Celiac disease (CD) is characterized by intestinal lesions induced by ingestion of gliadin, a gluten protein, in genetically prone individuals. There is a close relationship between CD and major histocompatibility complex genes (especially phenotypes HLA B8, HLA DR3, HLA DR7, and DQW2). A number of diseases have been described as associated with CD; most have an autoimmune origin and are associated with the same haplotypes as CD (1). This is a case report of an infant who presented with CD, as revealed by autoimmune hepatitis and erythroblastopenia.
The patient, a 30-month-old girl with no medical history, presented with fever and conjunctivitis. Superficial adenopathy less than 1 cm in diameter, firm hepatomegaly, and moderate splenomegaly were noted. Height and weight were normal. Blood cell count revealed a nonregenerative anemia (hemoglobin, 97 g/L; mean cell volume (MCV), 74 fL; reticulocytes, 30.8 × 10 9 /L) and 5.2 × 10 9 /L leukocytes, including 34% neutrophils, 51% lymphocytes, and 10% monocytes. Aminotransferase activities were elevated: aspartate aminotransferase, 181 IU/L (normal, 20–45); alanine aminotransferase, 166 IU/L (normal, 8–35). Gamma glutamyl transpeptidase (GGT) was 10 IU/L (normal, 10–33). Serum iron concentration was 10 μmol/L. Evaluation of the bone marrow revealed erythroblastopenia of 3.9% (normal > 20%); megakaryocytes were numerous and other lineages were normal. No cause was found. There was no drug intake, and tests for hepatitis A and B, Epstein-Barr virus infection, cytomegalovirus, and toxoplasmosis were negative. Serologic markers for herpes simplex virus 6 and parvovirus B19 were not obtained.
Three months later, hepatomegaly had decreased and biochemical abnormalities had improved: hemoglobin, 115 g/L; MCV, 82 fL; aspartate aminotransferase, 56 IU/L; alanine aminotransferase, 48 IU/L, and gamma glutamyl transpeptidase, 11 IU/L. Serum iron concentration decreased to 7.3 μmol/L and ferritin decreased to 10 ng/mL.
At aged 43 months, the child was hospitalized for respiratory infection. She had not gained weight during the preceding year. A firm 5-cm hepatomegaly and pallor were noted. Blood cell count revealed a nonregenerative microcytic anemia: hemoglobin, 7 g/L; MCV, 68.8 fL; and reticulocytes, 60.0 × 10 9 /L. Leukocyte and platelet counts were normal. Serum iron was 3 μmol/L and ferritin was 80 μg/L. Hemoglobin electrophoresis was normal and hemolysis investigation was negative. The bone marrow revealed only 7.4% erythroblasts and numerous megakaryocytes; other lineages were normal. Erythropoietic progenitor analysis revealed that there was no burst-forming unit–erythroid (BFU-E) without erythropoietin and slight growth with erythropoietin. Blood inhibitors were not found, nor was there any cytotoxic effect of CD8 lymphocytes.
Abnormal liver test results persisted: aspartate aminotransferase, 152 IU/L; alanine aminotransferase, 78 IU/L. Total bilirubin concentration was 5 μmol/L. The prothrombin time was 80 seconds. Blood immunoglobulins were increased: IgG, 24.9 g/L (normal, 5–14). α 1 -Antitrypsin, ceruloplasmin, and blood copper concentrations were normal. Serologic tests for hepatitis B and C and for HIV were negative. Serologic markers for herpes simplex virus 6 and parvovirus B19 were not performed. Antinuclear, anti–liver kidney microsome, antimitochondrial, and anticytosol antibody screening tests were negative. The titer of anti–smooth muscle antibodies tested by indirect immunofluorescence on sections of rat liver, kidney, and stomach were positive from 1:500. Antiactin titers were negative but antivimentin tests were positive from 1:500, and antireticulin antibodies were also positive from 1:500. Liver biopsy revealed active chronic hepatitis with severe lymphocyte inflammation, sometimes penetrating the limiting plate (Fig. 1). Portal inflammation was associated with lobe inflammation surrounding hepatic cells with fragmented necrosis.
