD4⁺ T-helper cell ligates with the B7 ligand on the surface of the APC, and the second co-stimulatory signal (2nd signal) is completed (upper right corner). The activated CD4⁺ T-helper cell can then differentiate and proliferate along type 1 and type 2 cytokine pathways (middle right). Deficiencies in the function or amount of cytotoxic T lymphocyte antigen 4 (CTLA4) can enhance the strength of the 2nd signal. Differentiation along the type 1 cytokine pathway can be promoted by polymorphisms of the tumor necrosis factor gene (TNFA*2) and tumor necrosis factor receptor superfamily gene (FAS), resulting in cell-mediated cytotoxicity by sensitized liver-infiltrating cytotoxic T cells and increased hepatocyte apoptosis (middle bottom). The apoptosis of hepatocytes can, in turn, release DNA cytosolic fragments that contribute to the production of diverse collateral autoantibodies (middle left). Autoantibody expression is, in part, host-dependent and influenced by the susceptibility alleles DRB1*03 and DRB1*07. Host genetic predispositions are also important in encoding the antigen-binding groove of the class II DR molecule through the actions of DRB1*0301, DRB1*0401, and DRB1*1301 alleles and in generating autoimmune promoters (cytokine and FAS polymorphisms) that enhance cell-mediated cytotoxicity and hepatocyte apoptosis. The enhanced expression of the anti-apoptotic protein (bcl-2) on the surface of cytotoxic T cells can protect them from programmed cell death and perpetuate their autoreactivity (middle). The cytokine pathways can be enhanced by deficiencies in the actions of T-regulatory cells (T-reg cells), which have suppressive effects that can be reversed by glucocorticoid treatment. Differentiation of B cells into plasma cells, via the type 2 cytokine response of activated CD4⁺ T-helper cells, can result in immunoglobulin production that, in turn, generates an antibody-mediated cellular toxicity. Natural killer (NK) cells with Fc receptors are directed against complexes of immunoglobulin with normal hepatocyte membrane constituents.

(Adapted from Czaja AJ. Autoimmune hepatitis—Part A: Pathogenesis. Expert Rev Gastroenterol Hepatol 2007; 1:113-128.)

Molecular mimicry of a foreign antigen and a self-antigen is the most common explanation for the loss of self-tolerance, but this mechanism has not been established for any autoimmune disease.^16–18 Genetic factors influence autoantigen presentation and CD4⁺ helper T cell recognition. The antigen-binding groove of the class II molecule of the major histocompatibility complex (MHC) is encoded by alleles that determine the groove’s configuration and ability to activate immunocytes. The susceptibility alleles of AIH in white North Americans and northern Europeans reside on the DRB1 gene and are DRB1*0301 and DRB1*0401 (see Fig. 88-4).^19–21

Different ethnic groups have different susceptibility alleles, a finding that supports a “shared motif hypothesis” of pathogenesis.^19–21 According to this hypothesis, the risk of disease relates to amino acid sequences in the antigen-binding groove of the class II MHC molecule, and multiple alleles encode the same or similar sequence (“shared motif”). The critical shared motif in white North Americans and northern Europeans with AIH is a six-amino-acid sequence represented by the code LLEQKR.^19–21 This sequence is located between positions 67 and 72 of the DRβ polypeptide chain of the class II MHC molecule, and lysine (K) in position 71 is the critical determinant of susceptibility. DRB1*0301 and DRB1*0401 encode identical amino acid sequences in the DRβ67-72 region and affect susceptibility similarly.

DRB1*0404 and DRB1*0405 are the susceptibility alleles in Mexican, Japanese, mainland Chinese, and Argentine adults and encode a similar sequence, except for arginine (R) instead of lysine (K) at the DRβ71 position.^20,21 Arginine is a positively charged amino acid that is structurally similar to lysine, and its substitution for lysine should not greatly alter the antigen-binding properties of the class II MHC molecule. By contrast, DRB1*1501 protects against AIH in white North Americans and northern Europeans, and this allele encodes isoleucine (I) instead of leucine (L) at position DRβ67 and alanine (A) instead of lysine (K) at position DRβ71. Alanine is a neutral, nonpolar amino acid that, when substituted for lysine, should greatly affect antigen presentation and immunocyte activation.

