A Population Pharmacokinetic Model for the Simultaneous Description of Linezolid Tissue and Plasma Disposition in Healthy Volunteers and Septic Patients
Plock, N.(1), C. Buerger(1), K. Kuester(1), C. Joukhadar(2), S. Kljucar(3), C. Kloft(1,4)
(1)Freie Universitaet Berlin, Institute of Pharmacy, Berlin, Germany; (2)University School of Medicine, Vienna, Austria; (3)DRK Kliniken Berlin Westend, Berlin, Germany; (4)Martin-Luther-Universitaet Halle-Wittenberg, Faculty of Pharmacy, Halle, Germany
Objectives: Linezolid, the first member of the oxazolidinones, has been approved for the treatment of severe gram-positive infections. The aim of the study was to characterise the subcutaneous (sc) and muscular (im) tissue pharmacokinetics (PK) of unbound drug in healthy volunteers and septic patients applying the microdialysis and population analysis technique.
Methods: 34 individuals (24 patients) were treated with 600 mg linezolid bid as iv infusion with the exception of oral multiple dosing for healthy volunteers. Samples were taken after single dose and at steady state over a period of 8 hours. Unbound concentrations were determined in plasma and tissue placing two microdialysis probes in each subject (sc and im). Population pharmacokinetic analysis was carried out using the software NONMEM.
Results: Linezolid disposition in ultrafiltrate was best described by a two-compartment model including a nonlinear influence on clearance (CL) implemented by an additional inhibition compartment. CL was estimated to be 11.5 L/h with a 57% noninhibitable fraction (VAR). Central and peripheral distribution volumes were estimated to be 20 L (V2) and 27 L (V3), respectively. Interindividual varibility (IIV) was established for CL, V2, V3, KA and VAR ranging between 20% and 79% (ωKA). Creatinine clearance, thrombocytes and weight were included as covariates on CL and weight on V3, thus reducing unexplained parameter variability by 7-41%. Unbound sc and im tissue concentrations were incorporated into the model using two additional compartments, each parameterised with assumed equal monodirectional rates into and out of the compartment. Parameters previously obtained for ultrafiltrate were fixed. Tissue penetration was assessed by using partition coefficients PC23 and PC24 for sc and im tissue, respectively, which were estimated to be 1.05 for PC23 and 1.07 for PC24. Between occasion variability on PC23 and IIV on PC24 yielded values of 44% and 28%. Parameters were generally estimated with good precision (RSE<47%).
Conclusions: Unbound linezolid pharmacokinetics in tissue and plasma of both populations were successfully described by the population pharmacokinetic modelling approach. In general linezolid penetrated well into tissue fluid but displayed high variability. Inclusion of covariates significantly reduced unexplained variability, a prerequisite for predicting unbound tissue concentrations and individualisation of antimicrobial therapy.