INTRODUCTION: Levofloxacin (LVX) is one of the most frequently used antibiotics in critical patients admitted to Spanish Intensive Care Units (ICU). Their use in community-acquired infections has been widely documented, while it is less frequent and known in nosocomial infections (NI). OBJECTIVE: To describe the indications and utilization patterns of LVX in the treatment of NI in patients admitted to Spanish ICU. MATERIAL AND METHODS: Open-label, retrospective, observational and multicenter study. All patients admitted to ICU and who were being treated for NI with LVX in the years 2004-2005 were included. A case report form (CRF) was drawn up and included demographic, infection, treatment, infectious process and patient development variables. NI-dependent LVX usage was described. A logistical regression analysis was carried out in order to identify the variables associated with a satisfactory response. Results are expressed by means of the odds ratio and a 95% confidence interval. RESULTS: A total of 949 patients who were given LVX for the treatment of 1,103 NI were recruited in 87 ICU: 460 (41.7%) with non-mechanical ventilation associated pneumonia, 256 (23.2 %) mechanical-ventilation associated pneumonia, 107 (9.7 %) with primary or vascular catheter-related bacteremia, 47 (4.3 %) with urethral catheter-related urinary infections, 42 (3.8%) with organspace or deep surgical infections and 191 (17.3%) who had other types of infection. An APACHE II upon admission of 19.6 (SD: 8) and severe sepsis or septic shock systemic response in 50.4% of all cases. On 776 (82.7%) occasions treatment was initiated on an empirical basis and in 589 (62.1%) cases the dose of choice was of 0.5 g/ 12 h, with a mean duration of 9 days. In 738 (77.8 %) patients, LVX was used in association with other antibiotics. The clinical response by treatment end was rated as satisfactory in 67.4 % of all NI. Factors related to a non-satisfactory response were as follows: APACHE II (OR: 1.05; 95% CI: 1.028-1.078); septic shock (OR: 2.62; 95 % CI: 1.623-4.219); the requirement for changes in treatment due to poor clinical progress (OR: 66.67; 95% CI: 15.384-250), the presence of non-covered microorganisms (OR: 6.58; 95% CI: 3.663-11.765), the appearance of new resistant pathogens (OR: 6.94; 95 % CI: 2.445- 19.608) or the diagnosis of a new infection (OR: 3.68; 95% CI: 1.504-8.929); solid neoplasm (OR: 1.98; 95% CI: 1.156-3.899); chronic liver disease (OR: 3.11; 95 % CI: 1.429-8.475) and the absence of etiology confirmation (OR: 2.39; 95 % CI: 1.624-3.510). One or more adverse events which were possibly or probably related to the use of LVX were detected in 104 (11.0%) patients. Total intra-ICU mortality amounted to 26.1%, while the accumulated in-hospital mortality was 33.8%. CONCLUSIONS: LVX is a common therapeutic option in the treatment of nosocomial infections in critical patients. It is predominantly used in an empirical manner, at a dose of 0.5 g every 12 hours and in combination with other antibiotics.