Genomic investigation of the vancomycin-resistant Enterococcus faecium in a tertiary hospital in China

Abstract

Objectives

In light of a VREfm spread in a large tertiary hospital in China, this study aimed to explore the transmission dynamics and genetic characteristics of these isolates within a global context.

Methods

A total of 27 VREfm were collected from a hospital in Guangdong province, China. Antimicrobial susceptibility testing, whole-genome sequencing, and transmission dynamics analysis were performed. Additionally, global sequence type 80 (ST80) data were gathered for genetic comparison, and complete plasmid sequences were reconstructed to characterize the genetic environment surrounding the vanA genes.

Results

The study identified VREfm ST80 as the main agent of the patient infection, with the exception of four ST789 isolates identified on environmental surfaces within one bed. The genomes displayed high clustering, with similar profiles of virulence factors and antimicrobial resistance genes. The study observed that genetically similar VREfm strains were distributed across different beds. Moreover, VREfm ST80 from the hospital clustered with other Chinese isolates and exhibited a higher number of VFs compared to other lineages. Although vanA genes were present in different STs, the vanA-carrying plasmids shared highly similar backbone structure, with Tn1546-associated elements detected in each.

Conclusion

This study sheds light on the genetic epidemiology of a VREfm spread in a Chinese hospital through comparison with global isolates. To control the spread of VREfm within hospitals, the implementation of active screening programs and rigorous infection control measures is essential.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13756-026-01755-2.

Introduction

Enterococcus faecium, a common species residing in the human gut, has increasingly become an important healthcare-associated pathogen in recent years, responsible for a wide range of infections [1]. The increasing prevalence of E. faecium poses a significant treatment challenge due to its inherent resistance to a broad spectrum of antibiotics. Historically, vancomycin served as the primary therapeutic agent for E. faecium infections [2]. However, the evolution of antibiotic resistance and horizontal gene transfer (HGT) via mobile genetic elements has led to the emergence of vancomycin-resistant E. faecium (VREfm). Currently, the most clinically significant resistance gene clusters are vanA and vanB, which are associated with the transposons Tn1546 and Tn1549, respectively [3].

E. faecium, particularly VREfm, is a leading cause of healthcare-associated infections, underscoring its increasing clinical burden [4]. For instance, in the Guangdong region of China, the prevalence of VREfm colonization among patients has risen sharply from approximately 5% before 2021 to 20–50% in 2023 [5]. This epidemiological shift is accompanied by a change in dominant clones. Specifically, the previously prevalent sequence types (STs) ST17 and ST78 have been largely replaced by ST80 since 2021 [5]. This trend is not confined to China, as a similar increase in VREfm colonization, exceeding 20% of patients, has been reported in German hospitals [6]. Although hospital-acquired VREfm may exhibit lower virulence than other multidrug-resistant organisms, its persistence in the hospital environment and specific metabolic adaptations contribute to its widespread dissemination [7]. Whole-genome sequencing (WGS) is a critical tool for characterizing the population structure and genomic features of VREfm in healthcare settings [8]. It enables the estimation of putative transmission links through methods such as multilocus sequence typing (MLST) and single nucleotide polymorphism (SNP)-based analysis [9, 10]. Furthermore, WGS facilitates the reconstruction and quantification of the most probable dissemination pathways of van gene clusters [11]. While previous studies have established the value of WGS for VREfm epidemiological surveillance and genetic features, the transmission dynamics of within-hospital spread, particularly those involving environmental reservoirs, and the global drivers of van gene dissemination remain to be fully elucidated.

In this study, we describe a spread of VREfm in the intensive care unit (ICU) of a Chinese hospital. Our aim is to elucidate the epidemiology and genomic characteristics using WGS within the context of the global VREfm landscape. By reconstructing complete plasmids, we investigated the transmission dynamics within the hospital and identified potential drivers of vancomycin resistance dissemination.

Methods

Sample collection

All samples and clinical data were collected between April and July 2023 at a large tertiary teaching hospital in Guangdong Province, China, which has a capacity of over 1000 beds. A spread of VREfm was identified in the hospital’s ICU, which comprises 12 beds with a nurse station responsible for patient care within the ICU. Since the first VREfm isolate was obtained from a patient’s urine in the ICU on April 13, 2023, there have been consistent reports of VREfm among patients in both the ICU and other departments. A total of 11 VREfm isolates were collected from 10 patients across various clinical samples, including blood, urine, wound swabs, secretion, and pleural effusion. The environmental samples are numbered according to the corresponding ICU bed. In instances where multiple VREfm isolates with identical susceptibility profiles were obtained from the same source in a single patient, only the first isolate collected was included in the genomic analysis.

