valvular heart disease but no rheumatic history. Generally, the main distinguishing characteristics of this subgroup are a more equal sex distribution, isolated mitral valve disease, a relatively short history of symptoms before surgery, and, in many cases, the need for surgery at a comparatively early age. These findings are difficult to equate with the generally accepted concept of the aetiology of acquired valvular heart disease. This traditional view is that virtually all cases are rheumatic. In patients with no history of rheumatic fever or chorea their valvular disease is usually attributed to subclinical rheumatic carditis. Since it is not possible to investigate patients at the time of onset of subclinical carditis, however, the assertion that all such cases are rheumatic is speculative. It could, nevertheless, be argued for the traditional view that patients with no rheumatic history-that is, those with allegedly subclinical rheumatic carditis-differ fundamentally from those who have had classical rheumatic fever and that this may, in some way, account for clinically observed differences between those with and those without a rheumatic history. In this context the AW19 antigens might be a factor predisposing to the development of subclinical, as opposed to overt, rheumatic carditis. This line of reasoning, however, does not explain why patients with AW19 antigens differ clinically from those other patients with no rheumatic history (group 2) who are also assumed to have had subclinical rheumatic carditis. The aetiology of these cases of valvular heart disease, if not rheumatic, is unknown. It has, however, been suggested that viruses may sometimes be implicated. In animals viruses can produce valve lesions with many of the characteristics of chronic rheumatic heart disease in man.' There is also evidence in man that viruses' 4 and similar organisms6 may cause valvular damage. Other suggestions have been made which postulate a link between HLA antigens and some virus infections. For example, the neurotropism of polio virus appears to be partly dependent on the presence of HLA antigens A3 or B7.7 In more general terms it has been suggested that these particular antigens might govern the response of the central nervous system to a common product of different viruses. Possibly the cardiotropism of viruses, such as those of Coxsackie group B, which is currently unexplained, has an analogous basis. This in turn could account for the association we have described between valvular disease and the HLA antigens A29 and AW30,131-that is, these HLA antigens might act as a conditioning factor8 which occasionally transforms a mild viral myocarditis-a common occurrence9into a severe pancarditis. Our study does not directly implicate viruses as a cause of acquired valvular heart disease. It does, however, provide further evidence which is inconsistent with the traditional "rheumatic" explanation. It also indicates a possible line of research into the aetiology of obscure heart disease which has not yet been explored. Qu...