FormalPara Key Summary Points

In a real-life, observational, multicenter study, cefiderocol was mostly administered as targeted therapy (72.5%), although the proportion of empirical therapy was non-negligible (27.5%).

Cefiderocol was mainly administered for lower respiratory tract infections and bloodstream infections caused by Acinetobacter baumannii, followed by Pseudomonas aeruginosa.

Notably, in a real-life scenario, cefiderocol was administered almost equally as monotherapy or as combination therapy (50.5% vs. 49.5%).

The almost equal distribution of cefiderocol monotherapy and cefiderocol-based combination therapies underlines the need for further study to ascertain possible differences in efficacy between the two approaches.

Introduction

Cefiderocol is a catechol-substituted siderophore cephalosporin showing in vitro rapid bactericidal activity against various carbapenem-resistant Gram-negative bacteria (CR-GNB) [1, 2].

Based on the results of phase-3 randomized controlled trials (RCT), cefiderocol was approved in Europe for the treatment of GNB infections with limited therapeutic options [3,4,5]. Various studies have described the use of cefiderocol for treatment of CR-GNB infections, including those caused by carbapenem-resistant Acinetobacter baumannii and CR-GNB producing metallo-β-lactamases (MBL) for which cefiderocol is usually among the very few active options [6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35]. However, such studies were designed mostly to evaluate cefiderocol for targeted treatment of specific infections and/or pathogens, while no data currently exist about the real-world use of cefiderocol in terms of types of therapy (e.g., empirical or targeted, monotherapy or in combination with other agents), indications, and characteristics of the treated patients. These data could be useful to improve the development of empirical and targeted therapeutic algorithms of cefiderocol in real life, ultimately aiming at better patient care.

In this multicenter, observational, prospective study, we primarily aimed at describing how cefiderocol is used in Italian hospitals, in terms of types of therapy, indications, and patient characteristics.

Methods

Setting and Objectives

The MULTI-SITA project is a novel platform developed by the Italian Society of Anti-Infective Therapy (SITA) and dedicated to conducting observational studies on invasive bacterial and fungal diseases. CEFI-SITA is an ongoing observational, prospective, multicenter study conducted in Italian hospitals within the MULTI-SITA project, registering use of cefiderocol in consecutive adult patients according to clinical practice. The prospective study period of the CEFI-SITA study is from 1 August 2022 to 31 December 2025. Here, in a pre-planned analysis, we report the preliminary descriptive results of the CEFI-SITA study. The primary objective of this preliminary analysis was to describe the characteristics of patients and infections treated with cefiderocol in the first 200 enrolled patients (see sample size calculation below) from 17 hospitals. Secondary objectives were: (1) to exploratorily describe factors associated with use of cefiderocol as part of a combined regimen versus monotherapy; and (2) to describe clinical cure rates and 30-day mortality in patients with GNB infections receiving targeted cefiderocol therapy. Patients receiving at least one dose of cefiderocol for any reason according to local practice were included in the study, in line with its observational nature. Exclusion criteria were (1) age less than 18 years and (2) already included in the study for a previous cefiderocol administration.

