Objectives: The European Society for Pediatric Gastroenterology, Hepatology, and Nutrition (ESPGHAN) 2012 guidelines, enabled for the first time, a nonbiopsy approach in the diagnosis of celiac disease (CD). We aimed to prospectively assess 4 tissue-transglutaminase (tTg) IgA assays of 4 random-access analyzers and examine their accuracy in diagnosing CD without a biopsy.
Methods: We enrolled 186 consecutive children referred to upper endoscopy and intestinal biopsy. One group included 109 patients with positive tTg that was referred for suspected CD. Another group included 77 patients with negative tTg referred because of other indications. All participants had a blood sample taken at the time of endoscopy. Samples were tested with 4 tTg IgA assays on automated analyzers and 1 Elisa kit. All intestinal biopsies were evaluated by a local pathologist, a central pathologist, and a CD expert blinded to each other. CD was diagnosed when full agreement was reached. Analytical performance of the assays included precision with controls and samples, lot to lot variation, and carryover.
Results: In our cohort, all tested tTg IgA-automated assays showed sensitivities above 98% and specificities above 99%. ROC analysis demonstrated AUC (area under the curve) >0.99 for all 4 analyzers. The positive-predictive values (PPV) were all >0.99 and negative-predictive values (NPV) were >0.97. The Elisa kit had sensitivity of 95%, specificity of 96%, AUC of 0.96, PPV of 0.98 and NPV of 0.93.
Conclusion: CD can be accurately diagnosed without biopsy based on tTg IgA levels at least 10 times the ULN using the 4 high-volume random-access analyzers used in our study.