A prospective, randomized, controlled monocentric trial was performed to evaluate the efficacy and safety of once daily ceftriaxone 2 g plus tobramycin 5 mg/kg in comparison to cefotaxime 2 g t.i.d. plus tobramycin 5 mg/kg qd in the treatment of neutropenic fever. In cases of fever > or = 38.5 degrees C and a neutrophil count below 1000/microliter, patients with hematological malignancies were assigned to ceftriaxone or cefotaxime, each with tobramycin. The primary endpoint was defined as defervescence < 37.5 degrees C on day 4-6 followed by at least 7 afebrile days. Secondary endpoints were overall response, defined as defervescence on day 25 and toxicity. There were 160 episodes of 114 patients included. Fever of unknown origin accounted for 79 episodes (51%), microbiologically defined infection for 36 (23%), clinically defined infection for 27 (17%), and both clinically and microbiologically defined infection for 14 episodes (9%). On an intent-to-treat basis 156 episodes could be evaluated for the primary endpoint. Ceftriaxone plus tobramycin and cefotaxime plus tobramycin resulted in a primary response in 46.9% and 45.3%, respectively. Overall response was achieved on study day 25 in 87.7% and 80%, respectively. No significant difference in toxicity was observed. Once-daily ceftriaxone plus tobramycin was not inferior to cefotaxime t.i.d. plus tobramycin qd in the empirical treatment of neutropenic fever.