Microbiological Procedures

Identification of bacterial isolates from clinical specimens was performed by means of matrix-assisted laser desorption ionization time of flight mass spectrometry (Vitek MS MALDI-TOF mass spectrometer, bioMérieux, Craponne, France; or MALDI Biotyper, Bruker Daltonics, Billerica, MA, USA) or automated systems, depending on standard local procedures. Antimicrobial susceptibility testing (AST) for antibiotics other than cefiderocol was also performed by means of automated systems (Vitek 2, bioMérieux; MicroScan, Beckman Coulter, Brea, CA, USA; or Phoenix, Becton Dickinson Diagnostics, Sparks, MD, USA) according to local standard procedures. Cefiderocol AST was performed by means of disk diffusion or broth microdilution methods, including either the reference broth microdilution MIC determination using iron-depleted cation-adjusted Mueller Hinton Medium [36] or commercial broth microdilution tests, according to local practice. Cefiderocol AST by gradient test was reported for one isolate. Results of susceptibility testing were interpreted in accordance with the European Committee on Antimicrobial Susceptibility Testing (EUCAST) clinical breakpoints, version 14.0 http://www.eucast.org/), also for cefiderocol. In more detail, resistance to cefiderocol was defined as: (1) for Enterobacterales, minimum inhibitory concentration (MIC) > 2 mg/L for broth microdilution and zone diameter < 23 mm for disk diffusion; and (2) for P. aeruginosa, MIC > 2 mg/L for broth microdilution and zone diameter < 22 mm for disk diffusion). For other GNB currently lacking sufficient evidence for defining cefiderocol breakpoints according to EUCAST, isolates were defined as non-wild-type based on EUCAST ECOFF breakpoints for cefiderocol (MIC > 0.5 mg/L for broth microdilution). For disk diffusion, zone diameters < 17 mm (for A. baumannii) and < 19 mm (for S. maltophilia), corresponding to MIC values below the EUCAST PK-PD breakpoint of cefiderocol, were also deemed as defining non-wild-type isolates in this study. An MIC > 2 mg/L and an MIC > 0.5 mg/L were also considered as the cut-off for defining resistant and non-wild-type isolates by gradient test, respectively. For the purpose of the study, carbapenem resistance was defined as resistance to any of the carbapenems not intrinsically inactive against the given GNB species. Presence of carbapenemase-encoding gene/s was detected using Verigene BC-GN (Nanosphere, Northbrook, IL, USA) or Xpert Carba-R (Cepheid, Sunnyvale, CA, USA), or inferred by means of the BioFire BCID2 or Pneumonia Plus syndromic panels (bioMérieux), whereas the production of carbapenemase/s was assessed by NG-test Carba 5 (NG Biotech, Guipry, France), depending on standard local procedures at the time the study was conducted.

Definitions and Data Collected for the Study

The following demographic and clinical variables were collected at the time of cefiderocol initiation: age in years; sex; previous hospitalization (within 6 months); admission from a long-term care facility (LTCF); diabetes mellitus; chronic obstructive pulmonary disease (COPD); previous myocardial injury; New York Heart Academy (NYHA) score; chronic liver disease (defined histologically as liver cirrhosis or in presence of a clinical diagnosis supported by laboratory, endoscopy, and radiologic findings [37]); chronic kidney disease (defined as estimated glomerular filtration rate < 60 mL/min/1.73 m2); chronic intermittent hemodialysis; solid neoplasm; metastatic solid neoplasm; hematological malignancy; previous hematopoietic stem cell transplantation (HSCT); previous solid organ transplantation (SOT); human immunodeficiency virus (HIV) infection; autoimmune disease; age-adjusted Charlson Comorbidity Index [38]; previous therapy with cefiderocol (within 6 months); previous antibiotic therapy with agents other than cefiderocol (within 6 months); previous chemotherapy (within 6 months); previous steroid therapy (within 6 months); previous therapy with other immunosuppressants (within 6 months); previous major surgery (within 3 months); previous isolation of carbapenem-resistant GNB (within 6 months); days from admission to cefiderocol initiation; intensive care unit (ICU) stay; sequential organ failure assessment (SOFA) score [39]; presence of central venous catheter (CVC); presence of urinary catheter; presence of septic shock [40]; presence of at least mild acute respiratory distress syndrome (ARDS) [41]; estimated creatine clearance (CLCr) and presence of at least stage 1 of acute kidney injury (AKI) according to the Kidney Disease: Improving Global Outcome (KDIGO) criteria [42]; concomitant coronavirus disease 2019 (COVID-19); total parenteral nutrition; neutropenia (defined as blood absolute neutrophil count < 500 cell/mm3); continuous renal replacement therapy (CRRT); extracorporeal membrane oxygenation (ECMO); absolute blood white cell count; serum C-reactive protein; serum procalcitonin; type of cefiderocol therapy, defined as monotherapy or combined therapy based on the anti-carbapenem-resistant GNB (CR-GNB) activity of agents administered concomitantly with cefiderocol (more in detail, combination of cefiderocol with at least one of the following agents was considered as a combined anti-CR-GNB therapy: aminoglycosides; fosfomycin; tigecycline, with the exception of targeted therapy of Pseudomonas aeruginosa infections; polymyxins; sulbactam or ampicillin/sulbactam, as empirical treatment or as targeted therapy for Acinetobacter baumannii infections); timing of cefiderocol therapy (empirical vs. targeted after identification of the causative agent); initial cefiderocol dosage according to estimated CLCr or hemodialytic treatment; type of infection treated with cefiderocol (according to the US Centers for Disease Control and Prevention/National Healthcare Safety Network [CDC/NHSN] surveillance definitions [43]; if CDC/NHSN criteria were not satisfied, local investigators indicated the diagnosis registered in the clinical chart, e.g., sepsis provided Sepsis-3 criteria were satisfied [40]); causative agent/s of the infection treated with targeted cefiderocol therapy. The following information was also collected during follow-up: causative agent/s of the infection initially treated with empirical cefiderocol, if subsequently identified; change of cefiderocol dosage during therapy due to changes in CLCr; clinical cure, defined as resolution of clinical signs and symptoms of infection, at day 7, 14, and 28 after cefiderocol initiation; cumulative mortality at day 30 after cefiderocol initiation; adverse events (AE), including serious AE (SAE), occurring during cefiderocol therapy, if any.

