In this report, the nature and importance of the variant forms of autoimmune hepatitis are discussed, and treatment options for these conditions are reviewed. Patients with variant disease are a consequence of progress in the classification of chronic liver diseases, and their number is destined to burgeon as the use of screening at the time of presentation is expanded and diagnostic criteria are rigidly applied.

Methods

Top Methods Summary Author & Article Info References

All English-language articles published between 1990 and 1995 that addressed the diagnosis of autoimmune hepatitis, atypical manifestations of the disease, and variations in response to conventional corticosteroid therapy were identified by a search of the MEDLINE database. Search terms included autoimmune hepatitis, primary biliary cirrhosis, primary sclerosing cholangitis, autoimmune cholangitis, chronic hepatitis B and C with autoimmune features, cryptogenic chronic hepatitis, and overlap syndromes. Each identified title and abstract were reviewed by the author. Articles that focused on diagnostic criteria for autoimmune hepatitis and those that discussed atypical presentations, mixed features, concurrent diseases, and overlap syndromes were scrutinized further. Complete copies of each of these articles were obtained and read by the author. Additional articles of potential relevance were identified from the bibliographies of the pertinent articles. Articles that reported substantial clinical experience or provided novel insights were abstracted and recorded in a computer database.

An extensive personal library of journals and reprints, maintained since 1975 and dedicated to the study of autoimmune liver diseases, was also searched. Pertinent articles found in this library were abstracted and entered in the database. A total of 548 articles were selected.

Autoimmune Hepatitis

Autoimmune hepatitis is an unresolving inflammation of the liver of unknown cause. It is characterized by the presence of hepatitis (piecemeal necrosis or interface hepatitis) that is periportal or more widespread on histologic examination, hypergammaglobulinemia, and autoantibodies in serum [5, 6] (Figure 1). The panels of the International Autoimmune Hepatitis Group, the World Congresses of Gastroenterology, and the International Association for the Study of the Liver have together standardized the definition of the disease [7-9], and the International Autoimmune Hepatitis Group [3] has promulgated criteria for definite and probable diagnosis.

View larger version (166K): [in this window] [in a new window]   Figure 1. Histologic features of autoimmune hepatitis. Mononuclear cells expand the portal tract, disrupt the limiting plate, and infiltrate the periportal region (ecemeal necrosis or interface hepatitis) (arrow). Foci of inflammatory cells are also present in the sinusoids of the lobule (lobular hepatitis). Plasma cells are an important component of the portal infiltrate. (Hematoxylin and eosin stain; original magnification, x 200.).

The definite diagnosis of autoimmune hepatitis requires the presence of compatible laboratory and histologic findings Table 1, the absence of viral markers, denial of excessive alcohol consumption, lack of exposure to blood products and hepatotoxic medications, and the presence of hypergammaglobulinemia ( globulin level > 1.5 times the upper limit of normal) Table 1 [3]. Lobular hepatitis is part of the histologic spectrum of autoimmune hepatitis if it is present in conjunction with periportal hepatitis Figure 1, but biliary lesions, copper deposits, and other changes suggestive of another cause of lobular hepatitis preclude the diagnosis of autoimmune hepatitis [3]. Antinuclear antibodies, smooth-muscle antibodies, or antibodies to liver/kidney microsome type 1 are required for the diagnosis and must be present in titers of at least 1:80 in adults and 1:20 in children Table 1 [3]. The definite diagnosis of autoimmune hepatitis requires satisfaction of all criteria shown in Table 1.

Findings that are compatible with autoimmune hepatitis but insufficient for a definite diagnosis of this condition justify the designation "probable autoimmune hepatitis" Table 1 [3]. Patients who receive this designation typically have histories of alcohol consumption or exposure to hepatotoxic drugs or have serum levels of globulin or autoantibody titers (or both) that are lower than those required for a definite diagnosis. Patients who are seronegative for antinuclear antibodies, smooth-muscle antibodies, and antibodies to liver/kidney microsome type 1 are also considered to have probable autoimmune hepatitis if they have other liver-associated autoantibodies of a less conventional nature, such as antibodies to asialoglycoprotein receptor, soluble liver antigen, liver-pancreas, or liver cytosol type 1 [3]. Patients with probable diagnoses must satisfy histologic criteria that are as rigid as those for patients with definite diagnoses (Table 1); (Figure 1). In this context, predominant bile duct lesions and cholestatic features preclude the diagnosis [3]. Patients who are deficient in one or more of the criteria for a definite diagnosis must be assigned a probable diagnosis (Table 1).

