Resistance profiles of carbapenemase-producing Enterobacterales in a large centre in England: are we already losing cefiderocol?

Abstract

Background

Carbapenemase-producing Enterobacterales (CPE) pose difficult therapeutic challenges. We aimed to characterize antimicrobial resistance profiles of CPE in our centre.

Methods

All non-duplicate CPE isolates between 1 August 2020 and 31 August 2023 in a large teaching trust in England were retrospectively studied. Cefiderocol antimicrobial susceptibility testing (AST) was performed using disc diffusion, ceftazidime/avibactam using disc diffusion and gradient diffusion, and ceftazidime/avibactam aztreonam synergy using the double disc diffusion method. EUCAST version 14.0 breakpoints were used.

Results

A total of 158 CPE from 136 patients were isolated. Most patients were colonized with CPE, but only 16.9% had active infections. Thirty-day all-cause mortality was 10.3%, increasing to 13% for patients with infections and to 18.2% for bacteraemias. OXA-48 was the most prevalent carbapenemase (48.1%), followed by NDM (38%). All isolates exhibited MDR profiles, with high levels of resistance to meropenem (41.1%). Resistance to cefiderocol was found in 69.7% of NDM-producing isolates, with a further 18.2% in the area of technical uncertainty. Ceftazidime/avibactam and aztreonam synergy was seen in 87.5% of isolates, whereas colistin and fosfomycin susceptibility remained high (98.1% and 97.2%, respectively). All OXA-48-producing isolates were susceptible to ceftazidime/avibactam, and 15.3% were resistant to cefiderocol. No patients had been exposed to cefiderocol beforehand, whereas three had been exposed to ceftazidime/avibactam. The most common risk factor for CPE isolation was travel and receiving healthcare abroad, especially in Asia.

Conclusions

We found high rates of resistance to cefiderocol in CPE isolates without prior cefiderocol exposure. Our results prohibit empirical use of cefiderocol for the treatment of CPE infections in our setting.

Introduction

Carbapenemase-producing Enterobacterales (CPE) are a global health concern due to their high resistance rates to many commonly used antimicrobials including carbapenems. Infections with CPE are frequently associated with significant treatment delays, and mortality has been reported to approach 40%.^1,2 Due to this fact, the WHO has listed CPE as critical priority pathogens for the development of new antimicrobials.³

Until recently, treatment options for CPE infections were sparse, and relied on a limited pool of agents, including amikacin, colistin, tigecycline and fosfomycin.⁴ These antimicrobials are frequently associated with significant toxicity, and their in vivo efficacy is often questioned, despite in vitro susceptibility.⁵ Since 2017, novel β-lactams and β-lactam/β-lactamase inhibitor combinations have entered the UK market, providing valuable treatment options for CPE infections. Ceftazidime/avibactam was the first novel antimicrobial to be approved, having activity against serine carbapenemases [OXA-48, Klebsiella pneumoniae carbapenemase (KPC)], as well as MBLs [New Delhi metallo-β-lactamase (NDM), Verona integron-encoded metallo-β-lactamase (VIM), imipenemase (IMP)] when combined with aztreonam.⁶ The combination of meropenem/vaborbactam followed in 2019, having high activity against KPC-producing strains, which are less common in the UK.⁷ Finally, in 2020, the siderophore cephalosporin cefiderocol was launched, which has activity against all carbapenemase classes, including MBLs.⁸ All three new agents have demonstrated promising results, showing superior efficacy in CPE infections compared with older agents.^9–11 For this reason, the most recent IDSA guidance and ESCMID guidelines recommend them as first-line treatments in monotherapy for CPE infections.^12,13

Despite the encouraging results, resistance to novel agents remains a significant concern. Reports of serine carbapenemase producers resistant to both ceftazidime/avibactam and meropenem/vaborbactam have emerged, and cefiderocol has been shown to be affected by a variety of resistance mechanisms.^14,15 For this reason, continuous surveillance of resistance trends to new agents is required, especially if they are to be used as monotherapy, in order to detect rapid rises in resistance and inform local guidelines and practice. Despite this need, large strain collections from England are lacking due to the low incidence of infections in each centre and lack of publicly available surveillance data.