Antigenic peptides are selected for display by the nature of the amino acids that interact with residues within the antigen-binding groove (see Fig. 88-4).^21,22 The critical six-amino-acid motif in AIH restricts the range of peptides that can be accommodated. Multiple self-antigens or foreign antigens may satisfy the minimal structural requirements and serve as immunogenic peptides. The ideal triggering epitope must have a negatively charged amino acid residue (aspartic acid or glutamic acid) at peptide position P4 to form a salt bridge with the positively charged lysine or arginine at DRβ71.²¹ Molecular modeling indicates that a negatively charged P4 residue in the antigenic peptide and the positively charged lysine or arginine at DRβ71 can form a P4-DRβ71 immunoreactive unit that is independent of the other residues within the antigen and antigen-binding groove. This minimal immunoreactive unit can be created by multiple antigenic peptides and class II MHC molecules, and the number of these units may affect susceptibility by a “dose effect.”

DRB1*1301 is associated with AIH in Argentine children²² and Brazilian patients^23,24 and encodes ILEDER at positions DRβ67-72. Glutamic acid (E), aspartic acid (D), and glutamic acid (E) are at positions DRβ69, DRβ70, and DRβ71, respectively, in the class II MHC molecule, and the presence of these critically located but negatively charged amino acid residues argues against the “shared motif” hypothesis of pathogenesis. The “molecular footprint” hypothesis of pathogenesis holds that susceptibility to AIH in different regions and ethnic groups relates to indigenous factors or agents favored by certain genetic phenotypes.^20,21 In South America, DRB1*1301 is associated with protracted hepatitis A virus infection, and persons with this allele may be “selected” from their environment to have prolonged exposure to viral and hepatic antigens that favor the development of AIH.²⁵ An understanding of the individual susceptibility allele in different geographic regions may allow use of this “footprint” to track the cause of the disease.

The “autoimmune promoter hypothesis” of pathogenesis complements the “shared motif” and “molecular footprint” hypotheses by proposing that genetic promoters inside and outside the MHC can affect disease occurrence, either in synergy (epistasis) with the principal susceptibility factors or in lieu of them.^18,20,21 Polymorphisms of the tumor necrosis factor (TNF)-α gene (TNFA*2),²⁶ the cytotoxic T lymphocyte antigen 4 gene (CTLA4),²⁷ and Fas gene promoter at position −670 (TNFRSF6)²⁸ have been associated with increased immunoreactivity, disease severity, and early progression to cirrhosis in white North American and northern European patients. Constellations of autoimmune promoters, as yet undefined, may individualize the disease by affecting its occurrence, clinical phenotype, and outcome.

Liver cell destruction is accomplished by either cell-mediated cytotoxicity, antibody-dependent cell-mediated cytotoxicity, or a combination of both mechanisms (see Fig. 88-4).^16–18 Cell-mediated cytotoxicity depends on the clonal expansion of CD8⁺ cytotoxic T cells that accomplish liver cell injury through the release of lymphokines. This mechanism is regulated by type 1 cytokines, and the −308 polymorphism of TNFA*2 may facilitate this pathway.²⁶ Antibody-dependent cell-mediated cytotoxicity is regulated by type 2 cytokines, and the natural killer cell accomplishes liver cell destruction by binding of its Fc receptor with an antigen-antibody complex on the hepatocyte surface.^16–18 The predominant mechanism depends on the phenotypic differentiation of the CD4⁺ helper T cell, which in turn reflects the cytokine milieu. The cytokine milieu may reflect polymorphisms of the cytokine genes that favor excessive production of some modulators, such as TNF-α, or deficient production of others.

