Susceptibility Testing Is Key for the Success of Cefiderocol Treatment: A Retrospective Cohort Study

Cefiderocol is a novel siderophore cephalosporin, which has proven in vitro activity against carbapenem-resistant (CR) Gram-negative pathogens and stability towards all carbapenemases. The aim of this study was to describe the first cases of prescriptions and the efficacy of cefiderocol for compassionate use in the 2 months following its access in France. We performed a national retrospective study of all patients who received at least one dose of cefiderocol from 2 November 2018 to 5 November 2019. We collected clinical characteristics and outcome through a standard questionnaire. Bacterial isolates from 12 patients were centralized and analyzed in the French National Reference Center for Antimicrobial Resistance, and sequenced using Illumina technology. Finally, 13 patients from 7 French university hospitals were included in the study. The main type of infection treated by cefiderocol was respiratory tract infections (RTI, n = 10). The targeted bacteria were Pseudomonas aeruginosa (n = 12), including carbapenemase-producing P. aeruginosa (n = 9), Acinetobacter baumannii (n = 2), Klebsiella pneumoniae (n = 1), and Enterobacter hormaechei (n = 1). Overall, of the 12 patients whose samples were analyzed, 5 P. aeruginosa strains were not susceptible to cefiderocol (4 categorized as resistant and 1 as intermediate) according to Clinical and Laboratory Standards Institute (CLSI) breakpoints. If considering susceptible strains, the cure rate was 6/7, while being 0/5 among not-susceptible strains. This study underlines the necessity to test strains in adequate conditions.


Introduction
Cefiderocol is a novel siderophore cephalosporin, with in vitro activity against multidrug-resistant (MDR) Gram-negative pathogens, and stability towards all carbapenemases, including metallo-β-lactamases (MBLs) [1]. However, few data on real-life use and the clinical efficacy of cefiderocol are available.
The objective of this study was to describe the efficacy of cefiderocol as compassionate use during the early access program in France.

Materials and Methods
We performed a national retrospective study of all adult patients who received at least one dose of cefiderocol from 2 November 2018 to 5 November 2019. Standardized questionnaires were sent to the prescribers to collect patients' baseline characteristics, infections type and management, microbiological data, reasons for cefiderocol use, doses and duration of cefiderocol treatment, concomitant antibiotic treatments, adverse events, and outcome. The research was conducted in accordance with the Declaration of Helsinki, and national and institutional standards. Patients were informed that cefiderocol was provided within a compassionate use program, and that their clinical data could be used, after anonymization, for research purposes.
Immunosuppression was defined as asplenia, neutropenia, agammaglobulinemia, organ transplant, hematologic malignancies, HIV infection with CD4 cells count <400/mm 3 , or end-stage liver disease. Immunosuppressive treatment was defined as systemic corticosteroids with a daily dose >20 mg of prednisolone equivalent during at least 3 weeks, cancer chemotherapy, or other immunosuppressive drugs (e.g., cyclophosphamide, azathioprine, cyclosporine, etc.).
Outcome was evaluated by the investigators at the patient's most recent visit after completion of cefiderocol treatment. Primary criteria was cure, defined as survival with no residual sign of infection, and pathogen eradication.
All bacterial isolates were centralized at the French National Reference Center for Antimicrobial Resistance (Bicêtre University Hospital, Le Kremlin-Bicêtre, France), where strain analyzing and drug susceptibility testing were performed by disk diffusion, and minimum inhibitory concentrations (MICs) were determined by broth microdilution (BMD) per the Clinical and Laboratory Standards Institute (CLSI) and European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines [2]. Cefiderocol MICs were also determined by BMD using iron-depleted and cation-adjusted Mueller-Hinton broth (Ther-moFisher, Waltham, MA, USA) as recommended by CLSI guidelines [3]. The breakpoints used were those defined by the CLSI, 2019 update. All bacterial isolates were sequenced using Illumina technology as previously described [4]. Resfinder server v3.2 (https://cge.cbs. dtu.dk/services/ResFinder/) and CARD database (https://card.mcmaster.ca) were used to identify the antimicrobial resistance genes. The MLST (MultiLocus Sequence Typing) was performed using the MLST 2.0 server (http://www.genomicepidemiology.org/).
Quantitative variables are presented as mean ± standard deviation (SD), or median and interquartile range (IQR). Qualitative variables are presented as number of occurrences and relative frequencies. All statistical analyses were performed using SPSS Statistics v.17.0 software (SPSS Inc., Chicago, IL, USA).
Among the 10 patients enrolled in the intensive care unit (ICU), the mean ± SD sequential organ failure assessment (SOFA) score was 6.9 ± 4.1, and 4 (30.8%) developed septic shock.
The most common cefiderocol regimen was 2 g tid (n = 6), while 3 patients with acute renal failure required a regimen of 750 mg bid (n = 1), or tid (n = 2). Cefiderocol was administered as 3-or 4-h infusions in 12 patients, and 1-h infusion in one patient.
Only 2 severe adverse events were reported. One patient suffered from renal failure, which resolved while cefiderocol treatment was not discontinued. One patient developed thrombocytopenia, which resolved after cefiderocol treatment discontinuation.
The most common pathogens targeted were Pseudomonas aeruginosa (n = 12), including 9 carbapenemase-producers, Acinetobacter baumannii (n = 2), Klebsiella pneumoniae (n = 1), and Enterobacter hormaechei (n = 1). Results of antimicrobial susceptibility testing, MLST typing, and antibacterial resistance genes at the National Reference Center for Antimicrobial Resistance are presented in Table 1, except for one patient with a P. aeruginosa strain (bacterial samples no longer available). Although 5 isolates targeted were finally reclassified as non-susceptible to cefiderocol (4 categorized  Analysis of the resistomes of these isolates indicated a variety of β-lactamase encoding genes. Among acquired carbapenemases identified, the bla VIM-2 (n = 3), bla VIM-4 (n = 1) and bla NDM-1 (n = 1) genes were identified in P. aeruginosa strains. The blaOXA-48-gene was identified in one isolate of K. pneumoniae, whereas the bla OXA-23 gene was identified in the two isolates of A. baumannii. In addition to these acquired carbapenemases, these isolates also produced their natural β-lactamases, bla PDC-, bla ADC-and bla ACT -like corresponding to class C β-lactamase identified in P. aeruginosa, A. baumannii and Enterobacter cloacae complex, respectively. However, no direct correlation was observed between resistance to cefiderocol and acquired carbapenemases, and likely resulted from the addition of different mechanisms beyond the production of a carbapenemase.

