In Vitro Activity of the Siderophore Cephalosporin, Cefiderocol, against Carbapenem-Nonsusceptible and Multidrug-Resistant Isolates of Gram-Negative Bacilli Collected Worldwide in 2014 to 2016

ABSTRACT The in vitro activity of the investigational siderophore cephalosporin, cefiderocol (formerly S-649266), was determined against a 2014–2016, 52-country, worldwide collection of clinical isolates of carbapenem-nonsusceptible Enterobacteriaceae (n = 1,022), multidrug-resistant (MDR) Acinetobacter baumannii (n = 368), MDR Pseudomonas aeruginosa (n = 262), Stenotrophomonas maltophilia (n = 217), and Burkholderia cepacia (n = 4) using the Clinical and Laboratory Standards Institute (CLSI) standard broth microdilution method. Iron-depleted cation-adjusted Mueller-Hinton broth (ID-CAMHB), prepared according to a recently approved (2017), but not yet published, CLSI protocol, was used to test cefiderocol; all other antimicrobial agents were tested using CAMHB. The concentration of cefiderocol inhibiting 90% (MIC90) of isolates of carbapenem-nonsusceptible Enterobacteriaceae was 4 μg/ml; cefiderocol MICs ranged from 0.004 to 32 μg/ml, and 97.0% (991/1,022) of isolates demonstrated cefiderocol MICs of ≤4 μg/ml. The MIC90s for cefiderocol for MDR A. baumannii, MDR P. aeruginosa, and S. maltophilia were 8, 1, and 0.25 μg/ml, respectively, with 89.7% (330/368), 99.2% (260/262), and 100% (217/217) of isolates demonstrating cefiderocol MICs of ≤4 μg/ml. Cefiderocol MICs for B. cepacia ranged from 0.004 to 8 μg/ml. We conclude that cefiderocol demonstrated potent in vitro activity against a 2014–2016, worldwide collection of clinical isolates of carbapenem-nonsusceptible Enterobacteriaceae, MDR A. baumannii, MDR P. aeruginosa, S. maltophilia, and B. cepacia isolates as 96.2% of all (1,801/1,873) isolates tested had cefiderocol MICs of ≤4 μg/ml.

Enterobacteriaceae were 1 and 4 g/ml, respectively. The cumulative percentage of isolates of carbapenem-nonsusceptible Enterobacteriaceae inhibited at various MICs of each agent tested is shown in Fig. 1. Enterobacter spp. demonstrated higher MIC 50 (2 g/ml) and MIC 90 (8 g/ml) values than the other genera/species of Enterobacteriaceae tested. The MIC range for cefiderocol for carbapenem-nonsusceptible Enterobacteriaceae was 0.004 to 32 g/ml, with 97.0% (991/1,022) of isolates having cefiderocol MICs of Յ4 g/ml (Fig. 2). The MIC 50 and MIC 90 values for isolates with concurrent carbapenem-nonsusceptible and ceftolozane-tazobactam-nonsusceptible phenotypes (n ϭ 1,005) were 1 and 4 g/ml, respectively, and for carbapenem-nonsusceptible and ceftazidime-avibactam-nonsusceptible phenotypes (n ϭ 235), they were 2 and 4 g/ml, respectively (  Fig. 3. The MIC 50 and MIC 90 values for cefiderocol against MDR P. aeruginosa were 0.25 and 1 g/ml, respectively (Table 3). Ceftazidime-avibactam, ceftolozane-tazobactam, and cefepime each demonstrated MIC 90 values of Ͼ64 g/ml against the same set of isolates of MDR P. aeruginosa. The MIC 50 and MIC 90 values for cefiderocol tested against isolates of P. aeruginosa with concurrent MDR and ceftolozane-tazobactam-nonsusceptible  phenotypes (n ϭ 199) and MDR and ceftazidime-avibactam-nonsusceptible phenotypes (n ϭ 167) were 0.25 and 2 g/ml, respectively, for both sets of isolates (Table 4). A total of 260 (99.2%) isolates of MDR P. aeruginosa exhibited cefiderocol MIC values of Յ4 g/ml, including 99.0% of ceftolozane-tazobactam-nonsusceptible isolates and 98.8% of ceftazidime-avibactam-nonsusceptible isolates. Cefiderocol MIC distributions for all MDR P. aeruginosa, ceftolozane-tazobactam-nonsusceptible MDR P. aeruginosa, and ceftazidime-avibactam-nonsusceptible MDR P. aeruginosa isolates were very similar (Fig. 4). The cumulative percentage of isolates of MDR P. aeruginosa inhibited at various MICs of each agent tested is shown in Fig. 5. The MIC 50 and MIC 90 values for cefiderocol against S. maltophilia were 0.06 and 0.25 g/ml, respectively. All S. maltophilia isolates tested had cefiderocol MIC values of Յ2 g/ml. The MIC 90 s for cefepime, ceftazidime-avibactam, ceftolozane-tazobactam, and meropenem were Ն64 g/ml, and they were Ͼ8 g/ml for colistin and ciprofloxacin. There are no published CLSI MIC breakpoints for S. maltophilia for any of the antimicrobial agents tested in this study. The cumulative percentage of isolates of S. maltophilia inhibited at various MICs of each agent tested is shown in Fig. 6.
