Many human diseases are the result of autoimmune attack, presumably related to a loss of tolerance to self. Autoimmune disease can be divided into either organ-specific illnesses, such as thyroid disease, type 1 diabetes, and mysasthenia gravis, or systemic illnesses, such as rheumatoid arthritis and systemic lupus erythematosus. The pathogenesis of autoimmune damage also segregates autoimmune disease in that some diseases or manifestations are mainly induced by autoantibodies. Pathogenesis may be mainly mediated by autoimmune T lymphocytes. Notwithstanding the underlying mechanism of disease, virtually all autoimmune diseases are associated with circulating autoantibodies, which bind self-protein. Furthermore, for many diseases these autoantibodies are found in serum samples many years before disease onset.

In the past several years a new autoantibody specificity has been identified in the sera of patients with rheumatoid arthritis. These autoantibodies bind citrulline, a post-translational modification of arginine. Markus Nielen and colleagues recently studied the presence of these autoantibodies and rheumatoid factor in blood donors who later developed rheumatoid arthritis (Arthritis Rheum 2004; 50: 380–86). About half the patients were positive for at least autoantibody at a median of 4·5 years before the onset of disease. The negative predictive value of these tests was high, while the positive predictive value was very high.

Autoantibodies might not be directly responsible for many of the manifestations of autoimmune disease, but they are markers of future disease in presently healthy individuals. Long-term large studies of outcome are needed to assess the use of assaying autoantibodies for prediction of disease. Such data could lead to intervention trials to prevent autoimmune disease, as are already underway in type 1 diabetes.