Objectives: To determine the pharmacokinetic characteristics of linezolid and teicoplanin in critically ill patients.Patients and methods: Serum was collected frequently during day 0 and then pre- and 1 h post-dose on days 1, 2, 3, 5, 7 and every third day thereafter during treatment. Serum linezolid concentrations were analysed using HPLC. Serum teicoplanin levels were analysed by fluorescence polarization immunoassay.Results: A two-compartment model was required to characterize linezolid pharmacokinetics (n=28) and account for the accumulation seen after multiple dosing. The estimated clearance was 0.049 +/-0.016 L/h/kg (+/-s.e.m. of estimate). At steady state (dosing interval 12 h), linezolid serum concentrations exceeded the breakpoint of 4 mg/L for 10.88 h (95% CI 10.09-11.66) after a 600 mg dose with an AUC/MIC of 92.4 (95% CI 57.2-127.7). Teicoplanin was best described by a two-compartment model (n=26). The clearance was 4.97+/-1.58 L/h. Serum levels exceeded the breakpoint of 4 mg/L for the entire dosing interval in all subjects (400 mg dose every 12 h) with an AUC/MIC of 399.3 (95% CI 329.6-469.0). However, only four of 14 exceeded trough serum concentrations of 10 mg/L. For both agents, trough levels were similar in those who survived and those who died.Conclusions: Linezolid dosage at 600 mg every 12 h was adequate in the critically ill without need for adjustment for renal function. For teicoplanin, further study is needed to confirm if a trough of 10 mg/L is associated with a higher rate of cure than 5 mg/L. If so, serum drug assays would be needed to ensure a therapeutic level.