Autoimmune diabetes: caught in the causality trap?

The lack of concordance between genotype and clinical diabetes has prompted a search for the infectious agent that precipitates this autoimmune disease. However, this approach may be misleading. It assumes that the disease-prone individuals that do not develop diabetes do not have autoimmunity. In the non-obese diabetic (NOD) mouse, the genotype is a primary determinant of autoimmunity. Not all animals of the disease-prone genotype develop clinical disease; however, all have autoimmunity. This is expressed as a destructive or non-destructive process. Multiple pathways are open to the immune system and whether or not the immune response is destructive and leads to the development of clinical disease, appears to be a random process. If this is the case, the most important questions relating to autoimmune disease are not those concerning the 'causative' agents. Instead we should be asking what are the differences between pathways open to the immune system and what factors affect the probability that one or another pathway is finally selected?