Objective: To assess the relative cost effectiveness of cephalosporin monotherapy options and aminoglycoside/ureidopenicillin combination therapy for the treatment of febrile episodes in adult patients with neutropenia. Design and setting: This was a retrospective cost-effectiveness analysis conducted from the institutional perspective. Methods: The analysis was based on 741 febrile episodes in adult patients with neutropenia enrolled in 5 randomised trials: 3 comparing monotherapy with ceftazidime or cefepime, and 2 comparing cefepime monotherapy versus aminoglycoside/ureidopenicillin combination therapy. Resource utilisation included costs for study antibacterials, treatment of adverse effects and failures, and hospitalisation. The primary end-point was the overall cost of treatment per patient. Cost-effectiveness ratios were also analysed. Results: No significant differences in clinical success rates were detected. Median per-patient costs in the monotherapy comparisons were $US7849 for cefepime and $US7788 for ceftazidime [1997 values; not significantly different (NS)]. Corresponding costs for the monotherapy versus combination therapy comparisons were $US9780 for cefepime and $US10 159 for gentamicin/ureidopenicillin (NS). Despite a higher acquisition cost for cefepime, there were no statistically significant differences in cost effectiveness compared with either ceftazidime monotherapy or gentamicin/ureidopenicillin combination therapy. Sensitivity analyses revealed that monotherapy can be cost effective compared with combination therapy in many situations. Conclusion: There were no economic differences between the 3 regimens tested. Therefore, drug cost should not be a deciding factor when choosing antibacterial therapy for the treatment of febrile episodes in adult patients with neutropenia.