Antibodies in the diagnosis of coeliac disease: a biopsy-controlled, international, multicentre study of 376 children with coeliac disease and 695 controls.

Diagnosis of coeliac disease (CD) relies on a combination of clinical, genetic, serological and duodenal morphological findings. The ESPGHAN suggested that biopsy may not be necessary in all cases. New guidelines include omission of biopsy if the concentration of CD-specific antibodies exceeds 10 times the upper limit of normal (10 ULN) and other criteria are met. We analysed the 10 ULN criterion and investigated multiple antibody-assays. Serum was collected from 1071 children with duodenal biopsy (376 CD patients, 695 disease-controls). IgA-antibodies to tissue transglutaminase (IgA-aTTG), IgG-antibodies to deamidated gliadin peptides (IgG-aDGL) and IgA-endomysium antibodies (IgA-EMA) were measured centrally. We considered 3 outcomes for antibody test procedures utilizing IgA-aTTG and/or IgG-aDGL: positive (>/=10 ULN, recommend gluten-free diet), negative (<1 ULN, no gluten-free diet) or unclear (perform biopsy). Positive (PPV) and negative (NPV) predictive values were based on clear test results. We required that they and their lower confidence bounds (LCB) be simultaneously very high (LCB >90% and PPV/NPV >95%). These stringent conditions were met for appropriate antibody-procedures over a prevalence range of 9-57%. By combining IgG-aDGL with IgA-aTTG, one could do without assaying total IgA. The PPV of IgG-aDGL was estimated to be extremely high, although more studies are necessary to narrow down the LCB. The proportion of patients requiring a biopsy was <11%. The procedures were either equivalent or even better in children <2 years compared to older children. All 310 of the IgA-aTTG positive children were also IgA-EMA positive. Antibody-assays could render biopsies unnecessary in most children, if experienced paediatric gastroenterologists evaluate the case. This suggestion only applies to the kits used here and should be verified for other available assays. Confirming IgA-aTTG positivity (>/=10 ULN) by EMA-testing is unnecessary if performed on the same blood sample. Prospective studies are needed.