A randomized trial of empirical antibiotic therapy with one of four beta-lactam antibiotics in combination with netilmicin in febrile neutropenic patients.

Over a two year period 174 evaluable episodes of fever in neutropenic patients were treated in a randomized study comparing four beta-lactam antibiotics, each given in combination with netilmicin. Exclusions included episodes due to viral or fungal infection, and trial violations. Most patients were receiving treatment for leukaemia, including 18% undergoing bone marrow transplantation. The overall response rate (EORTC criteria) was 66%, ranging from 56% for cefoperazone to 76% for mezlocillin. Microbial documentation was obtained in 31% of episodes; Gram-positive isolates were most frequent but Pseudomonas aeruginosa was found in 18 patients. In patients with microbiologically documented infection 70% improved, overall--from 40% with cefoperazone to 80% with piperacillin (P less than 0.05). Nephrotoxicity was seen in 6.7% and was associated with severe documented sepsis. Hypokalaemia was seen in 29% and was most marked in patients receiving ticarcillin. Rashes occurred in 6.6% overall, with no difference between the groups. Ototoxicity, shown by serial audiograms, was seen in 4.7% of patients. No evidence of vestibular dysfunction was seen in 62 patients studied. Of thirteen deaths due to the primary infection, seven were caused by Ps. aeruginosa and five by fungi.