An open, randomized, multinational, multicentre study was conducted to compare the efficacy, safety and tolerability of levofloxacin 500 mg twice daily with imipenem/cilastatin 1 g iv three-times daily in the treatment of hospitalized adult patients with clinically suspected bacteraemia/ sepsis. Levofloxacin patients could change from iv to oral administration after a minimum of 48 h iv treatment if clinical signs and symptoms of sepsis had improved. The primary efficacy analysis was based on the clinical and bacteriological response at clinical endpoint, A total of 503 patients were randomized and 499 included in the intent-to-treat population. The per-protocol population comprised 287 patients with bacteriologically proven infection. Clinical cure rates at clinical endpoint in the intent-to-treat population and per-protocol population were 77% (184/239) and 89% (125/140), respectively, for levofloxacin and 68% (178/260) and 85% (125/147), respectively, for imipenem/citastatin. At follow-up, the cure rates in the per-protocol population were 84% for levofloxacin and 69% for imipenem/cilastatin. The 95% confidence interval for both populations showed that levofloxacin was as effective as imipenem/citastatin. A satisfactory bacteriological response was obtained in 87% (96/110) of levofloxacin patients and 84% (97/116) of imipenem/cilastatin patients at clinical endpoint. Adverse events possibly related to the study drug were reported in 74 (31%) levofloxacin patients and 79 (30%) imipenem/cilastatin patients. There were no clinically appreciable differences between the treatment groups. Levofloxacin 500 mg twice daily, either iv or as sequential iv/oral therapy, was as effective and well tolerated as imipenem/cilastatin 1 g iv three-times daily in the treatment of hospitalized patients with suspected bacteraemia/sepsis.