OBJECTIVES: (1) To assess the predictive value of anti-tissue transglutaminase immunoglobulin A (IgA) antibodies (a-tTG) by chemiluminescence immunoassay (CLIA) related to duodenal histology in children with suspected celiac disease (CD) and (2) to determine the cut-off value of a-tTG by CLIA that allows diagnosis of CD avoiding biopsy. METHODS: Retrospective, descriptive, observational study in children between 1 and 16 years of age, studied for CD. Patients with IgA deficiency and those on a gluten-free diet were excluded. Sensitivity (S), specificity (Sp), positive predictive value (PPV), and negative predictive value (NPV) were calculated for ≥10, 30, and 50 times the normal a-tTG (normal value [NV]) compared with histology, evaluated by blinded pathologists. RESULTS: The total number of patients included was 262. The a-tTG IgA by CLIA with a cut-off point of 20 chemiluminescent units (CU) had a S of 99.5%, Sp, 10.26%, PPV, 86.38%, NPV, 80%. When a cut-off value of a-tTG ≥ 10 times NV (200 CU) was used, the S, Sp, PPV, and NPV were 65.4%, 69.23%, 94.2%, and 25.96%, respectively. Likewise, with a value ≥30 times NV (600 CU) the correlation with histology was 99.03%, reaching a PPV of 100% with a cut-off value ≥50 NV (1000 CU). Combining both methods a-tTG IgA + EMA IgA, we obtained similar results to the a-tTG IgA level for the proposed cut-off values. CONCLUSIONS: We suggest that the use of 30 times the NV cutoff values would be the best predictor of which patients do not need a biopsy.