A scoring system has also been proposed by the International Autoimmune Hepatitis Group [3] for the definite and probable diagnoses of autoimmune hepatitis (Table 2). This system promises to provide a template for the systematic assessment of autoimmune hepatitis and its variants. Each manifestation of the disease is graded in accordance with its presumed importance in establishing the diagnosis, and a composite score before and after corticosteroid therapy is calculated [3]. The final score reflects the diagnostic effect of consistent and inconsistent features and the net strength of the diagnosis. By grading the manifestations of each variant condition, the similarity of the variant to autoimmune hepatitis can be determined and the degrees of difference between variants can be evaluated. In this way, the variant can acquire a clinical identity that can affect disease management, facilitate classification schemes, and promote investigative studies and treatment trials. The scoring system requires validation in a prospective trial before it can be applied with confidence in clinical settings [10].

Autoimmune Variants

Patients with variant forms of autoimmune hepatitis have autoimmune features but do not satisfy the criteria for a definite or probable diagnosis of autoimmune hepatitis. In addition, they cannot be placed in another diagnostic category. Such patients have features associated with both autoimmune hepatitis and another type of chronic liver disease (overlap syndromes) or have findings that are incompatible with the diagnosis of autoimmune hepatitis according to the criteria codified by the international panels (outlier syndromes) (Table 3).

Implicit in the diagnosis of variant disease is a resemblance of the condition to autoimmune hepatitis [4, 5]. The variant condition is typically indolent and characterized by nonspecific systemic symptoms, including fatigue, arthralgias, and myalgias. A hepatitic biochemical profile coexists with or predominates over cholestatic laboratory changes, and extreme pruritus, xanthelasma, and hyperpigmentation are unusual. Patients may be of either sex and of any age, but they are commonly women 40 years of age or younger. Histologic findings are consistent with the diagnosis of autoimmune hepatitis, featuring periportal hepatitis with or without lobular hepatitis, but unusual morphologic features are often seen, including bile duct injury, steatosis, and portal lymphoid aggregates [4]. The results of liver biopsy are not diagnostic of an autoimmune variant, and they usually show only atypicalities that must be placed in the proper clinical context. Patients with autoimmune variants are either the result of overly precise diagnostic criteria, examples of the blurring of diagnostic distinctions at the fringes of disease categories, or reflections of concurrent disorders in an unfortunate host. As yet, no variant form of autoimmune hepatitis has been established as a separate entity. In the experience at the Mayo Clinic, at least 30% of adult patients in whom features of autoimmune hepatitis predominate have mixed or inconsistent findings that justify the designation of their conditions as variant forms (Figure 2).

View larger version (30K): [in this window] [in a new window]   Figure 2. Frequency of features shared by autoimmune hepatitis and other chronic liver diseases. Eight percent of patients with autoimmune hepatitis have antibodies to the E2 subunits of pyruvate dehydrogenase complex that are specific for primary biliary cirrhosis (PBC); 6% have histologic changes associated with cholangitis that suggest primary sclerosing cholangitis (PSC); and 13% have features that are typical of autoimmune hepatitis but lack autoantibodies (cryptogenic chronic hepatitis). Eleven percent of patients with predominant autoimmune features have hepatitis C virus RNA in serum, and they resemble patients with chronic viral hepatitis. In contrast, 10% of patients with these features have a fourfold or greater elevation of the serum alkaline phosphatase level, normal cholangiograms, and findings associated with autoimmune cholangitis.