To enhance infection control and prevent further spread of the VREfm outbreak within the hospital, an environmental screening was conducted on July 13, 2023. A total of 135 samples were collected from various sources, including patient monitoring equipment, bedside equipment, urinary catheterization devices, body surfaces, foot of the bed, bedside trash cans, injection pumps, and drawing curtains. Environmental surfaces were sampled using sterile cotton swabs, with each swab placed in 5 mL of streptococcal broth (Kailin, Guangdong, China) for enrichment. Following overnight incubation at 35 °C, the enriched broth was cultured onto Columbia agar (Dijing, Guangdong, China) and incubated aerobically for 24 h at 35 °C. Isolates were obtained and identified using VITEK^® 2 Compact system (BioMerieux, Marcy l’Etoile, Lyon, France), and their vancomycin resistance was assessed in accordance with the breakpoints outlined in the Clinical and Laboratory Standards Institute (CLSI-M100, 33rd edition) guidelines. All selected isolates were preserved and stored in the MicroBank at – 80 °C for further analysis.

Antimicrobial susceptibility testing

The minimum inhibitory concentration (MIC) for each isolate was determined using the VITEK^® 2 Compact system, following the CLSI-M100 guideline breakpoints [12]. The antibiotic panel tested included penicillin, ampicillin, high-concentration gentamicin, high-concentration streptomycin, erythromycin, linezolid, tigecycline, vancomycin, tetracycline, ciprofloxacin, levofloxacin, and nitrofurantoin. To confirm vancomycin resistance, E-test strips (Taikang, Wenzhou, China) were employed, following the CLSI M100 guidelines. The antimicrobial susceptibility testing control strain used was Enterococcus faecalis ATCC 29,212.

Genome sequencing and assembly

Bacterial genomic DNA was extracted using the STE method, as previously described [13]. For DNA library preparation, 0.2 µg of DNA per sample was utilized. The quality of the prepared libraries was assessed using the Agilent 5400 system (Agilent, USA). All libraries that met the quality criteria were pooled and subsequently sequenced on the NovaSeq 6000 platform (Illumina, USA) employing a paired-end 150 bp sequencing strategy. Quality control of the short reads was conducted with FastQC [14]. De novo assembly was performed using SPAdes v3.15.3 with the—isolate flag [15]. Assembly statistics and quality assessment were conducted using SeqKit v2.6.0 [16] and CheckM v1.16 [17], respectively. To further understand the genetic context surrounding the vanA genes, two isolates, VREfm1054 and VREfm10-3, were selected for long-read sequencing. These isolates represent two dominant STs involved in patient infections. DNA library preparation was conducted using the Native Barcoding Kit 24 V14 (SQK-NBD114.24) and sequencing was performed on the Nanopore MinION platform with flow cell R10.4.1 (Oxford Nanopore Technologies, UK). Long reads were trimmed using FiltLong (https://github.com/rrwick/Filtlong) and then hybrid assembled using Unicycler v0.5.1 [18]. The quality of the assembled genomes was assessed with CheckM v1.16 [17].

Split k-mer analysis to decipher putative transmission

SKA v1.0 [19] was utilized to generate k-mer files (k = 15) for each input isolate from the short-read data with default settings and without recombination filtering. Genome-to-genome pairwise single nucleotide polymorphism (SNP) distances were then calculated using ska distance, also with default settings and without any recombination filtering. In the context of SKA, a SNP is defined as the “number of split k-mers found in both samples where the middle base is an A, C, G, or T but differs between k-mer files.” Unlike traditional core genome alignment, SKA compares isolate pairs based on all genomic data, providing the high-resolution analysis necessary to infer patient-to-patient transmission [9, 20]. Following prior research on optimizing genomic approaches for identifying VREfm putative transmission in healthcare settings [9], we adopted an SKA threshold of ≤ 7 SNPs to ensure that no potential transmission pairs were overlooked. We validated this by calculating pairwise SNP distances across alternative thresholds, which revealed that the number of identified genome pairs stabilized at 7 SNPs (Fig. S1). Consequently, supported by both the literature and our comparative analysis, we defined ≤ 7 SNPs as the definitive threshold for genomic relatedness. However, relying solely on SNP distance for transmission inference is not robustly supported. Therefore, we further corroborated within-hospital transmission through three epidemiological links: (a) samples collected from the same bed, (b) samples from different beds with VREfm collected within a 30-day interval, and (c) samples from different beds managed by the same nurse. Because multiple environmental surface samples were collected within a single bed, analyzing them individually could overestimate putative transmission events. To minimize false-positive epidemiological links, we grouped all environmental surface samples from the same bed into a single unit, allowing us to more accurately estimate putative transmission events and to focus specifically on patient-to-patient and patient-to-environment transmission.

