Discriminant analysis for the diagnosis of childhood celiac disease.

A new statistical approach to the analysis of laboratory data has been introduced to optimize the use of absorption tests and gliadin antibody measurement for the diagnosis of childhood celiac disease. Serum antigliadin antibodies, as well as blood xylose, iron, and tryglycerides after oral load, were evaluated in 40 celiac children and 43 age-matched patients affected by other gastrointestinal diseases. Each test evaluated individually gave a considerable rate of false-positive and false-negative results. Discriminant coefficients produced for each test were used to compute a score that allowed correct classification of 99% of patients; 2.3% of false-positive and no false-negative results were recorded. This approach improves significantly the overall sensitivity and specificity for celiac disease of these laboratory tests and we propose its use for screening patients to be submitted to jejunal biopsy.