Defects in the counter-regulatory cytokine milieu may also reflect reduced numbers of intrahepatic natural killer T (NKT) cells and the failure of T-regulatory (T-reg) cells (CD4⁺CD25⁺ cells) to modulate CD8⁺ T cell proliferation and cytokine production.^18,29 Increased numbers of γδ T cells may also contribute to the cytodestructive process by recognizing antigens presented by the non-classic MHC molecules, and the recruitment and intrahepatic trafficking of cytotoxic T lymphocytes may be enhanced by the up-regulation of chemokines, such as CXCL16.¹⁸ Fibrogenesis, in turn, is fueled by the resulting inflammatory activity as perivascular hepatic stellate cells transform into myofibroblasts. The matrix proteins accumulate as tissue inhibitors retard the counteractive degradative actions of matrix metalloproteinases, and stellate cells continue to be activated in an autocrine fashion by transforming growth factor-β (TGF-β).³⁰ Glucocorticoid therapy can favorably alter the cytokine milieu, improve the number and function of the T-reg cells, impair activation of TGF-β, promote disappearance of the metalloproteinase inhibitors, enhance degradation of the fibrotic liver matrix, and foster apoptosis of the hepatic stellate cells (see also Chapter 2).³¹

CLASSIFICATION

Two types of AIH have distinctive serologic profiles. Neither has been ascribed a unique cause, specific treatment strategy, or special type of behavior (Table 88-3). The terms are useful as clinical descriptors and as research designations to ensure homogeneous study populations.

Table 88-3 Classification of Autoimmune Hepatitis Based on Autoantibodies

APECED, autoimmune polyendocrinopathy–candidiasis–ectodermal dystrophy; CYP2D6, cytochrome P450 2D6; pANCA, perinuclear anti-neutrophil cytoplasmic antibodies.

* Autoantibodies with an asterisk are investigational only and not available for routine clinical use.

VARIANT FORMS

Patients who have atypical features of AIH currently lack an official designation and confident treatment strategy.⁴¹ They may have manifestations of AIH and another type of chronic liver disease or findings that are incompatible with AIH by current diagnostic criteria (Table 88-4).^42,43

AUTOIMMUNE HEPATITIS AND CHRONIC HEPATITIS C

Eight percent of white North American adults with classic AIH have concurrent infection with hepatitis C virus (HCV), and 52% of patients with chronic hepatitis C have autoantibodies, concurrent immune diseases, or both.⁵⁶ Identification of these patients as belonging to one or the other group is important because interferon therapy can enhance the immune manifestations of persons with AIH and concurrent HCV infection, and immunosuppressive treatment can increase serum viral levels in persons with chronic hepatitis C and background autoimmune features.

The nature and degree of the associated immune manifestations distinguish AIH with coincidental HCV infection from chronic hepatitis C with autoimmune features (Fig. 88-5).^56,57 Concurrent immune diseases that reflect a cell-mediated response against autoantigens (autoimmune thyroiditis, Graves’ disease, inflammatory bowel disease) typify an autoimmune process and occur more commonly in AIH than in chronic hepatitis C.^56,57 They also are more likely to have clinical consequences if the immune diseases are exacerbated during antiviral therapy. Serum autoantibody titers can increase during antiviral treatment, but this increase is neither pathogenic nor associated with clinical deterioration. Patients with chronic hepatitis C and concurrent autoimmune features have immune complex diseases (vasculitis, glomerulonephritis, and symptomatic cryoglobulinemia) more frequently than do patients with classic AIH, and these findings support a predominantly viral disease. The immune manifestations are driven mainly by viral antigen.⁵⁶

Figure 88-5. Key features in determining the predominant autoimmune or viral nature of autoimmune hepatitis in patients with with concurrent hepatitis C infection. Autoimmune-predominant disease is characterized by high-titer (≥1 : 320) autoantibodies (Autoab), multiple autoantibody species, concurrent autoantigen (Autoag)-driven immune disease (autoimmune thyroiditis, Graves’ disease, or ulcerative colitis), and histologic features of interface hepatitis and portal plasma cell infiltration. Virus-predominant disease is characterized by low-titer autoantibodies, single autoantibody expression, concurrent immune complex diseases, and histologic features of portal lymphoid aggregates and hepatic steatosis. Immune complex diseases may include vasculitis, glomerulonephritis, and symptomatic cryoglobulinemia.