Dataset curation

Enterococcus faecium genomes were downloaded using the NCBI dataset command-line tools [21] with the—exclude-atypical parameter. Genomes lacking interpretable metadata for isolation source, date, or country were excluded from the analysis. Additionally, GenBank assemblies that duplicated those in the RefSeq assemblies were discarded. Further filtering of GenBank assemblies was performed using CheckM2 v1.1.0 [17], eliminating genomes with less than 95% completeness or greater than or equal to 5% contamination. The taxonomic check of all genomes was conducted using GTDB-Tk v2.5.2 [22]. The remaining 4232 ST80 genomes, which included our 27 genomes typed using mlst v2.23.0 [23], were dereplicated using Assembly-Dereplicator v0.3.2 (https://github.com/rrwick/Assembly-Dereplicator) at a Mash distance of 0.0015, which approximated to 99.85% ANI [24], balancing retention of unique strains against redundancy [25]. The dereplication process resulted in 696 representative genomes for further analysis.

Population analysis

The 696 representative genomes were clustered using fastbaps v1.0.8 [26] with the Dirichlet prior hyperparameters set to “optimised.baps.” Antimicrobial resistance genes (ARGs) were annotated using AMRfinderplus v4.0.23 [27]. Virulence factors (VFs) were identified using Abricate v1.0.1 (https://github.com/tseemann/abricate) against the VFDB 2022 [28] and a set of 35 curated hospital-associated virulence factor genes [29]. Gene enrichment analysis was performed using Scoary v1.6.16 [30] on major fastbaps clusters containing 69 or more representative genomes, constituting at least 10% of the total.

Phylogenetic analysis

The 696 genomes were aligned to the E. faecium KresVSE0002 reference genome (RefSeq accession: GCF_003408975.2) using Snippy v4.6.0 (https://github.com/tseemann/snippy). Recombination was subsequently removed with Gubbins v3.3.5 [31]. A core genome SNP phylogenetic tree was reconstructed using IQ-TREE v3.0.1 [32], employing the GTR + I+G substitution model. This was followed by 1000 ultrafast bootstrap replicates (-B) and optimization with nearest-neighbour interchange (-bnni). The phylogenetic tree was visualized using iTOL v6 [33], incorporating summarized metadata from the non-dereplicated dataset of 4232 genomes.

Plasmid analysis

The alignment of plasmids pVREfm1054 and pVREfm10-3 was visualized using Easyfig v2.2.2 [34]. To assess the presence of pVREfm1054-like and pVREfm10-3-like plasmids within the 696 dereplicated assemblies, these were aligned to pVREfm1054 and pVREfm10-3 following the previous published methods [35]. Briefly, the assemblies were aligned against the pVREfm1054 and pVREfm10-3 plasmids using ABRicate v1.0.1. For all multi-copy insertion sequence (IS) elements, BLASTn searches were conducted to query the plasmid sequences against themselves, identifying regions with 100% identity duplication. The hits from the Abricate output of assemblies against plasmids were cross-checked with these duplicated regions to identify any overlapping or missed hits. All mapping outputs were further concatenated and imported into an R script provided by Reid et al. [35]. The boundaries of IS elements on the linear map were further refined using ISEScan v1.7.3 [36]. Visualization was accomplished using ggtree v1.14.6 [37] in combination with linear maps of plasmids generated by Benchling (https://benchling.com/) and annotated with Bakta v1.11.4 [38].

