Celiac disease-related autoantibodies in end-stage autoimmune liver diseases: what is the message?


  • George N. Dalekos,

    1. Department of Medicine, Academic Liver Unit and Research Lab of Internal Medicine, University of Thessaly, Medical School, Larissa, Greece
    2. Research Group of Investigational Medicine, Centre for Research and Technology-Thessaly (CE.RE.TE.TH), Institute of Biomedical Research and Technology, Larissa, Greece
    Search for more papers by this author
  • Dimitrios P. Bogdanos,

    1. Liver Immunopathology, Institute of Liver Studies, Division of Gene and Cell Based Therapy, King's College London School of Medicine at King's College Hospital, London, UK
    Search for more papers by this author
  • James Neuberger

    1. Liver Unit, Queen Elizabeth Hospital, Birmingham, UK
    Search for more papers by this author

Celiac disease (CD), also known as gluten-sensitive enteropathy, is defined as a permanent intolerance to ingested gluten and is strongly associated with human leucocyte antigen (HLA) molecules DQ2 and DQ8. Immunoglobulin A (IgA) class endomysial antibodies (EMA) and IgA tissue transglutaminase antibodies (tTGA) have been considered as the laboratory serological tests of choice (sensitivity >93% and specificity 99–100 and 95% respectively) for the diagnosis of CD and can be used for screening or to determine the prevalence of CD in populations at high risk (1). However, the predictive value of these antibodies in the diagnosis of CD in those with liver disease does not appear to be as high as in those without liver disease, hence, histological confirmation is necessary before the patient is advised to adhere to a gluten-free diet.

The clinical presentation of CD can vary from a typical malabsorption syndrome or extraintestinal manifestations to a clinically silent disease, often detected by serological screening of those subjects at risk. The disease mainly affects the small intestine and also the bones, skin, nervous system and the liver. Hypertransaminasemia has indeed been reported in approximately 40% of adults and in 50% of children with a classical CD presentation at the time of diagnosis (2, 3); conversely, CD is diagnosed in up to about 9% of patients with unexplained hypertransaminasemia (4, 5). In a recent large general population-based study, CD patients had a two- to six-fold risk of developing liver disease, while prior liver disease was associated with a four- to six-fold increased risk of latent CD development. Very occasionally, CD may present with liver failure.

In this context, CD can co-exist with autoimmune liver diseases (6–11) (as with other autoimmune diseases such as thyroid disease) with reported prevalence rates of CD ranging from 0 to 6.4% (12–14) for autoimmune hepatitis (AIH), from 0 to 11% for primary biliary cirrhosis (PBC) and 1.6% (15) for primary sclerosing cholangitis (PSC). Large, controlled studies investigating the effect of a gluten-free diet in the outcome of the autoimmune liver diseases are currently missing (3). PBC appears to be three to 20 times more frequent in patients with CD compared with controls (7–11) while the presence of PSC in CD patients is increased at a four- to eight-fold compared with controls (6).

In this issue of the journal, Rubio-Tapia et al. (16) reported for the first time data on the prevalence of CD-related antibodies (EMA and tTGA) in a large series of patients with end-stage autoimmune liver diseases who underwent liver transplantation. These investigators have also studied the behaviour of EMA and tTGA antibodies following transplantation and the clinical outcome of patients with autoimmune liver disease and concomitant CD on a normal gluten-containing diet. They have found that CD-related autoantibody markers were more frequent in end-stage autoimmune liver disease patients compared with patients with non-autoimmune end-stage liver disease. Thus, the prevalence of tTGA and EMA antibodies was significantly higher in HLA-DQ2- or HLA-DQ8-positive patients with end-stage autoimmune liver disease compared with patients with non-autoimmune end-stage liver disease, 14.2 vs 5.4% (P=0.0001) and 4.3 vs 0.78% (P=0.01), respectively, while serological evidence of CD, defined as the co-occurrence of tTGA and EMA antibodies was five-fold increased in end-stage autoimmune liver disease patients in comparison with the controls (3 vs 0.6%). Unfortunately, this study was retrospective and intestinal tissues were unavailable for re-review (with the exception of only two patients). Hence, a definite diagnosis of CD could not be achieved in the majority of patients seropositive for CD-related autoantibodies. Of note, tTGA and EMA antibodies normalized in practically all of the patients (94 and 100% positive patients respectively), when serum samples were retested 6–12 or >24 months post-transplantation. This normalization has been achieved in the absence of gluten exclusion from the diet indicating that gluten may not immediately relate to autoantibody kinetics. The post-transplantation suppression of tTGA and EMA antibodies suggests that the lack of autoantibody seropositivity of post-transplanted sera cannot exclude a diagnosis of CD and may further support the need of pretransplantation screening in patients with end-stage autoimmune liver disease. Clinically, of the five patients with classical symptoms of CD, three improved post-transplantation but intestinal lymphoma was diagnosed 8 and 9 years after liver transplantation in the other two cases with CD-associated antibodies and clinically silent CD, and this cannot be neglected.

The reason(s) behind the significant and sustained decrease/normalization of CD-related antibody serology after liver transplantation, especially in the presence of gluten challenge, is obscure. Administration of immunosuppressive agents after liver transplantation can directly or indirectly inhibit autoantibody production, and this could be the case for EMA and in particular for tTGA. In fact, corticosteroids and azathioprine can induce normalization of CD-specific autoantibodies and lead to clinical, laboratory and histological response in refractory cases of CD, (17, 18) although in other presumed autoimmune diseases such as PBC, antimitochondrial antibodies remain and may increase in concentration. Multivariate analysis could point towards specific associations but the number of seropositive cases was too small (n=9) and the immunosuppressant schedules too diverse to allow such an investigation. Reduction of autoantibody levels may also relate to the scarcity of autoantigenic challenge or the elimination of liver-resident, CD-specific lymphocytes because of the removal of the diseased liver. The reason for the development of intestinal lymphoma in two cases with clinically silent CD remains enigmatic. Long-term immunosuppression could relate to the development of lymphoma but may also prevent recurrence of overt autoimmune liver diseases and concurrent CD and could be an alternative option in these patients for the control of CD (17, 18).

Pretransplant monitoring of CD-related autoantibodies maybe of help in HLA-DQ2 or HLA-DQ8-positive patients with end-stage autoimmune liver disease. Large, controlled, prospective, multicentre studies involving patients with autoimmune liver diseases with co-existing CD may provide novel information of considerable help for the best management of these cases. However, the diagnosis of CD in patients with liver disease, either before or after transplant, must be based on histological examination of the duodenum and the response to treatment.