Susceptibility to novel antimicrobial agents and molecular epidemiology of carbapenem-resistant enterobacterales: a 9-year study in Central China

Abstract

Purpose

This nine-year (2016–2024) surveillance study aimed to investigate the molecular epidemiology and antimicrobial susceptibility profiles of carbapenem-resistant Enterobacterales (CRE) in a tertiary hospital in Central China, with a focus on evaluating the in vitro efficacy of novel antimicrobial agents.

Methods

Antimicrobial susceptibility testing was conducted by broth microdilution to identify CRE isolates. From 1,403 non-duplicate CRE isolates(representing 13.4% of all Enterobacterales isolates recovered during the study period), 133 representative strains were selected for comprehensive molecular analysis. Carbapenemase genes were detected by PCR and sequencing, and multilocus sequence typing (MLST) was performed.

Results

Among the 1,403 CRE isolates, carbapenem-resistant Klebsiella pneumoniae (CRKP) predominated (65.5%). In the 133-strain subset, 97.7% harbored carbapenemase genes, primarily blaKPC-2 (82.7%). Cefiderocol and aztreonam-avibactam demonstrated 100% susceptibility against all tested strains. Against blaKPC-2 producers, susceptibility rates to ceftazidime-avibactam, imipenem-relebactam, and meropenem-vaborbactam were 92.7%, 96.4%, and 97.3%, respectively. The high-risk CRKP-ST11-blaKPC-2 clone dominated (81.2% of CRKP). Notably, 44.4% (8/18) of NDM-producing isolates and one IMP-4-producing isolate were susceptible to meropenem-vaborbactam, associated with low-level carbapenem resistance and absence of co-produced serine β-lactamases. Temporal variation in CRE proportion was observed, with a decline during 2020–2022 and an increase after 2023.

Conclusion

The CRKP-ST11-blaKPC-2 clone is the primary driver of CRE dissemination in our setting, highlighting the need for sustained infection control. Ceftazidime-avibactam, imipenem-relebactam, and meropenem-vaborbactam showed excellent activity against KPC producers, while cefiderocol and aztreonam-avibactam demonstrated potent activity against all CRE tested, including metallo-β-lactamase producers. The susceptibility of some NDM and IMP-4 producers to meropenem-vaborbactam underscores the heterogeneity within MBL-producing Enterobacterales and the importance of phenotypic susceptibility testing. Clinical studies are warranted to confirm the real-world efficacy and safety of these novel agents.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12866-026-05160-y.

Introduction

In 2024, the World Health Organization (WHO) updated its list of priority bacterial pathogens, categorizing carbapenem-resistant Enterobacterales (CRE) as a critical group requiring the highest priority for research and development [1]. Among CRE, carbapenem-resistant Klebsiella pneumoniae (CRKP) has emerged as a major public health challenge due to its multidrug resistance, global dissemination, and association with high morbidity and mortality, particularly among critically ill patients [2, 3].

In China, data from the Antimicrobial Surveillance Network (CHINET) indicate a dramatic increase in the resistance rates of K. pneumoniae to carbapenems, largely driven by the horizontal transfer of carbapenemase-encoding genes [4]. The primary mechanism of carbapenem resistance in CRE is the production of carbapenemases, with the class A serine β-lactamase KPC being predominant in many regions [5–7]. However, the dominant enzyme type exhibits geographical variation [8, 9].

For decades, therapeutic options for CRE infections have been severely limited, often relying on suboptimal agents such as polymyxins, tigecycline, and aminoglycosides, which are hampered by issues of efficacy and toxicity. The introduction of novel β-lactam/β-lactamase inhibitor combinations, such as ceftazidime-avibactam (CZA), imipenem-relebactam (IMR), and meropenem-vaborbactam (MEV), has improved treatment prospects for serine carbapenemase producers. Aztreonam-avibactam (ATA) was specifically designed to overcome metallo-β-lactamase (MBL)-mediated resistance, while cefiderocol (CDR), a siderophore cephalosporin, demonstrates broad activity against diverse carbapenemases [10].

Despite national surveillance data, detailed regional and hospital-specific epidemiological data from Central China, particularly regarding the impact of the COVID-19 pandemic and susceptibility to these novel agents, remain scarce. Therefore, this nine-year surveillance study aimed to elucidate the molecular epidemiology, resistance mechanisms, and antimicrobial susceptibility profiles, including to novel agents, of CRE in a large tertiary hospital in Henan, Central China, to inform local treatment and infection control strategies.

Materials and methods

Study design and bacterial isolates

This single-center retrospective study analyzed 1403 non-duplicated clinical isolates of CRE collected from inpatients at the Zhengzhou University Affiliated Chest Hospital between January 2016 and December 2024. According to the 2024 Clinical and Laboratory Standards Institute (CLSI) M100-S34 guidelines, CRE is defined as bacterial strains that exhibit resistance to at least one of meropenem, imipenem, or ertapenem [11].

Definition of non-duplicate isolates: For the purposes of this study, “non-duplicate” was defined according to standard epidemiological surveillance practices. Isolates were considered duplicate and excluded if they met any of the following criteria:

1. Same patient, same species, same admission: Multiple isolates of the same bacterial species from the same patient during a single hospital admission were considered duplicates, and only the first isolate was included .

2. Same patient, same species, 30-day window: For patients with multiple admissions or prolonged hospital stays, if the same species was isolated more than once within a 30-day period, only the first isolate was included .

3. Same patient, different species: Isolates of different species from the same patient (e.g., K. pneumoniae and E. coli) were considered independent isolates and both were included, as they represent distinct strain types and resistance mechanisms .

4. Surveillance vs. clinical distinction: Only isolates obtained for routine clinical diagnostic purposes (from infected or colonized sites) were included; active surveillance screening isolates (e.g., rectal swabs from asymptomatic patients) were not collected as part of this study .

This approach ensures that the 1,403 isolates represent the maximum one isolate per patient per species per 30-day period, minimizing bias from overrepresentation of patients with prolonged or recurrent infections while capturing the diversity of CRE species and resistance mechanisms encountered in clinical practice.

