Comparison of elimination and absorption pharmacokinetics of linezolid in cystic fibrosis patients by three nonlinear models
Schaeftlein, A.(1,2), Keel, R.A.(3), Kuti, J.L.(3), Kloft, C.(1)
(1) Department of Clinical Pharmacy, Martin-Luther-Universitaet Halle-Wittenberg/Freie Universitaet Berlin, (2) and Graduate Research Training program PharMetrX, Germany, (3) Centre for Anti-Infective Research and Development, Hartford Hospital, Hartford, CT, U.S.A.
Objectives: Linezolid, an oxazolidinone antibiotic used for the treatment of serious Gram-positive infections in cystic fibrosis patients (CFP), displays nonlinear pharmacokinetics (PK) by decreasing clearance over time. The PK of oral linezolid may be affected by gut absorption impairment, particularly in CFP. The objective of this analysis was to develop a population PK model adequately reflecting the elimination and absorption of this special population and to compare its predictive performance with other suggested nonlinear PK models for linezolid.
Methods: A concentration-time-dependent clearance inhibition (CTDCI) model , a Michaelis-Menten-type elimination model  and a linear/nonlinear elimination mixture model  were compared on their ability to describe and predict the elimination process of multiple dose linezolid in 8 CFP. All CFP received oral and intravenous linezolid 600 mg twice daily for 9 doses, separated by a 9 day washout . Visual inspection of the PK data revealed a delay in absorption after oral dosing in selected patient, which was investigated by implementing a lag time or absorption process via transit compartments (TCA)  in the final selected model. All data analyses were performed using the nonlinear mixed-effect modelling approach (NONMEMTM, Version VI). Model comparison was guided by the Akaike information criterion (AIC), goodness of fit (GOF) plots, and visual predictive checks (VPC).
Results: The data were best described by the CTDCI model resulting in the lowest AIC value (779) compared with the other elimination models (819-1080). GOF plots showed that the Michaelis-Menten and the mixture model were less capable of describing the data compared with the CTDCI model. The VPC of the CTDCI model best reflected the general trend of the time course and the variability observed in this population. The model with TCA was inferior to the one with lag time; the observed mean transit time was incorporated as prior knowledge for the lag time in the final model.
Conclusions: Among nonlinear models, the concentration-time-dependent inhibition model appeared to be most suitable to describe and predict linezolid elimination PK in CFP. The absorption process was well represented by the incorporation of the lag time model.
. N. Plock et al. Drug MetabDispos, 35:1816 (2007). S. Swoboda et al. Eur J ClinPharmacol, 66:291 (2010). AK. Meagher et al. Antimicrob Agents Chemother, 47:548 (2003). RA. Keel et al. 50th ICAAC, Boston, MA (2010). RM. Savic et al. J PharmacokinetPharmacodyn, 34:711 (2007)