Bacterial infection may be a life-threatening complication in the immunocompromised host, especially in the face of profound and persistent granulocytopenia induced by cytotoxicity. Under these circumstances, antibiotic therapy is started on an empirical basis; however, knowledge of the antibiotic susceptibilities of the offending pathogen may represent a useful guide for therapy adjustments in individuals who do not respond satisfactorily to initial antibiotics. Careful antimicrobial susceptibility testing may also represent the first step of epidemiology investigations of nosocomial outbreaks; moreover, knowledge of antibiotic susceptibility patterns of the prevalent pathogens in certain institutions may help clinicians to formulate empiric antimicrobial treatments for febrile granulocytopenic patients. Careful quantitative studies involving determinations of the minimal inhibitory concentrations may help to evidence early, potentially clinically significant decreases in susceptibility to first choice antibiotics for important nosocomial pathogens. Relationship between appropriate or inappropriate treatment, based on in-vitro susceptibility testing results, and clinical outcome may help to define the clinical significance of some emerging bacterial pathogens in immunocompromised patients.