Overlap Syndromes

Concurrent Features of Primary Biliary Cirrhosis

Autoimmune hepatitis can be distinguished from primary biliary cirrhosis with confidence in 92% of cases [11, 12]. The two disorders have distinctly different patterns of necroinflammation that are associated with characteristic laboratory and immunoserologic profiles. In 8% of patients, however, the patterns of hepatitis and cholestasis are not discriminating characteristics, and the disease of these few patients constitutes a variant syndrome. The most common findings in these patients are mitochondrial antibodies and histologic features of cholangitis (Table 3).

Mitochondrial antibodies are found by indirect immunofluorescence in 20% of patients with autoimmune hepatitis [11]. Titers of these autoantibodies characteristically range from 1:5 to 1:1280, and 88% of seropositive patients have titers of 1:160 or less [11]. The low titer reactivity is a background epiphenomenon that has no diagnostic relevance when associated with other features of autoimmune hepatitis [11]. Indeed, reinterpretation of liver biopsy specimens obtained from these patients has confirmed that 92% had histologic changes associated with autoimmune hepatitis [11]. Furthermore, the response to corticosteroid therapy is similar in patients with low titers of mitochondrial antibodies and in patients without mitochondrial antibodies: These patients have remission in 71% of cases [11].

The nonspecificity of low titers of mitochondrial antibodies has been further shown by assessments of immunoreactivity against the autoantigens specific for primary biliary cirrhosis. The mitochondrial antibodies detected in patients with features of both autoimmune hepatitis and primary biliary cirrhosis have a low frequency of reactivity against the trypsin-sensitive antigen on the inner mitochondrial membrane, which is specific for primary biliary cirrhosis [13-15]. These observations have been supplemented by immunoassays for antibodies to the M2 autoantigens of primary biliary cirrhosis [16-19]. Use of these assays has shown that no more than 8% of patients with autoimmune hepatitis have antibodies to the E2 subunits of the pyruvate dehydrogenase complex [12, 18, 20, 21]. Indeed, in one study [22], none of 30 patients with chronic hepatitis and mitochondrial antibodies had these antibodies.

Seropositivity for mitochondrial antibodies may also be a false-positive result. The patterns of mitochondrial antibodies and antibodies to liver/kidney microsome type 1 on indirect immunofluorescence are similar and can be confused [23]. Seroreactivity against the distal tubules of the murine kidney and the parietal cells of the murine stomach indicates the presence of mitochondrial antibodies. In contrast, seroreactivity against the proximal tubules of the murine kidney and the cytoplasm of murine hepatocytes indicates the presence of antibodies to liver/kidney microsome type 1 [23]. An exuberant pattern of immunofluorescence on the renal tubule can obscure distinctions and render the interpretation indeterminate or false. Retrospective analyses of patients with autoimmune hepatitis and mitochondrial antibodies [23] have indicated that this confusion has occurred in at least 27% of patients. An awareness of serologic misclassification and the availability of immunoassays for autoantibodies specific for the diagnosis of primary biliary cirrhosis have minimized this problem.

It is important to note that the specificity of the immunoreaction for mitochondrial antibodies is sometimes undeniable [12, 24]. Patients with this trait represent a true serologic overlap between autoimmune hepatitis and primary biliary cirrhosis. They may reflect the diagnostic insufficiency of the immunoserologic markers, concurrent disorders, an original misdiagnosis, or a transitional stage in the evolution of primary biliary cirrhosis. An empirical 3- to 6-month trial of corticosteroid therapy has been proposed to further characterize the predominant disorder Table 4 [25]. Responsiveness to corticosteroids justifies the presumption that autoimmune hepatitis is the predominant disease [11, 12, 25].

Histologic findings may also suggest concurrent primary biliary cirrhosis. Lymphoid, fibrous, pleomorphic, and destructive cholangitis are ancillary histologic manifestations that raise this specter (Table 3); Figure 3 [26]. Similarly, hepatic copper deposition is present in 19% of patients who have mitochondrial antibodies and clinical features of autoimmune hepatitis [11]. This histologic manifestation of cholestasis is absent in patients without mitochondrial antibodies, and it suggests an overlap variant [11, 27, 28]. In these patients, an empirical treatment trial of corticosteroids can be used to discriminate between primary biliary cirrhosis and autoimmune hepatitis Table 4 [11, 25, 27].

