The objective of this study was to evaluate the properties of ciprofloxacin in intensive care patients using a population approach. Seventy patients received ciprofloxacin. On Day 1, three to eight blood samples were taken over a 12-h period. Peak drug concentration (Cmax) and 24-h area under the concentration-time curve (AUC) were compared with the French breakpoint defining antibiotic susceptibility. A population pharmacokinetic modelling approach was then carried out. A two-compartment open model with a proportional error model best fitted the data. A relationship between the elimination constant rate and the Cockcroft creatinine clearance was found. Ciprofloxacin clearance was 13.6+/-5.8L/h, the volume of distribution was 62.0+/-10.7 L and the ciprofloxacin half-life was 3.7+/-1.8h. When the minimum inhibitory concentration (MIC) was equal to 1mg/L the inhibitory ratio (IR) was > or = 8 in only 10.8% of cases, and the AUC/MIC ratio (AUIC) was 42.0+/-36. In conclusion, this study highlights that the Cockcroft clearance significantly influences ciprofloxacin elimination. Target plasma concentrations for ciprofloxacin, the IR and AUIC were rarely reached with a standard dosing regimen. In critically ill patients, the observed pharmacokinetic variability is mainly responsible for the overly frequent low concentrations of ciprofloxacin, emphasising the need for therapeutic monitoring.