Human Leukocyte Antigen-DQ Allotypes in Adults Screened for Celiac Disease Reveal Historical Misdiagnosis - Results From the Trøndelag Health Study

BACKGROUND AND AIMS: The human leukocyte antigen (HLA) -DQ2.5, -DQ2.2 or -DQ8 allotypes are necessary for celiac disease (CeD) development. Their distribution may vary between adults with known, previously diagnosed CeD and newly diagnosed CeD detected through screening. METHODS: In the population-based HUNT4 study, 52,588 Norwegian adults were screened for CeD, identifying 846 (376 known, 465 new biopsy-confirmed and 163 potential) cases. Anti-TG2 seropositive subjects and a random sample of 1412 controls underwent HLA-DQA1- and HLA-DQB1-sequencing to infer relevant HLA-DQ allotypes. RESULTS: All new cases were DQ2 or DQ8 positive, while 21 known cases were DQ2 and DQ8 negative, suggesting misdiagnosis. Histological reassessment confirmed the diagnosis in 5/21. After adjusting known cases to 360, DQ2.5 frequencies were 87.8% in confirmed cases, 76.7% in potential cases, and 22.9% in controls. Among DQ2.5-negative subjects, DQ8 was most frequent, with 8.5% in confirmed cases and 10.4% in potential cases. DQ2.5/DQ2.5 carried the highest risk (odds ratio [OR] 16.7, 95% confidence interval 9.7-31.4), followed by DQ2.5/DQX (OR 5.7, 4.7-7.0), DQ8/DQ8 (OR 8.8, 4.6-16.8), and DQ8/DQX (OR 2.6, 1.7-3.8). ORs for DQ2.2/DQ2.2 and DQ2.2/DQX were 3.4 (0.5-15.0) and 0.8 (0.4-1.6), respectively. No clear differences were observed between known and new cases, however ORs for DQ2.5/DQX and DQ8 were higher in new cases. CONCLUSIONS: This study found a gene-dose effect for DQ2.5, DQ8, and DQ2.2 in both known and new cases, with a fairly high proportion of DQ8 for a Nordic population. A substantial proportion of DQ2 or DQ8 negative subjects among known cases, but absent in new cases, could be attributed to historical misdiagnosis.