Performance of disk diffusion, gradient test, and VITEK 2 for carbapenem susceptibility testing in OXA-48-like carbapenemase-producing <i>Enterobacterales</i> : a comparative study

This study aimed to compare the performance of disk diffusion, gradient test (ETEST), and VITEK 2 (AST-N223, AST-N428, AST-N432 cards) antibiotic susceptibility testing methods with the reference broth microdilution (BMD) for carbapenem susceptibility in OXA-48-like carbapenemase-producing <i>Enterobacterales</i> (CPE). A total of 107 CPE and 142 controls (<i>Enterobacterales</i> that do not produce any type of carbapenemases), all molecularly characterized by whole-genome sequencing, were tested for carbapenem susceptibility using BMD and derivative methods. Essential agreement (EA), categorical agreement (CA), major error, very major error, and bias were evaluated. In the OXA-48-like group, resistance frequencies by BMD for ertapenem, imipenem, and meropenem were 86.9%, 12.1%, and 10.3%, respectively. For OXA-48-like CPE, ETEST showed the highest EA among all methods for meropenem (100/107, 93.5%) and ertapenem (99/107, 92.5%), while EA for VITEK 2 cards were <90%. In contrast, for imipenem, VITEK 2 AST-N428 performed best with an EA of 95/105 (90.5%). CA was higher for ertapenem across all methods (93.5%-98.1%) compared to imipenem (59.8%-81.4%) and meropenem (78.8%-95.3%). The highest CA was achieved with ETEST for ertapenem and meropenem, and with VITEK 2 AST-N223 for imipenem. Significant variability was observed across different tests in resistance frequencies, MICs, EA, and CA for the OXA-48-like group. Ertapenem was the most useful carbapenem for detecting resistance in OXA-48-like CPE across all methods. Laboratories should be aware that susceptibility testing of imipenem leads to more erroneous results compared to the other carbapenems when using derivative methods. Additionally, most derivative methods tend to overcall carbapenem resistance in OXA-48-like CPE.IMPORTANCEOXA-48-like is the most frequent carbapenemase in western Europe, and both its rapid spread and its challenging-to-detect nature are a particular concern for adequate treatment and infection control purposes. Accurate determination of carbapenem minimal inhibitory concentrations (MICs) is of utmost importance, both for the selection of the best therapy and as a marker for carbapenemase detection. However, the performance of derivative susceptibility testing methods is unclear for OXA-48-like isolates. Our study reports on the varying performance of carbapenem susceptibility testing by disk diffusion, gradient test (ETEST), and VITEK 2 in OXA-48-like-producing <i>Enterobacterales</i>. The results of the present study can help to inform about the limitations of current susceptibility testing methods and serve to improve MIC determination in these challenging isolates.