In this study, we aimed to investigate the epidemiology and resistance profile of CPE infections in a large UK tertiary referral hospital over a period of 3 years. Our goal was to inform local treatment guidelines post the introduction of the three novel CPE treatments and describe the cohort of patients in whom CPE are isolated in our centre.

Methods

Ethical approval

This study was approved by the departmental audit lead and divisional director. As it is a retrospective review of anonymized clinical data, we had confirmation from UCLH Joint Research Office (Research Directorate) that further formal ethical approval was not required.

Study setting

This study was conducted in University College Hospitals NHS Foundation Trust (UCLH), a group of tertiary referral hospitals in central London, with a total inpatient bed capacity of 1022 beds. The Trust is a referral centre for haematology and haematopoietic cell transplantation in North and East London and West Essex, being one of the largest haemato-oncology treatment centres in Europe. The Hospital of Tropical Diseases and the National Hospital for Neurology and Neurosurgery are also part of the Trust.

Study population

All confirmed CPE isolated from cultures of clinical and screening samples from any anatomical site reported by the UCLH microbiology laboratory between 1 August 2020 and 31 August 2023 were eligible for inclusion in this study. Candidate isolates were identified by interrogating the local laboratory information system (WinPath Enterprise, Clinisys, Woking, UK). Duplicate isolates, defined as an isolate of the same species producing the same carbapenemase at any timepoint after the initial isolate, were excluded. No other exclusion criteria applied.

Isolate identification and antimicrobial resistance testing

The Microbiology laboratory is a joint venture between UCLH and Health Services Laboratories (HSL) and is United Kingdom Accreditation Service accredited.

During the study period, surveillance cultures for CPE colonization from rectal swabs were performed using the CHROMagar mSuperCARBA agar (CHROMagar, Paris, France). Enterobacterales from clinical specimens were screened for carbapenemase production using meropenem (cut-off MIC for investigation >0.125 mg/L or zone diameter <28 mm) as per the UK Standards for Microbiological Investigations.¹⁶ Species identification was performed with the MALDI-TOF MS method using the Bruker Biotyper^® system (Bruker Daltonics, Billerica, MA, USA). Suspected CPE isolates were tested using the RESIST-3 O.K.N. K-SeT (Coris BioConcept, Gembloux, Belgium), an immunochromatographic assay that detects three common carbapenemases (OXA-48, KPC and NDM), and from November 2021 with the NG-Test^® CARBA-5 (NG-Biotech Laboratories, Guipry, France), an immunochromatographic assay that detects the five most common carbapenemase producers (KPC, IMP, NDM, VIM and OXA-48).¹⁷ Samples with unexplained carbapenem resistance and a negative immunochromatographic assay were referred to the UK Heath Security Agency Antimicrobial Resistance and Healthcare Associated Infection reference laboratory, where they were further investigated with an in-house multiplex real-time PCR for carbapenemase genes.¹⁸

With regards to antimicrobial susceptibility testing (AST), blood, sterile site and rectal surveillance culture isolates were tested using the BD Phoenix automated identification and susceptibility testing (BD, Franklin Lakes, NJ, USA). Urine, wound and sputum isolates were tested using the EUCAST disc diffusion (DD) method. Once isolates were confirmed CPE, AST with second-line agents was performed. This was clinician driven in the beginning of the study period (therefore not universal), and later protocolized as follows: for OXA-48- and KPC-producing isolates, AST to cefiderocol and ceftazidime/avibactam was performed; and for MBL-producers, AST for cefiderocol, as well as ceftazidime/avibactam and aztreonam synergy, was performed. AST to colistin was performed for all amikacin-resistant isolates as well as if requested by clinicians. Cefiderocol susceptibility testing was performed using 30μg cefiderocol discs (Liofilchem, Abruzzi, Italy) on standardized supplemented Mueller–Hinton agar plates. Ceftazidime avibactam susceptibility testing was performed using either DD (Thermo Fisher Scientific Oxoid, Waltham, MA, USA) or MIC test strips (Liofilchem, Abruzzi, Italy). Ceftazidime/avibactam aztreonam synergy testing was performed using the double DD method (aztreonam discs from Thermo Fisher Scientific Oxoid, Waltham, MA, USA).¹⁹ Broth microdilution (BMD) was used to determine colistin MIC (ComASP^™® Colistin panel, Liofilchem, Abruzzi, Italy). Quality control strains for AST were used according to the 2024 EUCAST version 14.0 for all antimicrobials. In particular, for cefiderocol E. coli ATCC 25922 was used; for ceftazidime/avibactam E. coli ATCC 25922 and Klebsiella pneumoniae ATCC 700603 were used; and for aztreonam/avibactam synergy testing Klebsiella pneumoniae NCTC 13443 was used. Susceptibility results were interpreted retrospectively using 2024 EUCAST version 14.0 breakpoints and guidelines.

