David Khan (1), Pernilla Lagerbäck (2), Christer Malmberg (2), Otto Cars (2), Lena Friberg (1)
(1) Department of Pharmaceutical Biosciences, (2) Departments of Medical Sciences, Uppsala University, Sweden
Objectives: The developing problem of antibiotic resistance due to overuse of antibiotics is threatening public health. Valuable information on optimal dosing strategies can be obtained from in silico models based on in vitro time-kill curve experiments [1]. We have previously developed an in silico model for E. coli MG1655 and three mutants thereof exposed to ciprofloxacin [2]. This model has been expanded to fit three additional mutants and include filamentation effects. The aim of this work was to predict the outcome of competition experiments between E. coli MG1655 wild type (wt) and three of the mutants following ciprofloxacin exposure and to predict the time-course of bacterial kill for different dosing regimens.
Methods: Experimental data was obtained from 24h in vitro experiments with E. coli MG1655 (ΔaraB) in competition with different well characterized MG1655 mutants (MIC of ciprofloxacin 2-12 times the wt) in starting ratios of wt:mutant ranging from 10:1 to 10000:1. Ciprofloxacin concentrations were constant and chosen to be below, between and above the MICs of the wt and the mutant. Bacteria were quantified on MacConkey agar. The in silico model included compartments for susceptible, persister, preexisting resistant and filamented bacteria. The drug effect on the susceptible bacteria was described by an Emax-model. The competition experiments were predicted from the model and compared to the observed data. Additionally, predictions of different once- and twice-daily dosing regimens of ciprofloxacin were performed.
Results: The model successfully predicted the time course of bacterial kill and growth in the competition experiments except for the one experiment where the MIC of the wt and the mutant only differed two-fold, and the ciprofloxacin concentration was between the two MICs. For all mutants, dose sizes, half-lives, and wt:mutant ratios, the model predicted once-daily dosing regimens to be as efficient or superior compared to the twice-daily regimens of the same total dose amount in overcoming resistance and resulting in an overall bacteria kill.
Conclusions: The model was shown to adequately predict in vitro competition experiments and can thus be a valuable tool in the search for dosing regimens that minimize the growth of resistant mutants existing in a bacterial population. Over a range of doses and half-lives, once-daily dosing was found to be superior over twice-daily dosing by more efficiently suppressing emergence of resistance.
References:
[1] Nielsen EI et al., Pharmacokinetic/pharmacodynamic (PK/PD) indices of antibiotics predicted by a semimechanistic PKPD model: a step toward model-based dose optimization. Antimicrob Agents Chemother. 2011 Oct;55(10):4619-30.
[2] Khan D et al. PKPD-Modeling of time-kill curves from E. coli mutants exposed to ciprofloxacin. PAGE Abstract 2011.
Reference: PAGE 21 () Abstr 2525 [www.page-meeting.org/?abstract=2525]
Poster: Infection