Evaluation of expert rules for carbapenemase class identification in Enterobacterales isolates using the VITEK2 susceptibility testing platform

Prompt identification of enzyme class in carbapenemase-producing carbapenem-resistant Enterobacterales (CP-CRE) has critical implications for informing appropriate patient treatment, infection prevention, and public health. This study evaluates the performance of prototype bioMérieux Advanced Reporting Tool (bioART) expert rules to rapidly identify Klebsiella pneumoniae carbapenemase (KPC), metallo-β-lactamase (MBL), and OXA-48-like carbapenemase enzymes in CP-CRE. The bioART rules are designed for use with routine cards on the VITEK2 automated antimicrobial susceptibility testing (AST) platform. Results provided by the bioART rules should be interpreted in combination with the instrument Advanced Expert System (AES) to comprehensively evaluate phenotypic resistance phenotypes. Two hundred clinical isolates with varied β-lactam resistance profiles were enrolled, with 196 ultimately analyzed, including 115 CP-CREs. The AES alone detected CP-CRE isolates with a sensitivity of 83% and a specificity of 85%. The combined AES and bioART rules detected CP-CRE with a sensitivity of 96% and a specificity of 83%. Prediction of carbapenemase classes by the bioART rules was analyzed for a subset of 158 isolates, including only the species for which the rules have claimed indications. MBL, KPC, and OXA-48-like enzymes were detected with sensitivities of 97%, 76%, and 47%, respectively. Notably, sensitivity was 90% for single OXA-48-like producers, whereas OXA-48-like enzymes in dual New Delhi metallo-β-lactamase (NDM)/OXA-48-like-producing isolates were undetectable using phenotypic susceptibility patterns, and isolates were reported only as producing NDM enzymes. Overall, laboratories incorporating these tools into carbapenemase screening workflows may consider their utility in prompting confirmatory carbapenemase testing, guiding modifications to AST reporting, and/or prompting additional susceptibility testing. IMPORTANCE Carbapenem-resistant bacteria are a major public health concern due to their ability to spread in healthcare settings and cause infections that are difficult to treat with first-line antibiotics. Identification of the enzyme classes responsible for carbapenem resistance plays a crucial role in ensuring that patients receive effective treatments and controlling the spread of these bacteria. In this study, we evaluated the performance of a new approach to identify carbapenemase enzymes without additional hands-on testing. The method is designed for use with the VITEK2 automated susceptibility testing platform to recognize patterns of resistance to antibiotics and make predictions about the possible resistance mechanisms.