Patients with classic AIH typically express multiple antibodies in titers that exceed 1 : 320.⁵⁶ By contrast, patients with chronic hepatitis C typically express one type of autoantibody (ANA or SMA) in titers that are usually less than 1 : 320. Furthermore, patients with AIH have higher serum levels of AST, gamma globulin, and IgG than do patients with chronic hepatitis C. High serum titers (greater than 1 : 320) of autoantibodies, marked hypergammaglobulinemia, multiple autoantibody types, or the presence of autoantigen-driven concurrent immune diseases indicates an autoimmune-predominant syndrome. Further differentiation requires liver biopsy evaluation.

Patients with classic AIH more commonly have severe interface hepatitis, moderate to severe portal plasma cell infiltration, and panacinar hepatitis in liver biopsy specimens than do patients with typical chronic hepatitis C.⁵⁷ By contrast, patients with typical chronic hepatitis C have a higher frequency of portal lymphoid aggregates and steatosis than is seen in patients with classic AIH (see also Chapter 79). Histologic diagnoses based on these predominant patterns have high specificity (81% and 91%) and predictability (62% and 82%) for AIH and chronic hepatitis C, respectively.⁵⁸ Their sensitivity for each clinical diagnosis, however, is low (40% and 57%, respectively). Treatment of patients with mixed features of AIH and chronic hepatitis C should be appropriate for the predominant disease and must be based on the nature of concurrent immune diseases, number and titers of associated autoantibodies, and histologic pattern. Combined therapies with glucocorticoid and antiviral drugs should be avoided.

EPIDEMIOLOGY

The incidence of AIH among white northern Europeans is 1.9 cases per 100,000 persons per year, and its point prevalence is 16.9 cases per 100,000 persons per year.⁵⁹ In the United States, AIH affects 100,000 to 200,000 persons and accounts for 2.6% of the transplantations in the European Liver Transplant Registry and 5.9% in the National Institutes of Health Liver Transplantation Database. The frequency of AIH among patients with chronic liver disease in North America is between 11% and 23%.

The impact of genetic risk factors must be considered in assessing the occurrence of disease in different regions. The prevalence of AIH is greatest among northern European white persons who have a high frequency of HLA-DR3 and HLA-DR4, and AIH is found with similar frequency in the derivative populations of North America and Australia. The Japanese have a low frequency of HLA-DR3, and AIH in Japan is associated with HLA-DR4.^20,21 All populations are susceptible to AIH, which has been described in African Americans, South Americans, Arabs, Japanese, mainland Chinese, and subcontinental Indians. The prevalence of AIH among Alaskan natives (43 per 100,000 population) is higher than that reported in a white Norwegian population (16.9 per 100,000).⁶⁰

PROGNOSIS INDICES

The prognosis for AIH relates mainly to the severity of liver inflammation at the initial medical consultation, as reflected in the laboratory indices and the histologic findings. HLA status and ethnicity influence disease occurrence, clinical phenotype, and treatment outcome (Table 88-5). Certain serologic markers have shown promise as prognostic instruments, and the Model for End-stage Liver Disease (MELD) score may be useful for identifying problematic patients early.

CLINICAL FEATURES

The clinical features of AIH frequently reflect the inflammatory activity of the liver disease or the complications of cirrhosis (Table 88-6).^1,7 Cholestatic features may be present but do not dominate the clinical picture. Similarly, manifestations of liver decompensation, such as ascites, hepatic encephalopathy, and variceal bleeding, are uncommon findings at the initial medical consultation.

Table 88-6 Clinical Features of Autoimmune Hepatitis

AIH, autoimmune hepatitis; ANA, antinuclear antibodies; anti-LKM1, antibodies to liver-kidney microsome type 1; pANCA, perinuclear anti-neutrophil cytoplasmic antibodies; SMA, smooth muscle antibodies.

* Autoantibodies with an asterisk are investigational or are not available for routine clinical use.

