FormalPara Key Summary Points

Novel antibiotics for emerging multidrug-resistant (MDR) or extensively drug-resistant (XDR) bacteria, and for bacteria with difficult-to-treat resistance (DTR), are urgently needed.

Cefiderocol is an effective and safe rescue therapy for patients with infections due to difficult-to-treat pathogens.

Better results have been obtained in clinical practice when using cefiderocol in combination regimens, as they may increase its bactericidal potency.

Cefiderocol is well tolerated and has a favourable safety profile, although it is not exempt from the risk of selecting resistant mutants.

Introduction

Antimicrobial resistance is a global health emergency [1]. Novel antibiotics for emerging multidrug-resistant (MDR) and extensively drug-resistant (XDR) bacteria are urgently needed [2, 3]. Enterobacterales, Pseudomonas aeruginosa, and Acinetobacter baumannii are among the World Health Organization’s high-priority pathogens for research into and the development of new treatments [1, 3]. Another major concern is bacteria with difficult-to-treat resistance (DTR), defined as those strains that exhibit non-susceptibility to high-efficacy and low-toxicity agents, which implies resistance to all beta-lactam antibiotics (including carbapenems) and fluoroquinolones for gram-negative bacteria [4].

Cefiderocol is a novel cephalosporin chemically related to ceftazidime and cefepime [5], the main difference being a catechol moiety at the 3-position side chain [5,6,7], granting a unique mechanism for penetrating the bacterial cell wall as a siderophore. The periplasmic concentration of cefiderocol is increased after the hydroxyl groups of the catechol moiety form a chelate with ferric iron and then take advantage of the bacterial iron transportation systems [5,6,7]. Furthermore, this structure confers additional stability against all classes of carbapenemases, including serine carbapenemases (KPC, OXA), and metallo-β-lactamases (NDM, VIM, and IMP), preventing cefiderocol degradation and enhancing bacterial cell wall synthesis inhibition [5,6,7,8,9,10]. Cefiderocol showed potent in-vitro activity against a broad range of gram-negative bacteria, including those expressing extended-spectrum beta-lactamases (ESBL), carbapenemase-resistant Enterobacterales (CRE), and non-fermenting gram-negative bacteria such as DTR-Pseudomonas aeuruginosa (DTR-PA) and carbapenem-resistant Acinetobacter baumannii (CRAB) [7,8,9,10].

The efficacy of cefiderocol has been demonstrated in two phase III randomized clinical trials (RCT) [11, 12], showing encouraging results. However, clinical data assessing the efficacy and safety in real-life settings are scarce and limited to a few case reports and small case series [13,14,15,16,17,18,19,20,21,22,23]. In addition, there is a lack of available data on critically ill populations and patients requiring continuous renal replacement therapy (CRRT) or extracorporeal membrane oxygenation (ECMO). Furthermore, there are still some questions—such as whether to use it as a single agent or in combination therapy, and dose recommendations for real-world patients that require dosage adjustments outside of the package-insert-recommended dosing—which still need to be addressed.

Since cefiderocol is not yet commercialized in Spain, access in our country is available through a compassionate use program upon previous request to and acceptance from the Spanish Agency of Medicines and Medical Devices (AEMPS). Therefore, clinical experience in our country is reduced to a few patients with multi-resistant infections after the failure of previous treatments.

Case Presentations

Aim

The aim of this study is to describe our real-life experience with cefiderocol in a cohort of patients with infections caused by DTR-PA and other MDR and XDR pathogens unresponsive to other available treatment options.

Design

We conducted a retrospective, observational study in a tertiary-care hospital in Spain (Hospital Clinic de Barcelona). All patients treated with cefiderocol from March 2018 to April 2022 were included. Cefiderocol was provided by Shionogi & Co. Ltd. on a compassionate use basis after approval by the AEMPS. The first two patients in this series were previously described by other authors from our team [21]. The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of the Hospital Clinic de Barcelona (CIF-G-08431173-14/2022). Data from all patients were anonymously collected in an electronic dataset.