Discussion
Our study describes the early use of cefiderocol as salvage treatment against MDR Gram-negative pathogens, during the early access compassionate use program in France. Primary indication was respiratory tract infection due to carbapenemase-producing P. aeruginosa. Cure rate was 87.5% with cefiderocol-susceptible isolates confirmed per CLSI guidelines at the National Reference Center, vs. 0% with cefiderocol-resistant or -intermediate isolates (p < 0.01).
Carbapenem resistance in Gram-negative bacteria is a major threat worldwide [5], with increased mortality, especially among vulnerable patients [6]. Despite the recent development of various combinations of beta-lactams with beta-lactamase inhibitors, our armentarium against carbapenemase-resistant non-fermenting Gram-negative bacteria remains limited [7]. Carbapenem resistance mechanisms include both transmissible (e.g., carbapenemase production), and intrinsic resistance (e.g., porin loss and efflux pump), as in our study [5,8]. The prevalence of carbapenem-resistant non-fermenters now surpass that of Enterobacterales in many settings, representing a challenge for the management of most severe infections [7].
Previous studies showed a high rate of in vitro susceptibility to cefiderocol, prominently in P. aeruginosa, A. baumannii, and S. maltophilia, as compared with other broadspectrum antibiotics [1,9].
Furthermore, several multi-national surveillance studies demonstrated that cefiderocol exerts in vitro efficacy against 97% of carbapenem-resistant Enterobacterales, and more specifically against 91.1% of ceftazidime-avibactam resistant strains, using a resistance cefiderocol MIC breakpoint of 4 mg/L [9]. Using the same breakpoint, cefiderocol was also active in vitro against most clinical isolates of MDR P. aeruginosa and A. baumannii (99.2% and 89.7%, respectively).
However, these promising in vitro data were not confirmed in our study, with 5/16 strains not susceptible to cefiderocol at baseline, when analyzed at the National Reference Center. Those 5 strains were also resistant to all new beta-lactam/beta-lactam inhibitors combinations. It suggests that primary resistance to cefiderocol among MDR P. aeruginosa might not be rare. Indeed, elevated cefiderocol MICs have been reported for S. maltophilia and P. aeruginosa [10].
Therefore, it is of paramount importance to appropriately test cefiderocol susceptibility before prescription, especially when P. aeruginosa is involved. MIC determination must be conducted with iron-depleted, cation-adjusted Mueller-Hinton broths (as performed within the National Reference Center in our study) [1]. Disc diffusion might be a suitable alternative despite this method performing poorly with A. baumannii [11]. The provisional CLSI breakpoint of 4 mg/L seemed sufficient in our small study, but more data are needed, especially since CLSI, the Food and Drug Administration, and EUCAST breakpoints are different, leading to considerable variability when interpreting cefiderocol antimicrobial susceptibility testing [10,11]. Indeed, in our study, we used the CLSI breakpoints, which seem in line with clinical outcome. On the contrary, the use of EUCAST breakpoints, which is of 2 mg/L, would lead to a discrepancy between microbiological and clinical outcome regarding one patient (n • 4), infected with a P. aeruginosa strain which would be classified as resistant to cefiderocol, but had a favorable clinical outcome. On the other hand, patient n • 7, who was infected with a P. aeruginosa strain susceptible to cefiderocol according to CLSI breakpoints, presented with clinical failure. This could be explained by the patient's numerous comorbidities (including immunosuppressive treatments). Our study is, to the best of our knowledge, the first to provide exhaustive characterization of antimicrobial resistance genes among strains targeted with cefiderocol. Future studies are warranted to identify the relation with cefiderocol resistance.
Nonetheless, real-life clinical data regarding cefiderocol use are scarce. In a recent study, Falcone et al. described 10 critically-ill patients with either bacteremia or ventilatorassociated pneumonia caused by carbapenem-resistant A. baumannii, S. maltophilia, or NDM-producing K. pneumoniae treated with cefiderocol. All strains had baseline cefiderocol MIC < 2 mg/L. Cure rate and survival rate at 30 days were 70% and 90%, respectively [12].
These results are in line with our study, if we only consider patients with baseline isolates susceptible to cefiderocol according to CLSI guidelines, with a 6/7 (87.5%) cure rate. Results from the CREDIBLE trial suggest a lower clinical efficacy, around 50%, in a heterogeneous population of patients with various infections caused by carbapenemresistant gram-negative bacteria [13].
The main limits to our study are the small number of patients and the use of cefiderocol only as salvage therapy.

Conclusions
Early use of cefiderocol in France primarily targeted respiratory tract infections due to carbapenemase-producing P. aeruginosa. Cure rate was 87.5% with cefiderocol-susceptible isolates confirmed per CLSI guidelines at the National Reference Center, vs. 0% with cefiderocol-resistant or -intermediate isolates.