In vitro activity of cefiderocol against European Enterobacterales, including isolates resistant to meropenem and recentβ-lactam/β-lactamase inhibitor combinations

ABSTRACT Carbapenem-resistant Enterobacterales represent a major health threat and have few approved therapeutic options. Enterobacterales isolates were collected from hospitalized inpatients from 49 sites in six European countries (1 January–31 December 2020) and underwent susceptibility testing to cefiderocol and β-lactam/β-lactamase inhibitor combinations. Meropenem-resistant (MIC >8 mg/L) and cefiderocol-susceptible isolates were analyzed by PCR, and cefiderocol-resistant isolates by whole-genome sequencing, to identify resistance mechanisms. Overall, 1,909 isolates (including 970 Klebsiella spp., 382 Escherichia coli, and 244 Enterobacter spp.) were collected, commonly from bloodstream infections (43.6%). Cefiderocol susceptibility was higher than approved β-lactam/β-lactamase inhibitor combinations and largely comparable to cefepime-taniborbactam and aztreonam-avibactam against all Enterobacterales (98.1% vs 78.1%–97.4% and 98.7%–99.1%, respectively) and Enterobacterales resistant to meropenem (n = 148, including 125 Klebsiella spp.; 87.8% vs 0%–71.6% and 93.2%–98.6%, respectively), β-lactam/β-lactamase inhibitor combinations (66.7%–92.1% vs 0%–88.1% and 66.7%–97.9%, respectively), and to both meropenem and β-lactam/β-lactamase inhibitor combinations (61.9%–65.9% vs 0%–20.5% and 76.2%–97.7%, respectively). Susceptibilities to approved and developmental β-lactam/β-lactamase inhibitor combinations against cefiderocol-resistant Enterobacterales (n = 37) were 10.8%–56.8% and 78.4%–94.6%, respectively. Most meropenem-resistant Enterobacterales harbored Klebsiella pneumoniae carbapenemase (110/148) genes, although metallo-β-lactamase (35/148) and oxacillinase (OXA) carbapenemase (6/148) genes were less common; cefiderocol susceptibility was retained in β-lactamase producers, other than NDM, AmpC, and non-carbapenemase OXA producers. Most cefiderocol-resistant Enterobacterales had multiple resistance mechanisms, including ≥1 iron uptake-related mutation (37/37), carbapenemase gene (33/37), and ftsI mutation (24/37). The susceptibility to cefiderocol was higher than approved β-lactam/β-lactamase inhibitor combinations against European Enterobacterales, including meropenem- and β-lactam/β-lactamase inhibitor combination-resistant isolates. IMPORTANCE This study collected a notably large number of Enterobacterales isolates from Europe, including meropenem- and β-lactam/β-lactamase inhibitor combination-resistant isolates against which the in vitro activities of cefiderocol and developmental β-lactam/β-lactamase inhibitor combinations were directly compared for the first time. The MIC breakpoint for high-dose meropenem was used to define meropenem resistance, so isolates that would remain meropenem resistant with doses clinically available to patients were included in the data. Susceptibility to cefiderocol, as a single active compound, was high against Enterobacterales and was higher than or comparable to available β-lactam/β-lactamase inhibitor combinations. These results provide insights into the treatment options for infections due to Enterobacterales with resistant phenotypes. Early susceptibility testing of cefiderocol in parallel with β-lactam/β-lactamase inhibitor combinations will allow patients to receive the most appropriate treatment option(s) available in a timely manner. This is particularly important when options are more limited, such as against metallo-β-lactamase-producing Enterobacterales.

