Objective: The aim of this study was to summarise the evidence from 2004 to September 2009 on the performance of laboratory-based serological and point of care (POC) tests for diagnosing coeliac disease (CD) in children using histology as reference standard.
Patients and methods: We searched MEDLINE and EMBASE for studies reporting on children for tests based on IgA and IgG anti-gliadin (AGA), endomysial (EmA), anti-transglutaminase-2 (TG2), and anti-deamidated gliadin peptides (DGP) antibodies or POC tests. For inclusion, histological analysis of duodenal biopsies and sensitivity and specificity for index tests had to be reported. Data were pooled and summary measures calculated for sensitivity, specificity, positive and negative likelihood ratios ("LR+", "LR-"), and diagnostic odds ratios (DOR). In case of elevated statistical heterogeneity, studies reaching 90% sensitivity or specificity were reported.
Results: A total of 2510 articles were reviewed; 16 entered meta-analysis, reporting on 3110 patients (1876 with CD, 1234 without CD). For IgA-EmA, sensitivity was ≥90% in 7/11 studies and pooled specificity 98.2%. For IgA-anti-TG2, 11/15 studies yielded sensitivities ≥90% and 13/15 specificities ≥90%. For IgA-DGP, sensitivity ranged between 80.7% and 95.1% (specificity 86.3%-93.1%); for IgG-DGP between 80.1% and 98.6% (specificity 86.0-96.9%). IgA-EmA had the highest pooled DOR (554) and LR+ (31.8) for a laboratory test, followed by IgA-anti-TG2, IgG-DGP, IgA-DGP and IgA-AGA. POC tests showed a pooled sensitivity of 96.4% for IgA-TG2 (specificity 97.7%).
Conclusions: IgA-EmA and IgA-anti-TG2 tests appear highly accurate to diagnose CD. IgG-anti-DGP tests may help in excluding CD. IgA-AGA and IgA-DGP tests show inferior accuracy. POC tests may achieve high accuracy in the hands of experienced readers, but IgA-anti-TG2/EmA were superior.