Rapid antibiotic susceptibility testing directly from blood culture bottles

Bloodstream infections and sepsis are medical emergencies with high mortality rates, and the rapid initiation of appropriate antimicrobial therapy is critical. Conventional antibiotic susceptibility tests (AST) typically take 2-3 days, which can lead to delays in treatment decisions. The aim of this study was to comparatively evaluate the performance (speed, accuracy, reliability) of a rapid AST method applied directly from positive blood culture bottles using the disk diffusion method for various Gram-negative and Gram-positive bacterial species against the standard disk diffusion method. The study included Escherichia coli, Klebsiella pneumoniae, Pseudomonas aeruginosa, Acinetobacter baumannii, Staphylococcus aureus, Enterococcus faecalis, and Enterococcus faecium strains isolated from positive signal blood culture bottles. The rapid AST was performed by direct inoculation from the blood culture bottle onto Mueller-Hinton agar plates and placement of antibiotic disks. Inhibition zone diameters were read at 4, 6, 8, and 16-20 hours. These results were compared with the standard disk diffusion results (reference method) performed after 16-20 hours of incubation from subculture. Performance was assessed by calculating the Categorical Agreement (CA), Very Major Error (VME), Major Error (ME), and Minor Error (mE) rates. The developed rapid AST method provided stable and interpretable results for most tested bacterial species with 6 to 8 hours of incubation. This timeframe has the potential to provide antibiotic susceptibility test results approximately 36 hours earlier compared to conventional methods. Overall, acceptable levels of categorical agreement (above 90%) with the reference method were found in many bacteria- antibiotic combinations, particularly at the 8-hour readings. For Acinetobacter baumannii and Enterococcus faecalis isolates, 100% CA and zero error rates were achieved for all tested antibiotics. However, clinically critical and unacceptably high error rates were also identified: VME rates of 80% for clindamycin against Staphylococcus aureus, ME rates of 100% for amoxicillin/clavulanate against Escherichia coli, 46.7% for imipenem against Klebsiella pneumoniae, and 57.1% for ceftazidime against Pseudomonas aeruginosa were determined. High VME rates were also observed for Klebsiella pneumoniae trimethoprim/sulfamethoxazole and Escherichia coli ceftazidime combinations. The disk diffusion-based rapid AST method applied directly from positive blood culture bottles is a promising approach with the potential to significantly accelerate antibiotic susceptibility results. The method showed excellent performance for some pathogens such as Acinetobacter baumannii and Enterococcus faecalis. However, the unacceptable VME and ME rates found in certain critical combinations, such as Staphylococcus aureus clindamycin and Escherichia coli amoxicillin/clavulanate, indicate that the method is not reliable on its own for these specific scenarios. Therefore, the rapid AST results should be considered a preliminary report, and treatment decisions (especially de-escalation) must be confirmed by standard methods, particularly for combinations with a high risk of error. Further optimization, large-scale validation, and integration with rapid identification methods are recommended to enhance the clinical utility of the method.