Cefepime susceptibility testing in AmpC-hyperproducing Enterobacter cloacae complex: limitations of routine methods and a pragmatic diagnostic approach

We compared and evaluated three routine methods (MicroScan, gradient diffusion and disk diffusion) with broth microdilution (BMD) for cefepime against 200 AmpC-hyperproducing Enterobacter cloacae complex isolates. Reference BMD classified 40.0% as susceptible, 39.5% as susceptible with increased exposure, and 20.5% as resistant to cefepime, according to EUCAST criteria. Categorical agreement (CA) was 63.5% for MicroScan, 73.5% for gradient diffusion, and 69.5% for disk diffusion. MicroScan showed the highest sensitivity for detecting BMD-defined resistance (82.9%) but moderate specificity (71.0%), whereas gradient diffusion and disk diffusion showed lower sensitivities (46.3% and 53.7%, respectively) but higher specificities (95.6% and 92.5%, respectively). Similar method-dependent patterns were observed using CLSI criteria. Error-rate-bound analysis confirmed that these discrepancies were not restricted to isolates with MICs close to the clinical breakpoints. A modified EUCAST-based stepwise algorithm improved the balance of performance compared with MicroScan alone, mainly by reducing overestimation of resistance, but clinically relevant misclassification persisted. Overall, no routine method or algorithm fully replaced BMD, supporting confirmatory reference testing when accurate cefepime categorization is clinically critical.