This investigation was carried out within the Hematology Department of Fujian Medical University Union Hospital, Fuzhou, China, a major referral center with 240 beds dedicated to hematological conditions. The study population consisted of patients with hematological malignancies who were diagnosed with Enterobacterales BSIs between January 1, 2018, and December 31, 2022. The research protocol was approved by the Institutional Review Board of Fujian Medical University, and all procedures adhered to the ethical principles outlined in the 1964 Declaration of Helsinki.

In this study, a BSI caused by Enterobacterales was defined as the isolation of one or more of these bacteria in a single blood culture bottle during the study period. The term “carbapenem resistance” was used to describe bacteria that exhibited a minimum inhibitory concentration (MIC) of meropenem and/or imipenem equal to or greater than 4 mg/mL (Mermel et al., 2009). Septic shock was diagnosed based on clinical criteria, including hypotension requiring vasopressor therapy to maintain a mean blood pressure of 65 mm Hg or higher after appropriate fluid resuscitation (Durante-Mangoni et al., 2019). Neutrophil was defined as a neutrophil count less than 0.5 × 10⁹/L in the peripheral blood.

The clinical and microbiological data were meticulously extracted from the medical records of the patients. A detailed summary of the data collection process is presented in Figure 1 . A team of clinicians carefully reviewed each case to ensure that the primary infection sites met the criteria established by the National Healthcare Safety Network. Catheter-related bloodstream infections (CR-BSIs) were defined in accordance with the guidelines set by the Infectious Diseases Society of America (IDSA) (Shankar-Hari et al., 2016). These CR-BSIs were excluded from the study. In instances where no primary infection site met the specified criteria, the infection source was categorized as “unidentified”.

The bacterial isolates obtained from blood samples were cultivated on blood agar plates (Autobio Diagnostics CO.,Ltd, China). The bacterial species were identified using Bruker Biotyper (Microflex LT/SH Smart) MALDI-TOF MS (Bruker Daltonics GmbH, Bremen, Germany). Antimicrobial susceptibility testing was conducted using the Vitek-2 system (Vitek-AST-GN16) (BioMerieux, France), following the guidelines provided by the Clinical and Laboratory Standards Institute (CLSI, 2022). These methodologies were implemented in the Clinical Microbiology Laboratory of Fujian Medical University Union Hospital. All cases of Enterobacterales isolated from 2018 to 2022 were included into this study. For patients from whom many CRE were isolated, only a representative isolate was included in our study to avoid redundancy.

Genomic DNA from the CRE isolates was extracted using the TIANamp Bacteria DNA Kit (Tiangen, Beijing). Sequencing was conducted on the Illumina NovaSeq platform at Shanghai Personal Biotechnology Co., Ltd (Shanghai, China) in a paired-end mode with an insert size of 2 × 150 base pairs. The depth of coverage varied from 224-fold for 10,459,852 reads to 292-fold for 10,256,762 reads, with an average depth of coverage of approximately 260-fold. Gene content analysis was performed using tools provided by the Center for Genomic Epidemiology (https://cge.cbs.dtu.dk/services/). This included the identification of acquired resistance genes and known chromosomal mutations associated with antibiotic resistance using the ResFinder tool (Yu et al., 2019). Plasmid types were identified using PlasmidFinder (Zankari et al., 2012), and virulence genes were identified using VirulenceFinder (Carattoli et al., 2014). Multilocus-sequence typing (MLST) was conducted on the MLST website (https://cge.food.dtu.dk/services/MLST/). The genetic environment of the blaKPC-2 and blaNDM-5 genes was predicted using the RAST tool (https://rast.nmpdr.org/rast.cgi). Core-genome alignments were utilized for phylogenetic tree reconstruction using maximum likelihood estimation.

Differences between patient subgroups were statistically evaluated using the chi-square test, t-test, or nonparametric test, as appropriate for the nature of the data. GraphPad Prism version 9.5 was employed for all statistical analyses. A P value of less than 0.05 (two-tailed) was considered to indicate a statistically significant difference.

From January 1, 2018, to December 31, 2022, a total of 301 patients with Enterobacterales BSIs were enrolled in the study, including 65 cases (21.6%) with CRE and 236 cases (78.4%) with carbapenem-susceptible Enterobacterales (CSE). The patient cohort consisted of 172 males (57.1%) and 129 females (42.9%), with a wide age range from 1 to 80 years, and a median age of 46 years. The analysis of demographic data showed no significant differences in age, gender, disease subtypes, or the proportion of patients with peripherally inserted central catheters (PICCs) between patients with CRE and CSE BSIs ( Table 1 ).

The details of infection sources for the patients with Enterobacterales BSIs are summarized in Table 2 . The respiratory and digestive tracts were the most common sites of infection, accounting for 62.1% and 24.6% of cases, respectively. Notably, patients with CRE BSIs had a significantly higher proportion of perianal infections (29.2%) and multiple infection foci (52.3%) compared to those with CSE BSIs (13.1% and 38.1%, respectively). Additionally, the proportion of patients with an unidentified infection source was significantly lower in the CRE group (3.1%) compared to the CSE group (14.4%). The primary pathogens associated with Enterobacterales BSIs were carbapenem-resistant Escherichia coli (CREC), which accounted for 55.1% of cases, while carbapenem-resistant Klebsiella pneumoniae (CRKP) was the predominant CRE isolate, representing 69.2% of all CRE BSIs.

