Objectives: Although HLA class II genes are undoubtedly involved in the development of celiac disease, only 40% of the genetic susceptibility to celiac disease can be attributed to them. This implies that other genes, inside or outside the HLA region, play a role in the pathogenesis of the disease. Therefore, the main objectives of this study were to determine specific polymorphisms of the cytotoxic T lymphocyte-associated antigen 4 (CTLA4) and tumor necrosis factor α (TNF-α) genes, their relationship and interaction with HLA-DQA1 and -DQB1 genes, and the individual and mutual impact of polymorphisms on the predisposition and clinical expression of celiac disease. Patients and methods: The study included 202 patients diagnosed with celiac disease according to the revised ESPGHAN's (European Society for Pediatric Gastroenterology, Hepatology and Nutrition) criteria. The patients were clinically examined at the Department of Internal Medicine, School of Medicine Rijeka, and at the Department of Pediatrics, University Hospital Rijeka. The age of the onset of the disease, initial symptoms, type of disease, familial occurrence of the disease and the association with other autoimmune diseases were determined for each patient. The control group consisted of 400 healthy volunteer blood donors who were matched according to gender and place of residence. The samples were stored in the DNA bank of the Institute of Biology and Medical Genetics, Faculty of Medicine Rijeka. Polymerase chain reaction (PCR) and the subsequent restriction using appropriate restriction endonucleases were used for the identification of TNF-α gene promoter region polymorphisms -238 and -308 and CTLA4 gene polymorphisms +49 and CT60. HLA class II alleles genotyping was performed by PCR using allele-specific primers. Results: The study showed no statistically significant differences in the clinical presentation of the disease according to gender. Male subjects had a significantly higher (p=0.023) incidence of the familial type of the disease (42.5%) compared to female subjects (27.0%). There were no statistically significant differences in the age of onset of the disease between male and female patients (p=0.798). Hundred and six patients (52.5%) presented with the classical and ninety six (47.5%) with the nonclassical type of disease, with no significant differences between male and female subjects (p=0.321). Thirty seven patients (18.3%) suffered from other autoimmune diseases, while exocrine pancreatic insufficiency was found in 14 (14.9%) patients. At least one HLA-DQB1 risk allele was present in 97.1% of patients, while DQA1 risk alleles were present in 70.2% of patients. The presence of the CTLA4 +49 and CT 60 polymorphisms had no individual impact on the genetic susceptibility to celiac disease. Specifically, the frequency of A and G alleles and genotypes did not differ significantly between patients and healthy controls (p=0.206 and p=0.498, respectively), except between male subjects and controls regarding CTLA4 +49 alleles (p=0.027), genotype A/A (p=0.012) and A/G (0.048). However, the interaction of risk alleles for both CTLA4 polymorphisms showed significant impact on the occurrence of the disease, which was confirmed by the analysis of the haplotype (OR=4.66, p=0.0003). Although the impact of CTLA4 polymorphisms on the clinical expression of the disease did not show significant effects (p> 0.05), carriers of both risk alleles presented with the disease at an earlier age (4-5 years earlier) and suffered from exocrine pancreatic insufficiency more often (57.1% vs. 38.8%). The presence of the TNF-α gene -308 polymorphism strongly correlated with the presence of the disease at the significance level of p<0.001. TNF-α -238 polymorphism showed no correlation (p>0.05) with celiac disease, regardless of the type of disease or the gender of the subject. Examining the interaction of the TNF-α gene -308 and -238 polymorphisms has shown that TNF-α-308/-238 AG haplotype represents a risk factor for the development of the disease (OR = 2.53, p <0.0001). Statistical analysis has shown, however, that the presence of risk alleles had no impact on clinical parameters of the disease, both when observed individually or combined (p>0.05). The investigation of TNF-α and CTLA4 genes showed the importance of the interaction between TNF-α -308, CTLA4 +49 and CTLA4 CT60 polymorphisms. Together they contribute to the susceptibility or resistance to celiac disease depending on the presence of protective and/or risk alleles. Patients who are carriers of all three risk alleles had an associated autoimmune disease in 25.6% of cases, as compared to 17.6% of patients in whom associated disease did not occur. Homozygotes for the risk DQB1 alleles often had individual risk CTLA4 +49 G, CT60G CTLA4 and TNF-α-308A alleles present as well, which confirms their role and epistatic interactions in the development of celiac disease. Conclusion: In this study, for the first time epistatic influences of specific polymorphisms of CTLA4 and TNF-α genes were analyzed, as well as interactions with genes HLA-DQA1 and-DQB1, regarding the predisposition and the clinical expression of celiac disease. Research has shown that the interaction between CTLA4 +49 G, CT60G and TNF-α-308A polymorphism in addition with risk HLA DQA1 and-DQB1 genes, affect the predisposition to celiac disease in our patients, but not the clinical expression of the disease, except for certain trends which did not reach statistical significance. In addition to representing the first genetic study of celiac disease in Croatian patients, the study brings important information on the frequency of the studied polymorphisms in our population and indirectly contributes to the study of other complex polygenic diseases in this region.