Although it was once thought that autoantibodies were confined solely to patients with autoimmune diseases, it is now accepted that healthy individuals may express autoantibodies without necessarily developing’appropriate’clinical features. Autoantibodies are clearly part of the mosaic of factors involved in the development of autoimmune disease but on their own they are rarely truly pathogenic. They are often, however, very useful markers of a predisposition to develop certain clinical features. Among the relatives of patients with autoim-mune diseases it seems to be widely assumed that autoan-tibodies are detectable more often than in healthy controls. This is particularly true in the case of the relatives of patients with systemic lupus erythematosus (SLE). In this review we consider critically the data on which this premise is based and attempt to determine the likely frequency with which a variety of autoantibodies are actually present in healthy relatives.