A combination of immune, genetic, and metabolic markers potentially implicated in the development of insulin-dependent diabetes mellitus (IDDM) was studied in the general population. We screened 3,992 healthy schoolchildren, 12-18 years of age with no family history of IDDM, for islet cell antibodies (ICAs). Of the children, 69 (1.7%) were found to be ICA positive (ICA+), of whom 7 (0.17%) also were positive for insulin autoantibodies (IAAs). ICA+ children (group 1) were human leukocyte antigen (HLA) typed at the DQ locus along with 123 matched (group 2) and 235 random (group 3) control subjects (from the original cohort of 3,992). Of the ICA+ children, 28 underwent beta-cell function (beta-CF) studies. High-risk DQ types were surprisingly prevalent in all groups with 35.8% of random control subjects carrying DQB1*0302 and 8.9% carrying the highest risk HLA type for IDDM, DQB1*0302/*0201. Those individuals with higher ICA titer (> 19 Juvenile Diabetes Foundation units [JDF U]) had a significantly higher prevalence of DQB1*0302 than those with lower titer ICA or normal control subjects. Six of 7 individual positive for both ICA and IAA and typed at the DQ locus were DQB1*0302/*0201 heterozygotes or DQB1*0302 or DQB1*0201 homozygotes, representing three of the highest risk genotypes for IDDM. No correlation was observed between ICA titer or DQ type and beta-CF except that all those with beta-CF below the 5th percentile carried either DQB1*0302 or DQB1*0201. Prospective follow-up is underway to determine if any combination of DQ type and immune markers predicts decline in beta-CF and the development of IDDM.