Ready fatigability is the most common symptom (seen in 86% of adult patients) (see Table 88-6). Weight loss is uncommon, and intense pruritus argues against the diagnosis. Hepatomegaly is the most common physical finding (78%), and jaundice is found in 69% of patients. Splenomegaly can be present in patients with and without cirrhosis (56% and 32%, respectively), as can spider angiomata. As many as 34% of patients are asymptomatic at initial consultation, and 25% of adults have a normal physical examination.⁷ The discordance between the severity of inflammatory activity and the presence of symptoms is most common in children, whose clinical status frequently does not accurately reflect the severity of the underlying liver disease.

Hyperbilirubinemia is present in 83% of patients, but the serum bilirubin level is greater than three times the upper limit of normal in only 46%.⁴⁵ Similarly, the serum alkaline phosphatase level commonly is increased (81%), but elevations are two times the upper limit of normal in 33% and four times the upper limit of normal in only 10% of patients (see Table 88-6).⁴⁵

The hypergammaglobulinemia of AIH is polyclonal; the IgG fraction predominates. Paraproteins are common, and patients may have diverse, nonspecific serologic findings, including antibodies to bacteria (Escherichia coli, Bacteroides, and Salmonella species) and viruses (measles virus, rubella virus, and cytomegalovirus). Cryoglobulinemia may be present, but symptomatic cryoglobulinemia is rare.

Concurrent immunologic diseases are common and involve diverse organ systems, most frequently the thyroid.⁷ SMA, ANA, and anti-LKM1 are required for the diagnosis, and other autoantibodies may be present, as shown in Table 88-6,⁷ although these other autoantibodies do not have routine clinical applications.

TREATMENT

* Adapted from Czaja AJ. Safety issues in the management of autoimmune hepatitis. Expert Opin Drug Saf 2008; 7:319-33.

Treatment with azathioprine can be complicated by cholestatic liver disease, nausea, emesis, rash, pancreatitis, arthralgias, opportunistic infections, and cytopenias.⁷⁷ Five percent of patients treated with azathioprine develop early adverse reactions (nausea, vomiting, arthralgias, fever, skin rash, or pancreatitis), which warrant its discontinuation. The overall frequency of azathioprine-related side effects in patients treated with 50 mg daily is 10%, and the side effects typically improve after the dose is reduced or the therapy is discontinued.⁷⁷ Cytopenia is the most common consequence of treatment; bone marrow failure is rare. Cytopenia occurs in 46% of patients, and severe hematologic abnormalities occur in 6%.^77,78 These toxicities are not predictable by either genotyping or phenotyping for TPMT activity,^78–80 and the most common association with cytopenia in these patients is cirrhosis and presumed hypersplenism associated with portal hypertension (see Table 88-9).^78,80

Teratogenicity is a theoretical complication of therapy with azathioprine.⁷⁷ Azathioprine has been administered successfully in pregnant women with AIH, pregnant mothers with inflammatory bowel disease, and women who have conceived while taking azathioprine after liver transplantation.⁷⁷ Nevertheless, azathioprine has been associated with congenital malformations in pregnant mice, and these defects have included cleft palate, skeletal anomalies, hydrops fetalis, reduced thymic size, anemia, and hematopoietic depression. Furthermore, the placenta is only a partial barrier to the metabolites of azathioprine, and low levels of 6-thioguanine are detectable in the newborns of mothers treated for Crohn’s disease.⁷⁷ The odds ratio of having a child with congenital malformations while taking azathioprine for inflammatory bowel disease is 3.4. Azathioprine is not an essential medication in the treatment of AIH and can be discontinued during pregnancy, in which case the liver disease can be managed successfully by adjustments in the dose of prednisone.

Oncogenicity is another possible complication of therapy with azathioprine.⁷⁷ The incidence of extrahepatic neoplasms is 1 per 194 patient-years; the probability of tumor occurrence is 3% after 10 years; and the risk of malignancy is 1.4-fold greater than normal.⁸¹ The low but increased risk of malignancy does not contraindicate azathioprine therapy in AIH but emphasizes the importance of maintaining strict indications for treatment.

