Objective: To determine the prevalence of autoantibodies to IA-2 (IA-2Ab) and glutamic acid decarboxylase (GADAb) in type 2 diabetes, their relationship to disease duration, and their importance in management decisions. Methods: We undertook a study of 101 patients with type 2 diabetes (defined as nonketotic hyperglycemia at diagnosis) of varied duration (median, 4 years). Results were compared with those from 36 patients with type 1 diabetes also of varied duration (median, 2 years). IA-2Ab and GADAb were measured by radioligand-binding assays with use of in vitro-synthesized, 35S-labeled antigens. Results: Of the 101 patients with type 2 diabetes, 20 (20%) were positive for GADAb; only 4 of these 20 were positive for IA-2Ab: In comparison, 75% of patients with type 1 diabetes were positive for GADAb, IA-2Ab, or both (P<0.0001). The coincidence of IA-2Ab positivity in GADAb-positive patients with type 2 diabetes was significantly lower than in patients with type 1 diabetes (20% versus 73%, respectively; P = 0.002). All four IA-2Ab-and GADAb-positive patients with type 2 diabetes required insulin and were younger than those positive for GADAb alone (P = 0.018). GADAb positivity in patients with type 2 diabetes was highly associated with insulin requirement (P = 0.004), with an odds ratio of 5.8 in predicting insulin dependence. Among patients with type 2 diabetes receiving insulin therapy, disease duration was significantly shorter (P = 0.025) and body mass index was significantly lower (P<0.001) in GADAb-positive versus GADAb-negative patients. In contrast to type 1 diabetes, in which GADAb values were negatively correlated with disease duration (r = -0.34; P = 0.044), no significant correlation with disease duration was observed in type 2 diabetes (r = -0.166; P = 0.48). Conclusion: Irrespective of duration of disease, measurement of IA-2Ab and GADAb can help to identify those patients with type 2 diabetes most likely to require insulin therapy.