Imipenem versus targeted therapy in cancer patients.

In many instances, broad-spectrum antibiotics are initiated empirically in febrile cancer patients and continued for the whole duration of therapy. An alternative is to narrow the spectrum whenever the offending pathogen is identified. This study is aimed at comparing these two options. Non-neutropenic cancer patients with severe infections received empiric imipenem. After 72 h, those with microbiologically documented infection were randomized either to continue imipenem or to receive a targeted therapy. After 72 h of imipenem 76.1% were improved. After randomization, a higher efficacy was observed with imipenem (88.5 vs. 72.1%: P = 0.025). Bacterial and fungal superinfections were comparable. Costs were lower for targeted therapy in gram-positive infection and higher in gram-negative infection.