Data sources and measurement

Clinical data were collected from electronic medical records (Epic Systems Corporation, Verona, WI, USA). Ethnicity was recorded as white, Asian, black, mixed or other as defined in the 2021 UK Census. Comorbidities were recorded as outlined by the International Severe Acute Respiratory and Emerging Infections Consortium.²⁰ Immunosuppression was defined according to Chapter 6 of the UK Green Book.²¹

Infections were defined retrospectively by previously described standard criteria.²² Other positive cultures were considered colonization. Briefly, a patient with a CPE-positive culture was considered to have an infection if the isolate was from blood or any other sterile site. For patients with positive respiratory cultures, criteria outlined by the American Thoracic Society and the IDSA were used.^23,24 For infections in urine or surgical wounds, CDC/National Healthcare Safety Network criteria were applied.²⁵ Patients with cultures from non-surgical wounds were considered infected using previously published clinical criteria.²⁶ All rectal surveillance cultures were considered colonization. In UCLH, rectal surveillance cultures are routinely performed for CPE patient contacts, patients admitted under haematology and intensive care, as well as patients who are being transferred from other hospitals, in the UK or abroad, or who received healthcare in the UK or abroad within 12 months of admission. Ad hoc testing is also performed in patients with other risk factors or receiving broad-spectrum antimicrobials under the guidance of the local infection physicians.

Outcomes

The following outcomes were recorded for all patients with eligible CPE isolates: 30 day all-cause mortality, 90 day all-cause mortality, unsuccessful discharge at 30 days, readmission within 30 days, and readmission within 90 days. The following additional outcomes were recorded for patients with CPE infections: microbiological failure within 30 days, microbiological failure within 90 days, any stage acute kidney injury within 14 days, Clostridioides difficile infection within 30 days, ICU admission within 14 days, and effective empirical treatment. Microbiological failure was defined as re-isolation of the same CPE at least 72 h after the start of in vitro effective treatment. Empirical treatment was defined as the antimicrobials given prior to the time of the antibacterial susceptibility report, and effectiveness was extrapolated from in vitro susceptibility results.

Statistical analysis

Statistical analysis was performed using SPSS v29 (IBM Corp., Armonk, NY, USA). Continuous variables were expressed as medians with IQRs, categorical variables as percentages. Univariable comparisons of categorical variables were made using Pearson’s chi-squared test. We calculated 95% CIs using 10 000 bootstrap samples. The level of statistical significance was set at 0.05. No formal power analysis was performed as all available patients were enrolled. This study has been reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.²⁷

Results

Patient cohort characteristics and outcomes

During the study period, a total of 289 CPE were isolated in the study centre. After removal of duplicates, 158 samples from 136 unique patients were included in the final analysis. Fourteen patients isolated two different species of CPE, and four patients isolated three different species of CPE. Patient cohort characteristics are shown in Table 1. The median age of the cohort was 60 years (IQR 47–69) with an even sex distribution. White (47.8%) and Asian (41.2%) were the most common ethnicities. High rates of immunosuppression (36%), obesity (33.1%) and diabetes (27.2%) were seen, and many patients had solid (25%) or haematological (20.6%) malignancies. Patients with CPE were primarily admitted under intensive care (31.6%), surgery (29.4%), haematology (26.5%) or medicine (24.3%) (Figure S1; available as Supplementary data at JAC Online). Most patients were only colonized with CPE (83.1%), whereas 23 patients (16.9%) had active infections, including 11 bacteraemias (8.9%). Most CPE (65.2%) were detected in a sample taken more than 48 h after admission to hospital. The median length of stay in hospital for all patients was prolonged (median 21.5 days, IQR 9–41), and 32.6% remained inpatients at 30 days post the index CPE culture. Readmission rates were also high at 30 days (12.5%) and 90 days (21.3%). For the entire cohort 30 day all-cause mortality was 10.3%, increasing to 13% for patients with infections and to 18.2% for patients with bacteraemias. Of patients with infections, 17.4% were admitted to the ICU with a median length of stay of 5 days. The median time from sample collection to effective treatment was 57 h 27 min (IQR 16 h 31 min to 127 h 39 min). Empirical treatment contained at least one in vitro effective antimicrobial in 52.2% (12/23) of cases.

