Associations of anti-GAD antibodies with islet cell antibodies and insulin autoantibodies in first-degree relatives of type I diabetic patients

Sera from 114 first-degree relatives of insulin-dependent diabetes mellitus (type I diabetes) patients and 81 healthy individuals living in Germany were analyzed for antibodies to rat brain glutamic acid decarboxylase (GAD-ab) using an immunoprecipitation assay. The determination of GAD-ab in the 81 islet cell antibody (ICA) and insulin autoantibody (IAA) negative healthy individuals established a normal range (mean ± 2 SD); 2 healthy individuals (2.5%) possessed GAD-ab levels above this range, but became negative on follow-up. None of 86 ICA^-/IAA^- first-degree relatives had GAD-ab; whereas, 42.9% of 28 ICA⁺ and/or IAA⁺ relatives were positive for GAD-ab. Presence of GAD-ab was negatively correlated with IAA (P = 0.02) and positively with ICA (P = 0.0006). Follow-up samples were obtained from 25 of 28 ICA⁺ and/or IAA⁺ relatives with a mean (± SD) follow-up period of 20.6 ± 12.1 months. In these 25 relatives, GAD-ab were positive in 70% ICA⁺/IAA^-, 0% ICA^-/IAA⁺, and 57.1% ICA⁺/IAA⁺ relatives in the first sample and in 57.1% ICA⁺/IAA^-, 0% ICA^-/IAA⁺, and 70% ICA⁺/IAA⁺ relatives in the last sample. GAD-ab, once detected, persisted in 9 of 11 GAD-ab⁺ relatives. Of the relatives, 2 converted to GAD-positivity, concomitant with the appearance of ICA, and 2 others lost GAD-ab during follow-up. Of the 28 ICA⁺ and/or IAA⁺ relatives, 6 progressed to overt type I diabetes on follow-up, and GAD-ab were detectable in 4 of these relatives. In conclusion, this study indicates that GAD-ab in first-degree relatives are significantly correlated with ICA, are independent of the appearance of IAA, and are present in 66.7% of relatives who subsequently developed type I diabetes.