Meropenem-vaborbactam and imipenem-relebactam both act against isolates producing Class A and Class C β-lactamases; however, vaborbactam and relebactam are inactive against MBLs and oxacillinases (OXA) (18,23,24,39,40).Cefepime-taniborbactam and aztreonam-avibactam are currently in clinical development and have demonstrated in vitro activity against CREs, including isolates producing MBLs (41)(42)(43), but neither combination is currently approved for use in Europe (at time of manuscript acceptance).Susceptibilities of cefiderocol and approved BLBLI combinations have been compared against Enterobacter ales, including CR and BLBLI combination-resistant isolates, in the longitudinal surveillance studies SENTRY and SIDERO (44,45).Despite this, there remain limited data comparing these antimicrobials with developmental BLBLI combinations against Enterobacterales in Europe.
The aim of this study was to evaluate the in vitro activity of cefiderocol, merope nem, BLBLI combinations (approved and in development), and colistin against clinical Gram-negative isolates collected between 1 January and 31 December 2020 across six countries in Europe.Here, we report the results for Enterobacterales isolates, includ ing Klebsiella spp., Escherichia coli, Enterobacter spp., Citrobacter spp., Proteus spp., Serratia spp., and Morganella morganii.The results for Pseudomonas aeruginosa and Acinetobacter spp.isolates collected in this study are reported elsewhere (46).

Susceptibility profiles of isolates
Susceptibility to cefiderocol using European Committee on Antimicrobial Susceptibil ity Testing (EUCAST) breakpoints was 98.1% for Enterobacterales overall and ranged from 92.2%-100% by species, including 92.2% for Enterobacter spp.and 98.9% for Serratia spp.isolates; all Proteus spp.(n = 71) and M. morganii (n = 49) isolates were cefiderocol susceptible (Table 1).Susceptibility rates using Clinical and Laboratory Standards Institute (CLSI) breakpoints are also reported (Tables S3 and S4).Overall, Enterobacterales susceptibility to cefiderocol across European countries ranged from 96.9% in Italy to 99.0% in the United Kingdom (Table S5A through E).For all Enterobacter ales isolates, susceptibility was >90% to all BLBLI combinations, with the exception of ceftolozane-tazobactam (78.1%) (Table 1).For each drug, activity by species was broadly similar to the overall activity against all Enterobacterales, with Enterobacter spp.showing the most variability (largest variation was observed with ceftolozane-tazobactam: 61.1% against Enterobacter spp.vs 78.1% against all Enterobacterales) (Table 1).For colistin, 95.3% of isolates (using epidemiological cut-off values) could be considered wild type, i.e., those without phenotypically detectable acquired resistance mechanisms (Table 1).
Against cefiderocol-resistant Enterobacterales isolates (n = 37), existing BLBLI combinations had low to moderate activity, while both developmental combinations retained moderate to high activity (Table 2).In colistin-resistant Enterobacterales isolates (excluding Morganellaceae and Serratia spp.), all agents had good activity.All colistinresistant Klebsiella spp.were susceptible to cefiderocol.Susceptibilities to cefiderocol and BLBLI combinations, with the exception of ceftolozane-tazobactam, were compa rable among colistin-resistant Enterobacter spp.Susceptibilities for antibiotics against antibiotic-resistant Enterobacterales using CLSI breakpoints are reported in Table S4.
Most cefiderocol-resistant Enterobacterales [64.9% (24/37)] that were sequenced harbored mutations in all three groups of genes: those related to iron uptake, those encoding porins, and the PBP3-encoding gene ftsI.All isolates had mutations in at least one gene related to iron uptake (cirA-like, fiu-like, exbB, exbD, or tonB); two or more of these genes had mutations or were not detected in 91.9% (34/37) of isolates (Table 5).A high proportion [64.9% (24/37)] also harbored mutations in genes encoding porins and ftsI.A total of 18.9% (7/37) of cefiderocol-resistant Enterobacterales had ompC-like gene mutations or this gene was not detected; 43.2% (16/37) of isolates had ompF-like mutations or this gene was not detected.