Sample Size Calculation

A sample size of 200 patients was considered an acceptable compromise regarding the external validity of study results when considering the primary descriptive endpoint (i.e., use of cefiderocol in terms of indications and characteristics of treated patients). In more detail, by assuming normal distribution of the estimates of population parameters measured in the study sample (e.g., proportion of patients receiving cefiderocol as empirical therapy/all patients receiving cefiderocol), a sample size of 200 patients would guarantee a maximum error margin (confidence interval) of ± 7% with α = 0.05.

Statistical Analysis

The characteristics of patients and infections treated with cefiderocol were described through median (interquartile range, IQR) and absolute frequency (relative frequency, %) for continuous and categorical variables, respectively. The 95% confidence interval (CI) was calculated for both proportions [44] and median values [45] estimates. To assess factors associated with cefiderocol combination therapy, we performed Rubin’s multiple imputation as a preliminary step [46]. Then, the association of demographic/clinical variables with cefiderocol combination therapy (vs. cefiderocol monotherapy as reference) was first assessed through univariable logistic regression (LR) models. Subsequently, all variables showing a p value < 0.10 in univariable comparisons were included in an initial multivariable LR model, and further selected for inclusion in the final multivariable LR model (model A) by means of a stepwise backward procedure. Finally, as sensitivity analysis, variables included in model A were also included in an additional multivariable LR model that also included center as random effect (model B). The crude 30-day mortality from the initiation of cefiderocol therapy was summarized graphically using the Kaplan–Meier method in patients receiving (1) targeted cefiderocol therapy for Enterobacterales infection, (2) targeted cefiderocol therapy for Acinetobacter baumannii infection, (3) targeted cefiderocol therapy for Pseudomonas aeruginosa infection, and (4) targeted cefiderocol therapy for MBL-producing Gram-negative bacteria.

The analyses were conducted using SAS software (version 9.4; SAS Institute, Cary, NC, USA) and R Statistical Software (version 4.2.1; R Foundation for Statistical Computing, Vienna, Austria).

Ethical Considerations

The MULTI-SITA project was approved by the ethics committee of the coordinating center (Liguria Region Ethics Committee, registry number 390/2020). The amendment authorizing the conduct of the CEFI-SITA study within the MULTI-SITA project was approved by the Liguria Region Ethics Committee on 12 April 2022. The other participating centers followed the local ethical committees' requirements and started to prospectively enroll patients once activated. All conscious patients at time of enrollment signed an informed consent to participate in the study. A waiver of informed consent for data collection from unconscious patients at the time of enrollment due to severe clinical conditions was obtained within the ethics committee approval, in line with the observational nature of the analyses and in order not to bias research results towards high cure rates and low mortality prejudicing scientific validity.

