Infection remains a major complication of severe granulocytopenia. The administration of empiric antibiotic therapy at the first sign of infection or fever has significantly reduced the sequelae of infection in such compromised patients. Yet, the appropriate composition of the empiric antibiotic regimen continues to be debated. Antibiotic combinations have been the standard approach; multiple studies have confirmed the importance of antibiotic combinations in patients who are profoundly granulocytopenic and have gram-negative rod bacteremia. The presence of a synergistic antibacterial effect and high serum bactericidal activity of the antimicrobial regimen improves the response rate in such severe infectious episodes. However, the occurrence of such infections at certain centers is declining, and the introduction of new, highly active, broad-spectrum beta-lactam antibiotics has raised the question of using a single agent, or monotherapy, as empiric treatment of febrile granulocytopenic cancer patients. Approximately 20 studies have been completed, and although these trials have demonstrated support for this approach, caution must be exercised because of the inconsistent study designs, wide range of response definitions, small patient numbers, and limited frequency of treated documented infections. Therefore, at present, a broad-spectrum antibiotic combination still remains the treatment of choice for empiric therapy of febrile granulocytopenic cancer patients. Carefully designed, prospective, randomized, double-blind trials investigating new approaches, such as monotherapy, must continue.