AbstractINTRODUCTION The management of infections in cancer patients has evolved both conceptually and practically over the past twenty years. The advent of cytotoxic therapy opened new horizons on the road toward improved survival and even cure of some malignancies. With it, however, came an unprecedented and previously unrecognized array of abnormalities in host defense systems (Fig. 1), and consequendy, infectious complications that at first threatened to undermine any potential benefits of cancer therapy. The most profound of these alterations was granulocytopenia. Early on, it was recognized that successful management of the granulocytopenic patient required new and unique therapeutic guidelines. More traditional approaches for the treatment of infection in these patients, such as identification of a pathogen prior to institution of antibiotics, often led to disasterous consequences. Thus, the concept of empirical antibiotic therapy emerged, with routine empirical therapy now accepted as the standard of care when a granulocytopenic patient first becomes febrile. Early mortality from gram-negative sepsis has become relatively rare with this approach.