In this study a micro-ELISA (ELISA = enzyme-linked immunosorbent assay) was established and used to evaluate IgA and IgG antigliadin antibodies in 1,866 healthy adults. There was a covariation between the level of IgA antigliadin antibodies and the total serum IgA concentration, probably due to an increased IgA response in some healthy subjects. We could not find any correlation between the presence of IgG and IgA antibodies in the healthy population using the 97.5th percentile as a cutoff value. The specificity of various cutoff levels was compared with the sensitivity of the test in a population of 40 patients with coeliac disease. IgA antigliadin antibodies had a high specificity (95%) at a cutoff value giving a high sensitivity (80%). This was not possible with IgG antigliadin antibodies which had a low sensitivity (40%) when the cutoff value was selected to give a high specificity. Due to the low prevalence of coeliac disease, a decrease in the specificity of the test will have a pronounced effect on the positive predictive value. The results indicate that only IgA antigliadin antibodies are useful markers when screening subjects with few typical symptoms for biopsy when diagnosing coeliac disease, whereas IgG antibodies are of low value because of their low specificity.