Prospective randomized comparison of three antibiotic regimens for empirical therapy of suspected bacteremic infection in febrile granulocytopenic patients

Abstract

The standard regimen used by members of the European Organization for Research on Treatment of Cancer Antimicrobial Therapy Cooperative Group for empiric therapy of febrile neutropenic cancer patients has been treatment with ticarcillin plus amikacin. A three-arm prospective randomized controlled trial was performed to determine whether the extended-spectrum antipseudomonal penicillin azlocillin or the extended-spectrum cephalosporin cefotaxime had more or less efficacy than the beta-lactam in the ticarcillin-plus-amikacin regimen. A total of 742 patients from 22 institutions were evaluated. Single gram-negative rod bacteremias accounted for 83 episodes, and it was among these patients that the prognosis was least satisfactory, leading to a more intensive evaluation of this patient group. In these patients the azlocillin-plus-amikacin regimen resulted in a 66% response rate, compared with a 37% response rate for patients who received cefotaxime plus amikacin (P = 0.080) and a 47% response rate for patients who received ticarcillin plus amikacin (P = 0.207). The patients with gram-negative rod bacteremias and persistently profound granulocytopenia had substantially poorer response rates (37%) than the patients with rising granulocyte counts (73%; P = 0.004). A logistic regression analysis indicated that the following factors also affected infection resolution: beta-lactam utilization in the regimen (azlocillin was better than ticarcillin or cefotaxime), resolution of profound granulocytopenia (less than 100 cells per microliter) during therapy, and susceptibility to the beta-lactam antibiotic.