Evaluation of EUCAST Rapid Antimicrobial Susceptibility Testing for Gram-Negative ESKAPEEc Pathogens in Blood Cultures, with a Focus on Carbapenemase-Producing Klebsiella pneumoniae in a University Hospital in Palermo, Italy
- Autori: Cannella, S.; Pipito', L.; Piazza, M.; Graceffa, D.; Immordino, R.; Virruso, R.; Giammanco, G.M.; Cascio, A.; Bonura, C.
- Anno di pubblicazione: 2025
- Tipologia: Articolo in rivista
- OA Link: http://hdl.handle.net/10447/696583
Abstract
Background: Rapid antimicrobial susceptibility testing (RAST) allows early detection of resistance directly from positive blood cultures, potentially improving outcomes in bloodstream infections (BSIs). We evaluated the performance of EUCAST RAST for Gram- negative ESKAPEEc pathogens and characterized carbapenemase genes in carbapenem- resistant Klebsiella pneumoniae (CRKP). Methods: A total of 354 positive blood cultures were screened, including 51 monomicrobial Gram-negative ESKAPEEc isolates. RAST results at 4, 6, 8, and 16–20 h were compared with standard antimicrobial susceptibility testing (AST) obtained using the BD Phoenix™ system. Categorical agreement (CA) and error frequency were calculated. Multiplex PCR and Sanger sequencing were performed on 15 CRKP isolates to identify carbapenemase genes and allelic variants. Results: 51 Gram- negative ESKAPEEc isolates met the inclusion criteria for RAST (15 E. coli, 19 K. pneumoniae, 11 A. baumannii, and 6 P. aeruginosa). Overall performance varied markedly by species and antibiotic. E. coli showed frequent unreadable or ATU zones at early timepoints and wide CA variability (50–100%), with high very major error (VME) rates for AMP, TZP, and CAZ, particularly at 6–8 h. K. pneumoniae displayed consistently high CA (mostly 100%) for carbapenems, CAZ, and TZP. A. baumannii demonstrated excellent agreement (100% for most agents), except for GEN at 6–8 h. P. aeruginosa could be evaluated only at 16–20 h, showing high CA for AMK, CAZ, and CIP; lower CA for MEM (83%); non-calculable CA for IMI due to universal ATU readings; and a CA value of 0% for TZP due to the predominance of the ATU results. VMEs ranged from 0% to 26.1% across species and reading times, but carbapenems did not generate VMEs. Molecular analysis revealed blaKPC in 66.7%, blaNDM in 46.7%, and blaOXA-48 in 33.3% of isolates, with co-occurrence in several strains. Sequencing identified blaKPC-2 and blaNDM-1 as the predominant variants, with one isolate harboring blaNDM-5. Conclusions: EUCAST RAST markedly accelerates susceptibility reporting from positive blood cultures, but its accuracy is species- and time-dependent. Performance was excellent for K. pneumoniae (including CRKP) and A. baumannii and acceptable for P. aeruginosa at 16–20 h. In contrast, E. coli showed frequent ATU results at early timepoints and high ME/VME rates, making readings before 8 h unreliable for clinical decisions. Overall, RAST can effectively support rapid antimicrobial stewardship when species-specific limitations are recognized, and early-timepoint results are interpreted with caution.
