Abstract
The dosage of colistin for the treatment of enteric E. coli in animals necessitates considering the heteroresistant (HR) nature of the targeted inoculum, described by the presence of a major susceptible population (S1, representing 99.95% of total population) mixed with an initial minor subpopulation of less susceptible bacteria (S2). Herein, we report the 1-compartment population pharmacokinetics (PK) of colistin in chicken intestine (jejunum and ileum) and combined it with a previously established pharmacodynamic (PD) model of HR in E. coli. We then computed probabilities of target attainment (PTA) with a pharmacodynamic target (AUC(24h)/MIC) that achieves 50% of the maximal kill of bacterial populations (considering inoculums of pure S1, S2 or HR mixture of S1 + S2). For an MIC of 1 mg/L, PTA > 95% was achieved with the registered dose (75,000 IU/kg BW/day in drinking water) for the HR mixture of S1 + S2 E. coli, whether they harboured mcr or not. For an MIC of 2 mg/L (ECOFF), we predicted PTA > 90% against the dominant susceptible sub-population (S1) with this clinical dose given (i) over 24 h for mcr-negative isolates or (ii) over 6 h for mcr-positive isolates (pulse dosing). Colistin clinical breakpoint S ≤ 2 mg/L (EUCAST rules) should be confirmed clinically.