A population-based Norwegian study found that all patients newly diagnosed with celiac disease through systematic screening had known genetic risk types. In contrast, many patients diagnosed in the past who lacked these genetic types appeared to have been misdiagnosed. The findings suggest genetic testing may be especially helpful for ruling out celiac disease or reevaluating a questionable earlier diagnosis.
“Our findings highlight how improvements in diagnostic methods have increased the accuracy of celiac disease diagnosis over time,” first author Rebecka Hjort, PhD, of the HUNT Research Center at the Norwegian University of Science and Technology in Trondheim, Norway, told GI & Hepatology News. “The study also demonstrates the value of large population-based screening studies, both for identifying undiagnosed disease and for improving diagnostic quality. Genetic testing can be a useful tool in this process, but it must always be interpreted together with symptoms, blood tests, and biopsy findings.”
The study, published in Clinical Gastroenterology and Hepatology, examined human leukocyte antigen (HLA)-DQ allotypes in adults participating in Norway’s fourth Trøndelag Health Study, or HUNT4, conducted from 2017 to 2019. Researchers screened 52,588 adults for celiac disease and studied 376 patients with previously diagnosed disease, 465 newly diagnosed patients, 163 patients with potential celiac disease, and 1,412 randomly selected controls without celiac disease.
Screening used immunoglobulin A (IgA) and immunoglobulin G antibodies against transglutaminase 2 (TG2). Seropositive patients underwent repeat serology and upper endoscopy with small-intestinal biopsies. Two independent pathologists confirmed celiac disease when villous blunting corresponding to Marsh grade 3 was present. Patients with Marsh grades 1-2 were classified as having potential celiac disease.
Researchers analyzed two genes, HLA-DQA1 and HLA-DQB1, to identify genetic types associated with celiac disease. Every newly diagnosed patient carried at least one of three established genetic risk types: HLA-DQ2.5, HLA-DQ2.2, or HLA-DQ8. In contrast, 21 of the 376 previously diagnosed patients, or 6%, had none of these genetic types.
That difference led researchers to take another look at the 21 patients’ biopsy results. Only five had the intestinal damage needed to confirm celiac disease. The other 16, or 76%, were considered to have been misdiagnosed, usually because their biopsies showed little or no intestinal damage or because biopsy records were unavailable. After removing those cases, the previously diagnosed group fell from 376 to 360 patients.
Researchers also reviewed 40 randomly selected patients with a previous celiac disease diagnosis who did carry DQ2 or DQ8 allotypes. Three, or 7.5%, were classified as false positives. Based on that finding, the researchers estimated that about 43 of the original 376 previously diagnosed patients, or 11.4%, may have received a false-positive historical diagnosis. Additional analyses that accounted for these possible misdiagnoses showed a higher prevalence of the genetic types most strongly linked to celiac disease.
After confirmed misdiagnoses were removed, DQ2.5 remained the most common genetic risk type. It was found in 86% of previously diagnosed patients and 88% of newly diagnosed patients, compared with 22% of people without celiac disease. Having two copies of DQ2.5 was linked to the greatest risk, with 16.7 times the odds of celiac disease relative to patients without DQ2.5, DQ2.2, or DQ8. Another DQ2.5 combination, DQ2.5/DQX, was associated with 5.7 times the odds.
A gene-dose pattern also emerged for DQ8. After accounting for DQ2.5, patients with two copies of DQ8 had 8.8 times the odds of celiac disease, and those with one copy had 2.6 times the odds relative to patients without DQ2.5, DQ2.2, or DQ8. Patients with two copies of DQ2.2 had 3.4 times the odds, although the study did not find a clear overall association between DQ2.2 and celiac disease.
Subgroup analyses showed few major differences in HLA-DQ patterns between previously and newly diagnosed patients, and HLA-DQ patterns did not vary meaningfully with IgA TG2 antibody levels among newly diagnosed patients.
For clinicians, the findings suggest that HLA testing is more useful for ruling out celiac disease than confirming it. After confirmed misdiagnoses were excluded, the negative predictive value of testing for DQ2.5, DQ2.2, and DQ8 was 100% among newly diagnosed patients and 99% among previously diagnosed patients. The positive predictive value was 38% and 32%, respectively.
Limitations included the possibility that some previously diagnosed patients were still misclassified and that duodenal bulb biopsies may have contributed to false-positive diagnoses. The study also included patients from a single region of Norway, which may limit how broadly the findings apply.
The Research Council of Norway, the Liaison Committee for Education, Research and Innovation in Central Norway/Samarbeidsorganet, and the Norwegian Celiac Society funded the study. The authors reported no competing interests.
Expert Insight
Dr. Hjort elaborated on the findings with GI & Hepatology News.
Given the 11.4% estimated false-positive rate among previously diagnosed patients, should gastroenterologists consider reassessing older celiac diagnoses, particularly when the original serology or biopsy documentation is incomplete?
Dr. Hjort: Our findings suggest that some historical diagnoses may warrant re-evaluation, especially when the original test results are missing, difficult to verify, or were made using diagnostic methods that differ from today's standards.
Importantly, we do not suggest that all historical celiac diagnoses should be questioned. Most previously diagnosed patients in our study had findings consistent with celiac disease. However, among individuals lacking the necessary HLA risk alleles, a substantial proportion appeared to have been diagnosed on insufficient evidence.
For patients with incomplete documentation, uncertain histology, or atypical clinical histories, reassessment using current diagnostic tools, including HLA typing, may help clarify the diagnosis.
All newly diagnosed cases carried DQ2.5, DQ2.2, or DQ8, giving HLA testing a 100% negative predictive value in this group. How should that finding change the way clinicians use HLA typing when celiac disease remains diagnostically uncertain?
Dr. Hjort: HLA testing is particularly valuable when the diagnosis is unclear. For example, some patients start a gluten-free diet before they are properly tested, while others have inconclusive biopsy results or conflicting blood test results. In these situations, a negative HLA test can provide reassurance that celiac disease is not the explanation. A negative HLA result may spare patients from further invasive investigations and prompt clinicians to consider alternative explanations for symptoms.
You found clear gene-dose effects for DQ2.5 and DQ8, but HLA-DQ type was not associated with IgA TG2 levels in newly diagnosed patients. Does HLA genotype have any practical role in predicting disease phenotype or severity once celiac disease is established?
Dr. Hjort: Our study confirms that HLA genotype is strongly associated with disease susceptibility, but we found little evidence that it predicts serological activity once celiac disease has developed. While individuals carrying two copies of DQ2.5 had the highest risk of developing celiac disease, HLA genotype was not associated with anti-TG2 antibody levels among newly diagnosed patients.
These findings suggest that HLA typing is more useful for understanding disease risk than disease severity.