DISCUSSION
This study provides additional data on the in vitro susceptibilities of cefiderocol and BLBLI combinations, that have been approved or are in development, against a large collection of European Enterobacterales isolates.Data were collected from a greater number of sites per European country compared with the longitudinal surveillance programs SENTRY and SIDERO, which are more geographically spread (44,45).Susceptibility to cefiderocol was consistently high (>92%) across >1,900 clinical isolates of all Enterobac terales species collected from 49 sites in six countries, in line with previously reported rates in the SENTRY and SIDERO studies (44,47).BLBLI combinations had variable activity depending on species, but activity was generally high, with the exception of ceftolozanetazobactam.
Only a small proportion of tested Enterobacterales isolates exhibited resistance to one or more antimicrobials.Overall, cefiderocol and all tested BLBLI combinations consistently demonstrated activity against meropenem-resistant Enterobacterales [other than ceftolozane-tazobactam, which is not a recommended treatment for meropenemresistant Enterobacterales (34,35)], which largely comprised Klebsiella spp.Of the approved antimicrobials tested, cefiderocol demonstrated the highest activity against meropenem-resistant Enterobacterales (88%), including those with resistance to BLBLI combinations (61.9%-65.9%);high activity of cefiderocol against meropenem-resistant Enterobacterales (47,48), including those also resistant to ceftazidime-avibactam, has been reported previously (9,49).Other approved antimicrobials (ceftazidime-avibactam, ceftolozane-tazobactam, and meropenem-vaborbactam) demonstrated variable activity against meropenem-resistant Enterobacterales (71.6%, 0%, and 70.9%, respectively).The moderate-high activity of ceftazidime-avibactam, which ESCMID guidelines currently recommend for infections caused by carbapenem-resistant Enterobacterales (excluding MBL producers and if active in vitro), is in line with previous data (16,34,50).How ever, none of these approved BLBLI combinations demonstrated even moderate activity           Data were generated by whole-genome sequencing.A gene was considered to have a gross disruption if the coding sequence carried a nonsense mutation, frameshift, indels of >20 codons, or ablation of the canonical stop or start codons without a replacement immediately adjacent and in-frame.Genes were listed to be not found if a BLAST search with the reference gene yielded no hit with E-value <1E-25.C. koseri or Morganellaceae isolates for which no multilocus sequence typing scheme was available are indicated by "N/A." There were insufficient data available to confirm whether a gene in K. aerogenes with high identity to CMY/ACT/LAT was the chromosomal AmpC of K. aerogenes; these genes are indicated as "Unnamed Class C. " Non-β-lactamase genes that were either found to have gross disruptions or mutations or were not found are shown in gray.

