Background: Mechanism-based mathematical models (MBMs) address limitations of PK/PD indices, such as the time unbound antibiotic concentration exceeds the pathogen’s MIC (fT>MIC). Meropenem (MEM) is commonly used against Pseudomonas aeruginosa (PA) infections, however resistance mechanisms reduce its effectiveness. We developed a novel MBM to 1) characterise the full time-course of bacterial killing and resistance emergence in seven isogenic PA strains in response to MEM, and 2) determine the contribution of various resistance mechanisms present in these strains on bacterial outcomes.
Methods: We conducted 10-day hollow-fibre infection model (HFIM) studies using seven isogenic PA strains with OprD porin channel loss, MexAB-OprM efflux pump over-expression, AmpC β-lactamase over-expression, and the combinations thereof. The HFIM simulated MEM PK for critically-ill patients with normal renal function (t1/2,MEM=1.5h). All viable counts on drug-free, 3×MIC and 5×MIC MEM-containing agar across all strains, five clinically relevant regimens and control (n=90 profiles) were modelled simultaneously.
Results: fT>MIC could not explain the differences in bacterial response between strains. For example, regimens achieving ≥98% fT>1xMIC suppressed regrowth and resistance of one strain, while even 100% fT>5xMIC failed to achieve this against two other strains, despite all three of these having the same MIC. In contrast, the MBM well characterised all bacterial outcomes of all seven strains with the same model structure and without estimating strain-specific drug effect parameters (observed vs individual-fitted r2=0.98, observed vs population-fitted r2=0.96, Figure 1 – shows one strain). Three different resistance mechanisms present by themselves (single-mutant strains) and in combinations of two (double-mutant strains) were described by their effects on the estimated MEM concentration in the periplasmic space (Figure 2 – Model diagram).
Conclusions: fT>MIC could not explain the differences in bacterial response between strains. In contrast, the MBM well characterised all bacterial outcomes of seven isogenic strains simultaneously. The developed MBM is the first model to directly translate all major mechanisms of MEM resistance in PA and their complex interplay. This model represents a first necessary step towards personalised therapy adapted to the individual and the infecting pathogen.