The MICs of cefiderocol for the four isolates of B. cepacia tested in this study were 0.004, 0.008, 0.015, and 8 g/ml. Too few isolates were collected to generate MIC 50 and MIC 90 values.
If the entire data set is considered (n ϭ 1,873 isolates) and if species intrinsically resistant to colistin (B. cepacia and Serratia spp.) and species for which colistin MIC breakpoints are not available (S. maltophilia) are excluded (n ϭ 260), colistin nonsusceptibility was observed for 13.0% (209/1,613) of isolates tested (173 Klebsiella sp.

DISCUSSION
The current study demonstrated cefiderocol to be a more potent antimicrobial agent in vitro than cefepime, ceftazidime-avibactam, ceftolozane-tazobactam, ciprofloxacin, and colistin against a recent worldwide collection of clinical isolates of carbapenem-nonsusceptible Enterobacteriaceae, MDR A. baumannii, MDR P. aeruginosa, S. maltophilia, and B. cepacia (Tables 1 and 3). Cefiderocol maintained its potency against isolates of Gram-negative bacilli resistant to colistin and the ␤-lactam/␤-  A limited number of previous studies have determined the in vitro activity of cefiderocol against surveillance study isolates of Gram-negative bacilli as well as against Gram-negative bacilli with molecularly characterized ESBLs and carbapenemases and isolates resistant to carbapenems by mechanisms other than carbapenemases (13)(14)(15)(25)(26)(27)(28)(29). In a recent study, Falagas and colleagues tested cefiderocol by broth microdilution, using ID-CAMHB prepared according to the approved CLSI protocol, against a collection of 471 carbapenem-resistant isolates of Enterobacteriaceae, P. aeruginosa, and A. baumannii collected from inpatients in Greek hospitals (26). They reported MIC 90 s for cefiderocol ranging from 0.5 to 1 g/ml for individual species of Enterobacteriaceae and MIC 90 s of 0.5 g/ml for both P. aeruginosa and A. baumannii. In another recent study, in which cefiderocol was also tested by the broth microdilution method using ID-CAMHB prepared according to the approved CLSI protocol, a 2014 -2015 collection of clinical isolates of Gram-negative bacilli from North America and Europe was tested (27). In that study, MICs of cefiderocol were Յ4 g/ml for 99.9% of all Enterobacteriaceae (MIC 90 , 0.5 to 1 g/ml), for 97.0% of meropenem-nonsusceptible (MIC, Ն2 g/ml) Enterobacteriaceae (MIC 90 , 1 to 4 g/ml), for 99.9% of all P. aeruginosa isolates (MIC 90 , 0.5 g/ml), for 100% (353/353) of meropenem-nonsusceptible (MIC, Ն4 g/ml) P. aeruginosa isolates (MIC 90 , 0.5 g/ml), for 97.6% of all A. baumannii isolates (MIC 90 , 1 g/ml), for 96.9% of meropenem-nonsusceptible (MIC, Ն4 g/ml) A. baumannii isolates (MIC 90 , 1 g/ml), for 100% of isolates of S. maltophilia (MIC 90 , 0.25 to 0.5 g/ml), and for 93.8% of B. cepacia isolates (27). Slight geographic differences in susceptibilities to cefiderocol were identified in this study, with isolates from Europe demonstrating cefiderocol MIC 90 s that were one doubling dilution higher than those of isolates from North America for all Enterobacteriaceae (1 versus 0.5 g/ml) and meropenem-nonsusceptible P. aeruginosa (1 versus 0.5 1 g/ml) and two doubling dilutions higher for meropenem-nonsusceptible Enterobacteriaceae (4 versus 1 g/ml) (27). The cefiderocol MIC 90 was one doubling dilution higher for North American isolates of S. maltophilia than for isolates from Europe (0.5 versus 0.25 g/ml) (27).
