ABSTRACT
INTRODUCTION: Serological markers are of great interest in coeliac disease (CD), although intestinal biopsy is still the gold standard for establishing the diagnosis. Tissue transglutaminase IgA antibodies (AATGt-IgA) and antiendomysial antibodies IgA (AAE-IgA) are closely correlated to intestinal damage observed in biopsies. Villous atrophy (Marsh 3) plays a major role in CD diagnosis. Marsh 2 stage (crypt hyperplasia) as a CD marker is still under debate. OBJECTIVE: To ascertain an AATGt-IgA level that corresponds to a positive predictive value (PPV) of 100% for a histological CD diagnosis. MATERIAL AND METHODS: A series of 120 patients younger than 14 years, non- IgA deficient, who underwent an intestinal biopsy and were positive for both serological markers (AATGt-IgA and AAE-IgA). For AATGt-IgA, according to the manufacturer's recommendations, a value greater than 16 IU/mL is considered as a positive value. The PPV of AATGt was determined for different cut-off points. RESULTS: The histological findings distribution is directly correlated to the AATGt-IgA cut-off point. When the cut-off point is set above 7.5-10.6 times the commercial reference value, there is a 2.1% of Marsh 2 lessions and 93.4% of Marsh 3; above 10.6 times the reference value, all biopsies where Marsh 3 (100%). The PPV that considers Marsh 3 is (93.4%). The PPV, for considering Marsh 3 is low (55%) when AATGt-IgA serology is positive with levels between 16 and 67 IU/ml (1-4.2 times the cut-off point) and a higher value (92%) for concentrations between 68 and 118 IU/ml (4.3-7.4 times) and for cases with 69-170 IU/ml (7.5-10.6 times); above 170 IU/ml (>10.6 times) PPV is 100%. CONCLUSION: The use of values higher than the recommended cut-off point must logically improve specificity and PPV. In 31.6% patients positive for AATGt-IgA and AAE-IgA (38/120) it would have been possible to diagnose the disease without intestinal biopsy as of the PPV was 100%. It is not possible to standardise results as there are different commercial kits with variable cut-off points, so we must be cautious when setting recommendations based on AATGt-IgA.
