Objective: A prospective pharmacokinetic study was performed in Caucasian patients from an intensive care unit with respiratory support to evaluate the influence of this circumstance on the pharmacokinetic behaviour of levofloxacin. Patients and methods: A standard dosage regimen of 500 mg/day was administered to nine Caucasian patients included in the study, irrespective of their demographic characteristics. The experimental data on plasma concentrations were analysed by independent-modelling techniques to estimate the following pharmacokinetic parameters: area under the plasma concentration-time curve (AUC), volume of distribution at steady state (Vss), plasma clearance (CL), maximum plasma concentration at steady state (Cmax,ss) and elimination half-life (t1/2β). Multiple regression analysis was applied to establish the type of correlation between the pharmacokinetic parameters and patient characteristics; the Monte Carlo simulation technique was implemented for the pharmacokinetic/pharmacodynamic analysis based on the probability distribution of the values of AUC/minimum inhibitory concentration (MIC) and Cmax,ss/MIC observed in this group of patients. Results and conclusion: The results show that for AUC the simplest linear model with creatinine clearance as the only independent variable fits the data at a 99% confidence level, explaining more than 85% of the observed variability in this parameter. The volume of distribution, however, showed a statistical correlation with the severity of the illness (Simplified Acute Physiology Score II), although total bodyweight also explains a high percentage of variability of these parameters. Since the group of patients included in the study was small and also included obese individuals, it is difficult to estimate with precision the contribution of each circumstance (overweight or illness severity) to the pharmacokinetic behaviour of levofloxacin. © 2005 Adis Data Information BV. All rights reserved.