Direct antimicrobial susceptibility testing for acute urinary tract infection – comparison of AI-assisted zone interpretation and conventional susceptibility testing methods

Direct antimicrobial susceptibility testing (dAST) from urine samples using disc diffusion methods is in widespread use in some countries, allowing for the provision of results earlier than traditional AST methods, at a lower cost. However, it is a labour-intensive technique subject to reader variability and its use remains controversial, largely because there is no published method or standardized breakpoints.

This study examined the veracity of dAST from urine samples and investigated the capability of APAS Independence with the APAS-AMR analysis module (Clever Culture Systems) in providing a solution in lieu of manual plate-in-hand reading or minimum inhibitory concentration methodology using Vitek 2 (BioMérieux).

192 urine samples were inoculated directly onto test plates and incubated according to CLSI guidelines. Additionally, 82 significant isolates were identified and tested using traditional disc diffusion methods, as well as using the Vitek 2 as a reference. All plates were read by APAS and plate-in-hand methods, and results were examined to determine categorical agreement, minor errors, major errors, and very major errors. Differences in zone measurements between automated and manual reads were also analysed.

The performance of dAST from uncomplicated UTI samples was found to be comparable to testing isolated organisms in the Vitek system. Very major errors were largely due to poor lawn quality due to low inoculum levels, which the APAS-MAR system is able to detect and flag for microbiologist review. Furthermore, no significant difference was found between manual zone measurements compared to APAS for both direct, and traditional disc diffusion methods, demonstrating that together, dAST and APAS-AMR offer clear cost and time benefits for the laboratory.

Poster & Oral Presentation: by St Vincent’s Hospital, Melbourne, Australia

Conference: ECCMID 2023, Copenhagen, Denmark

Date: April 2023

Authors: Brenton, L., Hill, R., and Giglio, S.

Citation: Brenton, L., Hill, R., and Giglio, S. 2023. Direct antimicrobial susceptibility testing for acute urinary tract infection – comparison of AI-assisted zone interpretation and conventional susceptibility testing methods. 33rdECCMID. Copenhagen.