View larger version (159K): [in this window] [in a new window]   Figure 3. Histologic evidence of bile duct injury. Features of destructive cholangitis are present (arrow). A bile duct has been destroyed within the portal tract, and histiocytes indicate its previous location. This injury pattern is atypical of autoimmune hepatitis and suggests a diagnosis of primary biliary cirrhosis or autoimmune cholangitis. (Hematoxylin and eosin stain; original magnification, x 200.).

Concurrent Features of Primary Sclerosing Cholangitis

Patients with autoimmune hepatitis often have chronic ulcerative colitis (16%) and are at risk for concurrent primary sclerosing cholangitis [29-31]. In addition, patients with these diseases have similar HLA-related risk factors (HLA-B8, HLA-DR3, and HLA-DR4) [32-36]; immunocytes that react with antibody-coated target cells through the Fc receptors [37]; and antibodies to liver membrane protein complex in serum [37]. The similarities in genetic predisposition, cytotoxic effector mechanisms, and reactivities to a possible common target antigen facilitate the expression of both diseases in the same patient [29]. In patients with autoimmune hepatitis, the most common findings that suggest an overlap variant are inflammatory bowel disease, features associated with cholestasis, histologic findings of bile duct injury, and poor response to corticosteroid therapy (Table 3).

Cholangiography indicates the presence of primary sclerosing cholangitis in 42% of persons with autoimmune hepatitis and ulcerative colitis [29]. These persons are clinically and biochemically indistinguishable from patients with autoimmune hepatitis, ulcerative colitis, and normal cholangiograms and from patients with autoimmune hepatitis but not ulcerative colitis [29]. Diagnosis in these persons depends entirely on the results of cholangiography.

Cholestatic changes that suggest an overlap syndrome are pruritus, an atypical elevation of the serum alkaline phosphatase level, or both. Abnormal increases in the serum alkaline phosphatase level occur in 81% of patients with autoimmune hepatitis, but the increases are more than twofold greater than normal in only 33% of these patients and more than fourfold greater than normal in only 10% [5]. Patients with autoimmune hepatitis, pruritus, or elevated serum alkaline phosphatase levels at least fourfold greater than normal should be considered to have overlap variant and should be assessed by cholangiography.

Bile duct changes, such as fibrous cholangitis, fibrous obliterative cholangitis, ductopenia, portal edema, and bile stasis, are incompatible with the diagnosis of autoimmune hepatitis Figure 4 [3]. These changes are rare, and when they occur in a patient with autoimmune hepatitis, concurrent inflammatory bowel disease, or atypical elevations of serum alkaline phosphatase levels, they indicate a variant syndrome [29].

View larger version (157K): [in this window] [in a new window]   Figure 4. Histologic evidence of fibrous cholangitis. A concentric ring of fibrous tissue surrounds a bile duct (fibrous cholangitis) (arrow). This change, in conjunction with other clinical, laboratory, and histologic findings associated with autoimmune hepatitis, suggests an overlap with primary sclerosing cholangitis and justifies the use of cholangiography. (Hematoxylin and eosin stain; original magnification, x 100.).

Unfortunately, cholangiography may be insufficient for diagnosis in some cases. Histologic evidence of bile duct injury Figure 4 is more common in patients with autoimmune hepatitis who have ulcerative colitis than in patients with autoimmune hepatitis who do not have ulcerative colitis (41% compared with 0%; P = 0.02) [29]. Cholangiographic changes associated with primary sclerosing cholangitis, however, are absent in 14% of patients with typical histologic disease [29]. This discordance suggests that the overlap syndrome can involve only the intrahepatic bile ducts ("small-duct primary sclerosing cholangitis") [29, 38].