Table 1.

Patient characteristics (N = 136)^a

Age, y	60.3 (47.3–69.2)
Sex
Male	72 (52.9)
Female	64 (47.1)
Ethnicity
White	65 (47.8)
Asian	56 (41.2)
Black	13 (9.6)
Other	2 (1.4)
Comorbidities
Immunosuppression	49 (36)
Obesity	45 (33.1)
Diabetes	37 (27.2)
Solid malignancy	34 (25)
Chronic cardiovascular disease	32 (23.5)
Haematological malignancy	28 (20.6)
Chronic respiratory disease	15 (11)
Chronic renal disease	10 (7.4)
Chronic liver disease	9 (6.6)
Asthma	7 (5.1)
Rheumatological disease	6 (4.4)
Dementia	1 (0.7)
Pregnancy	1 (0.7)
Age-adjusted Charlson comorbidity index	4 (2–6)
Episode type
Rectal: colonization	101 (74.3)
Blood: infection	11 (8.1)
Urine: infection	9 (6.6)
Urine: colonization	9 (6.6)
Intra-abdominal: infection	2 (1.5)
Respiratory: colonization	2 (1.5)
Wound: colonization	1(0.7)
Bone: infection	1(0.7)
Carbapenemase type
OXA-48	62 (45.6)
NDM	53 (39)
NDM & OXA-48	12 (8.8)
IMP	5 (3.7)
KPC	3 (2.2)
VIM	1 (0.7)
Admission origin
Home	61 (44.9)
Hospital transfer from the UK	49 (36)
Tested as outpatients	15 (11)
Hospital transfer outside the UK	8 (5.9)
Long-term care facility	3 (2.2)
Disposition
Home	82 (60.3)
Death as inpatient	17 (12.5)
Remained outpatient	15 (11)
Discharged to long-term care facility	14 (10.3)
Hospital transfer within the UK	8 (5.9)
Length of stay in hospital (n = 121)b	21.5 (9–41.3)
Outcomes—all patients
30-day all-cause mortality	14 (10.3)
90-day all-cause mortality	29 (21.3)
Unsuccessful discharge at 30 days (n = 121)b	44 (32.6)
Readmission within 30 days	17 (12.5)
Readmission within 90 days	29 (21.3)
Outcomes—patients with infections (n = 23)
30-day all-cause mortality	3 (13)
90-day all-cause mortality	5 (21.7)
Unsuccessful discharge at 30 days	12 (52.2)
Readmission within 30 days	2 (8.7)
Readmission within 90 days	6 (26.1)
Remains on CPE treatment at 30 days	3 (13)
Microbiological relapse at 30 days	4 (17.4)
Microbiological relapse at 90 days	4 (17.4)
Acute kidney injury (any stage)	5 (21.7)
Clostridioides difficile infection	1 (4.4)
Intensive care admission	4 (17.4)

CPE, carbapenemase-producing Enterobacterales.

^aContinuous variables are presented as median (IQR), categorical variables as n (%).

^bExcludes patients tested as outpatients.

CPE characteristics and resistance profile

In the 158 isolated CPE (Table S1), Escherichia coli was the most common species (48.1%), followed by Klebsiella spp. (30.5%), Enterobacter cloacae complex (12%) and Citrobacter spp. (7.6%). OXA-48 was the most prevalent carbapenemase (48.1%), followed by NDM (38%), IMP (3.8%), KPC (1.9%) and VIM (0.6%). Twelve isolates tested positive for both NDM and OXA-48 enzymes (7.5%). The temporal distribution of the different samples is shown in Figure S2. No CPE outbreaks were declared during the study period.

