OBJECTIVE: To determine which serological marker(s) to use when screening for coeliac disease. DESIGN: In a population-based cross-sectional study we compared the use of antigliadin antibodies (AGA) of isotypes IgA and IgG, antiendomysial antibodies (AEA) of isotype IgA and antitransglutaminase antibodies (ATGA) of isotype IgA for detecting coeliac disease amongst adults. SETTING: Northern Sweden. SUBJECTS: A total of 1850 of 2500 (74%) invited adults (aged 25-74 years) who were randomly selected from the population register after stratification for age and sex. MAIN OUTCOME MEASURES: The sensitivity, specificity and predictive values of the AGA, ATGA and AEA tests. RESULTS: Nine cases of biopsy proven, previously undiagnosed coeliac disease were detected by screening. The sensitivity of both ATGA and AEA was 100% whilst AGA IgA and IgG both had a sensitivity of 89%. The AEA test had a specificity of 100% whereas the specificities of the ATGA, AGA IgA and IgG tests were 97, 96 and 78%, respectively. The positive predictive value for the AEA test was 100%, whereas it was considerably lower for the other tests (ATGA > AGA IgA > AGA IgG), with further decreases for all tests when shifting from a clinical to a screening situation. CONCLUSIONS: When screening for coeliac disease we suggest a serial testing approach, i.e. an initial ATGA test and, when positive, followed by an AEA test, provided that IgA deficiency has been excluded. However, assessment of the small intestinal mucosal morphology is still required to ascertain the diagnosis.