Among patients with autoimmune hepatitis, those who have ulcerative colitis respond less well to corticosteroids than do those without ulcerative colitis. Patients with ulcerative colitis also have remission less frequently (59% compared with 94% [P < 0.05]), have more treatment failures (41% compared with 6% [P < 0.05]), and progress to cirrhosis more often (75% compared with 25% [P < 0.05]) than do patients without inflammatory bowel disease [29]. Those persons with abnormal cholangiograms have the worst outcomes [29]. Demonstration of the characteristic cholangiographic features of primary sclerosing cholangitis in patients who do not respond to corticosteroids justifies discontinuation of treatment [29, 39]. Patients who do not respond to corticosteroids are candidates for investigational protocols, treatment of symptoms, or empirical therapy with ursodeoxycholic acid, if cholestatic symptoms are marked (Table 4).

Concurrent Features of Chronic Viral Hepatitis

Autoimmune hepatitis is by definition a nonviral disease [3]. Infection with one or more of the hepatitic viruses, however, can occur coincidentally [40]. The admixture of autoimmune hepatitis and chronic viral infection constitutes an important overlap syndrome.

Four percent of patients with autoimmune hepatitis are seropositive for antibodies to hepatitis C virus by second-generation immunoassay, and 4% have serologic markers of a previous infection with the hepatitis B virus [41]. Of patients whose presentation is atypical of autoimmune hepatitis or whose response to corticosteroid therapy is suboptimal, 11% have active hepatitis C viremia [42]. Conversely, 11% of patients with chronic viral hepatitis are seropositive for smooth-muscle antibodies and 28% are seropositive for antinuclear antibodies [40, 41, 43, 44]. Additional testing for thyroid antibodies and rheumatoid factor shows that the frequency of autoantibody expression in these patients is 62% [44, 45]. The frequent concurrence of viral infection and autoantibodies not only confounds the diagnosis but also complicates the treatment strategy.

Corticosteroid therapy in patients with predominant viral disease is unlikely to induce improvement and will increase the viral burden of the host [46, 47]. Similarly, interferon therapy in patients with predominant autoimmune features can intensify immune-mediated hepatocellular inflammation [48-50], transform chronic viral hepatitis into autoimmune hepatitis [51, 52], exacerbate a concurrent nonhepatic autoimmune disorder [53-57], or produce a plethora of autoantibodies of uncertain clinical significance [58-61].

Because treating patients who have mixed viral and autoimmune features can be harmful as well as ineffective, weighing all of the viral and autoimmune manifestations in each patient Table 3 is essential to determining the predominant diagnosis and instituting appropriate treatment (Table 4). The strength of immunoserologic reactions and the patterns of histologic injury are the most useful indices with which to separate patients.

Eighty-nine percent of patients with chronic viral hepatitis who are seropositive for smooth-muscle and antinuclear antibodies have serum titers of 1:80 or less [40, 44]. Only 11% of these persons have serum titers of 1:160 or more, and they rarely have titers that exceed 1:320 [40, 44]. In addition, concurrent seropositivity for smooth-muscle and antinuclear antibodies is rare [40, 44, 62]. In contrast, patients with definite autoimmune hepatitis have median serum titers of 1:160 for smooth-muscle antibodies and 1:320 for antinuclear antibodies [40, 43, 44]. Only 6% of these patients have isolated titers less than 1:80, and 60% are seropositive for both types of antibody [40, 43, 44]. By these standards, patients with smooth-muscle or antinuclear antibody titers of 1:320 or greater and persons with concurrent markers, regardless of titer, can be defined as having autoimmune-predominant disease [40]. Patients with true viral infection and either antibodies to liver/kidney microsome type 1 or smooth-muscle or antinuclear antibodies in titers less than 1:320 can be defined as having viral-predominant disease Table 3 [40].

Patients with autoimmune hepatitis have more moderate to severe plasma cell infiltration of the portal tracts (66% compared with 21%), more lobular inflammation (47% compared with 16%), and more piecemeal necrosis (23% compared with 0%) than do patients with chronic hepatitis C Figure 1 [63-65]. In contrast, patients with chronic hepatitis C have more portal lymphoid aggregates (49% compared with 10%), more steatosis (72% compared with 19%), and more bile duct damage or loss (91% compared with 20%) than do patients with autoimmune hepatitis Figure 5 [63-65]. The composite histologic findings for each disease have high specificity (0.81 for autoimmune hepatitis and 0.91 for chronic hepatitis C), and they should be carefully sought for their discriminative value [65]. Unfortunately, they have low sensitivity (0.40 for autoimmune hepatitis and 0.57 for chronic hepatitis C) and often cannot define the predominant disorder [65]. An important finding is that in patients with hybrid findings, the presence of moderate to severe piece-meal necrosis and marked plasma cell infiltration of the portal tracts constitute a compelling argument for an autoimmune-predominant variant (Figure 1). Alternatively, marked steatosis, bile duct changes, or prominent portal lymphoid aggregates compel the diagnosis of a viral-predominant variant Figure 5 [65].

