OBJECTIVES: While the algorithm to diagnose celiac disease (CD) in children with elevated anti-transglutaminase IgA (TGA-IgA) titers (>10 times upper limit of normal, ULN) is well defined, the management of children with low TGA-IgA values represents a clinical challenge. We aimed to identify the diagnostic value of persistently low positive TGA-IgA titers in predicting CD in children. METHODS: We retrospectively analyzed children with symptoms or signs of CD, not eligible for a no-biopsy approach. We included children with at least 2 TGA-IgA measurements, endomysial antibody (EMA) assessment and esophagogastroduodenoscopy with biopsies. TGA-IgA values were provided as multiples of ULN. Patients were classified in groups according to median TGA-IgA values: A (TGA-IgA>1 ≤ 5 × ULN; defined as "low-positive"), B (TGA-IgA > 5 < 10 × ULN; "moderate-positive"), and C (controls). RESULTS: Data of 281 children were analyzed. Of 162 children in group A, CD was diagnosed in 142 (87.7%), whereas normal duodenal mucosa was found in 20. In group B, all 62 children (100%) received a CD diagnosis. Group C included 57 controls. EMA were undetectable in 31 (15%) of mucosal atrophy cases. On the receiver-operating characteristic curve (area under the curve = 0.910), a mean value of 1.7 ULN showed a sensitivity of 81.4% and specificity of 81.8% to predict mucosal damage. CONCLUSIONS: Repeated low or moderate TGA-IgA values (<5 ULN or <10 ULN) are good predictors of a CD diagnosis. Symptomatic children with persistently low positive TGA-IgA titers should undergo esophagogastroduodenoscopy regardless of their EMA status.
OBJECTIVES: While the algorithm to diagnose celiac disease (CD) in children with elevated anti-transglutaminase IgA (TGA-IgA) titers (>10 times upper limit of normal, ULN) is well defined, the management of children with low TGA-IgA values represents a clinical challenge. We aimed to identify the diagnostic value of persistently low positive TGA-IgA titers in predicting CD in children. METHODS: We retrospectively analyzed children with symptoms or signs of CD, not eligible for a no-biopsy approach. We included children with at least 2 TGA-IgA measurements, endomysial antibody (EMA) assessment and esophagogastroduodenoscopy with biopsies. TGA-IgA values were provided as multiples of ULN. Patients were classified in groups according to median TGA-IgA values: A (TGA-IgA>1 ≤ 5 × ULN; defined as "low-positive"), B (TGA-IgA > 5 < 10 × ULN; "moderate-positive"), and C (controls). RESULTS: Data of 281 children were analyzed. Of 162 children in group A, CD was diagnosed in 142 (87.7%), whereas normal duodenal mucosa was found in 20. In group B, all 62 children (100%) received a CD diagnosis. Group C included 57 controls. EMA were undetectable in 31 (15%) of mucosal atrophy cases. On the receiver-operating characteristic curve (area under the curve = 0.910), a mean value of 1.7 ULN showed a sensitivity of 81.4% and specificity of 81.8% to predict mucosal damage. CONCLUSIONS: Repeated low or moderate TGA-IgA values (<5 ULN or <10 ULN) are good predictors of a CD diagnosis. Symptomatic children with persistently low positive TGA-IgA titers should undergo esophagogastroduodenoscopy regardless of their EMA status.