Pharmacodynamics of colistin resistance in carbapenemase-producing : the double-edged sword of heteroresistance and adaptive resistance.

The presence of heteroresistant subpopulations and the development of resistance during drug exposure (adaptive resistance) limits colistin's efficacy against carbapenemase-producing (CP-Kp) isolates.. The pharmacokinetic/pharmacodynamic (PK/PD) characteristics of both types of colistin resistance against CP-Kp are unknown. We therefore studied the PK/PD characteristics of colistin resistance in an PK/PD model simulating clinical colistin exposures. Two clinical isolates, one non-CP-Kp and one CP-Kp, with colistin MICs of 0.5-1 mg l at a final inoculum of 10 c.f.u. ml were used in an PK/PD dialysis/diffusion closed model simulating 4.5 MU q12h and 3 MU q8h clinical dosing regimens. Heteroresistant (HRS, bacteria with stable high-level resistance present before drug exposure) and adaptive resistant (ARS, bacteria with reversible low-level resistance emerging after drug exposure) subpopulations were measured and optimal PK/PD targets for reducing both ARS and HRS were determined. Cumulative fractional response (CFR) was calculated with Monte Carlo simulation for 9 MU q24h, 4.5 MU q12h and 3 MU q8h clinical dosing regimens. A 2-5 logc.f.u. ml decrease of the total bacterial population was observed within the first 2 h of exposure, followed by regrowth at 12 h. Colistin exposure was positively and negatively correlated with HRS and ARS 24-0 h c.f.u. ml changes, respectively. An optimal PK/PD (~0.5log increase) target of 35 fAUC/MIC (the ratio of the area under the unbound concentration-time curve to the MIC) was found for reducing both HRS and ARS of high-level resistance (MIC >16 mg l). The 4.5 MU q12h regimen had slightly higher CFR (74 %) compared to the other dosing regimens. High colistin exposures reduced high-level adaptive resistance at the expense of selection of heteroresistant subpopulations.