Cefiderocol was prescribed according to medical criteria by a multidisciplinary team of infectious diseases, critical care specialists, and clinical pharmacists, based on the patient's clinical situation and microbiological isolates. Cefiderocol was administered in a 3-h intravenous infusion (IV) at a standard dose of 2 g every 8 h diluted in 100 mL of normal saline solution, with dose adjustments for renal function made according to the manufacturer’s recommendations [24]. Patients with augmented renal clearance (ARC) (estimated creatinine clearance ≥ 130 mL/min, calculated by CKD-EPI) received 2 g IV every 6 h.

We collected demographic, clinical, and microbiological data from each patient up to 90 days after the end of cefiderocol treatment (ECT). Monotherapy and combination therapy were defined as cefiderocol used as the only intravenous anti-gram-negative agent and as the use of one or more active antibiotics in addition to cefiderocol, respectively.

Clinical cure (CC) was defined as the resolution of signs and symptoms of infection based on clinical/laboratory/image evidence of improvement, and microbiological cure (MC) was defined as the absence of the initially isolated microorganism in the same sample type, both assessed after 7 days from the ECT. Recurrence was defined as the appearance of the originally isolated microorganism (same susceptibility pattern), or the same microorganism with different susceptibility patterns (to include the development of cefiderocol-mediated resistance to other antimicrobials as a potential risk), in those patients who had previously reached MC. Susceptibility to cefiderocol was reassessed in the case of recurrence. Survival status was recorded at 30 and 90 days from ECT, differentiating between all-cause mortality and mortality due to the infection. Adverse events attributable to cefiderocol were recorded.

Antimicrobial susceptibility tests were performed in all patients before the start of cefiderocol treatment using disk diffusion and microdilution methods. Minimum inhibitory concentrations (MICs) were classified according to breakpoints established by the European Committee on Antimicrobial Susceptibility Testing (EUCAST, 2022) [25]. No EUCAST cefiderocol breakpoints for Acinetobacter baumannii and Achromobacter xylosoxidans are currently published.

Each isolated microorganism was classified according to previously defined criteria as MDR or XDR [2] and DTR [4]. Continuous variables are reported using median and interquartile ranges (IQRs), and categorical variables are described as frequencies and percentages.

Results

Ten patients treated with cefiderocol were included in the study. Nine (90%) patients were male with a median age of 61 (IQR, 47–68) years. The median hospitalization length of stay prior to cefiderocol initiation was 37 (IQR, 7–54) days. Seven (70%) septic patients were admitted to an intensive care unit (ICU) with median APACHE-II and SOFA scores of 16 (IQR, 15–17) and 6 (IQR, 4–8), respectively. Five patients had sepsis and two had septic shock. Five critically ill patients (5/7) underwent endotracheal intubation, and one required ECMO and CCRT.

The most common sources of infection were bacteremia and ventilator-associated pneumonia (VAP), accounting for 40% of the cases each. Regarding bacteremia, two patients had an endovascular prosthesis infection secondary to a transjugular intrahepatic portosystemic shunt (TIPS) device that could not be removed, while the other two patients had a catheter-related bloodstream infection. One of them also had severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) pneumonia. Concerning VAP, 1/4 occurred in a patient with acute respiratory failure also due to SARS-CoV-2. Among the two remaining patients without VAP or bacteremia, one had a skin and soft-tissue infection, and the other had spontaneous bacterial peritonitis. The patients’ characteristics, type of infections, isolated pathogens, and clinical and microbiological outcomes are summarized in Table 1.

Table 1 Description of patient characteristics, types of infections, isolated pathogens, combination therapies, clinical and microbiological outcomes, and mortality rates
Full size table

The most common causal pathogen was P. aeruginosa (70% of cases), of which six cases were DTR-PA. Three of these patients were coinfected with ESBL-NDM--Klebsiella pneumoniae and Candida auris (P5), Myroides odoratimimus (P6), and Enterobacter cloacae complex (P8). In the remaining patients, the isolated pathogens were OXA-23-producing CRAB, A. xylosoxidans, and ESBL-E. coli.