From this collection, 133 non-replicate CRE strains (applying the same definition as above: first isolate per patient per species per 30-day period) were selected for further molecular and susceptibility analysis based on the following criteria: (i) the first isolate from each patient per episode; (ii) representation of each year, major clinical departments, and specimen types; and (iii) confirmation of a carbapenem-resistant phenotype (MIC ≥ 4 µg/mL for meropenem or imipenem). All selected isolates were re-identified using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS, Zhongyuan, China) and stored at − 80 °C. To ensure the selected subset accurately reflected the overall epidemiology, a stratified random sampling strategy was employed. The total 1,403 isolates were stratified by year of isolation and species. From each stratum, isolates were randomly selected, with proportional representation of major clinical departments and specimen types. To enable robust molecular analysis of the dominant resistant clone, K. pneumoniae isolates were intentionally enriched (over-sampled) during the selection process. The representativeness of the final 133-strain subset was subsequently verified by comparing its distribution of species, years, departments, and specimen types with that of the entire 1,403 isolates (see Supplementary Table S2).

Definition of terms: Throughout this study, the term “CRE proportion” refers to the percentage of Enterobacterales isolates recovered from clinical specimens that met the CRE definition (resistance to meropenem, imipenem, or ertapenem). This is a laboratory-based proportion, not a population-based incidence rate (which would require denominators such as admissions, patient-days, or population size). Therefore, temporal changes in this proportion reflect shifts in the relative frequency of CRE among Enterobacterales isolates submitted for culture, which may be influenced by multiple factors including changes in patient volume, testing practices, healthcare-seeking behavior, and true changes in infection burden. The observational design of this study precludes causal inferences; any temporal associations discussed should be interpreted as ecological correlations requiring cautious interpretation.

Antimicrobial susceptibility testing

For the full cohort of 1,403 CRE isolates collected over the nine-year study period (2016–2024), antimicrobial susceptibility testing (AST) was consistently performed using the broth microdilution method to ensure methodological uniformity.The testing was performed using commercially prepared 96-well panels (e.g. DL-96 or BD-M50) containing serial two-fold dilutions of each antimicrobial agent, following the manufacturer’s instructions and CLSI standards for aerobic bacteria. The tested antibiotics included: aztreonam-avibactam (ATA), meropenem-vaborbactam (MEV), imipenem-relebactam (IMR), ceftazidime-avibactam (CZA), tigecycline (TGC),, colistin (COL), minocycline (MIN), fosfomycin (FOS), chloramphenicol (CHL), amikacin (AMK), imipenem (IPM), meropenem (MEM), ceftazidime (CAZ), aztreonam (ATM), cefepime (FEP), levofloxacin (LEV), ertapenem (ETP), ciprofloxacin (CIP), cefoperazone-sulbactam (CSL), and piperacillin-tazobactam (TZP). We conducted simultaneous susceptibility testing of both old and new drugs on the 133 CRE strains we selected, using the microbroth dilution method. Notably, the new drugs that warrant our special attention include aztreonam-avibactam (ATA), meropenem-vaborbactam (MEV), imipenem-relebactam (IMR), ceftazidime-avibactam (CZA) and cefiderocol (CDR).

Aztreonam-avibactam susceptibility testing

Avibactam was tested at a fixed concentration of 4 µg/mL in combination with two-fold serial dilutions of aztreonam (concentration range: 0.016–256 µg/mL) [12]. Aztreonam powder was obtained from Guangzhou GBCBIO Technologies Inc. (China), and avibactam powder was obtained from Nanjing Chemical Reagent Co., Ltd. (China). Stock solutions were prepared according to CLSI M07 guidelines. Avibactam was added to cation-adjusted Mueller–Hinton broth (CAMHB; BD-BBL, Becton-Dickinson, USA, supplied by BD Medical Equipment (Shanghai) Co., Ltd.) at a final concentration of 4 µg/mL immediately prior to panel preparation. Panels were prepared in-house, stored at -80 °C, and used within 30 days.

Breakpoint consistency and updates

A critical methodological consideration for longitudinal studies is the potential impact of changing CLSI breakpoints over time. The CLSI Subcommittee on Antimicrobial Susceptibility Testing publishes annual updates to the M100 supplement, which may include revised breakpoints for certain antimicrobial agents. To address this, we adopted the following approach: For each study year, MIC results were initially interpreted using the CLSI breakpoints in effect at the time of testing (e.g., CLSI M100-S26 [2016] through M100-S34 [2024]).To ensure comparability of resistance rates across the nine-year period, we subsequently re-interpreted all MIC data uniformly using the most recent CLSI M100-S34 breakpoints [11]. This retrospective re-interpretation eliminates the confounding effect of breakpoint changes and allows for valid temporal trend analysis.

Interpretive criteria were applied as follows: CLSI M100 2024 breakpoints for most agents [11]; FDA criteria for TGC [13]; EUCAST guidelines for COL [14]; For aztreonam-avibactam (ATA), in the absence of officially endorsed CLSI breakpoints during the study period, susceptibility was interpreted using aztreonam breakpoints, a common approach in recent literature prior to the availability of combination-specific breakpoints [15, 16]. This approach is supported by the mechanism of action of avibactam, which protects aztreonam from hydrolysis by serine β-lactamases, thereby restoring its intrinsic activity. The validity of this interpretation is further corroborated by the recent EUCAST-established ATA-specific breakpoints for Enterobacterales (susceptible ≤ 4 mg/L) [17, 18], against which our MIC distributions were retrospectively verified. Given the method-sensitive nature of cefiderocol susceptibility testing, specific attention was paid to the methodology for this agent. Cefiderocol minimum inhibitory concentrations (MICs) were determined by the broth microdilution (BMD) method using iron-depleted cation-adjusted Mueller–Hinton broth (ID-CAMHB), strictly following Clinical and Laboratory Standards Institute (CLSI) guidelines. Quality control strains included E. coli ATCC 25,922, P. aeruginosa ATCC 27,853, K. pneumoniae ATCC BAA-1705 (KPC-positive), and K. pneumoniae ATCC BAA-2146 (NDM-positive). All QC results were within CLSI-established acceptable ranges (e.g., for cefiderocol: 0.06–0.5 µg/mL for both E. coli ATCC 25922 and P. aeruginosa ATCC 27853) throughout the study period.

