Use of a threshold serum level of anti-gliadin antibodies improves diagnostic efficiency of the test in adult coeliac disease but is unreliable as a screening test.

Measurement of anti-gliadin antibodies is considered a highly sensitive test for coeliac disease in children. Specificity, however, appears to vary due to the presence of anti-gliadin antibodies in other diseases. Sensitivity and specificity of anti-gliadin antibody measurement for coeliac disease in adults has, thus, been assessed using the ratio of the densitometric unit test/the mean +3SD of densitometric unit values of a pool of sera of healthy biopsy-proven controls. Anti-gliadin antibodies-A and G were measured separately with an enzyme-linked immunosorbent assay in 64 coeliacs (20 males; 44 females; age range 14-71 years) with diagnosis confirmed at jejunal biopsy; and in 60 controls (25 males; 35 females; age range 16-69 years) with normal jejunal biopsy. Detection of anti-gliadin antibodies-A and G had a sensitivity of 58% and 61%, respectively, and a specificity of 85% and 94%. For the procedure a Receiver Operating Characteristic analysis was used. Considering anti-gliadin antibodies-A and G values of at least 0.9 densitometric unit, sensitivity was 50% and 60%, respectively, whereas specificity was 100% for both. These findings confirm the low sensitivity of these measurements in adult coeliacs and thus the unreliability for screening. The high specificity, when using a threshold value of 0.9 densitometric unit, may be useful in the evaluation of adults with suspected coeliac disease.