Comparison of Disk Diffusion, Gradient Test Strips and Broth Microdilution for Cefiderocol Susceptibility Testing on Multidrug-Resistant Gram-Negative Organisms

Cefiderocol is a last-resort option for infections caused by multidrug-resistant (MDR) Gram-negative organisms. However, susceptibility testing is technically demanding: variable performances have been observed for disk diffusion (DD), and gradient diffusion is not endorsed by EUCAST as an alternative to broth microdilution (BMD). Two hundred and five isolates of Enterobacterales (n = 106) and non-fermenters (n = 99), including 69 metallo-β-lactamase-producing Enterobacterales, were tested by DD, gradient diffusion and reference BMD in iron-depleted cation-adjusted Mueller–Hinton broth, applying EUCAST breakpoints where available. In this collection, 24.5% of Enterobacterales were resistant to FDC. The essential agreement between gradient diffusion and BMD was 58.5% overall and remained below the 90% threshold in all larger subgroups. For Enterobacterales, DD yielded 64.2% categorical agreement (CA), driven by major errors in 46.3% of susceptible isolates. Gradient diffusion reached 88.7% CA but missed 8/26 resistant isolates (mean bias—1.37 log2 dilutions). Among non-fermenters without clinical breakpoints, 17.6% (12/68) exhibited an MIC > 2 mg/L. Agreement relative to this cut-off exceeded 90% for both methods, yet DD and gradient diffusion failed to detect 5 and 3 of these 12 isolates, respectively. Error profiles differed by organism group: DD over-called resistance in Enterobacterales but missed non-fermenters above the EUCAST cut-off, while gradient diffusion underestimated Enterobacterales MICs.