The familial incidence of celiac disease (CD) confirms its genetic basis, although acquired factors are also involved. Many authors have reported a linkage between celiac disease and HLA antigens, but there are differences which depend on geographical areas, and nowadays the study must be done at the genetic level. Thirty-eight celiac children and 52 normal controls were included in this study. All individuals were chosen from the Castilla and Leon area. We used the reverse ¿dot block¿ technique, using sequence-specific oligonucleotide DNA probes (Cetus, USA) to determine the HLA-DQA1 alleles in DNA samples previously amplified by PCR (polymerase chain reaction). The different frequency of alleles in patients and controls was assessed by 3 statistical tests: chi square (chi(2)), relative risk (RR) and etiologic fraction (EF). A very high frequency of DQA1*0201 (chi(2):p <0.0001) and DQA1*0501 (chi(2): p <0.0001) alleles was observed in patients; all but one (97%) had the DQA1*0501 allele vs. 40% of controls (RR: 37.00; EF: 0.955). The DQA1*0201 allele also had a high prevalence in celiacs (58%)(RR: 1.375: EF:0.438). The DQA1*01 allele was only found in 10.5% of patients compared to 79% of controls (chi(2): p <0.0001) and the DQA1*03 allele was also decreased in celiacs. There was only one celiac girl without the DQA1*0501 allele. She had no other clinical or serological differences, as compared to the other patients. In the study of allele subtypes, among the DQA1*01 allele, 50% of patients were positive for DQA1*101 and the remaining 50% had DQA1*0102, but none of the individuals were positive for DQA1*0103. Among normal controls, 32 individuals (61.5%) expressed the DQA1*0102 subtype, 15 (28.9%) the DQA1*0101 subtype and 5 (9.6%) the DQA1*0103 subtype. All positive cases for DQA1-*05 belong to the DQA1* 0501 subtype, in both celiac and control groups. There were 10 possible combinations of HLA-DQA1 genes, but we found a very unequal distribution in both celiacs and controls. Only 4 genotypes were found in patients (DQA1*01/1*0501, DQA1* 0201/1*0501, DQA1*03/1*0501 and DQA1*0501/1*0501) and 8 in controls. The DQA1*0201/1*-0501 genotype was the most discordant, being positive in 55.3% of patients vs. 3.8% of controls (chi2:p < 0.001; RR: 1.235; EF:0.534). None of the 90 individuals studied expressed the DQA1*0311*03 genotype. The DQA1*0201/1*03 genotype was not shown by any control and by only 1 celiac patient. It is noteworthy that the DQA1*0201/1*0501 genotype was more frequent than the homozygous genotype DQA1*0501/1*0501. Our results do not suggest a dosage effect for the DQA1*0501 allele. The determination of the HLA-DQA1 gene is a helpful tool for the screening of individuals with a high risk of being celiacs.