Diagnosing Celiac Disease in 2002: Who, Why, and How?

In this issue of Pediatrics, Barera et al¹ expand on the significant, recognized relationship between type 1 diabetes and celiac disease (CD). The authors confirm the increased prevalence of CD in type 1 diabetes and characterize the timing of the identification of CD relevant to the onset of type 1 diabetes. The study shows that the diagnosis of CD may be made before, at the same time, or several years after the diagnosis of type 1 diabetes. Although most cases in this study were identified within 5 years of diagnosis of type 1 diabetes, the number of subjects tested declined over the 5-year period so that by the sixth year only 10% were actually retested. It is possible, therefore, that over time even more cases of CD/type 1 diabetes will be identified. As in most populations where screening is done to detect CD, most of the patients in this series were either asymptomatic or had unrecognized symptoms.

Why should we be concerned about the identification of apparently asymptomatic CD in patients with type 1 diabetes? First, CD is characterized by small intestinal inflammation with mucosal atrophy. Therefore, absorption of nutrients could be variable or overt malabsorption could occur, altering insulin requirements and affecting diabetic control. Improvement in diabetic control and a decrease in hypoglycemic episodes have been observed in children with CD and type 1 diabetes after institution of a gluten-free diet.² Second, CD can affect growth as demonstrated by the increase in body mass index in the patients after treatment. In the study by Barera et al, this improvement could be attributed, in part, to management of type 1 diabetes as well. Third, CD is associated with other diseases or conditions that are reduced or eliminated when individuals with CD follow a gluten-free diet. These include lymphoma, osteopenia, squamous cell carcinoma, and complications of pregnancy.³ The increased risk for lymphoma in individuals with untreated CD is particularly significant but seems to be reversed with long-term adherence to a gluten-free diet.⁴ Those with silent CD, identified through screening rather than by clinical presentation, may not have an increased risk for lymphoma, but additional study of this area is necessary.⁵ Metabolic effects of malabsorption such as anemia, short stature, vitamin deficiencies, and osteoporosis may develop in patients with gastrointestinal symptoms as well as silent CD cases.³ Although there are benefits in making the diagnosis of CD in children with type 1 diabetes, the ideal time for screening needs to be determined through additional experience and study. Considerations include minimizing the need for repetitive serologic testing and biopsy while making the maximum number of diagnoses. We also need to assess whether it is critical that asymptomatic CD be detected and managed from the time of diagnosis of type 1 diabetes or whether a period of family adaptation to living with type 1 diabetes is desirable before evaluation for CD. Of note is the decreased dietary compliance among adolescents identified through screening compared with those diagnosed because of symptomatic disease.⁶

Diabetes and CD are both autoimmune disorders, sharing the same high-risk HLA DQ2 genotype. One third of type 1 diabetes patients with the CD-associated HLA DQ2 genotype tested serologically positive for CD compared with <2% of patients lacking DQ2.⁷ Other autoimmune diseases associated with CD and type 1 diabetes include autoimmune thyroiditis, pernicious anemia, Sjogren syndrome, Addison’s disease, alopecia areata, and rheumatoid arthritis.⁸ The reasons for these associations may be that CD and these other disorders share a similar autoimmune pathogenic mechanism, or that the same gene is responsible for a proportion of these disorders.⁹ It is possible that chronic lymphocyte stimulation in the intestine in CD could result in an increase in autoantibody production and therefore stimulate the development of other autoimmune disorders.⁹ Older, untreated CD patients have a higher prevalence of autoantibodies than younger patients, suggesting that duration of gluten exposure increases the risk of developing autoantibodies.¹⁰ In some individuals with rheumatoid arthritis and pericarditis, the conditions disappeared when the patients followed gluten-free diets.⁹ Although type 1 diabetes is occasionally identified in previously diagnosed individuals with CD, in most cases CD antibodies are present at the time of or after diagnosis of type 1 diabetes. Diabetic autoantibodies generally precede the onset of clinical diabetes by an average of 3 years.¹¹ In one small series of adolescents with newly diagnosed CD and diabetes autoantibodies, treatment with a gluten-free diet was associated with disappearance of the diabetes-associated autoantibodies.¹⁰ There is, however, no evidence that treating CD with a gluten-free diet prevents type 1 diabetes. Additional research is needed to explore the possibility that CD could be involved in the etiopathogenesis of type 1 diabetes and other autoimmune diseases.

