The determination of glutamic acid decarboxylase and tyrosine phosphatase-like antibodies (GAD-AB and IA-2-AB) may be useful for the classification of diabetes, and in selected patient groups the measurement of these autoantibodies has been shown to be rather sensitive and specific.In this study we examined the use of these antibody determination in a clinical setting of 157 diabetic outpatients recruited randomly from our diabetes clinic. The prevalence of the different antibodies was set in relation to the clinically classified diabetes type and to diabetes duration.Among the patients with a clinical diagnosis of type 1 diabetes, the GAD-AB were clearly positive in 44% and borderline positive in 10%, whereas the IA-2-AB were positive or borderline positive in 36% of these patients. The prevalence of positive autoantibodies declined with increasing duration of type 1 diabetes. Among the patients with clinically diagnosed type 2 diabetes, the GAD-AB were clearly positive in 25.2% and borderline positive in 13.1%, IA-2-AB were only found in 4.7%. Patients with a clinical diagnosis of type 2 diabetes but positive for GAD-AB could not clearly be identified as having latent autoimmune diabetes in adults (LADA), since some of them did not need insulin therapy up to 10 years after the diagnosis of diabetes. The prevalence of GAD-AB in type 2 diabetic/LADA patients did not depend on diabetes duration.We conclude that the determination especially of GAD-AB may be useful for the classification of diabetes in clinically unclear cases. The additional determination of IA-2-AB appears to provide only limited additional information.
