Abstract The purpose of this research was to assess the role of antibodies to glutamic acid decarboxylase (GAD) in children with newly diagnosed type 1 diabetes in relation to other disease-associated autoantibodies and HLA-defined genetic disease susceptibility, to evaluate the role of GAD antibodies (GADA) in relation to clinical characteristics at the diagnosis of type 1 diabetes and to compare the frequency and levels of GADA between adult and childhood onset type 1 diabetes.The study population comprised altogether 999 children and adolescents with type 1 diabetes, 100 affected adult subjects and more than 370 non-diabetic controls. GADA were measured with a liquid radioligand assay, and a similar assay was used for the analysis of antibodies to the islet antigen 2 (IA-2) molecule. Islet cell antibodies (ICA) were determined with conventional immunofluorescence and insulin autoantibodies (IAA) with a liquid phase radioimmunoassay either in a tube or a plate format (microassay). GADA were detected at diagnosis in 68 to 73% of the children and adolescents with type 1 diabetes. GADA were more frequent in girls and in those older than 10 years of age at clinical disease manifestation. Subjects testing positive for GADA had higher levels of ICA and IAA than those negative for GADA. Multiple antibodies ( 2) were observed more often in girls and in children under the age of 5 years. Children with the HLA DR3/non-DR4 phenotype had the highest GADA levels, significantly higher than those seen in children with the DR4/non-DR3 combination. The highest prevalence of multiple autoantibodies was seen in subjects heterozygous for DR3/4. When studying HLA DQB1 genotypes those with the DQB1*02/y (y = other than *0302) genotype had the highest GADA levels as expected since DQB1*02 and DR3 are in strong linkage disease equilibrium. The same group had the lowest frequency of multiple antibodies among the children younger than 10 years of age.Patients diagnosed with type 1 diabetes before the age of 20 had a higher frequency of all four autoantibodies analysed than those presenting with clinical disease after the age of 20. The proportion of subjects testing negative for all four antibodies was substantially higher among adults than in those under the age of 20. The smallest age-related difference in antibody frequencies was observed for GADA, and the GADA-positive adult patients had on an average about three times higher antibody levels than the GADA-positive children. No association was observed between positivity for GADA and the degree of metabolic decompensation at the clinical presentation of type 1 diabetes. No significant differences were either seen between the subjects who tested positive for GADA at diagnosis and those who were negative in serum C-peptide concentrations, metabolic control or exogenous insulin requirement over the first 2 years of observation. The proportion of children in clinical remission was, however, lower among GADA-positive subjects than in GADA-negative patients at 18 months after the clinical manifestation. Positivity for multiple antibodies was associated with accelerated beta-cell destruction and increased exogenous insulin requirements over the 2-year observation period. The observations that GADA are related to female gender, older age and the HLA-DR3/ DQB1*02 haplotype suggest that a strong humoral immune response to GAD may reflect a propensity to general autoimmunity rather than specific beta-cell destruction.