Prosthetic valve endocarditis (PVE) has been traditionally divided into early (EPVE) and late (LPVE) forms, the division being made at 60 days after operation. Recent actuarial studies suggest that the risk of EPVE continues up to 12 months after operation. This new insight must increase the emphasis on perioperative prevention, including those measures taken at the time of operation, such as antibiotic prophylaxis, and particularly, the prevention of postoperative nosocomial bacteraemia which other recent studies suggest is a much more significant factor than previously appreciated. The application of DNA-based typing methods of the predominant causative organisms of EPVE [coagulase-negative staphylococci; (CNS)] can be increasingly expected to unravel the aetiology of EPVE and support more logical preventive measures. As with the prevention of native valve endocarditis, the prevention of LPVE currently relies on antibiotic prophylaxis at predictable times of bacteraemia. Epidemiological studies have shown that events currently recognized account for a very minor proportion of cases. The elucidation of the incidence, causes and potential preventive measures of the spontaneous bacteraemias responsible for most cases of LPVE remains a major task. The prevention of all forms of PVE is presently inadequate. Recent studies have not improved our abilities but have served to define areas in which existing measures should be re-emphasized and other areas in which more knowledge is urgently required.