c
A curated summary of the data is shown. d Gene was assumed to be non-functional and was not described as a gene in the results.
e Indicates a novel ST.
With the exception of aztreonam-avibactam and cefepime-taniborbactam, the activity of BLBLI combinations vs cefiderocol-resistant isolates was low; data on BLBLI combinations suggest they do not always demonstrate in vitro activity against a wide range of mechanisms of resistance, including certain β-lactamases (37,39,40).
All cefiderocol-resistant Enterobacterales isolates had mutations in at least one gene related to iron uptake.Mutations in genes encoding the siderophore receptors CirA and/or Fiu have previously been reported as key mechanisms of cefiderocol resistance in Enterobacterales (E.cloacae and K. pneumoniae), particularly in NDM-producing isolates (28,(61)(62)(63)(64)(65)(66)(67)(68)(69).Specifically, CirA dysfunction has been linked to impaired transport of cefiderocol by iron transport and decreased intracellular concentration in Enterobacterales (61,62); although interestingly, this has also been attributed to loss of fitness in a cefiderocol-resistant strain of NDM-producing K. pneumoniae (63).Mutations in the cirA-like gene were commonly observed in cefiderocol-resistant E. coli (3/3) and K. pneumoniae (8/8), although small proportions (1/3 and 2/8, respectively) were NDM producing.In addition, many cefiderocol-resistant Enterobacterales isolates harbored mutations in the ftsI gene for PBP3, although all K. pneumoniae isolates had wild-type ftsI genes.The relevant contribution of each of these mutations in genes related to iron uptake or PBP3, either alone or in combination with mutations in β-lactamase genes, is not known, nor is any potential impact on the competitive fitness or virulence of these strains.This study was not designed to investigate the mechanisms of resistance in detail, unlike a surveillance study design.Previous data suggest that individual mechanisms of resistance are not sufficient to confer a clinically relevant level of resistance, but that combinations of mechanisms can reduce cefiderocol susceptibility (59); these are consistent with the present study, where most cefiderocol-resistant isolates had multiple mechanisms of resistance.Further research on the effect of different combinations of resistance mechanisms is warranted.
The results from this study provide insights into the suitability of available therapeutic options for infections due to Enterobacterales with resistant phenotypes.Cefiderocol demonstrated high levels of in vitro activity against Enterobacterales (including those resistant to meropenem and BLBLI combinations, where approved BLBLI combina tions did not demonstrate activity) and isolates producing MBLs, which are the most therapeutically challenging carbapenemases to treat in meropenem-resistant Enterobac terales.Therefore, susceptibility testing of cefiderocol at the same time as existing BLBLI combinations would allow clinicians to appropriately choose from all available treatment options.
Furthermore, while colistin is an available treatment option, EUCAST does not currently recommend colistin as a monotherapy, and there is no clinical monotherapy breakpoint (70); therefore, data on susceptibilities to colistin are not reported in full in our study.Additionally, colistin treatment is often associated with high rates of nephrotoxicity (71, 72) and poor tissue penetration, particularly in the lungs (72,73), so it is less relevant than other agents for this isolate set, which included a high proportion of isolates from the respiratory tract.
There are several limitations to this study.As Enterobacterales isolates were only collected from six European countries, with up to 40 isolates per participating site, there were low numbers of some Enterobacterales with resistant phenotypes, particularly cefiderocol-resistant isolates; this precluded a clear analysis of potential mechanisms of cefiderocol resistance.Although this may be a result of the rarity of cefiderocol resistance in European Enterobacterales, this methodology was not designed to represent a surveillance study.Different methodologies and screening panels were used to analyze a finite selection of potential mechanisms of resistance in meropenem-and cefiderocol-resistant isolates.Although EUCAST recommends characterizing carbapenemases in meropenem-resistant Enterobacterales defined by a meropenem MIC >0.125 mg/L (vs MIC >8 mg/L used in this study) (74), this study characterized clinical isolates according to treatment decisions in the clinic, so potential mechanisms of resistance were only investigated in the set of meropenem-resistant isolates that would have been considered resistant in patients.This study did not characterize the expression levels of genes potentially conferring resistance, mutations in clinically relevant intrinsic β-lactamase genes, or potential mechanisms of resistance in meropenem-resistant Enterobacterales which were resistant to BLBLI combinations but susceptible to cefiderocol.Therefore, robust interpretations of the resistance mechanisms and the presence or absence of cross-resistance in isolates could not be made.Only the STs of cefiderocol-resistant isolates were identified, so clonal expansion of isolates with other phenotypes was not determined; however, these non-surveillance data would not have accurately reflected clonal epidemiology in Europe.Finally, it is important to note that in vitro data cannot replace clinical studies in patients, and in vitro activity may not reflect in vivo efficacy of a therapy in clinical practice.

Conclusions
This study builds on existing literature showing that cefiderocol remains the most broadly active approved β-lactam against a large proportion and range of Enterobac terales pathogens in Europe, including those resistant to meropenem or both merope nem and clinically relevant BLBLI combinations, where treatment options are limited.Cefiderocol should be tested in parallel with BLBLI combinations to ensure patients with infection due to CRE receive early treatment with the most appropriate treatment option.