Statistical analysis

Non-parametric analysis using Kruskal-Wallis test, followed by Dunn’s post-hoc test (Holm-adjusted), was performed using R v4.5.1 to compare differences in the number of VFs and ARGs among the major fastbaps clusters, with statistical significance set at p < 0.05. Non-parametric analysis was justified by violated normality (Shapiro-Wilk test) and homoscedasticity (Levene’s test) assumptions.

Results

Epidemiological investigation

On April 5, 2023, VREfm was isolated from a urine sample of an 83-year-old female patient (Patient 5, Fig. 1a) who had undergone surgical intervention in the ICU. Subsequently, consistent reports of VREfm-positive cultures emerged from both ICU patients and those in other hospital departments. In total, VREfm colonization was identified in ten consecutive patients, with one patient (Patient 4) exhibiting colonization at multiple sites, including the wound and urine samples. The demographic and clinical characteristics of the study population are detailed in Table S1. The median age at admission was 76 years (IQR: 60–83), and six patients were admitted to the ICU. The median length of hospitalization was 57 days (IQR: 14–90), with VREfm detection occurring a median of 27 days after initial admission. Prior to VREfm colonization, the majority of patients underwent surgical interventions (80%, 8/10), followed by the use of central venous catheters (50%, 5/10). All patients were subsequently discharged from the hospital or returned to a state of health. Among the patients, eight acquired infections, while the remaining two cases represented colonization. Following the identification of VREfm colonization within the hospital, the department of disease control and prevention initiated VREfm surveillance in the ICU. A total of 135 environmental samples were collected, of which 16 tested positive for VREfm. All 16 VREfm-positive samples were isolated from six ICU beds located near the nurse station (Fig. 1b). The positive environmental samples within ICU beds were medical monitoring equipment, bed, urinary catheterization device, trash can, injection pump and curtain (Table S2). In addition to vancomycin resistance, all isolates (n = 27) exhibited resistance to tetracycline and gentamicin, while remaining sensitive to tigecycline, as detailed in Table S2 (Fig. 2).

Fig. 1.

Demographics of VREfm colonization patients and environmental sampling. a The flowchart depicts the timeline of patient enrolment across various departments. Horizontal lines represent the enrolled departments, with red lines indicating ICU stays. The detailed temporal mapping of cases can be accessed at Table S2. b The diagram illustrates the layout of ICU beds within the hospital. Numbers indicate individual ICU beds, while letters represent the nurse responsible for each bed during the sampling period. The nodes show the ICU beds with positive VREfm colonization from environmental sampling

Fig. 2.

ARG and VF features of collected genomes. a The heatmap displays detected ARGs in VREfm, clustered based on antibiotic groups. b The heatmap presents detected VFs, clustered based on virulence categories. Font colour indicates the source of VREfm: green represents environmental samples, and red indicates human genomes

Putative transmission inference

To gain a comprehensive understanding of the transmission routes within our hospital, we quantified the genetic distances between isolates, employing a threshold of ≤ 7 SNPs to identify putative transmission. Our analysis revealed that 25 isolates were involved in potential intra-hospital transmission, resulting in a total of 177 genome pairs. Among these 177 genomic pairs, the majority occurred between patient and environmental sources, followed by environment-to-environment transmissions (Fig. 3a). Notably, three patients (P3, P6, and P90) who had not been admitted to the ICU did not exhibit evidence of putative transmission, as no genomic pairs were detected for these individuals. In the context of SNP-based transmission analysis, most VREfm isolates from patients exhibited close genetic proximity to those found in bed environments, with the exception of two distinct clusters (Fig. 3b). These clusters were observed exclusively within bed 10, characterized by ST789, and between patients P1 and P6, characterized by ST80. Although all VREfm isolates were collected within the hospital, relying solely on SNP distance to determine putative transmission events does not provide strong support. Therefore, we reinforced the evidence for putative transmission by incorporating three epidemiological connections, resulting in the identification of 51 putative transmission events. Six beds (B11, C5, C6, D6, D7 and E8) with positive VREfm presence were linked to infections in seven patients (P2, P4, P5, P6, P7, P8 and P10) residing within those beds.

Fig. 3.