Other studies have reported the in vitro activity of cefiderocol against isolates of Gram-negative bacilli harboring molecularly defined mechanisms of resistance. One of these studies reported that isolates of Enterobacteriaceae harboring ESBLs (e.g., CTXtype, SHV-type, and TEM-type), KPC-type carbapenemases, VIM-type and IMP-type carbapenemases, and OXA-type carbapenemases all had cefiderocol MICs of Յ4 g/ml as did 90% (44/49) of NDM-1-positive isolates (13). However, more recent studies from the United Kingdom (28) and France (29) using broth medium prepared with apotransferrin, a less potent and reliable iron chelator than Chelex, have reported that cefiderocol was less active than reported by Kohira et al. (13) against some isolates of Gram-negative bacilli carrying NDM-positive, OXA-type, and KPC carbapenemases. Another study reported MICs of cefiderocol of Յ2 g/ml for clinical isolates of P. aeruginosa (n ϭ 33) positive for GIM-1, IMP-type, or SPM-1 carbapenemases and that 87.5% (14/16) of VIM-positive isolates of P. aeruginosa had cefiderocol MICs of Յ4 g/ml (15). The same study also reported that IMP-1-, OXA-51-, and OXA-58-positive isolates of A. baumannii (n ϭ 29) had cefiderocol MICs of Յ4 g/ml while selected isolates harboring OXA-23 or OXA-24 were less susceptible to cefiderocol (15). Yamano et al. reported that no significant changes in the in vitro activity of cefiderocol were observed in isolates of P. aeruginosa with OprD deficiency or overproduction of efflux pumps or in isolates of K. pneumoniae with OmpK deficiency (30) while another study has reported cefiderocol MICs of Ն8 g/ml associated with certain isolates of Enterobacteriaceae demonstrating porin loss (28). A weakness in the current study was that molecular analysis of the isolates included was not performed to correlate cefiderocol MICs to genetic markers, and, therefore, the current results cannot be compared directly with other isolate data sets containing molecularly characterized isolates.
Cefiderocol is a promising, novel siderophore cephalosporin currently in clinical development and represents a potentially significant advance in the treatment options available to clinicians to care for patients infected with antimicrobial-resistant Gramnegative bacilli. The intent of the current study was to add to the limited amount of available in vitro data in which cefiderocol MICs were determined against Enterobacteriaceae and nonfermentative Gram-negative bacilli using the recently approved CLSI method for producing ID-CAMHB (17)(18)(19). Our testing demonstrated that cefiderocol possesses potent in vitro activity against carbapenem-nonsusceptible Enterobacteriaceae, MDR A. baumannii, MDR P. aeruginosa, S. maltophilia, and B. cepacia. Cefiderocol MICs were Յ4 g/ml for 96.2% (1,801/1,873) of all isolates of carbapenem-nonsusceptible and MDR Gram-negative bacilli tested in the current study. The potent in vitro activity of cefiderocol was maintained against both ceftazidime-avibactam-nonsusceptible and ceftolozanetazobactam-nonsusceptible isolates of Gram-negative bacilli, as well as against isolates nonsusceptible to colistin.

