Application of pharmacokinetic/pharmacodynamic modelling and simulation for the prediction of target attainment of ceftobiprole against meticillin-resistant Staphylococcus aureus using minimum inhibitory concentration and time–kill curve based approaches

Abstract

The purpose of this report was to compare two different methods for dose optimisation of antimicrobials. The probability of target attainment (PTA) was calculated using Monte Carlo simulation to predict the PK/PD target of fT>MIC or modelling and simulation of time–kill curve data. Ceftobiprole, the paradigm compound, activity against two MRSA strains was determined, ATCC 33591 (MIC=2mg/L) and a clinical isolate (MIC=1mg/L). A two-subpopulation model accounting for drug degradation during the experiment adequately fit the time–kill curve data (concentration range 0.25–16× MIC). The PTA was calculated for plasma, skeletal muscle and subcutaneous adipose tissue based on data from a microdialysis study in healthy volunteers. A two-compartment model with distribution factors to account for differences between free serum and tissue interstitial space fluid concentration appropriately fit the pharmacokinetic data. Pharmacodynamic endpoints of fT>MIC of 30% or 40% and 1- or 2-log kill were used. The PTA was >90% in all tissues based on the PK/PD endpoint of fT>MIC >40%. The PTAs based on a 1- or 2-log kill from the time–kill experiments were lower than those calculated based on fT>MIC. The PTA of a 1-log kill was >90% for both MRSA isolates for plasma and skeletal muscle but was slightly below 90% for subcutaneous adipose tissue (both isolates ca. 88%). The results support a dosing regimen of 500mg three times daily as a 2-h intravenous infusion. This dose should be confirmed as additional pharmacokinetic data from various patient populations become available.