Putative transmission networks identified in the hospital case study. a The pie chart shows the distribution of genomic relatedness, as determined by SKA analysis. b This panel presents a network diagram depicting genomic relatedness among all isolates. Each node represents the VREfm collection within the hospital, with node shapes distinguishing between isolates obtained from patient samples and those from environmental screenings. Connecting lines between nodes indicate a pairwise SKA distance of 7 or fewer SNPs. c This panel illustrates the network diagram of putative transmission corresponding to epidemiological connections. The line types differentiate various epidemiological connections, providing a detailed view of putative transmission pathways

Lineage analysis under global VREfm context

In the 4232 ST80 VREfm genomes studied here, the first isolate to be collected was in 1905, but 95% of the sequenced genomes were collected after 2014 (Table S3). At a mash distance of 0.0015, the genomes were dereplicated to 696 representative genomes. Fastbaps clustered these 696 representative genomes into 13 fastbaps clusters (FBCs), with approximately half of the genomes assigned to FBC5 (n = 247) and FBC3 (n = 132) (Table S3). The majority of ST80 VREfm genomes from China, including those from this study, are clustered closely under FBC9, which contain 623 non-dereplicated genomes (Fig. 4a; Table S3). Among the 623 non-dereplicated FBC9 genomes, 322 were from China. Of these, 12 environmental isolates in our study were obtained, which were the only environmental isolates classified within FBC9. Meanwhile, 263 isolates were associated with human disease, primarily isolated from urine samples. The source details for the remaining 47 genomes from human are unavailable (Table S3). Comparative genomic analysis revealed distinct profiles of ARGs and VFs among the FBC lineages. The FBC9 lineage harboured significantly fewer ARGs than FBC3 and FBC10, though its ARG content was comparable to that of FBC4, FBC5, and FBC7 (Fig. 4b). Conversely, FBC9 possessed significantly more VFs than FBC5, FBC7, and FBC10, but not significantly more than FBC3 and FBC4 (Fig. 4b). Gene enrichment analysis indicated that fluoroquinolone resistance in FBC9 is primarily associated with the gyrA S83I and parC S80R mutations, in contrast to the gyrA S83Y and parC S80I mutations observed in the other lineages (Fig. 4c; Table S4). Furthermore, daptomycin resistance in FBC9 was strongly linked to rpoC T641K and dltC S63C mutations. The dfrG gene was largely absent in FBC9. Finally, the virulence genes ecbA, tirE1, and tirE2 were also associated with FBC9, whereas capD was not.

Fig. 4.

Phylogeny and enrichment analysis of ARGs and VFs from global ST80 VREfm. a Recombination removed core genome SNP phylogenetic tree of 696 representative ST 80 genomes. Red triangles indicate isolates from this study. FBCs clades are coloured and numbered. Metadata of each node are summarized from the non-dereplicated 4232 genome datasets. Colour strips starting from inner circle to outer circle: median year of isolates, region of isolates, source of isolates, percentage of genomes with van ARGs, average number of ARGs. b Number of ARGs and VF genes in the 696 representative genomes from 5 major FBCs. Statistical significance (*p < 0.05, ***p < 0.0005) is shown only for comparisons with FBC9. c Gene enrichment analysis of ARGs and VF genes with Benjamini-Hochberg adjusted p-value < 1E-5

Plasmid analysis

To further elucidate the genetic context of the vanA genes, we selected two representative genomes, VREfm1054 (ST80) and VREfm10-3 (ST789). Through hybrid assembly, we successfully circularized the vanA-carrying plasmids: a 43,393 bp plasmid from VREfm1054, designated pVREfm1054, and a 52,046 bp plasmid from VREfm10-3, designated pVREfm10-3. Despite originating from different STs, these plasmids shared a highly conserved backbone, exhibiting nearly 100% nucleotide identity (Fig. 5a). The unaligned regions between the two plasmids primarily consisted of toxin-antitoxin systems and other hypothetical proteins (Table S5). The replicon (RepR) of these plasmids belonged to the Replicase/PriCT-1 superfamily domains. The majority of the content within these plasmids was associated with antimicrobial resistance, mediated by Tn4001 containing the aac(6’)-aph(2”), flanked by IS256, and Tn1546 containing the vanA gene clusters. To place the vanA-carrying plasmids in a global context, we mapped our plasmids to all 649 dereplicated ST80 genomes (Fig. 5b). Although our genomes clustered within FBC9, indicating close genetic relatedness, the vanA-carrying plasmids themselves were not uniformly similar. The plasmid identified in our hospital setting was notably conserved. Interestingly, while ST80 VREfm isolates carried different plasmid backbones, they shared the same transposons Tn1546. The conserved sequences of Tn1546 suggest it may have role in the dissemination of vanA genes.