Blood TPMT activity is significantly lower in patients with AIH and intolerance to azathioprine than in patients with uncomplicated courses of treatment.⁷⁹ Similar findings have been described in patients with inflammatory bowel disease and rheumatic conditions (see Chapter 111). These observations have suggested that routine screening of the blood TPMT level may identify persons with AIH at risk for azathioprine-related complications. The blood TPMT activity level, however, has not been predictive of toxicity in individual patients.^79,80 Genotypic and phenotypic screening for blood TPMT activity has not reduced the frequency of azathioprine-induced side effects in patients with AIH compared with unscreened patients,^78–80 nor has the occurrence of azathioprine-related side effects been associated with below-normal levels of TPMT activity.⁷⁸ Near-zero enzyme activity occurs rarely in otherwise normal individuals (0.3% to 0.5%), and the value of screening to detect this unusually low enzyme deficiency remains uncertain, especially if some patients with low levels do not exhibit azathioprine toxicity.^77,78 Testing for blood TPMT activity seems most appropriate in patients with pre-existing or progressive cytopenias and in those subjected to doses of azathioprine higher than the conventional schedule of 50 mg daily. Avoidance of azathioprine in patients with pre-existing or progressive cytopenias (leukocyte count below 2.5 × 10⁹/L or platelet count below 50 × 10⁹/L) and close monitoring (at three-month intervals) of the blood leukocyte and platelet counts in all patients taking the drug may be the best preventive strategy.^77,78

END POINTS

Glucocorticoid therapy is continued until remission, treatment failure, incomplete response, or drug toxicity occurs (Fig. 88-6).^1,75 Remission connotes absence of symptoms, resolution of all laboratory indices of active inflammation, and histologic improvement to normal liver tissue or inactive cirrhosis.⁸² Evaluation of liver tissue before drug withdrawal is essential to establish remission because histologic activity may be present in 55% of patients who satisfy other requirements for remission. Typically, histologic improvement lags behind clinical and laboratory resolution by three to eight months, and treatment should be extended for at least this period to compensate for the lag. Patients with normal serum AST and gamma globulin levels and normal liver biopsy findings immediately before drug withdrawal have a significantly lower frequency of relapse after drug withdrawal than patients with near-normal test levels and histology (60% vs. 90%, P < 0.001).⁸² Only 40% of treated patients, however, are able to improve fully, and complete resolution of the clinical and histologic manifestations of the disease does not preclude relapse after drug withdrawal.⁸²

Figure 88-6. Treatment algorithm for autoimmune hepatitis. Patients who satisfy indications for glucocorticoid therapy are given prednisone in combination with azathioprine or a higher dose of prednisone alone (see Table 88-8). Treatment is continued until the criteria for a treatment end point are met. Possible end points are remission, treatment failure, incomplete response, and drug toxicity. Therapy can then be discontinued, increased in dose, or reduced in dose. Responses to the dose adjustments determine the need for other actions. The principal indices of inflammatory activity are serum aspartate aminotransferase (AST) and gamma globulin levels.

Treatment failure connotes deterioration during therapy (see Fig. 88-6).^1,69,75 It is characterized by worsening of the serum AST or bilirubin levels by at least 67% of previous values, progressive histologic activity, or onset of ascites or encephalopathy. Conventional glucocorticoid therapy should be stopped, and a high-dose regimen should be instituted.

Incomplete response connotes improvement that is insufficient to satisfy remission criteria (see Fig. 88-6).^1,75 Failure to achieve remission within three years indicates that remission is unlikely and warrants discontinuation of conventional treatment.

Drug toxicity justifies premature withdrawal of medication or a reduction in dose (see Fig. 88-6).^1,75,77 Most side effects are reversible, and some consequences such as cataracts and osteopenia with vertebral compression have effective therapies. Weight gain, acne, edema, and diabetes may be consequences of the disease rather than the drugs.