Cefiderocol was never used as a first-line treatment. Before the use of cefiderocol, at least three anti-gram-negative regimes had failed in all patients, including a colistin-based one in 60% of them. Most patients received 2 g q8h of cefiderocol, and only one received 2 g q6h, which was subsequently switched to 8 g as a continuous infusion to optimize pharmacokinetic–pharmacodynamic (PK-PD) parameters. Three patients (P1, P2, and P5) needed dose adjustments due to renal impairment according to the manufacturer’s recommendations [24]. The median duration of cefiderocol treatment was 15 (IQR, 11–24) days. Cefiderocol was administered in combination with other anti-gram-negative agents in seven out of ten patients. The most frequent combinations were nebulized colistin (4/7) and I.V. amikacin (3/7), with the choice of antibiotic guided by the infection source and antibiotic susceptibility pattern. Antibiotic susceptibility, MICs, and the emergence of resistance after cefiderocol treatment are described in Table 2.

Table 2 Antibiogram of microorganisms before cefiderocol treatment and development of resistance to cefiderocol after treatment
Full size table

Patients’ Outcomes

Clinical cure was achieved in 90% (9/10) of the patients. Only one patient (P9) remained febrile under antibiotic treatment. Microbiological cure was observed in eight (80%) patients, although two of them had persistent rectal colonization by P. aeruginosa (P2) and E. coli (P10). Patient P4, with an unremovable TIPS and related bacteremia due to MDR P. aeruginosa, suffered a recurrence 40 days after the ECT, caused by a strain that developed resistance to cefiderocol. Conservative treatment was decided upon, and the patient died before the 90 days follow-up due to the infection.

Another patient (P9) with VAP never reached either clinical or microbiological cure. In this case, the infection was caused by a DTR-PA strain, which became resistant to cefiderocol while the patient was receiving it. P9 eventually died from the severe ongoing infection before the 90 days follow-up.

Only one patient (P3) died within 30 days after the ECT due to a worsening of the underlying disease (acute respiratory distress syndrome due to SARS-CoV-2 infection and severe myopathy of the critically ill patient), but without any evidence of DTR-PA infection. Four additional patients expired during the 31 to 90 days follow-up period, two of them (P4 and P9) due to the aforementioned complications and cefiderocol-resistant infections by MDR and XDR/DTR-PA, respectively. Overall, the 30-day and 90-day mortality rates were 10% and 50%, respectively. After more than 6 months of follow-up (201 [IQR, 152–347] days), five patients remained alive.

There was no evidence of mild-to-moderate side effects, except for one patient (P1) who developed thrombocytopenia on treatment day 14. However, the platelet count returned to baseline levels 3 days after the ECT. Two more patients (P4 and P6) developed thrombocytopenia while being on cefiderocol treatment; however, they did not recover after the ECT. Both patients were in the ICU because of septic shock and severe underlying diseases (liver and autoimmune disorder, respectively) that could have been the cause of thrombocytopenia. Hence, we cannot attribute the thrombocytopenia exclusively to the antibiotic. No other adverse events were reported.

Discussion

In this study, we report our early experience with cefiderocol as a last-resort antibiotic for difficult-to-treat infections due to MDR and XDR bacteria. This study is among the few case series describing the real-life use of cefiderocol and is the largest conducted in Spain. We included infections caused by different microorganisms (CRE, DTR-PA, CRAB, and A. xylosoxidans), in contrast with some previously described series where the main pathogen isolated was either A. baumannii [15, 16, 18, 19, 21], P. aeruginosa [14], or A. xylosoxidans [22].