Phenotypic and genotypic detection of carbapenemases

Carbapenemase production was phenotypically detected using the modified carbapenem inactivation method (mCIM) and EDTA-carbapenem inactivation method (eCIM) [11]. Genomic DNA was extracted, and major carbapenemase genes (blaKPC, blaNDM, blaIMP, blaVIM, blaOXA-48-like) were screened by multiplex PCR [19]. Positive results for blaKPC and blaNDM were confirmed by full-length gene amplification and Sanger sequencing using established primers to determine specific alleles (e.g., blaKPC-2, blaNDM-1) [20, 21]. Isolates that tested negative for all carbapenemase genes were not further investigated for alternative resistance mechanisms (e.g., porin mutations, efflux pumps), representing a limitation of this study.

Multi-locus sequence typing (MLST)

MLST was performed for K. pneumoniae, Enterobacter cloacae, and Escherichia coli according to schemes available on the Pasteur Institute(https://www.pasteur.fr/recherche/genopole/PF8/mlst/Kpneumoniae.html), PubMLST(https://pubmlst.org/ecloacae/), and EnteroBase(http://mlst.warwick.ac.uk/mlst/dbs/Ecoli) websites, respectively. Sequence types (STs) were assigned by comparing the sequences of seven housekeeping genes to the respective databases.

Statistical analysis

Minimum inhibitory concentration (MIC) data were analyzed using WHONET 5.6 software to calculate MIC₅₀ and MIC₉₀ values. Statistical analyses were performed using SPSS version 25.0. Trends in resistance rates from 2016 to 2024 were assessed using the Linear Regression trend test. Categorical variables were compared using the Chi-square test or Fisher’s exact test, as appropriate. A p-value < 0.05 was considered statistically significant.After the bacterial strain nucleic acid extraction, genomic DNA was sequenced on an Illumina HiSeq X Ten platform (Illumina Inc., San Diego, CA, USA), producing 150 bp paired-end reads and a raw read coverage of ≥ 200-fold. De novo assembly of the resulting short reads was performed with SPAdes (version 3.10.0).

Results

Demographic and clinical characteristics

The demographic and clinical characteristics of the 1,403 CRE isolates are summarized in Supplementary Table S1. Patients were predominantly elderly (> 60 years, 60.80%) and male (71.85%), with most isolates originating from the intensive care unit (44.55%) and respiratory department (26.02%).

Distribution and trends of CRE species

From 2016 to 2024, CRE accounted for 13.4% (1,403/10,480) of all Enterobacterales isolates. K. pneumoniae was the most prevalent species (65.5%, 919/1,403), followed by E. cloacae (12.3%), K. oxytoca (5.8%), and E. coli (5.3%) (Fig. 1).The proportion of Enterobacterales isolates that met the CRE criteria increased from 2016 to 2018, stabilized at approximately 20% in 2018–2019, then significantly declined during the COVID-19 pandemic (2020–2022), reaching a nadir of 6.83% in 2020. An increase was observed post-2022, with the proportion of CRKP among Enterobacterales rising from 10.84% in 2022 to 14.8% in 2023. Notably, the proportion of carbapenem-resistant E. cloacae (CRECL) increased sharply in 2024 (Fig. 2).

Fig. 1.

Distribution of major CRE species isolated from 2016 to 2024 (total N = 1,403). ‘Others’ category includes: Proteus mirabilis, Raoultella ornitholytica, Serratia plymuthica, Morganella morganii, etc

Fig. 2.

Temporal trends in the proportion of Enterobacterales isolates meeting carbapenem-resistant criteria, by species (2016–2024). Annual isolate numbers are provided in Supplementary Table S3

Antimicrobial resistance trends (2016–2024)

Among the 1,403 CRE isolates, resistance rates to individual carbapenems varied. While all isolates met the CRE definition (resistance to at least one carbapenem), resistance to meropenem ranged from 70.6% to 99.1% across the study period, and resistance to imipenem ranged from 70.9% to 93.4%. This reflects that some isolates were resistant to ertapenem (the most sensitive CRE marker) but remained susceptible or intermediate to other carbapenems. Ertapenem susceptibility testing was introduced in our laboratory in 2021, and from 2021 to 2024, ertapenem resistance among CRE isolates consistently exceeded 97% .Analysis of resistance trends revealed significant decreases (p < 0.05) in resistance rates to cefoperazone/sulbactam、ampicillin/sulbactam、piperacillin/tazobactam、cefazolin、cefuroxime、ceftazidime、ceftriaxone、imipenem、meropenem、gentamicin and aztreonam. In contrast, resistance rates to tigecycline, ceftazidime-avibactam, and colistin increased significantly (p < 0.05)(Table 1).

Table 1.

Temporal trends in resistance rates (%) of antimicrobial agents against CRE, 2016–2024

