Elisabet Nielsen (1), Anders Viberg (1), Otto Cars (2) and Marie Sandström (1)
(1) Department of Pharmaceutical Biosciences, Division of Pharmacokinetics and Drug Therapy, Uppsala University, Sweden; (2) Department of Medical Sciences, Section of Infectious Diseases, Uppsala University, Sweden
Objectives: In vitro time-kill studies are commonly used in the assessment of efficacy of antimicrobial agents. The aim of the present study was to develop a general semi-mechanistic Pharmacokinetic-Pharmacodynamic model that describes the killing of bacteria using data from time-kill experiments.
Methods: Time-kill curves were performed in which cultures of Streptococcus Pyogenes (M12 NCTC P1800) were exposed to constant concentrations of five different antibiotics; benzylpenicillin, cefuroxime, erythromycin, moxifloxacin, and vancomycin. The concentrations ranged from 0 to 64 times the minimum inhibitory concentration (MIC) and samples for viable count were taken at time points 0, 1, 2, 4, 6, 9, 12, 15, 18 and 24 h after start of experiments. In total 135 experiments were performed resulting in 2526 samples for viable counts. The data was modeled using NONMEM (version VI beta) using first-order conditional estimation method. All data was modeled simultaneously and the parameters in the model were specified as either bacterial or drug specific. Model performance was assessed by evaluation of diagnostic plots and precision of parameter estimates.
Results: A two compartment model was developed where the total bacterial population was divided into two subpopulations, one growing drug sensitive population (S) and one resting insensitive population (R). The total number of bacteria triggers the transfer rate from S to R. This model described well that bacteria exposed to low or none concentration of antibiotics will grow exponentially until reaching a maximum level of bacteria and then a stationary phase is reached. The tolerance development often seen in bacterial time-kill studies was also explained using the transfer between the different compartments.
Conclusion: The developed PK-PD model could describe both the early rapid killing and the late tolerance development seen after exposure of antibiotic to bacteria in an in vitro system. The model allows an efficient summarization of time-kill data and could be of use for prediction of better dosing strategies of antibiotics as well as for synergy studies of poly antibiotic treatment.
Reference: PAGE 14 () Abstr 768 [www.page-meeting.org/?abstract=768]
Poster: poster