CD previously was considered rare, with an estimated prevalence of less than 1 in 2000. However, with the use of serologic testing, information on prevalence has been more reliable than previous data that relied primarily on small intestinal biopsy of symptomatic cases. CD is now recognized as a common disease with a reported disease frequency based on serologic testing of 1 in 250.¹² In a recent report, nearly 1% of children in a US population had a positive serologic test for CD.¹³ In a study of children presenting to a pediatric gastroenterology clinic with symptoms or conditions associated with CD, the prevalence of CD was 1 in 57.¹⁴ Silent disease is frequently present with minimal symptoms or signs. The rate of symptomatic/asymptomatic cases in children is 1 in 5 to 1 in 7.¹⁵ The increased recognition of CD may be attributed to serologic screening tests identifying the asymptomatic or silent cases. There is an increased prevalence of CD in patients with Down syndrome and in relatives of individuals with CD.¹⁶ Both CD and Down syndrome have an association with autoimmune diseases, and the increased prevalence of CD in Down syndrome suggests that a common link may be an altered immune system and/or shared gene(s). Multiple-case families are common with an 8% risk to first-degree relatives.¹⁷

The expression of CD in pediatrics ranges from overt clinical symptoms to minimal or unrecognized symptoms. Classic clinical features include diarrhea, abdominal bloating, abdominal pain, weight loss, and growth failure. Subtle malabsorptive symptoms include anemia, bleeding, abdominal pain, neuropathy, and bone pain. The impact on growth may become apparent only after a child follows a gluten-free diet, with improvement in growth percentiles. CD is often compared with an iceberg, where the visible tip represents those with symptomatic CD.¹⁸ The majority of the iceberg represents those with silent and latent disease. People detected by screening who have histologic and/or serologic evidence of CD but no symptoms have silent CD. Those with latent CD have the potential for developing CD. For example, in the report by Barera and colleagues, nearly 40% of children ultimately diagnosed with CD were identified up to 5 years after testing negative for CD. The gold standard for diagnosis is considered an intestinal biopsy demonstrating some degree of villus atrophy. An accurate diagnosis requires pathologists who are aware of the spectrum of the mucosal abnormality and have experience in the pitfalls of biopsy interpretation. There may be only minimal changes on biopsy even when significant symptoms are present. The mucosal abnormalities, serology (if positive), and symptoms (if present) improve or normalize on a gluten-free diet. Biopsies should always be performed while ingesting a liberal unrestricted diet or after gluten challenge to ensure expression of the histologic abnormality. Follow-up biopsies are not necessary for those with classic symptoms, histology, and serology responding to treatment, but may be needed when the diagnosis is less clear. Additional research is needed to define appropriate treatment and follow-up of individuals with only minor mucosal abnormalities.

The development of serologic testing to evaluate people with suggestive symptoms and high-risk populations has greatly facilitated evaluation for CD. The presence of immunoglobulin A (IgA) endomysial or tissue transglutaminase antibodies is a very sensitive and specific predictor of histologic CD.¹⁹ These serologies are very sensitive tests that may detect affected individuals before intestinal injury. These serologic tests are, however, less sensitive in children younger than 2 years.²⁰ Some laboratories offer testing panels that include quantitative IgA level, serologic testing, and HLA DQ typing. As pointed out by Barera et al, screening may need to be repeated, but the frequency and intervals between testing for various high-risk groups are not established. Reasons for false-negative serologic results include IgA deficiency, age younger than 2 years, and probably the amount of gluten ingested in the diet. Serologic testing does not detect all CD cases, and anyone with symptoms should still undergo biopsy. Intestinal biopsy is recommended to confirm the diagnosis and the necessity of lifelong gluten restriction. Particularly in the United States, CD cases may lack biopsy confirmation because of expense and/or refusal to have the procedure performed.

CD is a more common disease than previously thought, and its diagnosis and treatment have much broader implications than management of a malabsorptive disorder. Untreated CD is associated with other autoimmune disorders, including type 1 diabetes, and can cause complications. Defining the CD phenotype is complex because individuals may or may not have symptoms. The histologic spectrum is variable, and symptoms do not necessarily correlate with the magnitude of the intestinal abnormality. Serologic testing may yield false-negative results in young children. We do not know whether screen-detected cases have the same risks of complications as cases presenting with symptoms. We need to understand how individuals with minor symptoms and mucosal abnormalities should be managed, the optimal timing for screening, and who should be screened. Pediatricians will continue to have the major role in implementing screening programs for CD and for identifying those at risk. More than a decade has passed since the criteria for diagnosis and management of CD has been formally addressed in pediatrics. There are ongoing efforts by organizations like the North American and European Societies for Pediatric Gastroenterology, Hepatology, and Nutrition to establish guidelines for these critical issues.

• Copyright © 2002 by the American Academy of Pediatrics