Antibiotics	2016		2017		2018		2019		2020		2021		2022		2023		2024		Total		Trend P-value
	n	R(%)	n	R(%)	n	R(%)	n	R(%)	n	R(%)	n	R(%)	n	R(%)	n	R(%)	n	R(%)	n	R(%)
Cefoperazone/sulbactam	79	79.7	99	89.9	111	97.3	113	95.6	78	92.3	125	96.8	172	86.6	342	83.1	263	82.1	1383	87.3	0.024
Ampicillin/Sulbactam	79	94.9	99	99	111	100	115	99.1	78	100	127	100	176	94.9	349	91.4	269	89.2	1403	94.7	0.001
Piperacillin/tazobactam	79	92.4	99	96	111	97.3	115	98.3	78	97.4	127	97.6	176	90.3	349	83.4	269	81.8	1403	89.7	0.006
Cefazolin	79	96.2	99	99	111	99.1	115	99.1	78	100	127	100	175	97.7	349	95.4	268	91.8	1400	96.6	0.012
Cefuroxime	79	97.5	99	99	111	99.1	113	99.1	78	98.7	85	100	176	97.2	349	93.1	269	90.3	1359	95.5	0.005
Ceftazidime	79	88.6	99	94.9	111	97.3	115	98.3	78	98.7	127	96.9	176	89.2	349	82.2	269	83.6	1403	89.4	0.046
Ceftriaxone	79	94.9	99	99	111	99.1	115	98.3	78	98.7	127	99.2	176	96.0	349	91.9	269	90.0	1403	94.9	0.012
Cefepime	79	89.9	99	91.9	111	97.3	115	96.5	78	93.6	127	96.1	176	81.8	349	80.8	269	81.0	1403	87.0	0.064
Cefoxitin	79	88.6	99	94.9	111	99.1	113	96.5	78	98.7	85	97.6	176	92.0	349	87.6	269	84.4	1359	91.1	0.177
Ertapenem	-	-	-	-	-	-	-	-	-	-	53	98.1	167	99.4	348	97.7	266	98.1	836	98.2	0.848
Imipenem	79	81	99	77.8	111	86.5	115	93.4	78	87.2	125	86.4	175	77.1	343	72.9	252	69.8	1377	78.4	0.028
Meropenem	79	92.4	99	98	111	99.1	115	96.5	78	98.7	126	85.7	175	77.8	345	72.5	267	70.0	1395	82.4	0.008
Amikacin	79	30.4	99	64.6	111	76.6	115	71.3	78	74.4	127	60.6	176	46.6	349	54.7	269	45.7	1403	56.3	0.739
Gentamicin	79	79.7	99	83.8	111	91.0	115	81.7	78	79.5	127	81.9	176	65.3	349	37.3	269	66.2	1403	73.8	0.043
Ciprofloxacin	79	78.5	99	86.9	111	91.0	115	95.7	78	87.2	127	92.9	176	88.6	349	80.5	268	86.2	1402	86.5	0.717
Levofloxacin	79	78.5	99	83.8	111	89.2	115	92.2	78	87.2	127	89.8	174	83.9	347	77.2	269	82.9	1399	83.6	0.571
Sulfamethoxazole/ Trimethoprim	79	82.3	99	60.6	111	59.5	115	49.6	78	52.6	127	70.1	176	67.6	349	53.6	269	71.7	1403	62.6	0.999
Chloramphenicol	77	72.7	97	30.9	111	54.1	110	52.7	77	49.4	126	59.5	172	42.4	341	41.9	257	38.5	1368	46.1	0.124
Minocycline	79	39.2	99	20.2	111	3.6	115	5.2	78	7.7	85	47.1	176	35.8	349	34.7	269	47.9	1361	30.9	0.393
Tigecycline	-	-	42	0	104	0.9	105	0.9	73	1.4	127	7.9	174	9.2	349	11.7	259	13.1	1233	8.4	0.001
Ceftazidime/avibactam	-	-	-	-	-	-	-	-	47	14.9	46	26.1	-	-	117	21.4	205	33.6	415	27.5	0.011
Aztreonam	-	-	-	-	-	-	-	-	23	91.3	124	89.5	167	86.2	349	83.4	369	78.1	934	83.3	0.009
Colistin	-	-	-	-	-	-	-	-	78	5.1	124	4	175	8.0	349	16.0	269	13.0	1004	11.6	0.009

All MICs were determined by the broth microdilution method, the reference standard for antimicrobial susceptibility testing. CRE is defined as resistance to at least one carbapenem (meropenem, imipenem, or ertapenem). Therefore, isolates may be resistant to ertapenem but remain susceptible or intermediate to meropenem or imipenem, explaining why meropenem and imipenem resistance rates are less than 100% among CRE isolates. Ertapenem is the most sensitive marker for CRE detection, and its testing was introduced in our laboratory in 2021

The trend P-value was tested using Linear Regression trend test;

The “-” indicates that the drug has not been used or tested in the current year Processing “-” data (only using years with available data)

ICU vs. non-ICU isolates

CRE isolates from the ICU showed significantly higher resistance to imipenem, meropenem, and colistin compared to non-ICU isolates (p < 0.05). Conversely, ICU isolates had significantly lower resistance to ciprofloxacin, levofloxacin, trimethoprim-sulfamethoxazole, and aztreonam (p < 0.05) (Table 2).

Table 2.

Comparison of resistance profiles for antimicrobial agents: ICU vs. non-ICU CRE isolates (2016–2024)

Antibiotics	ICU Isolates(n = 625)	Non-ICU Isolates(n = 778)	P.Value
	R(%)	R(%)
Cefoperazone/sulbactam	91.6	91.7	0.9999
Ampicillin/Sulbactam	93.6	95.6	0.098
Piperacillin/tazobactam	88.9	90.4	0.3732
Cefazolin	95.4	97.5	0.0542
Cefuroxime	94.3	96.5	0.0688
Ceftazidime	88.6	89.9	0.5119
Ceftriaxone	93.3	95.8	0.0585
Cefepime	85.0	88.6	0.0837
Cefoxitin	89.7	92.2	0.1294
Imipenem	81.8	75.7	0.0054
Meropenem	85.3	80.0	0.0089
Amikacin	54.6	57.7	0.2567
Gentamicin	71.2	75.8	0.0843
Ciprofloxacin	81.4	90.6	< 0.0001
Levofloxacin	78.2	87.9	< 0.0001
Sulfamethoxazole/Trimethoprim	58.7	65.7	0.007
Chloramphenicol	45.1	46.9	0.5276
Minocycline	27.3	33.8	0.0077
Tigecycline	7.2	9.1	0.2185
Ceftazidime/avibactam	27.6	26.8	0.7586
Aztreonam	79.9	85.3	0.0081
Colistin	14.9	9.6	0.0021

Representativeness of the selected CRE subset

To validate the selection process, the demographic and clinical characteristics of the 133 selected isolates were compared with the full cohort of 1,403 isolates (Supplementary Table S2). The annual distribution of the subset closely mirrored the yearly fluctuation of CRE incidence (P = 0.995), confirming unbiased temporal sampling. The departmental distribution (P = 0.947) and specimen type distribution (P = 0.434) also showed no significant differences between the subset and the total collection, indicating that the selected strains are representative of the overall clinical sources of CRE in our hospital. As expected, the species distribution differed significantly (P < 0.001), with a higher proportion of K. pneumoniae in the subset (88.0% vs. 65.5%), reflecting the intentional enrichment of this dominant species for detailed molecular analysis.

