Objective: The aim of the present study was to evaluate a panel of different antibody assays, including second-generation antigliadin kits, in a local paediatric population thought to be at risk for coeliac disease (CD).
Methods: Seventy-nine children, who tested positive for immunoglobulin A (IgA) antibodies to tissue transglutaminase (TG), underwent duodenal biopsy. At endoscopy, serum was collected from all of the patients, and 9 different coeliac antibody assays were performed, both as isolated assays and in combination. These included immunoglobulin A (IgA) anti-tissue transglutaminase (TGA), and IgA plus IgG anti-deamidated gliadin peptide (DGPAG). A diagnosis of CD was made if the biopsies showed Marsh grade 3 lesions.
Results: Twenty-four of 79 children had CD confirmed histologically. Only 39 of 79 were positive for Inova TGA, and 35 of 79 were positive for Inova DGPAG. Twenty-four of 39 who were TGA positive and 24 of 35 who were DGPAG positive had confirmed CD on biopsy. There was good correlation between TGA and DGPAG-positive predictive values. None of the modified gliadin tests produced false-negative results, and neither did the TGA.
Conclusions: The Inova DGPAG and TGA assays have similar use in predicting CD in a selected paediatric population; however, in children who are positive for TGA when screened for CD, more than half have negative TGA serology when repeat testing is done at the time of biopsy. Those with persistent TGA positivity have only a 61.5% probability of having histologic CD, compared with 68.6% of those children positive for DGPAG.
*King's College Hospital, Denmark Hill, London, UK
†Clinical Immunology, Auckland Hospital and Labplus, Auckland
§Department of Paediatric Gastroenterology, Starship Children's Hospital, Auckland
||Department of Pathology, Auckland Hospital, Auckland, New Zealand.
Address correspondence and reprint requests to Dr Simon Chin, Department of Paediatric Gastroenterology, Starship Children's Hospital, Auckland, New Zealand (e-mail: firstname.lastname@example.org).
Received 23 February, 2012
Accepted 1 February, 2013
The authors report no conflicts of interest.