View larger version (161K): [in this window] [in a new window]   Figure 5. Histologic evidence of steatosis and portal lymphoid aggregate. The hepatocytes adjacent to the portal tract contain fat (steatosis). In addition, the mononuclear infiltrate of the portal tract and the disruption of the limiting plate (ecemeal necrosis) are accompanied by a portal lymphoid aggregate (arrow). Steatosis and portal lymphoid aggregate are both atypical for autoimmune hepatitis and suggest a concurrent chronic hepatitis C infection. (Hematoxylin and eosin stain; original magnification, x 200.).

Corticosteroids are an option for patients with autoimmune-predominant disease Table 4 [40, 42, 66]. Treatment is empirical and is based in part on concerns about the deleterious effects of interferon. A 3- to 6-month treatment trial with a fixed maintenance dose of prednisone, 20 mg/d or 10 mg/d, in conjunction with azathioprine, 50 mg/d, is usually well tolerated and safe when monitored closely [40, 42, 66]. Fifty-seven percent of patients have improvement with this regimen, and the duration of treatment can be extended until full remission is achieved [42]. Patients who do not improve or who have relapse after withdrawal of corticosteroid therapy are candidates for interferon treatment. Anecdotal experiences have indicated that these persons can respond to antiviral therapy [46, 66].

Treatment with recombinant interferon, 3 million U subcutaneously three times a week for 6 months, is an option for patients with viral-predominant disease (Table 4). Progress can be evaluated and the benefit-risk ratio of the strategy assessed at 3 months. Exacerbations of the liver disease or the emergence of autoimmune-predominant manifestations compel discontinuation of this therapy.

Unfortunately, the prevalence of serious consequences in patients who receive either of these strategies is unknown. Treatment guidelines have been developed empirically in response to isolated case reports of adverse outcomes that may, in fact, be rare. Indeed, two recent studies [45, 67] have suggested that interferon can be administered with impunity to patients who have autoimmune markers. Other, similar experiences will undoubtedly change current management algorithms.

Outlier Syndromes

Autoimmune Cholangitis

Autoimmune cholangitis is a chronic inflammation of the liver, and patients with this condition have both hepatitic and cholestatic features. It is characterized by the presence of antinuclear or smooth-muscle antibodies and histologic evidence of bile duct injury (Table 3); Figure 3 [68, 69]. It has variously been designated as "mitochondrial antibody-negative primary biliary cirrhosis" [70-72] or as a variant of autoimmune hepatitis [69]. It cannot be classified under either rubric, however, because 1) the cholestatic findings associated with it preclude the diagnosis of autoimmune hepatitis [3] and 2) its lack of mitochondrial antibodies restricts its categorization as primary biliary cirrhosis [73]. Hence, it remains an outlier syndrome. Formally classifying this condition depends on distinguishing it from the disorders that it resembles. Comparisons of this condition with primary biliary cirrhosis [71-73] and autoimmune hepatitis [69] have yielded discrepant results, whereas preliminary studies [74] of the production of antibodies to carbonic anhydrase have suggested that autoimmune cholangitis is a distinct disorder.

Patients with autoimmune cholangitis are more often positive for antinuclear antibodies (100% compared with 18%; P < 0.001) and smooth-muscle antibodies (41% compared with 6%; P = 0.03) than are patients with primary biliary cirrhosis [72]. They also have lower levels of serum aspartate amino-transferase and IgM and a lower frequency of moderate to severe portal hepatitis on histologic examination (41% compared with 82%; P = 0.02) [72]. These observations suggest that autoimmune cholangitis and primary biliary cirrhosis are different disorders.