Carbapenemase genotypes

Among the 133 strains selected for molecular analysis, 130 (97.7%) were carbapenemase-positive. blaKPC-2 was the predominant gene (82.7%, 110/133), overwhelmingly found in K.pneumoniae (109 strains). blaNDM variants (NDM-1, -5, -9) were detected in 18 strains (13.5%), mainly in E.cloacae and K.oxytoca. One K.pneumoniae strain co-produced KPC-2 and NDM-1, and one produced IMP-4 (Table 3).

Table 3.

Detection of 133 CRE carbapenemase genotypes

Genotype	K.pneumoniae		E.cloacae	K.oxytoca		E.coli		R.ornitholytica	C.freundii	S.marcescens		Total n(%)
KPC-2		109								1	110(82.7)
NDM-1		1	6		4		1	1	1		14(10.55)
NDM-5		1					2				3(2.25)
NDM-9		1									1(0.75)
IMP-4		1									1(0.75)
KPC-2 + NDM-1		1									1(0.75)
Negative*		3									3(2.25)
Total		117	6		4		3	1	1	1	133

*Note: Among the 133 isolates, three (2.25%) tested negative for all screened carbapenemase genes (blaKPC, blaNDM, blaIMP, blaVIM, blaOXA-48-like). Resistance in these isolates may be mediated by alternative mechanisms, such as porin loss (e.g., OmpK35/OmpK36 mutations) combined with extended-spectrum β-lactamase (ESBL) or AmpC β-lactamase production, or overexpression of efflux pumps, which were not investigated in this study

Antimicrobial susceptibility profiles

The detailed MIC distributions are provided in Supplementary Table S4, and susceptibility rates by carbapenemase type are shown in Table 4. Cefiderocol and aztreonam-avibactam exhibited 100% susceptibility. Against blaKPC-2 producers, susceptibility rates were high for IMR (96.4%), MEV (97.3%), and CZA (92.7%). In contrast, NDM-producing strains showed poor susceptibility to IMR (0%) and CZA (0%), but remained susceptible to ATA (100%) and CDR (100%). Notably, 44.4% (8/18) of NDM-producing isolates and one IMP-4-producing isolate were susceptible to meropenem-vaborbactam (MEV).

Table 4.

Drug resistance and sensitivity rate of CRE strains with different enzyme producing types (%)

Antib-iotic	All stains (n = 133)					KPC-2 (n = 110)					NDM likes (n = 18)					KPC + NDM (n = 1)	IMP-4 (n = 1)	Others* (n = 3)
	R	I	S	MIC 50	MIC 90	R	I	S	MIC 50	MIC 90	R	I	S	MIC 50	MIC 90	R/S	R/S	R	I	S
CDR	0	0	100	2	2	0	0	100	2	2	0	0	100	1	2	S	S	0	0	100
ATA	0	0	100	0.25	0.5	0	0	100	0.25	0.5	0	0	100	0.125	0.25	S	S	0	0	100
IMR	13.5	6	80.5	0.5	4	0.9	2.7	96.4	0.5	0.5	72.2	27.8	0	4	16	R	S	100	0	0
MEV	5.3	7.5	87.2	1	8	0.9	1.8	97.3	1	2	22.2	33.3	44.4	8	16	R	S	66.7	33.3	0
TGC	3	0	97	1	2	2.7	0	97.3	1	2	5.6	0	94.4	0.5	2	S	S	0	0	100
CZA	21.1	0	78.9	2	> 256	7.3	0	92.7	1	4	100	0	0	> 256	> 256	R	R	0	0	100
MNO	15.8	39.1	45.1	8	16	11.8	40.9	47.3	8	16	38.9	27.8	33.3	8	32	R	R	33.3	0	66.7
COL	12.8	0	87.2	0.125	8	12.7	0	87.3	0.125	8	11.1	0	88.9	0.125	16	S	S	33.3	0	66.7
CHL	66.2	8.2	25.6	> 256	> 256	63.6	10	26.4	> 256	> 256	72.2	0	27.8	256	> 256	R	R	100	0	0
FOS	58.6	15.8	25.6	256	> 256	65.5	18.1	16.4	256	> 256	22.2	5.6	72.2	32	> 256	S	S	66.7	0	33.3
IPM	78.9	3.8	17.3	32	64	78.2	2.7	19.1	32	64	83.3	11.1	5.6	8	16	R	S	100	0	0
AMK	72.2	0.7	27.1	256	256	84.5	0	15.5	256	256	11.1	5.6	83.3	4	256	R	S	0	0	100
MEM	89.5	7.5	3	128	128	91.8	6.4	1.8	128	128	77.8	16.6	5.6	8	16	R	S	100	0	0
CAZ	94	0	6	256	> 256	92.7	0	7.3	256	> 256	100	0	0	> 256	> 256	R	R	100	0	0
FEP	96.2	3	0.8	128	256	96.4	2.7	0.9	128	256	100	0	0	64	> 256	R	I	100	0	0
ETP	97.7	0.8	1.5	128	256	98.2	0	1.8	256	256	100	0	0	16	64	R	I	100	0	0
CIP	98.5	0.7	0.8	64	128	99.1	0	0.9	64	128	94.4	5.6	0	32	128	R	R	100	0	0
LVX	96.2	3	0.8	32	64	99.1	0	0.9	32	64	88.9	11.1	0	16	64	R	I	66.7	33.3	0
CSL	99.2	0	0.8	> 256	> 256	99.1	0	0.9	> 256	> 256	100	0	0	> 256	> 256	R	R	100	0	0
ATM	94.7	0.8	4.5	> 256	> 256	100	0	0	> 256	> 256	61.1	5.6	33.3	32	> 256	R	R	100	0	0
TZP	99.2	0.8	0	> 256	> 256	100	0	0	> 256	> 256	100	0	0	> 256	> 256	R	I	100	0	0

