The Liver in Celiac Disease

Celiac disease (CD) presents with a spectrum of clinical disorders, mainly of autoimmune mechanism (1). A wide variety of liver injuries, particularly of the inflammatory type, have been reported in patients with CD (2). More rarely, other liver lesions, such as hepatic steatosis (3), which may be severe (4), cirrhosis and hepatocellular carcinoma, are found in the liver biopsies of children and adults with CD at the time of diagnosis (5). Among the inflammatory lesions, both mild liver injuries with limited fibrosis as well as severe and progressive liver diseases have been described (6). The first report of liver involvement in CD appeared in 1977. In this group of 74 newly diagnosed adults with CD, 40% had an elevation of serum aminotransferase activity, which normalized in most cases, during treatment with gluten-free diet. In addition, of 13 patients, 5 showed signs of reactive hepatitis and 7 (16%) showed histologic lesions of different types and severity ranging from steatosis with focal fibrosis to pronounced fibrosis and cirrhosis (7). Approximately 10 years later, in a study of children younger than 2 years of age with newly diagnosed CD and gastrointestinal symptoms, mild to moderate elevation of serum aminotransferase activity was observed in approximately two thirds of cases (8). The biochemical abnormalities, which usually resolved on a gluten-free diet, were associated with a histologic picture of mild mononuclear infiltrate of the portal tract with non-aspecific signs of liver cell damage and Kupffer cell hyperplasia (9). CD may present atypically as a cryptogenic liver disorder. The first description of this type of onset appeared in The Journal of Pediatric Gastroenterology and Nutrition in 1986 in a case report of a young girl with a persistent and cryptogenic elevation of serum aminotransferase and whose liver biopsy showed mild inflammation of the portal tract (10). The diagnosis of CD was suggested by a high titre of antireticulin antibody and was later confirmed by jejunal biopsy. A series of six children with persistent elevation of aminotransferase activity and with a histopathologic picture ranging from reactive hepatitis to moderately active chronic hepatitis was reported subsequently. These children were asymptomatic and jejunal histology typical of CD. In all children, aminotransferase activity normalized on a gluten-free diet. Normalization of histologic lesions was documented in two children, and in three others, a challenge with gluten was followed by an increase of serum aminotransferases (11). More reports in children and two retrospective studies in adults confirmed these observations (12–14) and suggested that as many as 9% of patients with persistent and cryptogenetic elevation of serum aminotransferases may be affected by asymptomatic CD (15). This condition, was once called gluten hepatitis (16). We now suggest defining the condition as celiac hepatitis, characterized by: Absence of hepatomegaly, splenomegaly, or both and of any clinical features suggesting chronic liver disease; Absence of hypergammaglobulinemia and of serum autoantibodies (excepted for antitransglutaminase-related autoantibodies); and Presence of mild lobular and portal tract inflammation, reversible on a gluten-free diet. A high prevalence of CD (4.3%), four to ten times higher than the expected, was recently found in Finland in a group of 185 adult patients who had undergone liver transplantation (17). The nature of the severe, end-stage liver disease leading to liver transplantation was, in most cases, autoimmune. The finding that the prevalence of CD is higher in adult patients with an autoimmune liver disease than in the general population was first reported by Lindberg et al. in 1979 (18). In this study, while systematically searching for CD in 16 adults with chronic hepatitis B surface antigen-negative active hepatitis, Lindberg et al. found three patients with CD, two of whom had no gastrointestinal symptoms. A recent study by Volta et al. (19) confirmed this observation, demonstrating a 4% prevalence of CD in a population of 181 patients with autoimmune hepatitis. A high prevalence of CD was reported recently in adults with autoimmune diseases of the bile ducts such as primary biliary cirrhosis, autoimmune cholangitis, and primary sclerosing cholangitis (20). Particularly strong evidence supports the association between primary biliary cirrhosis, an autoimmune liver disorder peculiar to adulthood, and CD. Prevalence of primary biliary cirrhosis has been reported to be as high as 3% in celiac patients, and about 6% of individuals with CD may be affected by primary biliary cirrhosis (21). A Scandinavian study has confirmed these data, finding a 20 fold increase in the prevalence of primary biliary cirrhosis in adults with CD (22). These studies suggest an active crosschecking of CD, by serum anti-transglutaminase antibodies, in patients with autoimmune diseases of bile ducts. In children, sporadic cases of autoimmune hepatitis or autoimmune sclerosing cholangitis associated with CD have been reported (16,23). In a multicenter study in Italy, autoimmune hepatitis occurred in 1.1% of 909 children with CD and in none of the healthy