The primary indications for cefiderocol were bacteremia and VAP. Regarding bacteremia, clinical and microbiological cure was reached in four out of four patients, while in patients with VAP it was reached in three out of four and in two out of four, respectively. Despite our moderate results in patients with VAP, sufficient lung penetration of cefiderocol has previously been demonstrated in mechanically ventilated critically ill patients with pneumonia, by measuring its concentration in the epithelial lining fluid [10, 26]. Overall, our results for clinical cure (90%) and microbiological cure (80%) are in concordance with previous case series in which clinical cure was achieved in 54 to 100% of the patients [14, 15, 18,19,20, 23, 27] and microbiological cure in 60 to 100% [14, 17, 18, 20, 23].

It is of note that cefiderocol has apparently performed better for both clinical and microbiological cure in real clinical scenarios than in RCTs. In the CREDIBLE-CR, clinical cure was achieved in only 53% of patients and eradication at the end of treatment in 48% [11], and in the APEKS-NP, the rates of clinical and microbiological cure were 65% and 41%, respectively [12]. This could be related to the high proportion of CRAB infections (46%) [11] and inadequate cefiderocol PK/PD target attainment, which could be solved by the implementation of a dosing strategy focused on achieving a Cmin/MIC ≥ 4 [16]. The better results obtained in clinical practice may also be due to the use of cefiderocol in combination regimens, which may increase its bactericidal potency [20]. The 30-day survival rate (90%) observed in the present study is also in line with the case series of Meschiari et al. (77%), Bleibtreu et al. (90%), and Falcone et al. (90%). However, we found a higher 90-day all-cause mortality (50%) compared with Meschiari et al. (35%) [14], although only two patients died due to the infection. The favourable outcomes are even more relevant since cefiderocol was used as at least a third-line treatment in all patients, and 70% of them were critically ill.

In our series, cefiderocol was used as a combination therapy in 7/10 (70%) of patients. Although there is still no consensus on the benefit of such an approach [12, 14], a trend towards better results with combination therapy [14, 19, 20, 28] compared to monotherapy [11, 15, 16, 18] has been observed in previously published cefiderocol series. Combination therapy offers several theoretical advantages; it can provide a synergistic effect and greater bactericidal action, and it may help to overcome the emergence of resistance [20], which is one of the main concerns when treating MDR, XDR, and DTR pathogens [1,2,3,4]. However, in our study, despite combined therapy with other active antibiotics, strains from P4 and P9 developed cefiderocol resistance during the course of treatment. Combination therapy may be of particular consideration when treating CRAB [16, 18, 23, 27], a microorganism with a high prevalence of heteroresistance [14], and also DTR-PA [4], which was isolated in P2, P3, P6, P7, P8, P9, but not P4, as in this case it remained susceptible to ciprofloxacin. However, combination therapy may be associated with a higher risk of toxicity and C. difficile disease [20]. In the four patients with VAP, combination therapy with nebulized colistin was used; however, the causal microorganisms were sensitive to colistin in only two of the patients (P7 and P9). In patients with bacteremia, no one antibiotic was used in combination more frequently than any other. Further studies are needed to determine the most appropriate companion drugs for optimizing the bactericidal action of cefiderocol-based regimens.

Therapeutic drug monitoring (TDM) could be used to guide optimal dosing, especially in patients who are critically ill due to pathophysiological changes in drug disposition and clearance [29]. To date, the main limitation of cefiderocol TDM is broad availability. Cefiderocol concentrations were measured in only three previously published case series [16, 17, 30]. Two cases of TDM have been reported in patients with meningitis; one due to DTR-PA [14] and the other due to CRAB [30]. Since cefiderocol is a time-dependent antibiotic, the optimal dose and administration for PK/PD target attainment in the licensed dosing regimens were based on maintaining the free plasma concentration above the MIC during 75–100% of the dosing interval (fT > MIC of 75–100% or a ratio fCmin/MIC of 0.75–1) [11, 12, 31]. However, Gatti et al. reported that fCmin/MIC < 4 was suboptimal, predicting a higher probability of microbiological failure in critically ill patients with VAP due to CRAB [16]. In the study by König et al. using the manufacturer’s recommendations for renal dose adjustment, all patients showed a prolonged cefiderocol half-life of at least 8 h, and trough concentrations were in the range between 25–70 mg/L, attaining targets above CMI 100% of the time in all cases [17].