Abbreviations: CDR Cefiderocol, ATA Aztreonam-avibactam, IMR Imipenem-relebactam, MEV Meropenem-vaborbactam, TGC Tigecycline, CZA Ceftazidime-avibactam, MNO Minocycline, COL Colistin, CHL Chloramphenicol, FOS Fosfomycin, IPM Imipenem, AMK Amikacin, MEM Meropenem, CAZ Ceftazidime, FEP Cefepime, ETP Ertapenem, CIP Ciprofloxacin, LVX Levofloxacin, CSL Cefoperazone-sulbactam, ATM Aztreonam, TZP Piperacillin-tazobactam, R Resistant, S Susceptible, MIC50/90 Minimum inhibitory concentration required to inhibit 50%/90% of isolates

The “Others” category includes three strains that tested negative for the carbapenemase genes screened

Molecular epidemiology and sequence types

MLST of 117 CRKP isolates identified 11 distinct STs. ST11 was dominant (81.2%, 95/117), and all ST11 isolates harbored blaKPC-2. Other STs included ST2237 (7.7%) and ST15 (2.6%). The single IMP-producing strain belonged to ST446, and the KPC-2 + NDM-1 strain was ST2237. E. coli and E. cloacae strains exclusively carried blaNDM and belonged to diverse STs (Table 5).

Table 5.

Carbapenemase distribution determined by CRE MLST

Organism	MLST	No.	KPC-2 n(%)	NDM-1n(%)	NDM-5n(%)	NDM-9 n(%)	IMP-4 n(%)	KPC-2+ NDM-1 n(%)	NA n(%)
Klebsiella pneumoniae (n = 117)	ST11	95	95(100)
	ST2237	9	8(88.9)					1(11.1)
	ST15	3	3(100)
	ST268	1	1(100)
	ST307	1	1(100)
	ST1998	1		1(100)
	ST784	1			1(100)
	ST1922	1				1(100)
	ST73	1							1(100)
	ST446	1					1(100)
	ST48	1	1(100)
	Not Typed	2							2(100)
Escherichia coli (n = 3)	ST6388	1		1(100)
	ST410	1		1(100)
	ST3489	1			1(100)
Enterobacter cloacae (n = 6)	ST171	2		2(100)
	ST418	1		1(100)
	ST25	1		1(100)
	ST336	1		1(100)
	ST78	1		1(100)

Abbreviations: MLST Multi-locus sequence typing, ST Sequence type

Characteristics of MEV-susceptible MBL-producing isolates

Among the MBL-producing isolates, one IMP-4-producing K. pneumoniae (ST446) was susceptible to meropenem-vaborbactam (MEV) with an MIC of 1 µg/mL, while remaining resistant to ceftazidime-avibactam (MIC = 256 µg/mL). This isolate also exhibited low-level carbapenem resistance, with meropenem and imipenem MICs of 1 µg/mL, and was susceptible to aztreonam-avibactam and cefiderocol (MICs ≤ 0.016 and 1 µg/mL, respectively). The complete susceptibility profile of this isolate is shown in Table 4 and Supplementary Table S5. Among the 18 NDM-producing isolates (excluding the KPC-2/NDM-1 co-producer), 8 (44.4%) were susceptible to meropenem-vaborbactam (MEV) with MIC ≤ 4 µg/mL, 6 (33.3%) were intermediate (MIC = 8 µg/mL), and 4 (22.2%) were resistant (MIC ≥ 16 µg/mL) (Table 4). The MEV-susceptible NDM isolates included diverse species and sequence types: E. cloacae ST171 and ST336, E. coli ST410 and ST3489, K. oxytoca (two isolates), K. pneumoniae ST1998, and C. freundii ST22 (Supplementary Table S5). All eight MEV-susceptible NDM isolates exhibited low-level meropenem resistance, with meropenem MICs ranging from 0.5 to 4 µg/mL, suggesting that NDM production alone, in the absence of additional resistance mechanisms, may not confer high-level carbapenem resistance.

Antimicrobial profile of ST11 vs. non-ST11 CRKP

ST11 CRKP isolates exhibited a significantly broader and more severe resistance profile compared to non-ST11 isolates. Resistance rates to fosfomycin, amikacin, meropenem, aztreonam, cefepime, and levofloxacin were significantly higher in ST11 strains (p < 0.05). Specifically, we have verified that ST11 isolates exhibited significantly higher susceptibility to imipenem-relebactam (97.9% vs. 68.2%, p < 0.05) and meropenem-vaborbactam (98.9% vs. 72.8%, p < 0.05) compared to non-ST11 isolates, consistent with the table data.(Table 6).

Table 6.

Comparison of resistance and susceptibility between ST11 and non-ST11 Klebsiella pneumoniae isolates

Antibiotic	ST11 CRKP(N = 95)			non-ST11 CRKP(N = 22)			*P -value
	R(%)	I(%)	S(%)	R(%)	I(%)	S(%)
IMR	0	2.1	97.9	27.3	4.5	68.2	0.000
ATA	0	0	100	0	0	100	1.000
CDR	0	0	100	0	0	100	1.000
MEV	1.1	0	98.9	13.6	13.6	72.8	0.021
TGC	3.2	0	96.8	0	0	100	1.000
MNO	6.3	42.1	51.6	36.4	45.5	18.2	0.001
CZA	8.4	0	91.6	22.7	0	77.3	0.068
COL	12.6	0	87.4	13.6	0	86.4	0.899
CHL	63.2	9.5	27.4	81.8	4.5	13.6	0.094
FOS	72.6	21.1	6.3	18.2	0	81.8	0.000
IPM	80	1.1	18.9	72.7	9.1	18.2	0.564
AMK	89.5	0	10.5	40.9	4.6	54.5	0.000
CAZ	91.6	0	8.4	100	0	0	0.349
MEM	97.9	2.1	0	59.1	27.3	13.6	0.000
ETP	97.9	0	2.1	95.5	4.5	0	0.468
ATM	100	0	0	90.9	4.5	4.5	0.034
FEP	100	0	0	77.3	18.2	4.5	0.000
CIP	100	0	0	95.5	4.5	0	0.188
LVX	100	0	0	86.4	13.6	0	0.006
TZP	100	0	0	95.5	4.5	0	0.188
CSL	100	0	0	95.5	0	4.5	0.188

