The emergence of antimicrobial resistance (AMR) has become a major public health threat (Chioro et al., 2015), with AMR incidence increasing globally due to inappropriate antibiotic use and poor infection control practices. Carbapenem-resistant Enterobacterales (CRE) are an especially concerning AMR threat due to their resistance to several last-resort antibiotics, leading to high mortality (World Health Organization, 2017). CRE infections are now among the most frequent and difficult-to-treat infections due to limited effective antimicrobial agent availability (Chotiprasitsakul et al., 2019, European Centre for Disease Prevention and Control (2018), World Health Organization, 2017). In 2015, it was estimated that over 2,000 patients died from infections caused by CRE, particularly bloodstream infections (BSI), in Europe, with the number of attributable deaths increasing more than six-fold between 2007 and 2015 (Cassini et al., 2019). Accordingly, the Centers for Disease Control and Prevention (CDC) assigned CRE the highest threat level, requiring public attention (Centers for Disease Control and Prevention, 2019).

Increased carbapenem use in hospitals has led to increased CRE prevalence, thus constituting a serious public health threat. A report published by the World Health Organization (WHO) and the European Centre for Disease Prevention and Control (ECDC) revealed that between 2017 and 2021, the prevalence of carbapenem resistance in Klebsiella pneumoniae isolates increased from 0% to 20% European Centre for Disease Prevention and Control and World Health Organization (2023). In Brazil, CRE prevalence increased from almost zero between 1995–1999, to >20% between 2015–2019 (Ferreira et al., 2020). In the United States, the prevalence of carbapenem resistance among Klebsiella pneumoniae, Enterobacter spp., and Escherichia coli (E. coli) bacteremia were 9.7%, 2.2%, and 0.1%, respectively (Satlin et al., 2017). The SENTRY study further demonstrated a statistically significant increase in CRE rates across Latin America over a 20-year period, increasing from 0.6 to 2.9% (Castanheira et al., 2019). According to the China Antimicrobial Surveillance Network (CHINET), the resistance rates of Enterobacterales isolates to imipenem and meropenem increased from 3.1% and 2.1% in 2005 to 11.1% and 10.8%, respectively, in 2024 (CHINET, 2024). BSI is one of the most severe nosocomial infections, with Enterobacterales being the most common associated pathogen (Zhou et al., 2024), and patients with CRE-BSI demonstrate significantly higher mortality (de Araujo et al., 2025; Trecarichi and Tumbarello, 2017), making it a critical clinical threat.

CRE develops antibiotic resistance through several mechanisms, with the production of carbapenemases—enzymes that degrade carbapenem antibiotics—being the primary mechanism (Lutgring and Limbago, 2016). Carbapenemases are primarily categorized into three classes. Class A features enzymes like the Klebsiella pneumoniae carbapenemase (KPC); Class B comprises metallo-β-lactamases, including the New Delhi metallo-β-lactamase (NDM) and imipenemase (IMP); and Class D is represented by oxacillinases, such as OXA-48. Antibiotic resistance in CRE is also mediated by two other mechanisms: the overexpression of efflux pumps and mutations in outer membrane proteins. KPC carbapenemase is widely found in the United States and Europe (Albiger et al., 2015; Iovleva and Doi, 2017), NDM carbapenemase is most commonly found in Indian, Romania, Spain, and Hungary (van Duin and Doi, 2017; Yong et al., 2009), and OXA-48-like carbapenemase is found in Turkey and surrounding countries (van Duin and Doi, 2017). In China, KPC-2 was the most prevalent carbapenemase among CRE strains isolated from adult patients (70.3%), and NDM among the CRE strains isolated from children (49.0%) (Han et al., 2020). Carbapenemase-encoding genes are often harbored on mobile genetic elements like plasmids, which facilitates their transfer between bacteria, significantly accelerating the spread of carbapenem resistance. Although novel antibiotics, such as ceftazidime-avibactam, have been used to treat CRE infection, they remain ineffective against NDM-producing CRE (Abe et al., 2022). Therefore, identifying and intervening in the risk factors for CRE-BSI is crucial to reducing its incidence among hospitalized patients and ultimately improving clinical outcomes.

