Mechanism-based Modelling of the Synergy of Colistin Combinations against Multidrug-Resistant Gram Negative Bacteria
J.B. Bulitta (1, 3), J. Li (2), A. Poudyal (2), H.H. Yu (2), R.J. Owen (2), B.T. Tsuji (3), R.L. Nation (2), A. Forrest (1, 3)
(1) Ordway Research Institute, Albany, NY; (2) Facility for Anti-infective Drug Development and Innovation, Monash University, Victoria, Australia; (3) School of Pharmacy and Pharmaceutical Sciences, Buffalo, NY.
Objectives: Design of rational antibiotic combination regimens is critical to treat infections by multidrug-resistant (MDR) A. baumannii (Ab), K. pneumoniae (Kp) and P. aeruginosa (Pa). As synergy of antibiotic combinations has been modelled empirically, we developed new mechanism-based models that specifically account for potential causes of synergy and that describe the rate of bacterial killing and regrowth for colistin alone or in combination with other antibiotics.
Methods: Checkerboard synergy panel studies were used to identify potent synergistic combinations of antibiotics that were then evaluated in time-kill studies. Combinations were: Colistin (C) & rifampicin (R) against Ab and Pa; C & meropenem (M) against Kp. Time-kill studies at an initial inoculum of 106 CFU/mL (8 samples up to 48 h per profile) included 8 concentrations each of C, R or M alone and 9 concentrations for combinations of C&R and C&M. Mechanism-based models with up to 4 populations with different susceptibility were fit to the time-kill data in NONMEM VI. An additive error model was used for viable counts on log-scale and a Poisson error was included to fit low viable counts (incl. observations with zero colonies on the agar plate).
Results: The combinations displayed more rapid killing and less regrowth at 24 and 48 h compared to each antibiotic alone. The curve fits for all three pathogens were unbiased and reasonably precise (r>0.93). The mechanisms of synergy were modelled as: R was estimated to substantially enhance (Emax = 244; EC50 < 0.1 mg/L for R) the rate of killing by C against Ab. C (EC50 1.1 mg/L for C) enhanced killing by M towards the meropenem-resistant population against Kp. R & C mutually enhanced the extent of killing of the other antibiotic against Pa.
Conclusions: The proposed mechanism-based models described the observed viable counts well and accounted for the presence of multiple bacterial populations. Such mechanism-based models that can propose and evaluate various mechanisms of synergy hold promise for rationally optimizing combination regimens in humans.