MATERIALS AND METHODS
Bacterial isolates. Isolates of Gram-negative bacilli tested in this study (n ϭ 1,873) were selected from the International Health Management Associates, Inc. ([IHMA] Schaumburg, IL), 2014 -2016 surveillance study frozen stock culture collection based on their known antimicrobial susceptibility testing phenotypes and/or their species identification; 413 isolates were selected from 2014, 1,123 isolates were from 2015, and 337 isolates were from 2016. Isolates of Enterobacteriaceae (n ϭ 1,022) were chosen based on their carbapenem-nonsusceptible phenotype (meropenem MIC of Ն2 g/ml) (5). Isolates of A. baumannii (n ϭ 368) and P. aeruginosa (n ϭ 262) were chosen because they demonstrated an amikacin-resistant (MIC, Ն32 g/ml), ciprofloxacin-resistant (MIC, Ն4 g/ml), and meropenem-resistant (MIC, Ն16 g/ml) MDR phenotypes (5,31). Isolates of S. maltophilia (n ϭ 217) and B. cepacia (n ϭ 4) were chosen based solely on their identities and the preponderance of each of these species to demonstrate MDR phenotypes. All isolates were originally grown from specimens of patients with a documented intra-abdominal, urinary tract, skin and soft tissue, lower respiratory tract, or bloodstream infection. Isolates tested in this study were limited to one per patient. The identities of all isolates were confirmed by IHMA using matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry (Bruker Daltonics, Billerica, MA). The 1,873 isolates of Gram-negative bacilli were collected by medical center laboratories in 52 countries. Specifically, 995 isolates were collected by medical laboratories in 24 countries in Europe, 399 isolates were from 10 countries in Latin America, 220 isolates were from 2 countries in North America, 155 isolates were from 8 countries in Asia, 61 isolates were from 3 countries in the South Pacific, 29 isolates were from 2 countries in Africa, and 14 isolates were from 3 countries in the Middle East.
The broth microdilution panels included growth control wells for both CAMHB and ID-CAMHB. The panels were incubated at 35°C for 20 h in ambient air before MIC endpoints were read. ID-CAMHB did not significantly affect the growth of any quality control or test organism. Reading the MIC of cefiderocol was contingent on the presence of strong growth in the ID-CAMHB growth control (i.e., a button of approximately 2 mm or greater). The cefiderocol MIC was read as the first panel well in which isolate growth was significantly reduced (i.e., a button of Ͻ1 mm or light/faint turbidity) relative to the growth observed in the ID-CAMHB growth control well. The method described here for reading MIC endpoints for cefiderocol was approved by the CLSI Subcommittee on Antimicrobial Susceptibility Testing but has not yet been published (17)(18)(19)32).
Cefiderocol currently does not have approved MIC interpretative breakpoints. In this study, cefiderocol MICs were analyzed by determining the numbers (percentages) of isolates with MICs of Յ4 g/ml. A concentration of cefiderocol of Յ4 g/ml was used to group isolates because in vitro pharmacokinetic/ pharmacodynamic and animal infection models that recreate human drug exposure have demonstrated that cefiderocol possesses bactericidal killing and clinical efficacy against isolates of Enterobacteriaceae, P. aeruginosa, and A. baumannii with cefiderocol MICs of 4 g/ml (33)(34)(35). Katsube and coworkers reported that the proposed human dose of cefiderocol of 2 g every 8 h, using 3-h infusions, maintained the free-drug concentration of cefiderocol in plasma above 4 g/ml for at least 75% of the dosing interval in patients with normal kidney function, resulting in Ͼ90% probability of target attainment and, therefore, probable clinical success (33). CLSI interpretive criteria, when available (5), and FDA interpretive criteria for ceftazidime-avibactam (36) were used to interpret MICs of the comparator agents tested. For colistin, CLSI interpretive criteria were used to interpret MICs for P. aeruginosa and A. baumannii. Colistin lacks CLSI or FDA breakpoints for Enterobacteriaceae; therefore, the European Committee on Antimicrobial Susceptibility Testing (EUCAST) MIC breakpoints for Enterobacteriaceae were applied to Enterobacteriaceae tested against colistin (susceptible, Յ2 g/ml; resistant, Ն4 g/ml) (37).

ACKNOWLEDGMENTS
The study was funded by Shionogi & Co., Ltd., Osaka, Japan, and funding included compensation fees for services in relation to preparing the manuscript.