DQA2 U allele: a genetic marker for relapse of Graves' disease.

The association of HLA-DR3 with Graves' disease in Caucasoids is well established but its significance is unclear and its clinical value as a predictive parameter for relapse after a course of antithyroid drug therapy is controversial. We have further investigated the predictive value at the genomic level in 51 patients with Graves' disease who were treated with a 6-month course of carbimazole and followed up for 2 years. Using DNA-restriction fragment length polymorphism (DNA-RFLP) allogenotyping, (i) complete concordance of HLA-DR assignment was observed between serological and DNA-RFLP analysis of all but one of 51 patients with Graves' disease; (ii) the DR beta 17(1)-DQ alpha 2-DQ beta 2a (a DNA-RFLP allogenotype of the classical Northern European haplotype of HLA-B8 DR3) was significantly (corrected P = Pcorr less than 0.02) associated with Graves' disease particularly in patients who relapsed (Pcorr less than 0.005); (iii) HLA-DR3 was highly associated with DQA2 U allele (chi 2 = 18.53, d.f.2, P less than 0.0005); (iv) a strong correlation between the DQA2 U allele and the outcome of the disease was observed. Relapse occurred in 91% (10/11) of the patients who were homozygous for the DQA2 U allele whilst only 65% (15/23) and 41% (7/17) of patients who were hetero or homo-zygous for the DQA2 L allele (DQA2 U/L and DQA2 L/L) relapsed within the same period of follow-up (chi 2 = 7.18, d.f.2, P less than 0.05). Though the relapse rate in patients with the DQA2 U/U genotype was not significantly higher than the relapse rate in patients with the DQA2 U/L genotype, it was significantly higher than the relapse rate in patients with DQA2 L/L genotype (P less than 0.0001) with a relative risk of 14.3.