Gram-positive bacteraemia in granulocytopenic cancer patients.

In the four EORTC International Antimicrobial Therapy Cooperative Group trials, the frequency of gram-positive isolates has increased significantly from 29% of single-organism bacteraemias in trial I (1973-1976) to 41% in trial IV (1983-1985). In trial IV febrile and neutropenic (less than 1000 polymorphonuclear lymphocytes per microliter) cancer patients were randomized prospectively to receive either azlocillin plus a long course (at least 9 days) of amikacin, or ceftazidime plus a short course (3 days) of amikacin, or ceftazidime plus a long course of amikacin. Without modification of the allocated antibiotics, the overall response rates for gram-positive bacteraemias were similar for all three regimens (19/37 [51%], 8/23 [35%] and 14/30 [47%]), respectively. However, in patients with prolonged and severe neutropenia, treatment with azlocillin plus amikacin was significantly more effective than with ceftazidime plus 3 days' amikacin (7/10 vs. 0/7). The overall response rate for these infections was significantly lower than that observed in trial I (46% vs. 74%), but this was not associated with increased mortality. The response to treatment was significantly influenced by the susceptibility of the infecting strain to the beta-lactam. Multivariate analysis revealed that increasing age, presence of a central venous catheter and resistance to beta-lactam adversely affected outcome. Future studies should be designed to improve the outcome of gram-positive bacteraemia in neutropenic patients with cancer.