The classification of diabetes mellitus into 2 main types, defined as Type 1 and 2 diabetes (T1DM, T2DM) relies mostly on the requirement of insulin therapy and on the presence of detectable immunologic abnormalities. However, this distinction is far from straightforward and there is considerable overlap between these 2 types of diabetes. Islet cell autoimmunity, which is characteristic of T1DM, appears in fact to be present in up to 10-15% of subjects diagnosed clinically with T2DM. In the UK Prospective Diabetes Study (UKPDS), it was reported that in patients diagnosed with in T2DM, the presence of autoantibodies to the enzyme glutamic acid decarboxylase (GAD) and cytoplasmic islet cell antibodies (ICA) were a predictor of insulin requirement as compared with patients not carrying these autoantibodies. These results are strikingly similar to a number of prospective studies carried out in childhood diabetes. If islet cell autoimmunity is truly present in 10-15% of subjects clinically diagnosed with T2DM, up to two million Americans might have an unidentified autoimmune form of T2DM, a prevalence similar to that of recent onset childhood diabetes. In addition, we found that in a subset of T2DM patients, a pronounced activation of the acute phase response that seems to be associated with islet cell autoimmunity. These results may in part explain the defect in insulin secretion as well as insulin resistance seen in T2DM. The identification of a subgroup of individuals at risk of developing T2DM using autoantibody as well as inflammatory markers is of public health interest, not only for the correct classification of diabetes, but also because immunomodulatory therapeutic strategies could potentially be instituted sufficiently early in a large number of patients diagnosed as having T2DM and most likely delay the onset of insulin requirement and the complications related with hyperglycemia.