controls or a control group with Crohn disease (24) In another study of 96 children with autoimmune hepatitis at King's College, London, the prevalence of CD was found to be significantly higher (3.4%) than expected (25). In the patients with CD, the autoimmune hepatitis had the serologic characteristics of type 1, which is associated with anti–smooth-muscle autoantibodies with anti-actin specificity (26). The features of the autoimmune liver disorders associated with CD were studied recently in a multicenter survey by the Italian Society of Pediatric Gastroenterology and Hepatology, which collected data on 14 patients with CD and an autoimmune liver disease (27). In most cases, the diagnosis of autoimmune liver disease preceded the diagnosis of CD, although in retrospect, some evidence of liver damage had been present in most patients since the onset. In some patients, however, an acute icteric hepatitis developed a few months after the diagnosis of CD and the start of a gluten-free diet. When an autoimmune liver disease developed in the setting of silent CD, the diagnosis of CD was suggested by sideropenic anemia or by systematic autoantibody screening. This study suggests that: All types of autoimmune liver disorder may be associated with CD, with autoimmune hepatitis the disorder most frequently found; All subtypes of autoimmune hepatitis can be associated with CD: type 1, as previously suggested (25), type 2 characterized by autoantibodies against liver or kidney microsomes (28), liver cytosol (29), or both, and even seronegative autoimmune hepatitis (30); Complete immunoglobulin A deficiency and the presence of an eosinophilic infiltrate of the portal tracts are a characteristic of patients with autoimmune liver disease and CD; and Reintroduction of gluten into the diet may be associated with a relapse in patients free from immunosuppression. The mechanism of liver injury in celiac patients is poorly understood. The two types of liver injury described may represent a spectrum of a single disorder where individual factors, such as precocity and duration of exposure to gluten, may influence the reversibility of liver damage (2–24). Autoimmune liver disorders and CD share common HLA class II aplotypes. In the white adult population, two aplotypes have been identified as susceptibility markers of autoimmune hepatitis: the complex HLA A1 B8 DR3 and the aplotype HLA DR4. (31). Similarly, specific HLA class II antigens such as HLA-DR3, particularly the HLA-DQ2 molecule and HLA DR4, confer susceptibility to CD (32). Patients with CD have an increased intestinal permeability to different substances that, absorbed in larger amounts through the damaged intestinal mucosa, may have a direct toxic effect on the liver. Alternatively, it may initiate an immunologic reaction toward liver antigens. An increased prevalence of organ-specific and non–organ-specific autoantibodies such as antithyroid or antipancreatic islet cell has been described in CD patients on a gluten-containing diet. These autoantibodies appear to be gluten dependent, because in most cases, they become undetectable on a gluten-free diet (33). So, increased permeability to intraluminal gut antigens could induce, in genetically predisposed individuals, an immune response both against antigens sharing common epitopes to self-liver-proteins and against usually cryptic antigens that are unmasked by the reaction with gliadin. It is known that mucosal damage in CD leads to exposure of tissue transglutaminase enzyme, the target antigen recognized by anti-endomysial antibody (34). The hypothesis that this autoantibody plays a role in extraintestinal manifestations of CD and particularly in liver injury is supported by the recent finding of extracellular deposition of immunoglobulin A class anti-tissue transglutaminase in liver biopsy of two patients with active CD (35). In conclusion, liver involvement in individuals with CD comprises: A so-called celiac hepatitis, which develops both in patients with intestinal symptoms and in individuals with asymptomatic CD. This condition is frequent, benign, clinically silent, possibly immune mediated, and resolves on a gluten-free diet. An autoimmune liver disease, which is usually an autoimmune hepatitis. These disorders are rare and in most cases are associated with clinical signs and symptoms of chronic liver disease. Moreover, they do not recover simply with the administration of a gluten-free diet, but need specific immunosuppressive therapy. A fortuitous association between CD and liver cannot be excluded because of the high prevalence of CD in the general population. These considerations recommend a rigorous search for evidence of liver disease among patients with newly diagnosed CD by monitoring aminotransferases activity. Likewise a search for CD should be initiated by obtaining anti-transglutaminase autoantibodies in children with any inflammatory liver disease of unknown cause. Attention should be paid in patients with chronic liver disease, to the risk of false-positive results of anti-transglutaminase assay because of the frequent contamination of guinea pig transglutaminase by hepatic proteins. The use of anti-human transglutaminase usually avoids this risk (36,37).