Serious infections in neutropenic patients may have mortality rates in excess 40%, even when properly treated. OBJECTIVE: This retrospective database analysis was performed to examine the mortality, length of stay and cost of imipenem/cilastin versus piperacillin/tazobactam in the treatment of seriously infected neutropenic patients. METHODS: The data were collected from July 1997 to June 1998 at 90 non-government non-specialty acute care hospitals over 100 beds throughout the United States. The data collected at discharge included UB-92, diagnostics, procedures, ICD-9's, DRG and drug dispensing data. All patients in the analysis were coded for a serious infection, primarily pneumonia (481 to 486.99) or septicemia (038 to 038.9), and neutropenia (284,284.4, 284.9 and 288). RESULTS: The database analysis identified 16,396 on mipenem/cilastin or piperacillin/tazobactam over 18 years of age. Of those treated with these agents, 2,563 were infected neutropenia patients. The length of stay was not statistically significant between the two groups. Difference in utilization patterns demonstrates the imipenem/cilastin group had a statistically significant (p<0.05) greater use as a third line (16%) agent and less as a first line agent (61%) than the piperacillin/tazobactam group (10% and 68% respectively). Mortality rates indicated imipenem/cilastin had statistically significant lower mortality rate (17.1%) versus the piperacillin/tazobactam (20.6%) group in the treatment groups. Overall cost was $1,130 lower in the imipenem/cilastin treated group despite the fact the drug cost was $267 higher. CONCLUSION: In this study population, imipenem/cilastin has a lower mortality rate, a higher drug cost and lower overall treatment cost in infected neutropenic patients when compared to piperacillin/tazobactam.