DNA from 164 Caucasian type I (insulin-dependent) diabetic patients and 200Caucasian nondiabetic control blood donors were analyzed by the polymerase chain reaction technique for HLA-DR4 and the associated Dw and DQB subtypes of DR4. The DQw8 subtype of HLA-DR4 was associated with type I diabetes in all DR4 subgroups (DR4/3 and DR4/non-3). Dw subtypes of DR4 other than Dw10 did not confer additional association with type I diabetes. Thus, the DQ region appears to provide the primary major histocompatibility association with type I diabetes in most DR4 patients. Diabetes 38: 1989 T he serological type HLA-DR4 is associated with five alleles of the DRB1 locus (Dw4, Dw10, Dw13, Dw14, and Dw15) and two alleles of the DQB1 locus (DQw7 and DQw8) (1,2). The DQw8 serological specificity, previously called DQw3.2, is associated with type I (insulin-dependent) diabetes, whereas the DQw7 specificity, previously called DQw3.1, is not associated with type I diabetes (3-9). Tait et al. (10) recently reported that the DQw8 association exists only in type I diabetic individuals who type as HLA-DR4/3. DR4 subtypes (Dw4 and Dw10) are also associated with type I diabetes (11-13); however, the role of other Dw specificities remains controversial (11)(12)(13)(14).To further clarify the specific major histocompatibility associations with type I diabetes, we typed 164 type I diabetic patients, of which 120 typed as HLA-DR4, and 200 blood donor controls, of which 55 typed as HLA-DR4, for DQB1 and DRB1 subtypes. The association of DQw8 and type I diabetes was seen in all DR4 subgroups analyzed (DR4/3 and DR4/non-3). The distribution of Dw subtypes, other than Dw10, was similar in our HLA-DR4 type I diabetic patients and nondiabetic control subjects. With these data, we calculated that the absolute risks for type I diabetes in unrelated individuals homozygous for DR4 (DQw8) or heterozygous for DR4 (DQw8) /DR3 (DQw2) are 6.7 and 8.5%, respectively, frequencies approaching those seen in family studies (14). Thus, the presence of DQw8 appears to be the primary association with type I diabetes in DR4 patients.
RESEARCH DESIGN AND METHODS