While a number of studies have explored risk factors associated with CRE infection, most have employed a case-control design with patients infected with carbapenem-susceptible Enterobacterales (CSE) as the control group. Directly comparing CRE cases to CSE may introduce selection bias, which can lead to the false identification of antibiotics as risk factors or an overestimation of the odds ratios of some antibiotics (Behar et al., 2008; Harris et al., 2002). To mitigate this bias, a case-case-control design was implemented, featuring two control groups: patients with CSE and those infected with non-Enterobacterales bacteria. This study sought to define risk factors for CRE-BSI, with the goal of informing clinical strategies for the early identification of high-risk patients and the implementation of timely measures.

Patients infected with CRE-BSI demonstrate increased morbidity and mortality, no effective antimicrobial treatment exists for these infections. Thus, identifying risk factors for CRE-BSI etiology and prognosis is essential for reducing incidence and informing optimal therapeutic strategies. The selection of an appropriate control group is critical for accurate assessment of risk factors for CRE-BSI. Using patients with CSE-BSI as controls, a practice common in previous studies, is methodologically problematic. This design inherently assumes a “replacement scenario” where patients are infected exclusively with either CRE or CSE, thereby failing to account for patients infected with other pathogens or those without infection. Patients with CSE constitute a minor fraction of all hospitalized cases and are thus not representative (Harris et al., 2001; Huang et al., 2023; Kaier and Frank, 2013). Furthermore, certain antibiotics may suppress or kill some CSE strains without affecting CRE strains. Consequently, the frequency of antibiotic use in CSE may be reduced, indirectly amplifying the risk factors for CRE (Ng et al., 2014). In addition, few studies employed patients without CRE infection as the control group, without distinguishing between susceptible and resistant groups, making it difficult to independently analyze the risk factors for Enterobacterales infection. This study employed a case-case-control design to comprehensively compare patient characteristics among those with CRE-BSI, CSE-BSI, and control group. It confirmed established risk factors while uniquely distinguishing those specific to carbapenem resistance from those common to Enterobacterales infection in general, thereby providing potential therapeutic options for CRE-BSI.

We found that Carbapenem-resistant Klebsiella pneumoniae was the predominant strain among cases with clinical BSI due to CRE (81.9%), which is slightly higher than that reported in other studies (Chen et al., 2022; Yan et al., 2020). CRE-BSI cases were most frequently identified in the ICU, with the hematology department being the second most common location. This distribution is attributed to the confined environment of the ICU, enhancing the risk of CRE transmission via airborne and contact transmission. Patients in the ICU often have severe comorbidities and require invasive procedures as well as treatment with broad-spectrum antibiotics, factors that contribute to an increased risk of CRE-BSI. Notably, patients with hematologic diseases have a unique vulnerability to CRE-BSI, characterized by prolonged hospital stays, broad-spectrum antibiotic use, chemotherapy, and immunodeficiency (de Souza et al., 2024).

Our findings demonstrate a markedly higher level of AMR in CRE-BSI compared to that in CSE-BSI. Given the high resistance of CRE-BSI to conventional antibiotics yet their preserved susceptibility to tigecycline and colistin, combination therapy emerges as the optimal strategy and should be tailored to the individual patient’s condition.

Statistically significant factors identified in the multivariate analysis of both the CRE-BSI and CSE-BSI groups represent risk factors for Enterobacterales-BSI. In contrast, factors unique to the CRE-BSI analysis constitute specific risk factors for CRE-BSI. Similarly, factors significant only in the CSE-BSI analysis define specific risk factors for CSE-BSI (Kaye et al., 2005). Further, exposure to specific antibiotics (third-generation cephalosporins, carbapenems, and quinolones) and glucocorticoids, along with a history of surgery and gastric tube insertion, emerged as independent risk factors for CRE-BSI. In contrast, surgical history was established as an independent risk factor for Enterobacterales-BSI.