The incidence of CRE infections among patients with Enterobacterales BSIs varied between 16.9% and 26.9% from 2018 to 2022, with the highest rate recorded in 2021 ( Figure 2A ). CRKP was the predominant CRE isolate identified during this period ( Figure 2B ). Genome sequencing was conducted on 35 CRKP cases, seven CREC cases, two carbapenem-resistant Enterobacter cloacae cases, and one carbapenem-resistant Klebsiella michiganensis case. Genetic analysis revealed that the predominant sequence type (ST) of CRKP was ST11, accounting for 68.6% of all cases, while CREC displayed a diverse range of ST types ( Figures 2C, D ). Enzymatic profiling indicated that 80% of CRKP carried the blaKPC-2 gene, and 71.4% of CREC carried the blaNDM-5 gene ( Figures 2E, F ).

As it was showed at Figure 3A , genetic background analysis of CRKP identified three distinct clades. The ST11-KL47-OL101 and ST11-KL64-O1/O2v clones were the predominant ones, with the ST11-KL47-OL101 clone possibly being involved in nosocomial infections in 2019. Additionally, 97.1% of CRKP strains carried multiple β-lactamase genes and plasmids. For CREC, analysis did not exhibit distinct clades, suggesting a high possibility of dissemination and transmission ( Figure 3B ). The distribution of virulence genes in CRKP and CREC was summarized in Figures 3C, D , respectively. Although these strains harbored a considerable number of virulence genes, none of the genes specifically associated with hypervirulent Klebsiella pneumoniae, including rmpA/rmpA2, peg-344, iroB, and iucA, which are particularly associated with CRKP, were detected.

The genetic environment upstream and downstream of the blaKPC-2 gene was shown in Figure 4A , with the core part of the gene environment primarily located between ISKpn27 and/or TnAs1. In Figure 4B , the upstream and downstream genetic environment of the blaNDM gene was illustrated, revealing significant differences in the genetic environments of different NDM subtypes, suggesting different plasmids or origins for these subtypes. Figure 4C depicted the genetic environment of the blaNDM-5 gene, which was the predominant subtype of blaNDM. The core region of blaNDM-5 exhibited a similar genetic environment, predominantly containing the structure of IS5-blaNDM-5-bleMBL-trpF-dsbD. However, multiple IS sequences were observed near blaNDM-5 in some strains, indicating a high possibility of horizontal transfer and dissemination, as indicated by strain numbers 413255 and 409497.

A total of 34.6% of patients had been treated with carbapenems within the three months preceding sample isolation, with no significant difference between the CRE and CSE groups. Notably, compared to the CSE group, patients in the CRE group had a significantly higher rate of histories of multiple hospitalizations (70.8% vs. 34.3%, P <0.0001), ICU exposure (18.5% vs. 2.1%, P <0.0001), and prior CRE infection (18.5% vs. 3.8%, P = 0.0002). The CRE group primarily consisted of patients with newly diagnosed (43.1%) and relapsed/refractory diseases (46.2%), with a small proportion (10.8%) of patients in remission status ( Figure 5 ). In contrast, the CSE group had a relatively even distribution among these categories, with proportions of 37.3%, 33.1%, and 29.7% respectively. Administration of glucocorticoids, transplantation, and immunosuppressants was similar between the CRE and CSE groups. No significant differences were observed in the proportion of patients with neutropenia, or in the levels of C-reactive protein (CRP), procalcitonin (PCT), and interleukin-6 (IL-6) between the two groups. The detailed clinical characteristics of CRE BSIs are summarized in Table 3 .

In the CRE group, infection-related septic shock was more common compared to the CSE group (43.1% vs. 19.9%, P = 0.0003), and the all-cause mortality rate within 30 days was significantly higher in the CRE group (56.9% vs. 24.6%, P <0.0001). Age was a factor that influenced mortality, with a rate of 45.5% for those under 46 years and 68.8% for those 46 years or older (P = 0.0804). The disease subtype and status had a significant impact on the occurrence and survival of CRE BSIs in patients with malignant hematologic disorders. For example, the mortality rate for patients with acute leukemia was 45.7%, which was lower than the rate for patients with lymphoma, multiple myeloma (MM), or myelodysplastic syndrome (MDS) (84.2%, P = 0.0056), and no difference was observed between acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL).

Additionally, strains carrying blaKPC-2 or blaNDM-5 were associated with higher mortality rates compared to those without these carbapenemases. The administration of tigecycline and/or polymyxin B could reduce mortality, with the combination therapy of tigecycline and polymyxin B exhibiting the lowest mortality rate at 34.5%. Notably, the mortality rate among patients who developed septic shock was 78.6%, which was significantly higher than the 40.5% mortality rate among those who did not develop septic shock (P = 0.0026). The prognostic factors for patients with CRE BSIs are summarized in Table 4 .