TDM could be particularly useful when there is a need for extracorporeal techniques, such as CRRT or ECMO. Cefiderocol CRRT dose adjustments are described by the manufacturer according to PK/PD studies [24, 31]. Cefiderocol optimal dosing on ECMO is still unknown, and whether there is a different pharmacokinetic profile or significant drug sequestration in the circuit is currently being explored in a RCT (NCT04995835). König et al. observed an increase in the volume of distribution in two patients on ECMO, resulting in lower peak concentrations with the usual 2 g q8h dose [17]. Gatti et al. used 2 g q8h in three patients under ECMO and 1.5 g q8h in another one (ECMO and CRRT), obtaining ratios of fCmin/MIC between 2.4 and 26.7 [16]. Finally, Meschiari et al. described the use of 2 g q6h in one patient with ECMO [14]. In our case series, P5 initiated cefiderocol while being on veno-venous ECMO (3200 rpm, blood flow 4 L/min initially, changed according to evolution) at a dose of 2 g q8h, which was subsequently reduced to 1 g q8h because of additional CRRT needs (effluent flow rate [EFR]: 3.0 L/h); this led to a good outcome. There is a lack of data from patients undergoing ECMO and CRRT simultaneously. In the CREDIBLE-CR and APEKS-NP RCTs, 1.5 g q12h was used under CRRT [11, 12]. However, Wei et al. and Wenzler et al. provided cefiderocol dosing recommendations according to EFR in patients receiving CRRT (1.5 g q8h for EFR 3L/h) [32, 33]. P2 and P4 had impaired hepatic function and carried a TIPS; dosing for these two patients was not adjusted, as this condition plays a negligible role in the PK of cefiderocol [24].

Cefiderocol is generally well tolerated and has a favourable safety profile [11, 12]. In our case series, three patients developed thrombocytopenia (P1, P4, and P6), although we can only attribute this adverse effect to cefiderocol in one (P1). In this patient, treatment was stopped and the platelet count recovered after 3 days [21]. Bleibtreau et al. also described thrombocytopenia as an adverse event, showing the same resolution after discontinuation of cefiderocol [28].

Cefiderocol is not exempt from the risk of selecting resistant mutants, as was the case with two of our patients (P4 and P9). Karakonstantis et al. elucidated that a combination of different mechanisms, which may become synergistic, are necessary to develop resistance to cefiderocol. The main one seems to be the co-expression of various β-lactamases in combination with permeability defects in cefiderocol-resistant isolates. Regarding mutations affecting the expression/function of siderophore receptors, the most involved genes were cirA and fiu in Enterobacterales, piuA and piuD in P. aeruginosa, and pirA and piuA in A. baumannii. Moreover, if the MIC is close to the PK/PD breakpoints, it is easier for additional mutations to raise the MIC above them [27].

Overall, this study has some limitations. First, due to the limited number of patients and the descriptive nature of the study, no definitive conclusions can be drawn. Secondly, an analysis of the mechanisms of resistance to cefiderocol was not performed. Third, TDM to guide optimal dosing was not available in our center. Fourth, the majority of the patients had normal to mild renal dysfunction, which may underrepresent critical care patients. Fifth, cefiderocol was never used as first-line treatment, only as rescue therapy, which may affect the effectiveness data. Nevertheless, real-world data may help clinicians to correctly place cefiderocol in the broad setting of antibiotics to use when treating multiresistant infections.

Conclusion

In conclusion, our experience provides real-life data on a set of patients with difficult-to-treat gram-negative bacterial infections that were unresponsive to other available treatment options, in which cefiderocol achieved a clinical cure rate of 90%, a microbiological cure rate of 80%, and a low rate of recurrence. In the current scenario, where multiresistant infections are on the rise and the development of new antibiotics is slow and limited, the results are encouraging, although data from a larger number of patients are needed.