Choice of antibiotic therapy for the management of infection in the neutropenic patient continues to challenge the clinician. The shift toward grampositive organisms and the continuing need to provide gram-negative coverage demands the use of an agent or agents that provide coverage for the spectrum of potential infecting organisms. Cefepime is an extended-spectrum fourth-generation cephalosporin that has good activity against gram-positive and gram-negative organisms; in addition, it resists degradation by Bush group 1 β-actamases. These properties make this agent a promising candidate for empiric therapy with febrile neutropenic patients. Data presented in this article are from febrile neutropenic cancer patients enrolled into two randomized, prospective, nonblinded comparative U.S. clinical trials. Patients were randomized to receive cefepime (2 g thrice daily) or a comparator regimen of either ceftazidime (2 g thrice daily) or piperacillin + gentamicin (3 g every 4 hours +1.5 mg/kg every 8 hours). When indicated, vancomycin was added to the regimen. A total of 109 febrile episodes were treated with cefepime and 107 episodes were treated with the comparator regimens. Neutropenia (≤500 PMNs/mm³) persisted for ≥10 days in >40% of episodes and severe neutropenia (≤100 PMNs/mm³) in >25%. More than 40% of the total number of episodes were documented bacterial infections. These characteristics did not differ among treatment groups. Duration of therapy was similar in both groups (median: cefepime, 9 days; comparators, 11 days). In >40% of episodes, patients received study therapy without addition of other antibacterials (cefepime, 46%; comparators, 41%). Vancomycin was added in almost half of all the episodes (cefepime, 45%; comparators, 53%). Patients became afebrile by the fourth day of study therapy in approximately 60% of episodes (cefepime, 58%; comparators, 60%). In approximately 75% of the episodes, patients had a satisfactory response at the end of therapy (cefepime, 74%; comparators, 76%); and following approximately 90% of episodes, patients survived for >30 days (cefepime, 90%; comparators, 92%). Eradication rates were similar for all pathogens for cefepime and comparator agents. There were similar numbers of superinfecting organisms in each treatment arm; most involved gram-positive organisms. These multiple measures of efficacy suggest that initial empiric cefepime monotherapy is comparable to the pooled experience with standard therapies and that antibacterial modifications occur with similar frequency for cefepime compared with standard empiric regimens.