In UC, as in many other diseases with a suspected autoimmune etiology or pathogenesis, an association with certain HLA class II specificities has been investigated. In the Japanese, several studies have shown a positive association with DR2, but in Caucasian populations the results are conflicting. Therefore we undertook a study of HLA class II gene association with UC in a large group of white Madrid patients and ethnically matched controls, using molecular biology techniques to investigate whether any allelic subspecificity within the HLA-DR2 group is associated with susceptibility to or protection against UC. Patients with ulcerative colitis (n = 107) and 200 controls were typed using molecular, DNA-based techniques for HLA-DRB1, DQA1 and DQB1 alleles. Those HLA-DR2+ were then specifically typed for the individual alleles within the HLA-DR2 group. We observed a positive association with HLA-DR15 (P=0.021) and its subtype DRB1*1501 (P=0.018). HLA-DRB1*1502 was also increased, although its frequency both in patients and controls was very low. When the HLA-DR2+ population was studied, HLA-DRB1*1601 was significantly decreased in patients (P=0.026). Both HLA-DR3 (P=0.002) and HLA-DQB1*02 (P=0.001) were also negatively associated with the disease, the latter especially with pancolitis. Therefore, HLA class II association with UC is complex, and separate alleles confer either susceptibility or resistance. Conflicting results with HLA-DR2 appear to be due to the presence in this group of both positively associated (HLA-DRB1*1501 and DRB1*1502) and negatively associated (HLA-DRB1*1601) subspecificities. Moreover, HLA-DR3 and HLA-DQB1*02 are associated negatively.