Primary systemic vasculitis affecting smaller vessels is usually associated with antineutrophil cytoplasmic antibodies (ANCA). The ANCA-associated vasculitides include Wegener's granulomatosis, microscopic polyangiitis, Churg Strauss syndrome and renal limited vasculitis. There is considerable evidence that genetic factors influence susceptibility to ANCA-associated vasculitis, including reports of familial cases, differences in racial incidence, and associations with polymorphic variants of proteins such as alpha-1-antitrypsin. There is mounting evidence, from clinical and in vitro studies, that ANCA may be pathogenic. However, it is also clear that autoreactive T cells are likely to be involved, by providing T cell help for ANCA production and possibly by producing cell-mediated immune injury. Indeed, T cells from patients with vasculitis have been shown to proliferate in vitro in response to the target antigens of ANCA - proteinase 3 and myeloperoxidase. In most T-cell-dependent autoimmune diseases there are clear positive and/or negative associations with HLA genes. These genes are encoded in the major histocompatibility complex (MHC) and their products, the HLA molecules, play a central role in the generation of T cell responses. For this reason, many investigators have looked for HLA associations in ANCA-associated vasculitides. Problems in analysing these reports include the definition of the diseases concerned, and the varying methodology of HLA typing. A number of positive and negative associations with HLA genes have been reported in systemic vasculitis. However, it is striking that no consistent association has been identified in different series. In recent studies there have been positive associations with HLA DR1, DQw7 or DR8, negative associations with DR3 or DR13, or no significant associations. This lack of an obvious and consistent HLA association is extremely interesting, and suggests that the T cell response in vasculitis may be very heterogeneous, or that a genuine strong association has yet to be identified. Further investigation of this problem is clearly needed to improve our understanding of the pathogenesis of ANCA-associated vasculitides.