Results

The demographic and clinical characteristics at time of cefiderocol initiation of the 200 included patients (enrolled from August 2022 to September 2023) are displayed in Table 1. Their median age was 66 years (IQR 52–75) and 76% were male (152/200). The characteristics of cefiderocol treatments are reported in Table 2. Overall, cefiderocol was started as empirical and targeted therapy in 27.5% (55/200) and 72.5% (145/200) of cases, respectively.

Table 1 Demographic and clinical characteristics of adult patients treated with cefiderocol
Full size table
Table 2 Characteristics of cefiderocol therapy
Full size table

Empirical Therapy

As shown in Supplementary Table S1, the most frequently reported indications for empirical cefiderocol therapy were sepsis (33/55, 65.5%) and lower respiratory tract infection (10/55, 18.2%). In 28/55 cases of empirical therapy (50.9%), GNB grew from cultures collected at the time of treatment initiation, of which 14/26 (54%, missing = 2/28) were CR-GNB, mostly A. baumannii, P. aeruginosa, and Klebsiella spp. (all 4/14 each, 28.6%). Among identified CR-GNB, production of carbapenemase/presence of carbapenemase-encoding genes was assessed in 12/14 cases (85.7%), resulting positive in 5/12 cases (41.7%), of which 3/5 were positive for MBL (60.0%). Identified MBL-producers were 1 VIM-producing Escherichia spp., 1 NDM-producing Escherichia spp., and 1 NDM-producing K. pneumoniae. The other identified carbapenemase-producing organisms were 1 KPC-producing K. pneumoniae and 1 OXA-producing A. baumannii. Cefiderocol AST was performed among 8/14 identified CR-GNB (including 2/3 MBL-producers), with one A. baumannii isolate being non-wild-type (production of carbapenemase/presence of carbapenemase-encoding genes investigated but not detected). The other 7 isolates were all wild-type/susceptible to cefiderocol.

Targeted Therapy

Regarding targeted cefiderocol therapy (Supplementary Table S1), the most frequently reported indications were lower respiratory tract infection (63/145, 43.4%) and bloodstream infection (56/145, 38.6%). Overall, 170 GNB causative agents were retrieved from 145 infections treated with targeted cefiderocol therapy (see details in the legend of Supplementary Table S1). Infections treated with targeted cefiderocol therapy were most frequently monomicrobial (122/145, 84.1%), mainly caused by A. baumannii (67/122, 54.9%), P. aeruginosa (25/122, 20.5%), and Klebsiella spp. (18/122, 14.8%). Rates of carbapenem resistance among these infections were 91.9% (57/62 tested isolates), 95.5% (21/22 tested isolates), and 93.8% (15/16 tested isolates) for A. baumannii, P. aeruginosa, and Klebsiella spp., respectively. Production of carbapenemase/presence of carbapenemase-encoding genes was assessed in 121/170 isolates (71.2%), resulting positive in 51/121 cases (42%), of which 26/51 (51.0%) were MBL-producing (21.5%), mostly VIM-producing P. aeruginosa (12/26, 46.2%), followed by NDM-producing Klebsiella spp. (5/26, 19.2%) and VIM-producing Enterobacter spp. (4/26, 15.4%). Other detected carbapenemases were KPC-type (n = 11), OXA-type (n = 11), and not specified (n = 3). Cefiderocol AST was performed in 74/170 isolates (43.5%) from infections receiving targeted cefiderocol therapy (including 13/26 MBL producers, 50.0%), with 3/57 tested A. baumannii isolates (5.3%) being non-wild-type (production of carbapenemase/presence of carbapenemase-encoding genes was investigated but not detected in all three isolates). No other isolates were resistant/non-wild-type. The demographic and clinical characteristics at time of targeted cefiderocol initiation, stratified according to the different causative agents of infection, are available in supplementary Table S2.