With regards to AST (Figure 1a, Table S2), all isolates exhibited an MDR profile with high levels of resistance to meropenem (41.1%, 65/158) and piperacillin/tazobactam (100%, 158/158). Second-line agents were tested for most but not all isolates, reflecting local testing policy as described in the Methods section (Figure 1a–c). Resistance profiles of NDM-producing and dual NDM- and OXA-48-producing agents were analysed together as they exhibited similar resistance rates (Figure 1b). In NDM-producing strains, high levels of resistance to cefiderocol were noted (69.7%), and an additional 18.2% of strains tested within the area of technical uncertainty (ATU) of 21 and 22 mm. Therefore, only 12.1% of NDM-producing isolates were susceptible to cefiderocol using the 2024 EUCAST version 14.0 DD breakpoint. Ceftazidime/avibactam and aztreonam synergy was seen in 87.5% of isolates tested. Susceptibility to colistin (98.1%) and fosfomycin (97.2%) remained high, as did susceptibility to nitrofurantoin (88.6%). Resistance levels to amikacin and tigecycline were higher, and only 68.1% and 52.9% of strains, respectively, were susceptible, although many isolates were not tested to tigecycline (52.8%). Overall, resistance rates did not differ by infection site.

Figure 1.

AST results of study isolates. (a) All isolates (N = 158); (b) NDM-producing and NDM- and OXA-48-producing isolates (n = 72); (c) OXA-48-producing isolates (n = 76). Numbers represent percentages of all tested isolates for that antimicrobial and do not take into account non-tested isolates. Ceftazidime/avibactam and aztreonam S and R represent presence or absence of in vitro synergy, respectively, rather than formal susceptibility breakpoints.¹⁹ AMK, amikacin; ATM, aztreonam; CAZ, ceftazidime; CIP, ciprofloxacin; CRO, ceftriaxone; CST, colistin; CZA, ceftazidime/avibactam; CZA & ATM, ceftazidime/avibactam and aztreonam; ETP, ertapenem; FDC, cefiderocol; FEP, cefepime; FOF, fosfomycin; GEN, gentamicin; I, susceptible, increased exposure; IPM, imipenem; MEC, mecillinam; MEM, meropenem; NIT, nitrofurantoin; R, resistant; S, susceptible; SXT, trimethoprim/sulfamethoxazole; TGC, tigecycline; TMC, temocillin; TZP, piperacillin/tazobactam; U, area of technical uncertainty. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

OXA-48-producing isolates (Figure 1c) were universally susceptible to ceftazidime/avibactam in our cohort (100%). Resistance to cefiderocol in these strains was again observed, with only 74.4% testing susceptible and an additional 10.3% in the ATU. Amikacin exhibited higher susceptibility rates (94.7% susceptibility) compared with colistin (90.5% susceptibility). Many OXA-48-producing isolates were susceptible to meropenem and third- and fourth-generation cephalosporins (Figure 1c), yet the clinical efficacy of these antimicrobials for the treatment of OXA-48 infections remains to be determined. Susceptibility results by species are shown in Table S3.

Cefiderocol DD diameters for all CPE, NDM- and OXA-48 producing isolates are shown in Figure 2(a–c). If the 2023 FDA/CLSI DD breakpoints were applied only 2.7% (3/113) of all isolates would have tested resistant to cefiderocol (4.5%, 3/66 of NDM-producers; and none of the OXA-48 producers), and a further 11.5% (12/113) would have tested in the ATU (16.7%, 11/66 of NDM producers; and 2.6%, 1/39 of OXA-48 producers) (Figure 2a–c).

Figure 2.