Unfortunately, the differences that have been described in large groups of patients are not useful for differentiating between the disorders in individual patients, and studies at this level indicate that the diseases are indistinguishable [70, 71]. Even mitochondrial-antibody status does not define valid clinical subtypes if patients have similar histologic findings [71]. These observations suggest that autoimmune cholangitis and primary biliary cirrhosis are the same disease and that a distinction made on the basis of mitochondrial-antibody status is superfluous [71].

Patients with autoimmune cholangitis resemble patients with autoimmune hepatitis in that they invariably have smooth-muscle or antinuclear antibodies, or both. These autoantibodies are nonspecific in nature, and their presence is insufficient for the diagnosis of autoimmune hepatitis [75]. Nevertheless, their presence, combined with anecdotal reports of limited responses to corticosteroid therapy [69, 73], have justified the subclassification of autoimmune cholangitis as a variant of autoimmune hepatitis [69]. The histologic findings of bile duct proliferation, interlobular bile ductopenia, periportal deposition of copper, pseudoxanthomatous changes, or granulomatous cholangitis in the context of a normal cholangiogram are more than sufficient to eliminate the consideration that autoimmune cholangitis is a subclassification of autoimmune hepatitis Figure 3 [73].

Patients with autoimmune cholangitis have antibodies to carbonic anhydrase more often than do patients with primary biliary cirrhosis or autoimmune hepatitis [74]. Carbonic anhydrase is present in high concentrations in bile duct cells, and antibodies to this enzyme are a disease-specific marker of injury to the biliary epithelium [74]. The unique pattern of immunoreactivity seen in patients with autoimmune cholangitis implies that disease-specific pathogenic mechanisms exist in these patients. The enzyme may be exposed or expressed on the surface of injured bile duct cells as a consequence of etiologic factors associated only with autoimmune cholangitis [74]. This display may, in turn, trigger production of autoantibodies, which characterizes the disease. These observations support the classification of autoimmune cholangitis as an distinct outlier syndrome.

Persons with manifestations of autoimmune cholangitis have variable responses to corticosteroids and ursodeoxycholic acid Table 4 [69, 70, 73]. Preliminary experiences suggest that these therapies can improve the clinical and laboratory features but not the histologic changes associated with the disease [69, 73]. In patients with symptomatic disease, empirical maintenance therapy with prednisone alone (20 mg/d) for 3 to 6 months is appropriate Table 4 [69, 73]. If symptoms are not alleviated or laboratory manifestations do not improve, trial therapy with ursodeoxycholic acid (13 to 15 mg/kg of body weight per day) for a similar period is justified Table 4 [70, 73]. Expectations of success must be guarded, and the duration of treatment must be adjusted according to realistic assessments of response.

Cryptogenic Chronic Hepatitis

Thirteen percent of adults with nonviral chronic hepatitis satisfy criteria for autoimmune hepatitis but lack the characteristic autoantibodies Table 3 [4]. These patients are currently designated as having cryptogenic (autoantibody-negative) chronic hepatitis, and they are frequently excluded from therapies that may benefit them. Their similarity to persons with autoimmune hepatitis and their frequent seropositivity for novel autoantibodies suggest that they have an outlier syndrome [4].

Patients with cryptogenic chronic hepatitis are similar to patients with autoimmune hepatitis in age (mean ±SD, 42 ± 6 years compared with 45 ± 2 years); sex (67% compared with 78% were female); frequency of concurrent immunologic diseases (25% compared with 45%); and serum levels of aspartate aminotransferase, bilirubin, globulin, and IgG [76, 77]. In addition, HLA-B8 (75% compared with 50%), HLA-DR3 (71% compared with 51%), and HLA-A1-B8-DR3 (57% compared with 38%) occur as often in these patients as in patients with autoimmune hepatitis [77]. Patients with cryptogenic chronic hepatitis also have histologic findings that are indistinguishable from those of patients with autoimmune hepatitis Figure 1 [65, 77]. Most important, autoantibody-negative patients with cryptogenic chronic hepatitis respond as well to corticosteroid treatment as do patients with autoimmune hepatitis [76, 77]. They also enter remission (82% compared with 78%), have relapse after withdrawal of drug therapy (67% compared with 81%), sustain remission after treatment (33% compared with 19%), deteriorate during therapy (9% compared with 12%), and die of liver failure (9% compared with 2%) as often as do patients with autoimmune hepatitis [77].

