The prevalence of celiac disease (CD) in Sweden is about 4 cases per 1,000 people. Screening for CD with serological tests indicates similar high prevalences in many other countries. Between 1 November 1992 and 30 April 1995, 133 children (9 months to 16.7 years of age) with suspected CD were studied. The predictive value (PV) of immunoglobulin A antigliadin antibodies (IgA-AGA) in the serum as assayed with two new commercial automated immunoassays--the Pharmacia CAP System Gliadin IgA FEIA (CAP) and the UNICAP-100 (UNICAP)--and with three "in-house" methods was evaluated using assessment of the small intestinal mucosa morphology as the "gold standard." All serum samples were analyzed for total serum IgA. At presentation the diagnostic sensitivities and specificities of the different tests varied from 0.72 to 0.88 and 0.67 to 0.87, respectively. All methods showed a higher sensitivity for CD in younger children. The area under each assay's receiver operating characteristic curve was calculated and varied between 0.82 and 0.89. The positive and negative PVs for the CAP and UNICAP, which were assays with a high sensitivity and a high specificity, respectively, were estimated. In the clinically selected population (prevalence of CD, 1 in 3) the positive PV was about 55%, and in the general population (prevalence, 1 in 250) it was about 1%. The negative PVs for both CAP and UNICAP were close to 100%; thus, when the AGA test was negative, the risk for CD was small. Interestingly, five children had serum IgA levels below the detection limit (<0.07 g/liter) when on a gluten-free diet, whereas they had normal levels at the time of the first biopsy. In conclusion, the automated immunoassays--based on ImmunoCAP technology--for analysis of IgA-AGA had a reliability comparable to that of the in-house methods.