Clinical isolates
Between 1 January and 31 December 2020, Gram-negative clinical isolates from hospitalized inpatients were collected at 49 sites across Austria, France, Germany, Italy, Spain, and the UK (see Table S8 for details of participating centers).Each site was requested to collect 20 Klebsiella spp.and 20 other Enterobacterales [methods and results for P. aeruginosa and Acinetobacter spp.isolates are reported elsewhere (46)].Isolates included those from all infection sources, with the exception of the urinary tract.Only one isolate of the same genus and species was allowed per patient.Matrix-assis ted laser desorption/ionization-time of flight mass spectrometry was used for species identification at IHMA Europe Sàrl (Monthey, Switzerland).
International Organization for Standardization 20776-1 susceptibility testing standards and EUCAST guidance were followed for the preparation of antimicrobials for testing, with the exception of ceftazidime-avibactam, and MIC determinations (75,76); tryptic soy agar plates containing 5% sheep blood were sourced from Liofilchem (Roseto degli Abruzzi, Italy; product code: 11037), cation-adjusted Mueller-Hinton broth (CAMHB) was sourced from Becton Dickinson (Franklin Lakes, NJ, USA; product code: 212322), and iron-depleted CAMHB (used for cefiderocol testing) was prepared by IHMA Europe Sàrl.Sensititre freeze-dried panels (Thermo Fisher Scientific Inc., Waltham, MA, USA) were used in the preparation of ceftazidime-avibactam for testing, as MIC values were only available for this agent when this validated commercial method was used.All antibiotics were tested daily using the quality control strains K. pneumoniae ATCC 700603, K. pneumoniae ATCC BAA-2814, and E. coli ATCC 25922 (recommended by EUCAST) (77).The MIC values for each tested antibiotic were manually read as the lowest concentration inhibiting visible growth.
For cefiderocol and meropenem, MIC values were determined more than once; a third MIC determination was carried out if MIC values differed by >1 dilution, when the geometric mean was reported.

Analysis
Antimicrobial susceptibility results were interpreted in accordance with EUCAST clinical breakpoints (v.14.0, 2024) (78) (see Table S3A).Analyses in accordance with CLSI breakpoints (2023) (79) were also conducted and are presented in the supplementary material.Meropenem resistance was defined using a breakpoint of MIC >8 mg/L, relating to high-dose, extended-infusion (2 g, 3-h infusion) meropenem; similarly, isolates with a meropenem MIC >8 mg/L when tested with a fixed vaborbactam concentration of 8 mg/L were considered resistant to meropenem-vaborbactam.Aztreonam-avibactam and cefepime-taniborbactam do not currently have approved EUCAST MIC breakpoints; breakpoints for aztreonam and cefepime alone were used (Table S3A) (78).

Identification of β-lactamase genes in meropenem-resistant isolates
Isolates with a meropenem MIC >8 mg/L and a cefiderocol MIC ≤2 mg/L were analyzed by PCR (performed by IHMA Europe Sàrl) to identify the presence of β-lactamase genes that may confer meropenem resistance (see Table S10 for genes and primers used).Data on β-lactamase genes in isolates that were meropenem resistant (MIC >8 mg/L) and cefiderocol resistant (MIC >2 mg/L) were generated by WGS (see below).
DNA extraction was performed from a single colony obtained from a fresh tryptic soy blood agar culture for each isolate, using the QIAGEN TissueLyser II instrument (Hilden, Germany) as per the manufacturer's instructions.All isolates then underwent PCR amplification and sequencing to screen for the presence of genes encoding clinically relevant β-lactamases: extended-spectrum β-lactamases (ESBLs: bla SHV , bla TEM , bla CTX-M , bla VEB , bla PER , and bla GES ), AmpCs (bla ACC , bla CMY I/MOX , bla CMY II , bla DHA , bla FOX , and bla ACT-MIR ), and carbapenemases (bla KPC , bla OXA , bla NDM , bla IMP , bla VIM , bla SPM , bla GIM , and bla GES ).Amplicons were sequenced by Fasteris (Geneva, Switzerland) and then analyzed using SeqScape Software 3 (Thermo Fisher Scientific Inc.).Limited sequencing was used to screen bla TEM and bla SHV to identify TEM-type and SHV-type enzymes containing amino acid substitutions common to ESBLs (bla TEM : amino acids 104, 164, 238, and 240; bla SHV : amino acids 146, 179, 238, and 240), and to screen bla CTX-M (groups 1, 2, 8, 9, and 25) to identify CTX-M-type enzymes containing the D240G amino acid substitution associated with elevated ceftazidime MICs.Genes encoding SHV-type and TEM-type enzymes were reported as ESBL or original-spectrum β-lactamase genes.The 16S ribosomal DNA in all isolates was also amplified by PCR and sequenced for bacterial identification.