Monotherapy and Combined Therapy

Cefiderocol was administered as monotherapy and as combined therapy for CR- GNB infections in 101/200 (50.5%) and 99/200 (49.5%) cases, respectively (Table 2). Supplementary Table S3 reports the results of univariable analyses of factors associated with administration of cefiderocol as a combined therapy for CR-GNB infections, whereas factors retaining an independent association with combined therapy for CR-GNB infections in the final multivariable models are presented in Table 3. In model A, previous isolation of carbapenem-resistant Acinetobacter baumannii [odds ratio (OR) 2.56, 95% CI 1.01–6.46, p = 0.047) and previous hematopoietic stem cell transplantation (OR 8.73, 95% CI 1.05–72.54, p = 0.045) were associated with administration of cefiderocol within a combined regimen, whereas chronic kidney disease was associated with cefiderocol monotherapy (OR 0.38 for combined regimen, 95% CI 0.16–0.91, p = 0.029). In model B, including the same variables of model A plus center as a random effect, the direction of fixed effects was the same registered in model A.

Table 3 Multivariable analysis of factors associated with use of cefiderocol in combination with other anti-CR-GNB agents a
Full size table

Cure Rates and Tolerability

Cure rates at days 14, 21, and 28 are descriptively summarized in supplementary Table S4, divided by pathogen and combination or monotherapy. As shown in Fig. 1, the cumulative 30-day mortality was 19.8% (95% CI 7.0–37.4) in patients receiving targeted cefiderocol therapy for Enterobacterales infection (panel A), 45.0% (95% CI 32.4–56.8) in those receiving targeted cefiderocol therapy for A. baumannii infection (panel B), 20.7% (95% CI 7.3–38.7) in patients receiving targeted cefiderocol therapy for P. aeruginosa infection (panel C), and 22.7% (95% CI 8.0–41.9) in patients receiving targeted cefiderocol therapy for MBL-producing GNB (panel D). Overall, 4/200 patients (2.0%) experienced a suspected drug-related AE during cefiderocol administration. Two patients developed mild skin rash, one patient experienced hyperchromic urine and moderate increase in liver enzymes values, and one patient developed status epilepticus categorized as SAE. Cefiderocol was discontinued in 2/4 patients (50.0%) experiencing AE.

Fig. 1
figure 1

Cumulative mortality up to day 30 in patients receiving targeted cefiderocol therapy for Enterobacterales infection (panel A), Acinetobacter baumannii infection (panel B), Pseudomonas aeruginosa infection (panel C), and MBL-producing Gram-negative bacteria (panel D). MBL metallo-β-lactamases. Analyses limited to Infection by only one Gram negative genus (with the exception of Enterobacterales infection, for which concomitant infection by more than one member of the Enterobacterales order was also considered). The time of origin was set at the day of cefiderocol initiation. Death was the event of interest and right-censoring was applied at the end of follow-up (hospital discharge or day 30, whichever came first). Site/s of Enterobacterales infection: bloodstream infection (n = 12); lower respiratory tract infection (n = 7); urinary tract infection (n = 2); skin and soft tissue infection (n = 1); intra-abdominal infection (n = 1); intra-abdominal infection plus bloodstream infection (n = 1); lower respiratory tract infection plus bloodstream infection (n = 1); urinary tract infection plus bloodstream infection (n = 1). Site/s of P. aeruginosa infection (n = 25): lower respiratory tract infection (n = 12); bloodstream infection (n = 6); urinary tract infection (n = 2); bone and joint infection plus bloodstream infection (n = 1); intra-abdominal infection (n = 1); lower respiratory tract infection plus bloodstream infection (n = 1); skin and soft tissue infection (n = 1); skin and soft tissue infection plus bloodstream infection (n = 1). Site/s of A. baumannii infection (n = 67): lower respiratory tract infection (n = 30); bloodstream infection (n = 29); bone and joint infection (n = 2); urinary tract infection (n = 2); intra-abdominal infection (n = 1); lower respiratory tract infection plus bloodstream infection (n = 1); skin and soft tissue infection (n = 1); site/s not reported (n = 1). Site/s of MBL-producing Gram-negative infection (n = 22): lower respiratory tract infection (n = 8); bloodstream infection (n = 5); urinary tract infection (n = 3); intra-abdominal infection (n = 2); skin and soft tissue infection (n = 2): lower respiratory tract infection plus bloodstream infection (n = 1); skin and soft tissue infection plus bloodstream infection (n = 1). Type of MBL enzyme (n = 20): NDM (n = 12); VIM (n = 19); NDM (n = 3). Type of MBL-producing causative agent: P. aeruginosa (n = 12); Enterobacterales (n = 10)

