RESUMO
Objectives: Colistin is a last-resort antibiotic against the critical-status pathogen Pseudomonas aeruginosa. There is still uncertainty regarding how to accurately measure colistin susceptibility in P. aeruginosa. Evaluation of antimicrobial susceptibility testing (AST) methods is largely hampered by the lack of resistant isolates and those around the susceptibility breakpoint. The aim of this study was to generate such strains in a morbidostat device for use in AST method evaluation. Methods: A morbidostat device was used to cultivate susceptible clinical strains into isolates with a wide range of colistin MICs. Subsequently, five commercial AST methods were compared against the gold standard broth microdilution (BMD) method: MICRONAUT-S, SensiTest, Sensititre, Rapid Polymyxin Pseudomonas and Etest. Results: A total of 131 P. aeruginosa isolates were used for colistin susceptibility test evaluation (100 colistin susceptible and 31 colistin resistant). The 31 colistin-resistant isolates evolved resistance in the morbidostat to different MIC ranges (4-512 mg/L, 100% resistance generation efficacy). The categorical agreement (CA) rates for MICRONAUT-S, SensiTest and Rapid Polymyxin Pseudomonas were 94.7%, 93.9% and 92.4%, respectively. The Sensititre achieved the highest CA score (96.9%), whereas the Etests had the lowest CA score (84%). The very major discrepancy (VMD) rates for all tests were between 3.2% and 67.7%. Conclusions: The morbidostat device can efficiently provide laboratories with colistin-resistant strains for test evaluation. Although CA rates were high for commercial AST methods except for Etests, none met the ≤1.5% CLSI limit for VMD rates. Performance was generally inferior when using isolates with low-level resistance.