Poster Presentation International Celiac Disease Symposium 2026

Are tissue transglutaminase IgA results reliable at the 10x upper limit of normal? An important clinical decision point in diagnosing paediatric no-biopsy coeliac disease (#262)

Shaun SC Ho 1 2 , Michael Fernandez 3 , Vraj Evans 3 , Yang Tran 4
  1. The Royal Children's Hospital Melbourne, Parkville, VIC, Australia
  2. Murdoch Children’s Research Institute, Parkville, VIC, Australia
  3. The Royal College of Pathologists of Australasia Quality Assurance Programs, St Leonards, NSW, Au
  4. St Vincent’s Pathology Melbourne, Fitzroy, VIC, Australia

Aim: ESPGHAN 2020 criteria permit a diagnosis of no-biopsy coeliac disease in children when tissue transglutaminase immunoglobulin A (tTG-IgA) is ≥10 times the upper limit of normal (ULN) with a confirmatory positive endomysial antibody (EMA). This study evaluated the variability of tTG-IgA immunoassays between methods at the 10x ULN threshold.

 

Methods: Retrospective review of The Royal College of Pathologists of Australasia Quality Assurance Programs (RCPAQAP) Coeliac Disease Antibodies surveys from 2022 to mid-2026. Each QAP survey distributed a single-donor serum sample to participating laboratories (n=62-64). Method-group medians were expressed as multiples of ULN stated by the manufacturers and local laboratory reference interval for each reagent.

 

Results: Thirty-six QAP reports were reviewed. Six tTG-IgA reagents returned sufficient quantitative data. The reports included 10 normal, 22 elevated, and four ‘no target set’ (<80% consensus for qualitative interpretation). Seven reports showed concordant ≥10x ULN results across all method-groups, 11 returned consistent results between 1 and 10x ULN, and four were re-issued distributions of a single donor sample (35-year-old male with coeliac disease, EMA positive) over three years, which showed variability around the 10x ULN threshold. For example, in survey 25-01, tTG-IgA expressed as multiples of ULN varied from 2.9x to 14.1x ULN between various reagents and manufacturer/local laboratory reference-interval sources. Thirteen surveys (36%) had no consensus in EMA target result.

 

Conclusion: Quantitative tTG-IgA results are not interchangeable between immunoassays, and expression as multiples of ULN does not reliably harmonise them. Discordance at the 10x ULN clinical decision threshold affected 11% of surveys (4/36). EMA consensus was absent in a third of QAP reports. The real-world discordance may be more prevalent. Prospective multicentre studies are required to validate the 10x ULN threshold between assay methods and histology. Laboratory internal and external quality assurance programs should monitor tTG-IgA ≥10x ULN variations within and between methods.