To study the optimal cut-off value for anti-tissue transglutaminase type 2 IgA antibodies (TG2A) in serum to select for diagnostic small bowel biopsies for celiac disease in children with type 1 diabetes mellitus.

Children with type 1 diabetes mellitus with elevated TG2A titers and duodenal biopsies performed during the course of their diabetes treatment were included. Anti-endomysial antibodies were recorded if present. The optimal TG2A cut-off value, expressed as the ratio between obtained value and upper limit of normal (ULN), was determined using receiver operating characteristic curve analysis and compared with the cut-off value used in the European Society for Pediatric Gastroenterology, Hepatology and Nutrition guidelines in terms of sensitivity, specificity, positive and negative predictive value.

We included 63 children. The optimal cut-off value for performing biopsies is demonstrated to be 11 times the ULN. Raising the cut-off value from 3 times the ULN to 11 times the ULN changed sensitivity from 96% to 87% and increased specificity from 36% to 73%, increased the positive predictive value from 88% to 94% and lowered negative predictive value from 67% to 53%. The percentage of normal histology was decreased from 12% to 6%.

Increasing the TG2A cut-off value for performing duodenal biopsies in children with type 1 diabetes mellitus and suspected celiac disease leads to a substantial reduction of unnecessary biopsies. We advocate to adapt the European Society for Pediatric Gastroenterology, Hepatology and Nutrition 2012 guidelines for this group of children, including monitoring patients with TG2A levels of less than 11 times the ULN over time.