Established practice dictates that physicians and dentists give antimicrobial agents at the time of procedures associated with bacteremia in an effort to prevent IE in persons with underlying cardiovascular diseases. Although prospective controlled study has not been established efficacy, very few cases of IE have been reported in patients at risk who have been treated with regimens recommended by the American Heart Association. Furthermore, studies in rabbits in recent years have shown that IE can be prevented by antimicrobials. The results in rabbits have raised some theoretical questions regarding efficacy of currently recommended regimens for man and have led some investigators to propose revisions of the American Heart Association recommendations. We feel that direct translation of results from experimental IE in rabbits into recommendations for prophylaxis of IE in man is subject to question. It is unlikely that a prospective study comparing placebo with antimicrobials will be undertaken, but it is possible and reasonable to undertake studies to document the occurrence of IE in patients at risk who are given established antibiotic regimens. With such data, modifications of established regimens could be proposed and tested. Regimens that are found to be totally effective in a large number of patients at risk might be modified toward simpler programs--for example, modifying parenteral therapy to oral therapy or changing multiple doses of antibiotic to a single dose. If regimens fail, even once in several thousand patients at risk, the factors responsible might be determined and a clear rationale for modification achieved.