Full size image

Discussion

In this multicenter study, cefiderocol was mostly administered as targeted therapy (72.5% vs. 27.5% as empirical therapy), mainly for lower respiratory tract infections and bloodstream infections caused by A. baumannii, followed by P. aeruginosa. Cefiderocol was almost equally administered as monotherapy or as combination therapy (50.5% vs. 49.5%).

Most available studies on the use of cefiderocol in real life are focused on the targeted treatment of infections caused by various or specific CR-GNB [7, 10,11,12,13,14,15, 17, 18, 21,22,23,24, 26,27,28,29,30,31,32,33,34,35]. While useful for reporting cure rates in specific infections, these studies were unable to generally depict how cefiderocol is currently used by physicians, having been designed for other purposes. In our opinion, this topic is of interest considering the following: (1) current international and national guidelines/guidance documents provide recommendations on the use of cefiderocol for targeted treatment, while there is still no clear guidance regarding its empirical use [47,48,49,50,51]; and (2) various studies have reported use of cefiderocol either in combination or as monotherapy, thus we deemed it of interest to descriptively explore their relative frequency [30, 52]. Regarding the first point, in our study, cefiderocol was mainly administered as targeted therapy. However, the proportion of empirical therapies was not negligible, representing more than one-quarter of all cefiderocol prescriptions. This could reflect the physicians’ willingness to prescribe an active therapy without delay in patients with severe infections and risk factors for CR-GNB, in areas where the prevalence rate of CR-GNB, especially MBL producers, is high. Overall, these results suggest the need for standardized consensus regarding the use of cefiderocol within empirical therapeutic algorithms, in order to maximize proper indications in line with antimicrobial stewardship principles, but also to not perilously delay treatment in patients with infections caused by cefiderocol-susceptible CR-GNB.