Abbreviations: IMR Imipenem-relebactam, ATA Aztreonam-avibactam, CDR Cefiderocol, MEV Meropenem-vaborbactam, TGC Tigecycline, MNO Minocycline, CZA Ceftazidime-avibactam, COL Colistin, CHL Chloramphenicol, FOS Fosfomycin, IPM Imipenem, AMK Amikacin, CAZ Ceftazidime, MEM Meropenem, ETP Ertapenem, ATM Aztreonam, FEP Cefepime, CIP Ciprofloxacin, LVX Levofloxacin, TZP Piperacillin-tazobactam, CSL Cefoperazone-sulbactam, R Resistant, I Intermediate, S Susceptible

*P-values were calculated using the Chi-square test or Fisher’s exact test where appropriate (e.g., when expected cell counts were < 5)

The shaded area denotes the period of the COVID-19 pandemic.Values represent the percentage of Enterobacterales isolates (by species) that met CRE criteria each year. These are laboratory-based proportions, not incidence rates; temporal associations should be interpreted cautiously due to potential confounders (see Discussion). CRKP, carbapenem-resistant Klebsiella pneumoniae; CRECL, carbapenem-resistant Enterobacter cloacae; CRKO, carbapenem-resistant Klebsiella oxytoca; CRECO, carbapenem-resistant Escherichia coli.

Discussion

This study describes temporal trends in the proportion of CRE among Enterobacterales isolates in a single center over nine years in Central China. It is important to emphasize that these are laboratory-based proportions, not population-based incidence rates, and the observational design precludes causal inferences. The temporal associations discussed below should be interpreted with caution, as multiple unmeasured confounders may influence the observed trends.

The 13.4% CRE proportion in our center exceeds the CHINET national average, underscoring a serious local resistance challenge [4, 22]. CRKP predominance (65.5%) aligns with global trends [1, 23]. CRE proportion declined during 2020–2022, coinciding with China’s dynamic zero-COVID policy. This temporal association likely reflects multiple concurrent factors: enhanced infection control, reduced high-risk admissions, altered healthcare-seeking behavior, and changed antibiotic prescribing [6, 24]. However, without standardized denominators (admissions, patient-days) or antibiotic consumption data, these remain plausible explanations rather than proven causes. The post-2023 increase coincided with pandemic restriction relaxation and normalized healthcare utilization, potentially reflecting both endemic clone resilience and confounding from resumed procedures and increased admissions. Given the observational design and multiple confounders, these associations should be interpreted as ecological correlations. Studies with standardized denominators and multivariable analyses are needed to elucidate drivers of CRE trends.

The molecular epidemiology reveals clonal expansion as the primary driver of CRE spread in our setting, with ST11-blaKPC-2 accounting for over 80% of CRKP. This mirrors the epidemiological landscape in many Chinese hospitals but contrasts with the ST258 dominance often seen in the western countries [25–27]. The association of ST11 with a significantly broader multidrug-resistant profile underscores its success and the challenge it poses for therapy. The presence of a KPC-2 + NDM-1 co-producing strain, though rare, signals the concerning convergence of multiple resistance mechanisms and underscores the need for rapid diagnostics to guide therapy.

Our susceptibility data offer critical insights for treatment. An important methodological consideration in this study pertains to the interpretation criteria for aztreonam-avibactam (ATA). During the initial analysis period, officially endorsed CLSI breakpoints for ATA were not available. Therefore, we employed aztreonam breakpoints for susceptibility interpretation, a pragmatic approach adopted by several recent studies investigating ATA activity prior to the establishment of combination-specific breakpoints [15, 16]. The rationale for this approach is grounded in the mechanism of action: avibactam inhibits a broad range of serine β-lactamases (including ESBLs, KPC, and AmpC), thereby protecting aztreonam from hydrolysis and restoring its intrinsic activity against Enterobacterales [16, 18]. Consequently, in the presence of avibactam, the susceptibility of an isolate to the combination is essentially determined by its intrinsic susceptibility to aztreonam, assuming no other resistance mechanisms are present.

This methodological approach has since been validated by the recent establishment of official ATA breakpoints. In 2024, EUCAST assigned clinical MIC breakpoints for ATA against Enterobacterales: susceptible ≤ 4 mg/L and resistant > 4 mg/L [17, 18, 28]. These breakpoints were supported by comprehensive analyses integrating microbiological surveillance data (99% of > 100,000 Enterobacterales isolates had ATA MICs ≤ 8 mg/L), pharmacokinetic/pharmacodynamic target attainment simulations, and clinical outcomes from phase 3 trials (REVISIT and ASSEMBLE) [18, 29]. Retrospective verification of our MIC data against these EUCAST breakpoints confirmed that all 133 isolates in our study (MIC range 0.125-1 mg/L) would be categorized as susceptible, consistent with our original findings using aztreonam breakpoints. This concordance underscores the validity of our approach and reinforces the conclusion that ATA exhibits exceptional in vitro activity against CRE in our setting, including both KPC- and NDM-producing strains. The upcoming CLSI M100 Ed36 is expected to include key recommendations for ATA testing, which will further standardize susceptibility interpretation for this important novel agent [30].

In contrast, cefiderocol(CDR), as an iron carrier cephalosporin, has a unique ‘Trojan horse’ mechanism that may give it PK/PD advantages in the treatment of infections such as lung infections. Two pivotal international randomized controlled trials (e.g., CREDIBLE-CR and APEKS-NP) have demonstrated that cefiderocol exhibits clinical cure and microbiological eradication rates comparable to the best available therapy in treating Gram-negative bacterial infections, including those caused by carbapenem-resistant Klebsiella pneumoniae (CRKP) [31, 32]. Although no resistant strains were detected in the present study, there is a growing number of global reports documenting cefiderocol resistance, with identified mechanisms including mutations in iron uptake systems, loss of outer membrane porins, and overexpression of β-lactamases [33–36].

