Real-Life Case Finding of Undiagnosed Celiac Disease: Exposing the Celiac Ugly Truth?

Abstract

Introduction: Celiac disease (CD) is a common but underdiagnosed disorder. Case-finding (testing at-risk groups) is the preferred strategy for clinical detection. The primary aim was to estimate the frequency of indications for clinical testing in adults with undiagnosed CD and seronegative matched referents. The secondary aim was to estimate the frequency of actual clinical testing. Methods: Stored blood samples from 35,299 adults (age 18.1-87.7 years) with no prior diagnosis of CD were tested for tissue transglutaminase antibodies (tTGA). Samples with a positive/equivocal tTGA, defined as ≥3.0 U/ML, were tested for endomysial antibodies (EMA). Undiagnosed CD was defined by a positive EMA. 408 cases with undiagnosed CD were identified. From the tTGA negative individuals, 408 age and gender matched referents were selected. Medical records were systematically reviewed by a medical professional blinded to serology status, for indications for clinical testing, and evidence of clinical testing with serology and/or biopsies. Indications for clinical testing were consistent with the 2013 American College of Gastroenterology guidelines. Conditional logistic regression was used to test the association of case status and indications for clinical testing. Results: Eight matched pairs were excluded due to lack of data. 800 individuals were included (400 with undiagnosed CD and 400 referents), 61% female, median age of 44.2 years. The frequency of any indication for clinical testing was 159 (40%) among undiagnosed CD and 147 (37%) among referents (P=0.32). There was no difference in frequency of most indications for clinical testing among those with and without undiagnosed CD (Table 1). Hypothyroidism, dyspepsia, and chronic diarrhea were exceptions, with the former found more frequently among undiagnosed CD, and the latter two among referents. Frequency of clinical testing was low (2.3% among undiagnosed CD and 6.5% among referents; P=0.005).Table 1: Unadjusted Conditional Logistic Regression of Undiagnosed Celiac Disease and Indications for Clinical TestingConclusion: This is the largest study on real life case-finding in North America. Our study strongly suggests that the current paradigm for clinical detection of CD may not be effective, as most of the generally accepted indications for clinical testing failed to discriminate among persons with undiagnosed CD and matched referents. The study also confirms the low frequency of clinical testing for CD in the community. Clinical testing for CD, particularly with biopsies, was higher in referents, perhaps due to their higher rates of dyspepsia and chronic diarrhea.