Regarding the use of cefiderocol as monotherapy or within combined regimens, our results could reflect the current lack of solid evidence on this aspect. Against this background, the possible perception of some clinicians of potential reduced activity of cefiderocol in patients with severe A. baumannii infections could have prompted to consider combination therapy in some cases. This is in line with the independent association we found in multivariable models between previous isolation of carbapenem-resistant A. baumannii and use of cefiderocol in combined regimens, although it is also of note that no association was found between targeted treatment of A. baumannii infections and use of cefiderocol within combined regimens. Furthermore, in our opinion, this controversial point merits some additional considerations. First, it should be considered that the apparently higher mortality rates (50% [21/42] vs. 18% [3/17] for best available therapy, as 49-day mortality) in patients with A. baumannii infections treated with cefiderocol in the CREDIBLE-CR RCT were possibly confounded by factors such as severity of presentation and baseline diseases (e.g., severe renal dysfunction, ongoing shock, shock 31 days before randomization, and ICU stay at randomization were more frequent in patients treated with cefiderocol than in those receiving best available therapy) [5]. A similar situation was present in our study, in which mortality of A. baumannii infection was higher than mortality of infections caused by other organisms. Indeed, severe clinical presentation was more frequent in patients with A. baumannii infections than in those with infections caused by other organisms (e.g., septic shock was present in 29% of patients receiving cefiderocol for the targeted treatment of A. baumannii infections vs. 15.4% and 16.0% in patients receiving cefiderocol for the targeted treatment of Enterobacterales infections and P. aeruginosa infections, respectively, as shown in supplementary Table S2). Second, it should also be noted that, although mortality rates of cefiderocol-treated A. baumannii infections in some other studies were in line with our results [13, 15, 18, 23, 24, 33], in many other studies, including the APEKS-NP randomized controlled trial, they were far lower, ranging from 18 to 37% [3, 7, 12, 14, 16, 27, 29, 32, 53]. Third, lower mortality of cefiderocol-treated A. baumannii infections in comparison with non-cefiderocol-based regimens was suggested by recent meta-analyses including data from both CREDIBLE-CR and observational studies [54,55,56]. Lastly, among isolates subjected to cefiderocol AST, non-susceptibility (i.e., non-wild-type) was exclusively detected in A. baumannii, a finding consistent with that of a recent meta-analysis reporting the highest prevalence of cefiderocol non-susceptibility for A. baumannii, compared to other major Gram-negative pathogens [57]. Overall, considering all the above, in our opinion, the relevant research question for future studies should regard which subgroups of patients with A. baumannii infections could benefit the most from cefiderocol monotherapy versus combination therapy (in turn possibly influencing the choice of combined regimens). Regarding other results, the significant association between previous HSCT and administration of cefiderocol within a combined regimen deserves further investigations. Indeed, while the perception of severe baseline conditions and increased mortality due to immunosuppression connected to baseline disease and/or its treatment could have played a role in influencing the choice of combination, also for targeted treatment, the subgroup of patients who underwent HSCT was small in this preliminary analysis (n = 9), thus a spurious association due to chance alone cannot be definitely ruled out pending further dedicated data. By contrast, the association between chronic kidney disease and cefiderocol monotherapy could reflect the physicians’ decision, at least in some cases, not to administer cefiderocol together with potentially nephrotoxic agents (e.g., polymyxins, aminoglycosides) in patients with already impaired renal function, or the fact that these patients are at increased risk of urinary tract infections (for which a trend towards preference of monotherapy was observed in our study, albeit not statistically significant, as reported in Supplementary Table S3).

This present preliminary analysis of the CEFI-SITA study has some limitations to be acknowledged. The first is that, while we reported cumulative mortality in patients receiving cefiderocol in subgroups according to different causative organisms, the analysis was not primarily designed with this aim, thus the resulting unadjusted estimates (e.g., not adjusted for appropriateness of targeted therapy based on in vitro activity) should be interpreted with caution pending further data. However, the low cumulative mortality registered in infections by Enterobacterales, P. aeruginosa, and MBL producers is worth mentioning, and is in line with our previous findings of the possibly changing landscape in the treatment of CR-GNB registered in the past few years [58]. A second limitation is connected to the small sample size of patients treated with empirical cefiderocol. Indeed, although the registered 27.5% proportion of patients receiving empirical cefiderocol is solidly based on sample size estimates, subgroup proportions (within empirical therapy) have a larger degree of uncertainty and may require confirmation in further dedicated studies. Of note, this also includes the analysis of either crude or adjusted mortality in patients receiving empirical cefiderocol with subsequent isolation of CR-GNB as etiological agents, which has been deferred to a later phase of CEFI-SITA due to the limited sample size of this subgroup in this preliminary analysis. Third, no standardized microbiological approach was used for cefiderocol AST across participating centers, due to the observational, descriptive representation of daily routine practice. Fourth, owing to the expression of genes not commonly included among targets of rapid molecular tests, production of carbapenemases might have not been thoroughly evaluated in some cases (e.g., OXA-23 in A. baumannii). Additional studies employing WGS are therefore needed to decipher the molecular bases of carbapenem resistance in study isolates.

Conclusions

Cefiderocol is mainly used for targeted treatment in Italian hospitals, although empirical therapies account for more than 25% of prescriptions and should require dedicated standardization and guidance. The almost equal distribution of cefiderocol monotherapy and cefiderocol-based combination therapies underlines the need for further study to ascertain possible differences in efficacy between the two approaches.