Evaluating the adequacy of intervals for repeat antimicrobial susceptibility testing

다국어 초록 (Multilingual Abstract)

Background: The optimal frequency for repeat antimicrobial susceptibility testing (AST) of isolates of the same species remain undefined despite its importance in antimicrobial therapy. This study assessed suitable repeat AST intervals by analyzing time-dependent changes in susceptibility patterns. Methods: This retrospective observational study analyzed laboratory AST data collected over 10 years (2010–2019) from a tertiary hospital in Seoul, Korea. We examined AST results for “paired” (same patient, same day) and “successive” (same patient, different times: 1–7, 8–30, and 31–365 days) isolates of common bacterial species: Staphylococcus aureus, Enterococcus faecalis, Enterococcus faecium, Escherichia coli, Klebsiella pneumoniae, Acinetobacter baumannii complex, and Pseudomonas aeruginosa. The essential agreement (EA), categorical agreement (CA), essential MIC increase (EMI), and change from nonresistant to resistant (CNRR and CNRR+EMI) rates were calculated. Results: Paired isolates showed > 90% CA for most species, although P. aeruginosa had a lower EA (71.5%) and CA (69.6%). For successive isolates, the EMI, CNRR, and CNRR+EMI generally increased over time. The S. aureus and E. faecalis species demonstrated high stability with low CNRR+EMI rates. Conversely, P. aeruginosa exhibited the most notable changes: its CNRR+EMI increased from 4.1% (1–7 days) to 6.6% (31–365 days), and its EMI remained the highest across all time periods, starting at 7.1% within 1–7 days. Conclusion: The stability of antimicrobial susceptibility patterns varies considerably by bacterial species and time interval. For bacteria with low CNRR+EMI rates, such as S. aureus and E. faecalis, the frequency of repeat AST should be reduced, whereas for bacteria with variable susceptibility patterns, such as P. aeruginosa, more frequent retesting should be performed. The study findings could help laboratories optimize repeat AST protocols and balance timely resistance detection with optimal resource management.