RESULTS

Prednisone alone or in combination with azathioprine induces clinical, biochemical, and histologic remission in 65% of patients within three years.^1,75 The average treatment interval until remission is 22 months. Therapy improves survival. The 10-year life expectancies following treatment for patients with and without cirrhosis at the time of the initial medical consultation are 89% and 90%, respectively. The overall 10-year survival rate is 93% and is comparable to that of an age- and sex-matched cohort from the population at large (94%). Patients who have histologic cirrhosis respond as well as do noncirrhotic patients and should receive similar treatment, with the same expectation of success. Twenty-one percent of patients managed with conventional regimens of prednisone alone or in combination with azathioprine sustain their remission for a median of 76 months after drug withdrawal, and an effort should be made to withdraw all patients from initial therapy after criteria for remission have been satisfied.⁸³

Treatment with glucocorticoids also may reduce hepatic fibrosis.⁸⁴ Fibrosis scores have improved in 56% of patients followed for 55 ± 9 months, and fibrosis did not progress in 33% of patients followed for 62 ± 14 months. Histologic activity indices decreased concurrently, and patients in whom the histologic activity indices improved had a higher frequency of improvement in the fibrosis scores (80% vs. 25%, P = .002). These findings suggest that improvement in hepatic fibrosis occurs in conjunction with reduction in liver inflammation and that glucocorticoids may facilitate the disappearance of fibrosis by suppressing inflammatory activity.^30,84 Small case studies have also suggested that cirrhosis can disappear during treatment, but this possibility must await confirmation by assessments more reliably reflective of cirrhosis than conventional needle biopsy of the liver.

RELAPSE

Patients who enter remission commonly experience an exacerbation after drug withdrawal.^1,75,83 Relapse is defined as the reappearance of histologic disease after discontinuation of drug therapy. An increase in the serum AST level to more than three-fold the upper limit of normal after discontinuation of drug therapy is invariably associated with interface hepatitis on histologic examination. This biochemical change is sufficient to diagnose relapse without requiring liver tissue assessment.

Relapse occurs in 50% of patients within six months of discontinuing therapy, and most patients (70% to 86%) experience an exacerbation within three years.^82,83 Reinstitution of the original treatment induces another remission, but relapse commonly recurs after termination of therapy. Repeated relapse and re-treatment is associated with a cumulative morbidity and mortality. Cirrhosis develops more commonly (38% vs. 4%, P = 0.004); death from hepatic failure or need for liver transplantation occurs more often (20% vs. 0%, P = 0.008)⁸⁵; and drug-induced side effects are more frequent (70% vs. 30%, P = 0.01) in persons who relapse than in those who sustain remission after drug withdrawal.⁷⁷ The frequencies of each complication increase with each subsequent relapse and re-treatment. The optimal time to interrupt this sequence is after the first course of treatment and relapse, and the preferred treatment strategy is to implement maintenance therapy with azathioprine.⁸⁶

After the initial relapse, therapy with prednisone and azathioprine is re-started and continued until clinical and laboratory resolution is achieved again (see Fig. 88-6).⁸⁶ The dose of azathioprine is then increased to 2 mg/kg daily as the dose of prednisone is reduced. Azathioprine is then continued indefinitely as a chronic maintenance regimen. Eighty percent of patients are able to sustain remission in this fashion over a 10-year period of observation. Azathioprine-induced cytopenias compel dose reduction in 9% of patients; glucocorticoid-related side effects improve; and arthralgias associated with glucocorticoid withdrawal eventually resolve (but may be protracted). Malignancies of various types develop in 7% of patients, but an association with azathioprine is disputed.⁸⁶

An alternative management strategy after the first relapse is to maintain suppression of inflammatory activity using daily, low-dose prednisone (see Fig. 88-6).⁸⁷ Eighty-seven percent of patients can be managed long term on prednisone at less than 10 mg per day (median dose, 7.5 mg per day). The dose is titrated to the lowest level needed to prevent symptoms and to maintain serum AST levels below three times the upper limit of normal. Side effects attributable to previous glucocorticoid therapy resolve in 85% of patients; the immediate survival rate is comparable to that for persons who have relapsed and been treated with a conventional regimen of prednisone alone or a lower dose of prednisone and azathioprine (91% vs. 90%), and new complications do not occur. The azathioprine and low-dose prednisone maintenance regimens have not been compared directly, but the azathioprine schedule has intuitive appeal because it eliminates the need for prednisone.

