Abstract Background Celiac disease (CeD) is a chronic autoimmune disorder triggered by the ingestion of gluten in persons with the genetic risk factors human leukocyte antigens (HLA)-DQ2 and DQ8. Diagnostic evaluation of patients with suspected CeD often includes serology tests for tissue transglutaminase (tTG) antibodies. While the presence of HLA-DQ2 or DQ8 is required for the initiation and development of CeD, the genetic risk associated with the various alleles is not equivalent. Previous studies in patients with normal total IgA demonstrated the risk gradients of various HLA-DQ2 and DQ8 heterodimers in CeD. However, the impact of both gene variants in IgA deficient patients has not been investigated. In this study, HLA-DQ2 and HLA-DQ8 as risk factors for tTG-IgG positivity in IgA deficient patients were evaluated. Methods Retrospective data was collected from a cohort of individuals (n=3,515) with IgA concentrations below the age-dependent reference interval with suspicion for CeD who were tested for HLA-DQ genotyping and tTG-IgG (reference interval < 6 U/mL) at Mayo Clinic Laboratories. Total IgA was measured using quantitative nephelometry (Siemens BN™ II System); HLA typing was performed using the reverse SSO DNA typing assay (LABType SSO kits). tTG-IgG was determined by enzyme-linked immunosorbent assay (QUANTA Lite R h-tTG, Werfen Diagnostics) on the Dynex Agility System. The prevalence and frequency distribution of HLA-DQ2 and DQ8 haplotypes were determined and the odds ratio (OR) was calculated to evaluate the risk gradient for tTG-IgG positivity based on different HLA-DQ variants. Results Among IgA deficient patients, 58% (n=2,041) expressed HLA-DQ2 or DQ8 haplotypes. Relative to patients without permissive alleles, patients homozygous for HLA-DQ2.5 exhibited a significantly increased risk for tTG-IgG positivity (OR = 10.98, 95% CI: 6.56-18.18, p < 0.001) while those homozygous for HLA-DQ2.2 and DQ8 did not (DQ2.2/DQ2.2_OR = 0.72, 95% CI: 0.04-3.44, p = 0.75; DQ8/DQ8_OR = 0.96, 95% CI: 0.05-4.64, p = 0.97). Patients heterozygous for HLA-DQ2.5 (DQ2.5/x) displayed a lower odds ratio relative to homozygous patients (OR = 3.31, 95% CI: 2.29-4.82, p < 0.001); however, the risk for tTG-IgG positivity increased in the presence of HLA-DQ2.2 (DQ2.5/DQ2.2_OR = 7.59, 95% CI: 4.48-12.65, p < 0.001) or HLA-DQ8 (DQ2.5/DQ8_OR = 6.26, 95% CI: 3.55-10.73, p < 0.001) alleles. Conversely, in the absence of HLA-DQ2.5, HLA-DQ2.2 heterozygotes (DQ2.2/x_OR = 2.80, 95% CI: 1.77-4.39, p < 0.001) and HLA-DQ8 heterozygotes (DQ8/x_OR = 0.99, 95% CI: 0.52-1.77, p = 0.99) displayed lower correlations with tTG-IgG positivity, suggesting a potential risk elevation for tTG-IgG positivity is associated with the combination of DQ2.5 and DQ2.2 or DQ8. Conclusions Our study indicates the risk gradient of HLA-DQ genes in IgA deficient patients. Overall, compared to patients without HLA-DQ2 or DQ8 haplotypes, the homozygous HLA-DQ2.5 genotype had the highest correlation with tTG-IgG positivity, followed by HLA-DQ2.5/DQ2.2 and HLA-DQ2.5/DQ8 heterodimers. The results suggest that HLA-DQ2.2 or HLA-DQ8 alone does not confer risk, but when in the presence of HLA-DQ2.5 these alleles may increase the risk for tTG-IgG positivity. Results from this study shed light on the risk of CeD in patients with IgA deficiency.