An unexpected finding in this study was that 44.4% (8/18) of NDM-producing isolates and the single IMP-4-producing isolate were susceptible to meropenem-vaborbactam (MEV). While vaborbactam does not inhibit metallo-β-lactamases (MBLs), susceptibility to MEV in MBL producers has been increasingly recognized and can be explained by several mechanisms [37–40]. For the NDM-producing isolates, susceptibility to MEV was associated with low-level carbapenem resistance (meropenem MICs 0.5–4 µg/mL), Porin alterations: Mutations in porin genes (e.g., OmpK35/OmpK36) can reduce carbapenem uptake, lowering intrinsic MICs even in MBL producers and the absence of co-produced serine β-lactamases [37–39]. Notably, the MEV-susceptible NDM isolates exhibited diverse species and MLST types (ST171, ST336, ST410, ST1998, etc.), indicating that this phenotype is not restricted to a single clone. The IMP-4-producing isolate (ST446) merits special attention. Its susceptibility to MEV (MIC = 1 µg/mL) is consistent with early reports by Chu et al., who observed that some IMP-4 carriers remained carbapenem-susceptible due to loss of gene expression [41]. More recent studies have demonstrated that MEV exhibits differential activity against individual MBLs, with the order IMP > VIM > NDM-1 [40, 42]. This hierarchy reflects intrinsic differences in enzyme kinetics: IMP-4 has a lower catalytic efficiency against meropenem (relative hydrolysis rate only 8% of that for benzylpenicillin) compared to NDM-type enzymes [43]. The low meropenem MIC (1 µg/mL) of this isolate suggests that IMP-4 production alone, in the absence of additional resistance mechanisms, may not confer high-level carbapenem resistance.

These findings have important clinical and research implications. First, they highlight the heterogeneity within MBL-producing Enterobacterales and underscore the limitations of relying solely on carbapenemase gene detection for predicting susceptibility to novel β-lactam/β-lactamase inhibitor combinations. Second, the differential activity of MEV against IMP-type enzymes suggests that this agent may have a role in treating infections caused by IMP producers with low-level carbapenem resistance, although clinical studies are urgently needed. Third, the relative susceptibility of IMP enzymes to vaborbactam-containing combinations provides a rationale for future drug development, potentially by optimizing the boronic acid scaffold to enhance activity against a broader range of MBLs [40].

While our in vitro findings are encouraging, they should not be directly extrapolated to clinical recommendations without supporting evidence from prospective studies. The correlation between in vitro susceptibility and clinical outcomes for these novel agents in CRE infections requires further investigation, particularly in patient populations with complex underlying conditions.

This study has several limitations. First, its single-center design may limit the generalizability of findings to other regions or hospital settings. Second, the lack of environmental sampling and asymptomatic colonization data prevents a comprehensive understanding of transmission reservoirs and dynamics. Third, and most importantly, our molecular analysis was limited to PCR-based detection of major carbapenemase genes. This approach left the resistance mechanisms of three carbapenemase-negative isolates unexplained; potential mechanisms such as porin deficiencies (e.g., OmpK35/OmpK36 mutations) or efflux pump overexpression, often in concert with ESBL or AmpC production, were not investigated [44, 45]. Future studies employing WGS are essential to provide deeper genomic resolution, elucidate the complete resistance and virulence landscapes, and inform more precise infection control and treatment strategies.

Conclusion

In conclusion, the CRKP-ST11-blaKPC‑2 clone is the cornerstone of CRE endemicity in our hospital. The temporal decline in CRE proportion during the COVID-19 pandemic highlights complex interactions between healthcare practices and resistance, though causal conclusions are precluded by the laboratory-based data and multiple confounders. Sustained infection control measures targeting successful clones remain essential. The exceptional in vitro activity of cefiderocol and aztreonam-avibactam against CRE, including metallo-β-lactamase producers, supports their therapeutic potential. The observation that 44.4% of NDM-producing isolates were susceptible to meropenem-vaborbactam highlights the complex interplay between carbapenemase production and other resistance mechanisms, and underscores the importance of phenotypic susceptibility testing for guiding therapeutic decisions.However, clinical studies with patient outcome data are needed to confirm real-world efficacy and safety. Future research should integrate genomic surveillance with clinical trials to translate these findings into evidence-based strategies.

Authors’ contributions

Conceptualization: Y.H., Q.W.; Methodology: Y.H.; Software: Z.L.; Validation: H.C.; Formal Analysis: Y.H., Y.Z., H.C., X.Q.; Investigation: Y.W., S.X., W.H.; Resources: Y.W., S.X., W.H.; Data Curation: Z.L., J.L.; Writing – Original Draft: Y.H.; Writing – Review & Editing: Q.W., Y.H., W.W.; Visualization: J.L.; Supervision: Y.Y.; Project Administration: W.W.; Funding Acquisition: Y.H. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the Medical Science and Technology Project of Henan Province (grant No. LHGJ20190759).

Data availability

The datasets generated and/or analysed during the current study are available in the NCBI BioProject repository under accession numbers PRJNA1460529, PRJNA1460326, andPRJNA1460451.These can be accessed at: https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1460529, https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1460326, https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1460451.

Declarations

Ethical approval and consent to participate

All procedures performed in this study involving human participants were in accordance with the ethical standards of the institutional research committee, the Declaration of Helsinki (as revised in 2013), and the Measures for Ethical Review of Biomedical Research Involving Human Subjects issued by the National Health Commission of China (2016). This study was reviewed and approved by the Institutional Ethics Committee of the Affiliated Chest Hospital of Zhengzhou University (Approval No: 202007033). Owing to the retrospective nature of the study and the use of anonymised clinical and laboratory data that posed no risk to participants, the requirement for written informed consent was waived by the Institutional Ethics Committee of the Affiliated Chest Hospital of Zhengzhou University. All patient identifiers were removed from the dataset prior to analysis, and data were accessed only by the authorised research team.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Clinical trial number

Not applicable.