Bacterial antibiotic resistance studies using in vitro dynamic models: Population analysis vs. susceptibility testing as endpoints of mutant enrichment

Emergence of bacterial antibiotic resistance is usually characterised either by population analysis or susceptibility testing. To compare these endpoints in their ability to demonstrate clear relationships with the ratio of 24-h area under the concentration-time curve (AUC24) to the minimum inhibitory concentration (MIC), enrichment of ciprofloxacin-resistant mutants of four clinical isolates of Pseudomonas aeruginosa was studied in an in vitro dynamic model that simulates mono-exponential pharmacokinetics of ciprofloxacin over a wide range of the AUC24/MIC ratios. Each organism was exposed to twice-daily ciprofloxacin for 3 days. Amplification of resistant mutants was monitored by plating on media with 2×, 4×, 8× and 16× MIC of ciprofloxacin. Population analysis data were expressed by the area under the bacterial mutant concentration-time curve (AUBCM). Changes in P. aeruginosa susceptibility were examined by daily MIC determinations. To account for the different susceptibilities of P. aeruginosa strains, post-exposure MICs (MICfinal) were related to the MICs determined with the starting inoculum (MICinitial). For each organism, AUC24/MIC relationships both with AUBCM and MICfinal/MICinitial were bell-shaped, but the latter were more strain-specific than the former. Using combined data on all four isolates, AUBCM showed a better correlation than MICfinal/MICinitial (r2 = 0.75 vs. r2 = 0.53). The shift of MICfinal/MICinitial relative to AUBCM vs. AUC24/MIC curves resulted in a weak correlation between AUBCM and MICfinal/MICinitial (r2 = 0.41). These data suggest that population analysis is preferable to susceptibility testing in bacterial resistance studies and that these endpoints should not be considered interchangeable.