Celiac disease: the new proposed ESPGHAN diagnostic criteria do work well in a selected population.

BACKGROUND: The need for an early and accurate diagnosis in celiac disease (CD) has focused attention on new diagnostic approaches, based on the efficiency of serological markers and the high negative predictive value of human leukocyte antigen (HLA) non-DQ2/8. METHODS: We performed a retrospective review of all of the patients suspected of having CD who had undergone a small bowel biopsy in our gastroenterology unit. All symptomatic children with serological marker at time of biopsy (immunoglobulin A-tissue transglutaminase antibody, endomysial antibody, and HLA genotype) were included. The triple test (TT) was positive if immunoglobulin A-tissue transglutaminase antibody was 10 times the upper limit of normal, plus positive endomysial antibody plus human leukocyte antigen-DQ2/DQ8. RESULTS: A total of 150 patients met the inclusion criteria and were enrolled in the study. One hundred sixteen were positive for the TT; 113 of 116 (97.4%) had a Marsh 2/3 histological lesion and had been considered to have CD. Thus, positive predictive value of the TT was 97.4%. The other 3 cases (2.6%) had Marsh 0/1 lesion, so we consider them to be false-positives for the TT; however, on follow-up, all 3 children developed histological damage after a gluten challenge. Finally, the positive predictive value of the TT was 100%. Thirty-four patients were negative for the TT: 22 patients are celiac, 3 are celiac but challenge gluten diet is pending, and the 9 patients left have other gastrointestinal disorder. CONCLUSIONS: Our study supports the view that in selected children who are symptomatic and positive for the TT, CD diagnosis could be established independent of histological findings.