Identification of β-lactamase genes and other potential resistance mecha nisms in cefiderocol-resistant isolates
Isolates with a cefiderocol MIC >2 mg/L were analyzed by WGS to identify possible mechanisms of resistance.DNA isolation was performed using the QIAGEN QIAamp DNA Mini kit, and library preparation was performed using the Illumina DNA Prep kit (San Diego, CA, USA) at International Health Management Associates Inc. (Schaumburg, IL, USA).Libraries were then shipped to Azenta (South Plainfield, NJ, USA), where short-read WGS (2 × 150 base pairs; paired-end) was performed on an Illumina HiSeq platform to a 100× depth of coverage.Quality control was performed using the CheckM lineage workflow (80)(81)(82) to ensure low contamination (≤5%) and completeness of assemblies (≥95%) were achieved.Multilocus sequence typing was used to determine the related ness of isolates; the Achtmann scheme was used for analyzing E. coli.
Genomic assemblies were created using the QIAGEN CLC Genomics workbench (v.21.0.5).In order to identify β-lactamase genes of interest, assemblies were queried using the ResFinder database (83) with coverage and identity thresholds of ≥35% and ≥72%, respectively.Genes identified with <100% identity or coverage were evaluated for a variant by pairwise alignment to a reference sequence using the ResFinder database (83).Variants were defined using the Bacterial Antimicrobial Resistance Reference Gene Database from the National Center for Biotechnology Information (BioProject 313047).
Non-β-lactamase genes of interest included those encoding PBP3 (ftsI), porins [ompC-like and ompF-like (in K. pneumoniae: ompK36 and ompK35, respectively)], and those related to iron acquisition (cirA-like, fiu-like, exbB, exbD, and tonB).Genes were analyzed by pairwise alignment and classified as wild type if they had 100% amino acid sequence identity to the species-specific reference sequence (Table S11).Genes were also screened for gross disruption vs species-specific reference sequences (Table S11) and were considered to have gross disruption if the coding sequence carried a nonsense mutation, frameshift, indels of >20 codons, or ablation of the canonical start or stop codons without a replacement immediately adjacent and in-frame.Genes were not considered disrupted if there were ablated start or stop codons immediately adjacent to intact, in-frame start or stop codons.Genes were listed to be not found if a BLAST search with the reference gene yielded no hit with E-value <1E-25.Reference genes for several siderophore uptake genes in Citrobacter koseri, Klebsiella oxytoca, or Serratia marcescens were not available, so these genes were not analyzed. 382

TABLE 2
In vitro activity of cefiderocol, BLBLI combinations, and other relevant antibiotics against Enterobacterales isolates with resistant phenotypes a,b Results are not reported for isolates tested against antibiotics to which they had an expected resistance phenotype.Susceptibility was assessed according to EUCAST breakpoints (including high dosage breakpoints and breakpoints for the agent without inhibitor, where necessary), except colistin, for which ECOFF values were used to define the proportion of isolates that could be considered wild-type strains, i.e., those without phenotypically detectable acquired resistance mechanisms.Data are shown where n ≥ 20 isolates were available.Data on susceptibility to colistin are shown in parentheses as colistin is not recommended for monotherapy and is not associated with a clinical monotherapy breakpoint (as per EUCAST v.14.0 guidance). b

TABLE 5 β
-Lactamase genes and other potential resistance mechanisms identified in cefiderocol-resistant Enterobacterales isolates (n = 37) a,b

TABLE 5
β-Lactamase genes and other potential resistance mechanisms identified in cefiderocol-resistant Enterobacterales isolates (n = 37) a,b (Continued) (Continued on next page)

TABLE 5 β
-Lactamase genes and other potential resistance mechanisms identified in cefiderocol-resistant Enterobacterales isolates (n = 37) a,b (Continued on next page)