In the absence of the classic markers of autoimmune hepatitis and the presence of antireticulin antibodies, underlying digestive disease was suspected. Digestive endoscopy did not reveal macroscopic abnormality. Histology revealed total villous atrophy with an increase in intraepithelial lymphocytes; inflammatory colitis was present, predominantly lymphocytic without gland atrophy and basal thickening. Antigliadin antibody concentrations were 43 U (IgG) and 146 U (IgA) (positivity threshold = 20 U), and antiendomysial antibody titer was also at 1:500. This confirmed celiac disease. Human leukocyte antigen typing was A1, A10, B8, DR3, and DQ2.
A gluten-free diet was started along with corticoid therapy with prednisone 2 mg · kg −1 · d −1 . Table 1 details the evolution of clinical and laboratory findings. The outcome was favorable: weight gain, hemoglobin, hepatic function, and immunoglobulin concentration returned to normal. Corticosteroids were gradually reduced while the gluten-free diet was maintained. After 6 months on prednisone 0.6 mg/kg every other day, the child was doing well. Her height and weight increased. Moderate hepatomegaly persisted. Hemoglobin transaminases and all antibody titers normalized. Colony-forming units–erythroid (CFU-E) and BFU-E progenitor growth was back to normal, and no inhibitor was found. However, blood IgG again increased. Corticosteroid therapy was raised to 1 mg · kg −1 · d −1 and combined with azathioprine 2 mg · kg −1 · d −1 . Immunoglobulin concentration then returned to normal and corticosteroid therapy was gradually reduced. Three years after initiation of treatment, the child is receiving prednisone (0.5 mg/kg every other day) and azathioprine 2 mg · kg −1 · d −1 . All clinical and biologic data are normal (Table 1).
In this case, CD was suspected because of an isolated poor weight gain. This presentation mode is not uncommon and digestive signs often are missing at this stage. Diagnosis was confirmed according to the European Society of Paediatric Gastroenterology and Nutrition's 1990 criteria (2), based on the presence of total villous atrophy, as detected from small intestine biopsy; positive specific antibody screening (antigliadin, antireticulin and antiendomysium); and by the restoration of weight gain and negative antibodies on a gluten-free diet. Celiac disease is associated with a number of autoimmune diseases (insulin-dependent diabetes mellitus, autoimmune thyroiditis, arthritis, and connective disease) (1). Human leukocyte antigen haplotypes B8 and DRw3 are often found in CD patients and in those with autoimmune diseases. The originality of this case lies in the triple association of CD, autoimmune hepatitis, and erythroblastopenia.
The diagnosis of active autoimmune hepatitis in this child was confirmed by the presence of polyclonal hypergammaglobulinemia, histologically detectable lesions of active chronic hepatitis, and the autoantibody antivimentin, in the absence of markers of B and C virus infection. There are two main types of autoimmune hepatitis (3,4) : type 1 hepatitis (the most frequent form) is characterized by the presence of smooth-muscle Ab (3). These antibodies are directed against microfilaments by the presence of actin, against intermediate filaments by the presence of vimentin, and against microtubules by the presence of tubulin (5). Type 2 hepatitis is characterized by the presence of antiendoplasmic reticulum Ab (anti–liver kidney microsome Ab type 1 or type 2 (4). In our case, the autoimmunity markers found were anti–smooth muscle Ab, not of the antiactin type but of the antivimentin type, positive from 1:500. In the Dighiero et al. study (5) involving 54 serum specimens from patients with anti–smooth muscle Ab-positive autoimmune hepatitis, the authors found antiactin Ab in 51.9% of cases, antimyosin Ab in 31.5%, antitubulin Ab in 35.2%, antitropomyosin Ab in 34%, antitroponin Ab in 11.3% and antivimentin Ab in 22.6% of cases. That study thus confirmed the clear predominance of antiactin Ab in type 1 autoimmune hepatitis. The autoimmune character of chronic hepatitis in our patient was further suggested because liver function and immunoglobulin concentration returned to normal with corticosteroid therapy and increased again when corticosteroid was reduced. Antivimentin Abs are described in viral infections, systemic autoimmune diseases, Sjögren syndrome, inflammatory bowel disease, Behçet syndrome, and graft-versus-host disease (6) but not classically in CD.
