The objective of this presented prospective randomized study was to compare the efficacy of empirical antimicrobial monotherapy with imipenem 3 x 0.5 g per day to 3 x 1.0 g per day for treatment of infections in neutropenic patients. A total of 192/220 febrile episodes were evaluable for clinical efficacy. The overall response rate was 53/93 (57%) vs. 57/99 (58%). Of the different infection types, fever of unknown origin (FUO) showed the best response, with defervescence in 29/41 (71%) and 36/42 (86%) cases, respectively (not significant). Unfavourable results were found in pneumonias [5/20 (25%) vs. 4/23 (17%)]. The median time until persistent defervescence was equal in both groups (2 days), likewise the median duration of imipenem therapy in responders (7 days). The most frequent micro-organisms were Gram-negative, documented in 22% of the febrile episodes in the lower dosage group vs. 17% of all episodes in the patients with imipenem 3.0 g per day (Gram-positives 17% vs. 14%, fungal 5% vs. 8%). In the lower dosage group, fever with abdominal symptoms occurred less frequently (8% vs. 15%), and significantly more patients tolerated imipenem without any side-effects (95.8% vs. 79.4%), especially regarding severe nausea/vomiting (2.1% vs. 11.8%). Of the initial non-responders, 35/40 (88%) vs. 41/42 (98%) were cured after therapy modification. There was no significant difference in the use of further antibiotics such as aminoglycosides, glycopeptides, ceftazidime or amphotericin B, except a marginally higher use of metronidazole in patients with imipenem 3.0 g per day (3% vs. 10%). Overall, we found no significant differences in efficacy between the two study groups, but more frequent side-effects with imipenem 3.0 g per day.