Seventy-eight Japanese diabetics were HLA-typed, with special reference to age at onset, insulin dependency, and family history. HLA-A9, B5, and BW40 were increased, but A1, A3, and B8, which are found frequently among Caucasians, were almost absent among Japanese healthy controls as well as diabetics. J-1, a Japanese specific subclass of BW22, was significantly increased in juvenile-onset diabetics as compared with controls or diabetics with late onset. J-1 was also increased in the diabetics with insulin dependency and/or positive family history. But the association of J-1 with juvenile-onset diabetes mellitus was found to be the strongest. A tendency to a decrease in B5 was also observed in Japanese diabetics with juvenile onset, but this did not reach statistical significance as far as corrected P was concerned. These results showed that genetic markers for diabetes mellitus, especially that with juvenile onset, were different among Japanese from those found among Caucasians. There is ample evidence to indicate a race specificity in HLA phenotypes in diabetics as well as in controls. These findings also strongly suggest that juvenile-onset diabetes mellitus is a disease entity in itself and different from late-onset diabetes mellitus in origin and pathogenesis.

Thirty-three Japanese patients with Graves' disease and 106 healthy controls living in the Kagoshima area, the southernmost part of the Japanese mainland, were HLA typed by the NIH method. None of them were related to each other. The only antigen showing an increased frequency in Japanese patients with Graves' disease was HLA-BW35 (corrected P < 0.02). A decreased frequency of B5 in the patients was also statistically significant (corrected P < 0.02).

No abstract

The frequency of HLA BW54 and B5 in Japanese patients with JOD is increased and decreased, respectively. In JOD patients without a family history of MOD, the frequency of BW54 is significantly increased, whereas in JOD patients with a positive family history the frequency was not increased in a statistically significant manner.

No abstract