The increase in transaminase level is not uncommon in CD (7,8). It is transient, is not associated with symptoms, and returns to normal in 2 to 8 weeks with gluten-free diet alone. Unlike autoimmune hepatitis it does not require immunosuppressive treatment. The cytolysis so defined is also called “gluten hepatitis.” The frequency of normal transaminase has been estimated between 39% and 47% in adults. Data are much less precise in children. Bonamico et al. (9) found cytolysis in 59.9% of cases in a population of 65 children aged 6 months to 18 years. There are two pathophysiologic hypotheses to this CD/hepatic cytolysis association. The most likely and most frequently invoked hypothesis is that of a dysregulated immune process that would induce liver damage by autoantibody synthesis. The hypothesis implicating endogenous or exogenous hepatotoxic substances absorbed by the abnormal digestive mucosa has also been mentioned (10).
Nonspecific reactive hepatitis or chronic hepatitis has been recorded in a smaller proportion of atypical, often silent, CD cases. The 6 children described by Vajro and al. (11) had gluten-induced hypertransaminasemia caused by asymptomatic celiac disease. Hepatic damage may be another atypical form of celiac disease in children.
The CD/autoimmune hepatitis association is far less common in children. A few cases of autoimmune hepatitis have been reported in the literature associated with silent celiac disease (11–14). Vajro et al. (11) in 1993 reported three cases of nonspecific hepatitis (two of persistent chronic hepatitis and one of active chronic hepatitis). No autoimmunity was documented in these three children. In an Italian, multicenter study coordinated by Maggiore et al. (14) during a 5-year period, 27 children out of 2,000 with CD presented with hypertransaminasemia. Histologic tests performed in 9 children, revealed 3 cases of cirrhosis (2 of these patients had antiendoplasmic reticulum Ab). The potential outcome of these hepatitis cases is poor, carrying a risk of cirrhosis that requires immunosuppressive treatment, independent of a gluten-free diet. As in most autoimmune diseases, haplotype A1/B8/DR3 is frequently found in patients with autoimmune hepatitis, even in the absence of CD. One third of patients have that haplotype. Patients with the B8 haplotype appear to respond better to corticosteroid therapy (15).
The patient in this case had B8 haplotype and responded well to corticosteroid therapy. Her liver parameters returned to normal after 2 months on a gluten-free diet and corticosteroid therapy. In this case, the child also presented with erythroblastopenia. The condition did not appear to have a nutritional origin. Iron deficiency causes inefficient erythropoiesis with only average erythroblastic hyperplasia (16). Chronic erythroblastopenia persisting for 1 year, with underlying CD, suggested an immunologic disorder. In such a situation, immunosuppressive therapy has proven useful, furthermore when BFU-E are present in culture (17). Co-culturing of bone marrow precursor cells in the presence of the patient's serum and T lymphocytes did not elicit a cytotoxic effect nor did it reveal the presence of any blood inhibitors. Therefore, although no circulating Ab could be found, the erythroblastopenia appeared to be part of an autoimmune disease, especially because it improved during corticosteroid treatment. To our knowledge, three cases of CD associated with erythroblastopenia in children have been reported. La Placa et al. (18) described a case of a 2-year-old with erythroblastopenia when CD was diagnosed. In the two cases reported by Dufour et al (19), a lymphocytic cytotoxic effect was observed. They suspected an autoimmune origin. Trinchet et al. (20) described a case of erythroblastopenia associated with active chronic hepatitis in an adult with anti–smooth muscle antibodies. In vitro addition of the patient's serum or her lymphocytes had no inhibiting effect on erythroblast differentiation. Cyclophosphamide treatment corrected anemia in 6 weeks.
Early screening and management of children with CD are necessary. An Italian study coordinated by Ventura (21) revealed increased prevalence of autoimmune diseases in CD patients, especially in subjects older than 4 years at the time of diagnosis. The risk of developing an autoimmune disease is closely linked to prolonged exposure to gluten. This case also suggests that CD must be screened for in patients with atypical chronic hepatitis of suspected autoimmune origin.