Cefiderocol disc diffusion diameter results for (a) all study isolates (N = 113); (b) NDM-producing and NDM- and OXA-48-producing isolates (n = 66); and (c) OXA-48-producing isolates (n = 39), according to 2024 EUCAST breakpoints version 14.0. 2023 FDA/CLSI breakpoints are also shown for comparison. Green: susceptible; Orange: area of technical uncertainty; Red: resistant. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

With regards to risk factors for CPE acquisition (Figure 3), almost half of the patients (49.3%) had travelled outside the UK in the last 12 months, most of them receiving healthcare in their destination (36.8%). Destinations of travel (n = 69) included countries in Asia (72.5%), Africa (15.9%), Europe (10.1%) and Latin America (1.5%). The most common countries of travel included Kuwait (18.8%), Turkey (14.5%), India (13%) and Bangladesh (7.2%). Patients with NDM isolates were more likely to have received healthcare abroad compared with patients with OXA-48 isolates (47.7% versus 31.1%, P = 0.04). One-quarter of cases (27.9%) had been admitted to the ICU over the last 12 months or had been admitted to hospital at least three times over the same time period (15.4%). Only 17.6% of cases had documented exposure to carbapenems over the last 12 months (Figure 3). Very few patients had received ceftazidime/avibactam (n = 3) or colistin (n = 2) over the last 12 months, and none had received cefiderocol.

Figure 3.

Risk factors for carbapenemase-producing Enterobacterales isolation in study patients (N = 136). Univariable comparisons of categorical variables were made using Pearson’s chi-squared test; 95% CIs were calculated using 10 000 bootstrap samples; *P = 0.04. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

Discussion

In this study, we investigated a recent unselected cohort of CPE from a single centre in a low-incidence setting for antimicrobial resistance (AMR). Our results revealed very high resistance rates to cefiderocol in strains from cefiderocol-unexposed patients, whereas ceftazidime/avibactam and ceftazidime/avibactam with aztreonam demonstrated good in vitro activity against OXA-48- and NDM-producing isolates, respectively.

Our results have significant implications. Cefiderocol is recommended as a first-line treatment option for severe infections with MBL-producers both by IDSA and ESCMID.^12,13 Yet, empirical use would be inappropriate in our setting given our findings and highlights the importance of monitoring local and national epidemiology. Despite that, CPE resistance rates for England are not routinely published by the UK Health Security Agency.²⁸ High levels of cefiderocol resistance in English strains have been found before in a highly selected collection of isolates reported to a central reference laboratory between 2008 and 2018.⁸ Our collection is much more recent and includes consecutive strains, making it more representative. Other studies found lower resistance rates.²⁹ Globally, concerns have been expressed regarding increasing cefiderocol resistance, especially in CPE, with a recent meta-analysis showing a 40% resistance rate, which aligns with our findings.³⁰ Monitoring resistance to cefiderocol in the future will be important, especially in England, where the drug is currently reimbursed via a delinked subscription payment model.³¹

Our study also highlights critical discrepancies in the current cefiderocol resistance testing landscape between EUCAST and FDA/CLSI, especially with regards to DD testing. Using the 21 mm 2024 EUCAST version 14.0 breakpoint, 69.7% of NDM-producing strains were resistant to cefiderocol, compared with 4.5% using the 9 mm FDA/CLSI 2023 breakpoint. This is a significant variation that is likely to lead to differences in the use of cefiderocol in countries adopting breakpoints from one or the other regulator. It should be noted that the EUCAST breakpoint was informed by a 2022 independent EUCAST-led study of 263 Enterobacterales, including many resistant isolates.³² The study showed that most isolates with a DD diameter below 21 mm were categorized as resistant after BMD testing, although some susceptible strains were observed within the 18–20 mm range.³² No strain under 18 mm was sensitive when tested by BMD.³² On the contrary, the FDA/CLSI breakpoint has not been reviewed since February 2021 and is primarily informed by the results of the cefiderocol licensing studies, where very few resistant strains were observed, as well as pharmacokinetic/pharmacodynamic data from the neutropenic mouse model, and in vitro data submitted by the manufacturing company.^33–36 This suggests that EUCAST recommendations for cefiderocol DD testing might be more up to date.³⁷ Our work shows that further studies are needed to develop clinical consensus on the breakpoints used for cefiderocol resistance testing.³⁸