Fig. 5.

Alignment of ST80 genomes to pVREfm1054 and pVREfm10-3. a Alignment of pVREfm1054 and pVREfm10-3. b Heatmap of nucleotide identity across 100 bp segments of pVREfm1054 and pVREfm10-3. Light blue arrow in linear map indicates IS elements coding sequences, boundary refined by ISEScan and ISfinder

Discussion

In this study, we provide a comprehensive genomic analysis of a VREfm spread in an ICU, offering high-resolution insights into the transmission dynamics and dissemination of vanA gene clusters. Our findings identify VREfm ST80 as the primary agent responsible for the hospital spread, with the exception of four ST789 isolates found on environmental surfaces within a single bed. Colonization was observed not only in patients but also on environmental surfaces, underscoring the critical importance of stringent hygiene practices in hospital ICUs. WGS was performed on all collected isolates, revealing similar profiles of VFs and ARGs. Given these highly similar genomic features, we conducted a transmission analysis using SNP analysis with SKA, as adopted from previous studies [9, 39]. This analysis revealed that all genomes were highly clustered, indicating that the same bed could serve as a potential reservoir for VREfm colonization and persistence. In a global context, we observed that the majority of VREfm ST80 isolates from China clustered under FBC9 and harbored more VFs compared to other lineages based on in silico gene presence/absence analysis. The mode of vanA gene dissemination was estimated through complete sequencing of two representative genomes from ST80 and ST789. Despite the vanA genes being carried by different STs, the vanA-carrying plasmids were highly similar, suggesting the possible dissemination of Tn1546 across different STs.

Numerous reports have documented the high burden of infections in ICUs, which are recognized as important reservoirs for antimicrobial-resistant pathogens [40, 41]. A previous study conducted in a hospital in Guangdong demonstrated that 52.9% of VREfm isolates were obtained from the ICU, with a significantly higher prevalence compared to other departments [42]. The increased prevalence of VREfm within hospitals may be linked to the relaxation of restrictions following the COVID-19 pandemic and the associated stress on antibiotic usage [43]. A healthcare-associated outbreak requires three or more common-source cases within seven days. Because our genetically related isolates were detected intermittently over several months, epidemiological evidence suggest persistent environmental contamination and may contribute to subsequent transmission. While many surveillance efforts focus on patient-related transmission, our study extended the analysis to include ICU environmental sampling, revealing potential persistent colonization on surfaces. Following the environmental screening, the hospital implemented rigorous infection control measures. These included enforcing strict hand hygiene protocols for healthcare workers, patients, and visitors, comprehensive environmental cleaning and disinfection, antimicrobial stewardship, and visitor restrictions. As a result of these measures, ICU infections have been effectively controlled, with no VREfm isolates detected after July 2023. This underscores the critical importance of hygiene and infection control measures in managing and preventing within-hospital infections. In Guangdong Province, China, the presence of VREfm has been consistently reported [5, 42]. However, the prevalence of VREfm dramatically increased from approximately 20% to 50% between 2021 and 2023 [5]. Within Guangdong, there has been a notable lineage shift in VREfm infections from ST17 and ST78 to ST80. In our hospital, most VREfm isolates were identified as ST80, while ST17 and ST78 were not detected. Additionally, ST789 has been documented in colonization of environmental surfaces carrying vanA genes in our study. Although ST789 is not a predominant lineage responsible for patient infections and reports of ST789 remain limited, with detections primarily in Algeria [44] and UK [45], its capacity to carry vanA genes suggests a potential role in mediating vancomycin transmission and serving as a reservoir. Our findings regarding VREfm ST80 reinforce its predominant role in this region, underscoring the need for stringent infection control measures to manage hospital outbreaks. The epidemiological data, encompassing both patient-related and ICU environmental factors, provide valuable insights into developing effective management strategies to control VREfm infections.

