Article Text

Download PDFPDF

Antigliadin antibody testing for coeliac disease in children under 3 years of age is unhelpful
  1. S Holding1,2,
  2. M Abuzakouk1,
  3. P C Doré1,3
  1. 1
    Immunology Department, Hull Royal Infirmary, Hull, UK
  2. 2
    Postgraduate Medical Institute, University of Hull, Hull, UK
  3. 3
    Hull-York Medical School, University of Hull, Hull, UK
  1. Stephen Holding, Immunology Department, Hull Royal Infirmary, Anlaby Road, Hull HU3 2JZ, UK; steve.holding{at}

Statistics from

It has been our practice to perform immunoglobulin (Ig)A anti-tissue transglutaminase (tTG), IgA/IgG antiendomysial (EMA) and IgA/IgG antigliadin (AGA) antibody tests for children under 3 years of age being investigated for coeliac disease. This was based on data showing that EMA may be negative in children. In a study of 277 children under 2 years of age who had biopsy-proven untreated coeliac disease, 88% were EMA positive, 100% were IgG-AGA positive, and 84% were IgA-AGA positive. Eleven per cent of patients were positive for AGA only (though 25% of these were positive for IgG-AGA only).1 The overall AGA specificity is poor, especially that of IgG-AGA.2 Importantly, over a third of AGA-positive young children became spontaneously negative over the following 3 years.3

Data on using AGA alongside recombinant human tTG-based tests are limited and study sizes are small. Equivalent sensitivity of tTG and EMA has been shown with AGA in 27 children with untreated coeliac disease (mean age 5.0 years),4 and a study of 30 children (median age 2 years; range 8 months to 12 years) showed superior sensitivity of tTG (93%) over EMA (90%) and AGA (60%/93% for IgA/IgG).2 Day et al analysed 36 antibody-positive samples (IgA-EMA and/or IgG/IgA-AGA) in children aged 6 months to 15 years. Seven had biopsy-proven coeliac disease, of which two were positive by AGA (IgA and IgG) only.5

We therefore audited our use of AGA in children to determine if it added any value to the protocol.

The results of AGA tests performed between 1 January 2006 and 23 September 2007 in children under 3 years of age without a previous diagnosis of coeliac disease were collected. Until February 2007, all samples had IgG-EMA, IgA-EMA, IgG-AGA and IgA-AGA tests performed. From February 2007 IgA-tTG was performed in addition to the above on all samples.

All AGA-positive samples without other serological evidence of coeliac disease were identified. Case notes were retrieved for these patients and the diagnoses were established. Specific checks of histopathology records to 31 December 2008 were made to identify those that had had a duodenal biopsy performed.

One hundred and sixty-seven AGA tests were performed in children under 3 years of age, of which 24 had results above the 99.5 percentile of normal (2 IgA only; 18 IgG only; 4 IgA and IgG). Sixteen of these patients had no other serological evidence of coeliac disease (2 were positive for IgA-AGA only, and 14 for IgG-AGA only). Only 1 of these 16 children (with an established diagnosis of primary hypogammaglobulinaemia) went on to have a duodenal biopsy, which was normal. These 16 children with no serological evidence of coeliac disease apart from raised AGA were followed up for a median of 23 months (range 17–36 months). None had a final diagnosis of coeliac disease. Most were being investigated for poor weight gain that resolved on follow-up. Three had food allergy, one had fetal alcohol syndrome, one had irritable bowel syndrome, and one had type 1 diabetes. In all the other cases a failure to thrive resolved on follow-up and the patients were discharged, including one patient who underwent a trial gluten-free diet and reintroduction to no effect.

If gliadin testing had not been offered, no diagnoses of coeliac disease would have been missed during the 21 month period. Based on the results of our audit, it appears that all AGA-positive results in the presence of negative tTG/EMA were false positives. It could be argued that the failure to identify any patient with coeliac disease in the group was due to lack of confidence in the AGA assay with a resultant failure to perform a biopsy. However, all of these patients had a clear final diagnosis or resolution of the problems that precipitated investigations for coeliac disease. Though coeliac disease was not specifically excluded by biopsy, the follow-up period and clinical outcomes were sufficient to make this diagnosis unlikely.

There are few studies of the role of AGA in screening programmes for young children in parallel with human recombinant tTG-based assays. The published literature shows that a small number of young children with early coeliac disease are positive for AGA only. However, this must be balanced against the large number of false-positive AGA results seen in this group of patients.

These data support the view that AGA testing of children under the age of 3 years being investigated for coeliac disease is not justified. In our cohort, the use of only tTG and/or EMA in these children would have reduced the false-positive rate without affecting the detection rate.


View Abstract


  • Competing interests: None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.