Treatment after relapse does not need to be indefinite.⁸³ The probability of ultimately achieving inactive disease that does not require continuous drug therapy is 28%. This possibility justifies an effort to withdraw treatment in all patients who have relapsed previously and who have stably maintained inactive disease on maintenance therapy for at least one year. This effort can be repeated if earlier withdrawal attempts have failed.⁸³

TREATMENT FAILURE

Nine percent of patients deteriorate during glucocorticoid therapy (treatment failure) (see Fig. 88-6).^1,69,75 High doses of prednisone alone (60 mg per day) or prednisone (30 mg per day) in conjunction with azathioprine (150 mg per day) constitute standard treatment in this group.^1,75,88 Each schedule induces clinical and biochemical improvement in 70% of patients within two years. Histologic resolution, however, occurs in only 20%, and long-term therapy is frequently necessary. These patients are at risk of liver failure and serious drug toxicity. Liver transplantation must be considered at the first sign of hepatic decompensation. The development of ascites typically heralds the need for a transplant evaluation (see Table 88-5).

Alternative management strategies for treatment failure have included the administration of cyclosporine, tacrolimus, mycophenolate mofetil, ursodeoxycholic acid, budesonide, 6-mercaptopurine, methotrexate, and cyclophosphamide.^75,89 In each instance, experience has been limited, and in most reports the preliminary results have been encouraging but uncorroborated. Among the new drugs used in treatment failure, only ursodeoxycholic acid has been evaluated by a randomized controlled clinical trial, and the findings did not support its use.⁷⁵ Site-specific interventions will be possible as soon as the pathogenic mechanisms are clarified.⁷⁵ These therapies may include peptides to block autoantigen display within the class II MHC molecules, agents such as CTLA4 to temper immunocyte response, T-cell vaccination, oral tolerance regimens, cytokine manipulations, mesenchymal stem cell transplantation, and gene therapies that can offset the over-expression of certain cytokines, limit fibrosis, and promote regeneration.

INCOMPLETE RESPONSE

In 13% of patients, clinical improvement is observed during therapy without achieving remission criteria (see Fig. 88-6).^1,75 The diminishing benefit-to-risk ratio of protracted therapy justifies an alternative strategy. The administration of azathioprine (2 mg/kg daily) as the sole drug⁸⁶ or a low-dose prednisone regimen⁸⁷ is a reasonable approach. The goal of treatment is to reduce and stabilize disease activity on a drug schedule that is well tolerated.

DRUG TOXICITY

For patients with serious side effects from therapy, treatment usually can be continued with the single tolerated drug (prednisone or azathioprine) in an adjusted dose (see Fig. 88-6).^1,75,77 Cyclosporine, 6-mercaptopurine, and cyclophosphamide also have been used successfully after drug toxicity in isolated cases.^1,75

LIVER TRANSPLANTATION

Liver transplantation is effective in the treatment of decompensated AIH (see Chapter 95).^1,75 Five-year survival rates for patients and grafts range from 83% to 92%, and the actuarial 10-year survival rate after transplantation is 75%. AIH recurs in at least 17% of patients, and AIH develops de novo in 3% to 5% of patients who undergo transplantation for nonautoimmune liver disease. Acute rejection, glucocorticoid-resistant rejection, and chronic rejection occur more commonly in patients undergoing transplantation for AIH than for other conditions, and patients with AIH are more difficult to withdraw from glucocorticoids.⁹⁰

Recurrent AIH typically is mild and develops in patients who are inadequately immunosuppressed.⁸⁹ Dose adjustments usually are sufficient to suppress the disease, but progression to cirrhosis and graft failure have been reported. De novo AIH is a clinical syndrome that affects both children and adults who undergo transplantation for nonautoimmune liver disease.⁹⁰ Patients with de novo AIH in whom glucocorticoid therapy fails, experience worsening hepatic fibrosis with possible graft loss, and those who do not receive glucocorticoids progress to cirrhosis, require retransplantation, or die of liver failure. De novo AIH in some adults has been associated with severe centrilobular necrosis, and adult patients have been reported to express an atypical anti-liver-kidney cytosolic antibody of uncertain pathogenic significance.⁹⁰ AIH should be included in the differential diagnosis of allograft dysfunction in all transplant recipients