Celiac disease in pediatric patients according to HLA genetic risk classes: a retrospective observational study

Background: Celiac disease (CD) is an autoimmune enteropathy in which HLA-DQ haplotypes define susceptibility. Our aim was to evaluate if belonging to a certain HLA-DQ class risk could be associated to the clinical, serological and histological presentation of CD. Methods: We performed a retrospective observational monocentric study including all 300 patients diagnosed with CD, who underwent HLA typing. Clinical, serological and histological data was collected from clinical records and their association with HLA-DQ class risk was verified through statistical tests. Results: In our sample mean age at onset was 6.7 ± 4.2 years, with a prevalence of females (n = 183; 61%), typical symptoms (n = 242; 80.6%) and anti-tTG IgA ≥ 100 U/mL (n = 194; 64.7%). Family history was present only in 19% (n = 57) of patients, and it was not significantly associated with any of the clinical and demographical data analyzed or the belonging to a certain HLA-DQ class risk. We found in the male population more frequently a coexistence of CD and atopic syndrome (males: n = 47; 40.2%; females: n = 50; 27.3%; p = 0.020). Early age of onset, instead, was associated with typical symptoms (m = 6.4 ± 4; p = 0.045) and elevated liver enzymes (m = 5 ± 3.8; p < 0.001), while later age of onset was associated with presence of other autoimmune diseases (m = 8.2 ± 4; p = 0.01). We observed statistically significant influences of HLA class risk on antibodies and liver enzymes levels: G1, G4 and G2 classes showed more frequently anti-tTG IgA ≥ 100 U/mL (n = 44; 80%, n = 16; 69.6%, n = 48; 67.6% respectively; p-value = 0.037), and in patients from G2 class we found enhanced liver enzymes (n = 28; 39.4%; p-value = 0.005). HLA class risk was still significantly associated with anti-tTG ≥ 100 (p = 0.044) and with hypertransaminasemia (p = 0.010) after a multiple logistic regression adjusted for the effect of gender, age at onset and family history. Conclusions: We failed to prove an association between HLA-DQ genotypes and the clinical features in our CD pediatric patients. Although, our results suggest an effect of the DQB1–02 allele not only on the level of antibodies to tTG, but possibly also on liver involvement