Carbapenems versus other beta-lactams in treating severe infections in intensive care: a systematic review of randomised controlled trials.

Carbapenems have not been comprehensively compared in clinical trials with fourth-generation cephalosporins (4GC) and antipseudomonal penicillins (APP) in the treatment of severe infections (SI) and febrile neutropenia (FN). A systematic review of CENTRAL, EMBASE, MEDLINE and JICST-EPlus for randomised controlled trials was conducted to establish the currently available evidence. Database searching was supplemented by hand searching and contacting conference organisers. Searching was completed in November 2006 and no restriction was placed on the language of publication. Data were extracted on clinical response, bacteriologic response, all-cause mortality and adverse events. Of the 265 papers identified, 12 were appropriate for meta-analysis (four 4GC and eight APP). The results showed that carbapenems are associated with a significant reduction in all-cause mortality (relative risk 0.62, 95% confidence interval: 0.41 to 0.95; p=0.03) compared to APP in the treatment of SI, and withdrawals due to adverse events (RR 0.65, 95% CI: 0.45 to 0.96; p=0.03) are also less common. When compared in the treatment of FN, carbapenems are associated with a significant increase in clinical response during the initial 72 h of treatment (RR 1.37, 95% CI: 1.09 to 1.74; p=0.008) and bacteriologic response (RR 1.73, 95% CI: 1.03 to 2.89; p=0.04). For all other outcomes, including all comparisons with 4GC, there were no significant differences between treatments. The use of carbapenems rather than APP could reduce mortality and, by simplifying treatment decisions, reduce the time before patients receive appropriate antibiotic treatment. The currently available evidence is insufficient for distinguishing between carbapenems and 4GC.