Large-scale programs designed to assess risk for type 1 diabetes through serologic assessment of autoantibodies to recombinant beta-cell autoantigens are hampered by several limitations, including the methods for sample collection and assay performance, as well as the volume required for autoantibody determinations. The present study was designed to develop a low sample-volume, primary screening method for autoantibody detection of high specificity and sensitivity, and to determine the feasibility of dried blood spots collected on filter paper in serving as vehicles for such determinations. Autoantibodies to glutamic acid decarboxylase (GAD) and ICA512bdc (IA-2), both individually and in combination, were determined in persons with type 1 diabetes, healthy controls, or individuals with other autoimmune disorders. Autoantibody results for serum, plasma, and dried blood spots were compared. GAD, IA-2, and combined GAD/IA-2 autoantibodies were concordant in their measurement from minimal volumes of serum, plasma, and whole blood extracted from dried filter paper. The autoantibody levels from the dried blood spots were, however, lower than corresponding serum samples, and, as currently designed, failed to detect low-titer autoantibodies. Despite this limitation, screening for diabetes risk can be performed using small volumes of whole blood, serum, or plasma collected onto filter paper. These methodological improvements should simplify matters, reduce costs, and increase the efficacy of screening programs for type 1 diabetes. Further development of better substrates/methods for blood-specimen collection seems necessary to exploit the full potential of this and other autoantibody measurement strategies for screening large populations.