We tested the combination of ceftazidime/avibactam and aztreonam as a treatment option for MBL-producers and found that 87.5% of isolates showed evidence of synergy. Although no universally agreed methodology for testing this combination exists,³⁹ and synergy does not necessarily equal susceptibility, we used a standardized double DD approach for all our strains through the study period.⁴⁰ Our results suggest that the ceftazidime/avibactam combination with aztreonam is likely to be effective for the treatment of MBL-producing CPE in our setting, as suggested by the IDSA guidance.¹² Yet it is limited by the lack of standardized breakpoints, and a lack of synergy has been found in NDM-producing isolates with PBP3 mutations.⁴¹ The synergistic combination has been shown to be a good predictor of susceptibility to the soon-to-be-marketed aztreonam/avibactam, which might represent an important treatment option for CPE in the near future.⁴²

Finally, our study generated some important data for infection prevention and control of CPE in hospital settings. Most of our patients isolated CPE as inpatients and had prolonged hospital stays, with a median of 21 days. Readmission rates at 30 and 90 days were also high, and a significant proportion were admitted to intensive care. The prolonged healthcare contact of CPE-colonized patients increases the risk of transmission in healthcare settings and subsequent invasive infections, requiring significant efforts from medical and nursing teams to control them, and straining isolation room capacity.⁴³ We also found a significant association of CPE acquisition, especially NDM-producers, with travel and healthcare contact, especially in some Asian countries. Although this might partly reflect our testing policy, CPE colonization has been described to be more common in travellers to Asia and highlights the need for global solutions to effectively tackle AMR.⁴⁴

Our study has notable strengths. It is a real-world study where AST results to last-line agents for CPE were performed according to the latest EUCAST recommendations and were used to inform clinical management of CPE infections in real time. The study investigated a very recent cohort of unselected consecutive isolates (thus minimizing selection bias), starting from the period when cefiderocol was first introduced to the UK in autumn 2020. The cohort of patients was very well characterized, especially with regards to risk factors for CPE isolation, which was facilitated by availability of electronic patient records. Cefiderocol DDs were recorded and results were adjusted to the latest EUCAST guidelines to reflect the most recent evidence. DD results were presented in detail so future adjustments can be made if breakpoints change further.

Our study also has limitations. We did not perform BMD to confirm our findings from DD, as this is not part of our routine lab testing. Existing commercially available assays have high rates of inaccuracy, and for this reason EUCAST recommends the use of DD testing for cefiderocol, which is what our lab has used.⁴⁵ This is the method all labs in the UK are currently using for cefiderocol susceptibility testing, and most of Europe to our knowledge. Interpretation of cefiderocol BMD results can be challenging, and a gold standard is not universally agreed. In addition, it is worth noting that reproducibility of DD results is a concern with Acinetobacter spp. and Pseudomonas aeruginosa, but not with Enterobacterales.³⁸ Our study is a clinical microbiology study from a ‘real-life setting’ using methods validated in our lab for clinical use. For this reason, we believe our findings are highly relevant to all the settings that also use DD for cefiderocol AST and have implications for the clinical use of cefiderocol in these settings, even if the reference AST method was not used.

Other limitations include that we did not test all isolates to all antimicrobials, which might skew results towards more resistant strains. Despite this, we did test more than 90% and 80% of NDM producers for cefiderocol and ceftazidime/avibactam with aztreonam, respectively, which is a meaningful sample proportion. We cannot exclude that high resistance levels are due to a few circulating mobile genetic elements causing the same resistance profile in different species in our centre, as we did not sequence the isolates, yet the temporal association of our isolates was not suggestive of transmission events. Finally, carbapenem exposure outside of our centre was not consistently documented in the medical record and this may have led to an underestimation of this risk factor.

Conclusion

In summary, we found high rates of cefiderocol resistance using DD in a recent cohort of cefiderocol-unexposed patients with CPE. Monitoring of resistance rates to novel antimicrobials and harmonizing cefiderocol AST will be crucial in the global fight against AMR.

Funding

This study was carried out as part of our routine work.

Transparency declarations

I.B. reports investigator-initiated research funding from Shionogi B.V. for research on Gram-negative infections outside the submitted work; honoraria from Shionogi B.V. and Menarini UK; and participating in an advisory board for GSK on the UK subscription model for antimicrobials outside the submitted work. All remaining authors report no conflicts of interest.

Supplementary data

Figures S1 and S2 and Tables S1 to S3 are available as Supplementary data at JAC Online.