Management of Febrile Neutropenia: Single Agent or Combination Therapy?.

The best strategy for the empirical treatment of fever in profoundly neutropenic patients (pts) remains unclear. We retrospectively compared the efficacy and safety of 2 antibacterial regimens consecutively used in our center for febrile neutropenia (FN) in pts undergoing intensive chemotherapy for acute leukemia or autologous stem cell transplantation. From 1/99 to 6/01, 374 episodes of FN in 179 pts were treated with piperacillin-tazobactam plus amikacin (PTA, 4.5 g q6h and 20 mg/kg q24h respectively), then from 7/01 to 12/03, 428 episodes in 223 pts were treated with cefepime (CEF, 2g q8h). Any modification of the initial regimen was evaluated as a failure. Pts were nursed in HEPA-filtered rooms and received prophylactic aciclovir (and ciprofloxacin until 12/99) with no antifungal primary prophylaxis. There were no significant differences between the PTA and CEF episodes regarding median age of the pts, median days with neutrophils <500/μl or presence of an intravenous central catheter. The proportion of documented bacteremias was similar (33/30%) with a Gram negative predominance in both periods (60/66%), mainly due to E. coli (61/64%) and P. aeruginosa (14/14%). As for Gram positive bacteremias, there was an increase in S. aureus (18% in PTA vs 27% in CEF) and a marked decrease in S. viridans (24/4%). Resistance of Gram negative bacteria to the protocol drugs remained low (7% to PT and 4% to A in the first period, 5% to CEF in the second); resistance of coagulase negative staphylococci to ciprofloxacin dropped from 33 to 15%. Success rates were 35% for PTA and 32% for CEF. The reasons for modification of the initial empirical regimen were similar in the 2 periods, the most frequent being persistent fever (56/59%) followed by progression of infection (20/18%), relapsing fever (19/14%), withdrawal due to toxicity (3/6%) and resistant pathogen (1/3%). The addition of an antifungal was the most usual modification, occuring in 44% of PTA and 39% of CEF episodes. Vancomycin (for which a restriction policy had been introduced in 1995 after an outbreak of vancomycin-resistant enterococci) was added in 17 and 15% of the episodes. Amikacin was added in 38% of the CEF episodes. The occurence of further infections did not differ between the 2 periods (26 and 29%) nor did the mortality due to infection (7.7 and 6%). The proportion of severe adverse events (CTC-NCI grades 3/4) was low in both periods (4 and 1%). Cutaneous allergic reactions were less frequent with PTA than with CEF (5 vs 9%). We conclude that monotherapy with CEF was as effective as the combination of PTA for empirical treatment of FN. Both regimens resulted in a low infectious mortality and a very low emergence of resistance, and both were well tolerated.