Comparison of cefiderocol in-vitro susceptibility testing modalities

Sir,

Cefiderocol (FDC) is a siderophore cephalosporin shown to have promising clinical outcomes against multidrug-resistant (MDR) Acinetobacter baumannii (AB) and Pseudomonas aeruginosa (PA) in a clinical study [1]. However, significant challenges exist in susceptibility testing as reliable, reproducible results require the use of iron-depleted cation-adjusted Mueller-Hinton broth (ID-CAMHB) [2] Unfortunately, ID-CAMHB requires advanced customized ordering and in-house preparation is too labour-intensive for clinical microbiology laboratories to undertake. Currently, FDC susceptibility is not performed onsite in our institution, rather samples are sent to a third-party laboratory. In the best-case scenario, clinicians will have FDC susceptibilities approximately 5 days from sampling. Delays in obtaining susceptibility results may prolong the time before initiating appropriate antibiotics and adversely affect patients. There are several ready-to-use FDC susceptibility testing methods available, but low reliability on some products have been reported, particularly for non-lactose fermenting Gram-negative organisms [3,4]. To guide local antimicrobial stewardship efforts, we undertook a pilot study to compare different quantitative FDC susceptibility testing methods.

Nineteen clinical isolates (AB, n = 10 and PA, n = 9) resistant to ≥3 antibiotic classes (including carbapenems) were obtained between December 2021 and November 2022 from 17 unique patients in our hospital. These 19 isolates were retrospectively selected based on whether the patient received FDC therapy and/or if the isolate demonstrated FDC non-susceptibility. The clonality of the isolates was determined using Fourier-transform infrared spectroscopy (Bruker IR Biotyper^®), which was reported to be a reproducible and discriminatory screening tool of clinical isolates compared with whole genome sequencing [5]. Cefiderocol minimum inhibitory concentrations (MICs) were determined by (1) gradient strips (FDC range 0.016–256 μg/mL; Liofilchem^®), (2) broth microdilution (BMD) panel (FDC range 0.008–128 μg/mL; ComASP^®, Liofilchem^®), and (3) BMD in standard MHB (FDC range 0.06–32 μg/mL; Becton Dickinson). These findings were compared to those obtained by the reference method in ID-CAMHB (FDC range 0.06–32 μg/mL). All susceptibilities were ascertained in duplicates and the reference method results were confirmed by two independent investigators (N.S.T. and L.V.) on different days. A standard laboratory reference (e.g., PA ATCC 27853) was used as the control. Breakpoints established by the Clinical and Laboratory Standards Institute were applied. In addition to categorical correlations, comparisons were analyzed utilizing linear regression of log-2 transformed MICs. With the best-performing method, 27 additional non-MDR PA isolates (recovered from unique patients in 2023) were tested to strengthen our conclusion.

Over 80% of MDR isolates were clonally unique. A series of 3 PA isolates were sequentially obtained from a single patient while on FDC therapy (Appendix 1). Based on the reference method, only one isolate was FDC-resistant. The remainder of the isolates were distributed within the susceptible category (FDC MIC range 0.06–4 μg/mL). Categorically, BMD in standard MHB and the gradient strip methods agreed with the reference method in 80–90% of PA isolates. Notably, susceptibilities obtained by BMD in standard MHB, and the gradient strip methods overestimated three isolates (two and one, respectively) as intermediate. The BMD panel incorrectly estimated the single resistant isolate as susceptible to FDC. Investigation into errors revealed 2 minor, 1 minor, and 1 very major error for PA isolates tested by BMD in standard MHB, gradient strip, and BMD panel, respectively. Categorically, all methods agreed with the reference when testing the AB isolates without minor/major errors.

When assessing FDC MICs of the MDR PA isolates quantitatively, the gradient strips demonstrated a reasonable, positive correlation when compared with the reference method (R² = 0.85, Fig. 1b), while the alternative methods demonstrated weaker correlations (R² < 0.5). Furthermore, the gradient strips categorically agreed with the reference method in all 27 additional isolates (MICs were within 4x difference of the reference method in 24/27 isolates). For the AB isolates, despite demonstrating categorical agreement across all modalities, the testing methods exhibited poor correlations with the reference method (R² < 0.3) resulting from low accuracy in determining the MICs seen with the reference.

Fig. 1.

Comparison of cefiderocol MICs by different methods. Comparisons of alternative cefiderocol susceptibility testing modalities (a) broth microdilution panel, (b) gradient strips, (c) standard MHB for Pseudomonas aeruginosa; (d) broth microdilution panel, (e) gradient strips, (f) standard MHB for Acinetobacter baumannii clinical isolates. All MICs were log-2 normalized. The dotted red lines represent the lines of identity. Each coloured line represents the line-of-best-fit with the equation and R² correlation for each comparison. ID-MHB, iron-depleted cation-adjusted Mueller Hinton broth; MHB, Mueller Hinton broth; BMD, broth microdilution; MIC, minimum inhibitory concentration. ^∗Gradient strip MICs were rounded to next highest MIC, if necessary, to conform to 2-fold dilutions in other methods.

When utilizing FDC, clinicians often empirically treat serious Gram-negative infections without the knowledge of in-vitro susceptibility. Our results suggest that categoric susceptibility was reasonably well estimated by the BMD in standard MHB and gradient strip methods. For determining MICs, the gradient strip method may be an effective alternative to the reference standard when assessing PA susceptibility to FDC. As a result, implementation of the gradient strip may reduce send-out testing and the time until FDC susceptibilities are known. We believe that this finding may be beneficial for clinicians taking care of our patients. However, it is important to note that currently available alternatives have demonstrated low accuracy and their use has been discouraged [4]. The search for an expeditious and accurate FDC susceptibility testing methods remains a priority. Additional studies with larger samples are warranted to confirm our findings.