A systematic review and meta-analysis by Zhu et al. showed that common risk factors for CRE infection included severe underling diseases, prior antibiotic use (carbapenems, aminoglycosides, quinolones, vancomycin, glycopeptides), ICU admission, invasive procedures (central venous catheter use, mechanical ventilation, tracheostomy, urinary catheter use), and longer LOS (Zhu et al., 2020). We confirmed that antibiotics use, surgery history, and gastric tube insertion increased the risk of CRE infection, while identifying carbapenemase genes (notably blaKPC-2 and blaNDM) as the dominant resistance mechanism among CRE strains in China. Genes encoding carbapenemases are often located on mobile genetic elements, and exposure to carbapenems can promote the emergence of these resistance genes (Han et al., 2020). In contrast to some previous findings, our results demonstrated that third-generation cephalosporin use was a significant risk factor for CRE-BSI, this suggests a broader selective pressure for CRE, indicating that resistance can be promoted by exposure to non-carbapenem antibiotics. Thus, enhancing inpatient antibiotic stewardship and implementing high-dose therapy for a controlled duration is a superior strategy for limiting infection risk (Jeon et al., 2008). Although numerous studies have investigated the risk factors for CRE infection, prior use of quinolones is an independent risk factor remains unclear (Satlin et al., 2017; Stuever et al., 2022; Zuo et al., 2021). Exposure to quinolones promotes bacterial selective pressure, driving resistance not only to this drug class but also to carbapenems via multiple mechanisms. Quinolones can upregulate the expression of the MexEF-OprN efflux pump, triggering a deficiency in OprD—an outer membrane porin—which induces resistance to multiple antibiotics, including carbapenems (Wang et al., 2018). Furthermore, plasmids that mediate quinolone resistance may concurrently carry carbapenemase genes, such as KPC, further contributing to the emergence of carbapenem-resistant strains (Zhu et al., 2020). Unlike previous reports, our study indicated that exposure to glucocorticoid was a significant risk factor for CRE-BSI, possibly due to patients with CRE-BSI requiring glucocorticoid therapy for severe underlying diseases. By disrupting the intestinal microenvironment through eliminating susceptible bacteria and promoting resistant strains, glucocorticoid treatment thereby promotes the evolution of opportunistic microbes into pathogens. Therefore, strengthening antibiotic management measures and regular drug resistance monitoring are crucial for avoiding unnecessary antibiotic exposure. Surgical history was identified as a risk factor for CRE-BSI and Enterobacterales-BSI compared to control. Surgery often prolongs LOS, thereby increasing the risk for nosocomial infection. Furthermore, open wounds serve as portals that can elevate the risk of any type of infection (Lavigne et al., 2021), further highlighting the critical role of aseptic techniques in patient care. Gastric tube insertion, another significant risk factor, can provide a suitable place for CRE colonization, and can damage the gastrointestinal mucosa, thereby enhancing the potential for opportunistic bacterial invasion (Tischendorf et al., 2016). Furthermore, gastric tube insertion is often indicated for patients with gastrointestinal dysfunction and compromised nutritional status. This high-risk profile may promote bacterial translocation from the gut, thereby increasing susceptibility to pathogens such as CRE.