Patients with cryptogenic chronic hepatitis can frequently be reclassified into the autoimmune diagnostic category through the use of novel immunoassays. Fourteen percent of these patients have antibodies to soluble liver antigen [12, 78, 79], and 33% have liver-pancreas antibodies [80]. Longitudinal studies [81] have also indicated that some patients manifest conventional autoantibodies later in the course of disease. These observations underscore both the variability of autoantibody expression in individual patients and the importance of serial testing.

Patients with cryptogenic chronic hepatitis are candidates for corticosteroid therapy and should receive the same therapy that their autoantibody-positive counterparts receive for 3 to 6 months Table 4 [76, 77]. Evidence of clinical and laboratory improvement during this period justifies continuation of the regimen until remission, treatment failure, incomplete response, or drug toxicity occurs [81, 82]. Candidates for treatment must have evidence of severe inflammatory activity, and they should not be confused with persons who have cryptogenic chronic liver disease. The latter patients typically have inactive cirrhosis that is best managed symptomatically or with liver transplantation [83-85].

Summary

Top Methods Summary Author & Article Info References

Standardization of the diagnostic criteria for autoimmune hepatitis has focused attention on patients whose conditions are now unclassifiable. The presence of mitochondrial antibodies, histologic changes associated with cholangitis, disproportionate elevation of the serum alkaline phosphatase level, diverse laboratory features associated with cholestasis, and concurrent true viral infection and the absence of conventional autoantibodies in persons whose disease otherwise typifies the syndrome of autoimmune hepatitis constitute variant forms of autoimmune hepatitis. The uniqueness of these conditions must be recognized, and their distinct classifications must be preserved.

"Forced" diagnoses or unwarranted assimilations of patients into the classic categories must be avoided to prevent both contamination of the rigorously defined homogeneous disorders and cloaking of the variant subgroups. Patients with variant disease may share certain pathogenic mechanisms, genetic predispositions, or environmental factors that promote concurrence of certain diseases. Careful analyses of these subgroups may elucidate common threads of pathogenesis that assessments of pure syndromes could not identify. In addition, full characterization of these subgroups will permit their classification, formalize their clinical existence, and promote trials of treatment for them. The current perception of their rarity may, in fact, already reflect their submergence into the conventional categories, ignorance about the legitimacy of their existence, or both.

Treatment strategies are empirical in all instances, but they must be logical and carefully monitored. The predominant manifestations of the variant must be identified and treated with an established regimen. Therapeutic flexibility is essential; alterations in treatment are dictated by changes in the predominant character of the disease and by the adequacy of the patient's response. A scoring system promises to standardize assessments and determine degrees of similarity between autoimmune hepatitis and its variants.

The general internist must understand that variant forms of autoimmune hepatitis do not have established natural histories or therapies. Accordingly, each patient must be evaluated regularly and must receive individualized care. Unconventional testing is not necessary because peculiarities of diagnosis are defined by standard studies. Assays for antibodies to smooth muscle, nucleus, mitochondria, and liver/kidney microsome type 1 are sufficient to catalogue the presence and nature of immunoreactivity. Laboratory measurements of serum aspartate aminotransferase, bilirubin, globulin, and alkaline phosphatase (liver isoenzyme) levels are sufficient to compare hepatitic and cholestatic features, and assessments of hepatitis B surface antigen and antibodies to hepatitis C virus by commercial second-generation immunoassays are adequate for evaluating viral components. Liver biopsy is required in all patients, and it is important to review each biopsy specimen for mixed (atypical) features. In addition, a low threshold for cholangiography should be used. The absence of a precise diagnosis does not indicate misdiagnosis as long as this absence reflects the application of current codified diagnostic criteria for conventional syndromes.