1. Collin P, Reunala T, Pukkala E, et al. Coeliac disease associated disorder and survival. Gut 1994; 35: 1215–8.
2. Walker-Smith JA, Guandalini S, Schmitz J, et al. Revised criteria for diagnosis of coeliac disease. Report of working group of European Society of Paediatric Gastroenterology and Nutrition. Arch Dis Child 1990; 65: 909–11.
3. Maggiore G, Veber F, Bernard O, et al. Autoimmune hepatitis associated with anti-actin antibodies in children and adolescents. J Pediatr Gastroenterol Nutr 1993; 17: 376–81.
4. Maggiore G, Bernard O, Homberg JC, et al. Liver disease associated with anti-liver kidney microsome antibody in children. J Pediatr 1986; 108: 399–404.
5. Dighiero G, Lymberi P, Monot C, et al. Sera with high levels of anti-smooth muscle and anti-mitochondrial antibodies frequently bind to cytosqueleton proteins. Clin Exp Immunol 1990; 82: 52–6.
6. Whittingham S, Mackay IR. Smooth muscle autoantibodies. In: Peter JB, Shoenfeld Y, eds. Autoantibodies. New York: Elsevier Science BV; 1996: 767–73.
7. Leonardi S, Bottaro G, Patane R, et al. Hypertransaminasemia as the first symptom in infant celiac disease. J Pediatr Gastroenterol Nutr 1990; 11: 404–6.
8. Renaud C, de Boissieu D, Dupont C, et al. Hypertransaminasémie révélatrice ou associée à une maladie coeliaque. Arch Pediatr 1994; 1: 528–9.
9. Bonamico M, Pitzalis G, Culasso F, et al. Il danno epatico nella malattia celiaca del bambino. Minerva Pediatr 1986; 38: 959–62.
10. Linberg T, Berg NO, Borulf S, et al. Liver damage in coeliac disease or other food intolerance in childhood. Lancet 1978; 18: 390–1.
11. Vajro P, Fontanella A, Mayer M, et al. Elevated serum aminotransferase activity as an early manifestation of gluten-sensitive enteropathy. J Pediatr 1993; 122: 416–9.
12. Caiulo VA, Averbene C, Olmi S, et al. Associazione tra malattia celiaca ed epatite cronica autoimmune. Descrizione di un caso. Minerva Pediatr 1993; 45: 511–3.
13. Maggiore G, de Giacomo C, Scotta MS, et al. Celiac disease presenting as chronic hepatitis in girl. J Pediatr Gastroenterol Nutr 1986; 5: 501–3.
14. Maggiore G, Ceccarelli M, Colombo C, et al. Hepatic lesion in childhood celiac disease: a multicentric retrospective study. J Pediatr Gastroenterol Nutr
15. Johnson PJ, Mc Farlane IG, Eddleston ALWF. The natural course and heterogeneity of autoimmune-type chronic active hepatitis. Semin Liver Dis 1991; 11: 187–96.
16. Bernard J, Levy JP, Viret B. Anémies sidéropéniques par carence martiale. Hématologie. Collection médico-chirurgicale. Flammarion
17. Charles RJ, Sabo KM, Kidd PG, et al. The pathophysiology of pure red cell aplasia: implications for therapy. Blood 1996; 87: 4831–8.
18. La Placa G, Arlati S, Verdura C, et al. Descrizione di un caso di malattia celiaca associata ad eritroblastopenia transitoria in eta pediatrica. Pediatr Med Chir 1998; 20: 153–4.
19. Dufour C, Pesce F, Micalizzi C, et al. Two cases of childhood acquired pure red cell aplasia associated with celiac disease: remission after gluten withdrawal and in vitro evidence of cell-mediated inhibition of marrow erythroid cells by patient T-lymphocytes. Eur J Pediatr 1995; 154: 763–4.
20. Trinchet JC, Lacombe C, Islam S, et alR. Association of autoimmune chronic active hepatitis and acquired erythroblastopenia cured by cyclophosphamide. Gastroenterol Clin Biol
21. Ventura A, Magazzu G, Greco L. Duration of exposure to gluten and risk for autoimmune disorders in patients with celiac disease. Gastroenterology 1999; 117: 297–303.