To gain a deeper understanding of the transmission dynamics of VREfm within the hospital, we employed SKA to assess genetic relatedness. While most studies [46, 47] traditionally use MLST for pairwise genome comparisons, recent advancements in genomic techniques for VREfm putative transmission suggest that SKA provides the high-resolution insights necessary to accurately infer putative transmission events between patients. In contrast, MLST-based core genome alignments often fail to capture the precise genetic relatedness among isolates, potentially leading to incorrect classification of isolates as either involved or not involved in putative transmission events [9, 48]. Integrating insights from previous transmission analyses [9, 11, 47], we established a threshold of ≤ 7 SNPs to reliably identify putative transmission pairs. To reinforce the genomic connections identified, we supplemented them with epidemiological data. A close SNP distance can indicate high genetic homogeneity, and epidemiological connections can strengthen putative transmission links. We observed that many genetically close genomic pairs spanned different beds within a 30-day interval. Our analysis underscores the critical importance of stringent infection prevention protocols and rigorous hygiene practices among healthcare workers to mitigate the spread and putative transmission of VREfm.

van-positive VREfm is not only a concern in China but also a worldwide issue, with considerable variation in dominant clones observed across geographical regions. For instance, VREfm ST192 and ST17 were reported as predominant clones in Germany [49], while ST80 and ST17 are leading VREfm in Thailand [50] and Taiwan [51], respectively. To gain a comprehensive understanding of ST80 VREfm, we compared all ST80 VREfm isolates within a global context. ST80 is not only an endemic clone in China but has also been identified as a high-risk clone responsible for VREfm outbreaks in the United States [52], Australia [53], and Europe [54]. However, the ST80 VREfm isolates from our study were genetically distinct from those found in other countries, as they clustered into FBC9. Previous analyses of ST80 in Guangdong revealed a similar pattern, with limited identical ST80 isolates identified in India, although not as the predominant type [5]. Furthermore, we observed that the plasmid backbone in our hospital isolates differed from those in other studies, although all vanA genes were located on Tn1546. Interestingly, similar to previous study [5], no conjugation-related genes were detected, and no transconjugants were obtained from these plasmids. This suggests that vanA dissemination is primarily driven by clonal spread or Tn1546 integration. A potential explanation for this phenomenon may be the high fitness cost associated with Tn1546, which imposes significant growth pressure on the bacteria [55]. As a result, conjugation elements may be selectively lost in favour of maintaining vancomycin resistance. In summary, these findings reveal that ST80 VREfm exhibits regional specificity, indicating a distinct evolutionary pattern. The presence of Tn1546 highlights its potential as a putative driver of vanA gene dissemination.

Our study has several limitations. The primary limitation is the relatively small sample size. Although we conducted a global genomic comparison of VREfm, the inclusion of additional data from other hospitals, supported by more comprehensive epidemiological information, could enhance our findings. Another limitation is that all environmental VREfm were collected exclusively from ICU surfaces. To gain a more comprehensive understanding of transmission dynamics within hospitals, future studies could include samples from healthcare workers and other areas of the hospital environment. The virulence of VREfm carrying more VFs could also be validated by phenotypical measurements in future studies. In addition, because sampling was conducted at a single time point, an important limitation of this study is the inability to infer persistence over time or the directionality of contamination.

In conclusion, our study elucidated the genetic epidemiology of a VREfm spread in a hospital in China by comparing local isolates with global counterparts. We further explored the transmission dynamics of VREfm across ICU beds. The reconstruction of plasmids revealed a unique plasmid backbone specific to our hospital, though it carried the common Tn1546 element, which may be responsible for vanA gene dissemination. To mitigate the spread of VREfm within hospitals, active screening programs and stringent infection control measures should be implemented.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

S.J. and Y.L. designed the study and methodologies. H.Z. and Y.D. conducted epidemiological investigation and collected the data. YX.C. performed the bioinformatics analysis and figure plot. Q.W., Y.X., D.J., C.C., and Z.H. collected the data and revised manuscript. S.J. and YX.C. wrote the original draft with contributions from all the authors.

Funding

This work was supported by the Guangdong Basic and Applied Basic Research Foundation (2024A1515010319).

Data availability

The genomic data are available from NCBI Bioproject under accession number: PRJNA1356268.

Declarations

Ethics approval and consent to participate

This research received ethical approval from the Ethics Committee of Shenzhen Traditional Chinese Medicine Hospital (IRB No. K2025-119-01). The ethics committee waived informed consent due to this study’s observational nature.

Competing interests

The authors declare no competing interests.