CRE are of epidemiological interest for their association with poor outcomes and ability to spread rapidly throughout hospitals (Van Duin et al., 2020). Patients with CRE-BSI demonstrated a significantly higher 30-day mortality than those with other infections. S Sabino et al. reported that the 30-day mortality for CRE-BSI was 63.8% (Sabino et al., 2019), which was higher than that in our study (25.7%). Several factors were independent risk factors for 30-day mortality, including patient group, age ≥ 65 years, respiratory diseases, and digestive system diseases. Several studies have reported that the overall mortality rate in patients with CRE-BSI ranged from 37% to 65%, and some studies have shown that CRE infection was an independent risk factor for mortality (Kedišaletše et al., 2023; Li and Ye, 2017; Sabino et al., 2019). Similarly, patients with CRE-BSI exhibited higher mortality and longer hospital stays in our study, consistent with a previous study (Strich et al., 2020), we observed that the 30-day mortality of CRE-BSI was frequently associated with underlying respiratory diseases in patients. Patients with respiratory diseases are susceptible to both airway colonization and subsequent CRE-BSI (Liu et al., 2012). Patients with CRE-BSI who died within 30 days were significantly older than survivors. Potential reasons for this may include more frequent healthcare exposure and antibiotic use among older adults. Patients with digestive system diseases—who typically exhibit gastrointestinal dysfunction and compromised nutritional status—may require gastric tube insertion. In critically ill patients, this procedure can damage the mucosal barrier, promote bacterial translocation from the gut, and thereby increase the risk of CRE-BSI (Correa et al., 2013; Munoz-Price et al., 2010). Thus, it is imperative to implement stricter infection prevention and control strategies to prevent transmission within hospitals and employ robust antimicrobial stewardship to ensure appropriate antibiotic use, thereby reducing the selection pressure for resistant organisms (Pallares et al., 2022).

Arterial catheterization emerged as a significant risk factor for in-hospital mortality among CRE-BSI patients. This finding reinforces that invasive procedures increase mortality by damaging the mucosal barrier and enabling bloodstream invasion, underscoring the necessity of aseptic technique for infection prevention.

Several studies have established CRE colonization as an independent risk factor for subsequent CRE infection. Therefore, it is essential to strengthen active screening for CRE to facilitate early detection, early isolation, and early intervention in high-risk departments and patients (Chen et al., 2022; Sun et al., 2021). However, the WHO and the US CDC recommend screening for CRE based on local epidemiological conditions (Centers for Disease Control and Prevention, 2015; World Health Organization, 2017). Nosocomial CRE-BSI cases in our hospital were concentrated in the ICU and among patients with specific risk factors, thus active CRE screening should be implemented for these high-risk departments and patient populations.

Clinical outcomes vary considerably among different antimicrobial regimens, and an effective therapeutic approach for CRE-BSI remains elusive. Studies have demonstrated that combination antimicrobial therapy is preferred over monotherapy, particularly in severely-ill patients (Kumar et al., 2010). In our study, no difference was observed between survivors and non-survivors with CRE-BSI treated with monotherapy and combination therapy, consistent with a previous study (Xiao et al., 2018). This is primarily attributable to the complex causes of mortality in patients with CRE-BSI and the lack of an optimal treatment regimen for CRE-BSI. In administering anti-infective therapy, clinicians should consider patients’ individual conditions (including underlying diseases) and their risk factors for mortality, to guide adequate early empirical antibiotic therapy—a critical factor in reducing mortality.

This study is subject to several limitations. Its single-center, retrospective design may affect the generalizability of the findings and could introduce potential information and selection biases. Furthermore, the absence of carbapenemase gene detection means some carbapenem-susceptible carbapenemase-producing isolates may have been misclassified as CSE. Future studies employing molecular analyses will help draw more precise conclusions.

This matched case-case-control study delineated distinct risk factor profiles for bloodstream infections. Surgical history was established as an independent risk factor for Enterobacterales-BSI, while exposure to specific antibiotic classes (third-generation cephalosporins, carbapenems, and quinolones) and glucocorticoids, along with surgical history and gastric tube insertion, were independently associated with CRE-BSI. Furthermore, several factors predicted increased mortality, including CRE-BSI itself, respiratory diseases, age ≥ 65 years, digestive system diseases and arterial catheter are more likely to have higher mortality. Notably, no significant differences in outcomes were observed across various antimicrobial regimens. Consequently, early identification of these risk factors is crucial, and primary prevention strategies—including antimicrobial stewardship, reduced hospitalization, and minimized invasive interventions—are essential to reducing the incidence and mortality of CRE-BSI. These findings provide critical insights to guide effective antimicrobial therapy development and prevention of CRE-BSI.