Acute myeloid leukaemia (AML) is a malignant proliferation of haematopoietic cells. The massive proliferation of abnormal blasts and immature cells in the bone marrow prevents normal haematopoiesis. Thus, the bone marrow and peripheral blood are characterised by an increase in immature leukocytes, mainly blasts (Assi et al., 2019; Wrighton, 2019). Immature cells accumulate in the bone marrow and replace normal bone marrow cells, megakaryocytes, and red blood cells, leading to loss of normal bone marrow function, bleeding, anaemia, and infection-related complications (McKenzie, 2005). Several factors related to the prognosis of patients with AML, such as infection, disease relapse, and haemorrhage. In addition, haematopoietic stem cell transplantation, chemotherapy regimens, and duration of agranulocytosis are closely related to patient prognosis. Among these, the incidence of bloodstream infections (BSIs) in haematological malignancies is 4–90% (Tumbarello et al., 2012; Samonis et al., 2013; Kolonen et al., 2017). Furthermore, the widespread utilisation of antibiotics has led to a gradual increase in the incidence of drug-resistant bacteria. Inappropriate empirical antimicrobial therapy or repeated infections contribute to the rise in the incidence of multiple drug resistance and mortality rates in patients (Moghnieh et al., 2015; Tang et al., 2020). Fortunately, by taking prompt preventive measures, we can significantly reduce the incidence of BSIs and lower the occurrence of related complications. This underscores the critical importance of early intervention. Such measures not only contribute to enhancing patient survival rates but also notably improve their overall quality of life, thereby creating a more conducive environment for their treatment. Hence, the significance of early BSIs prevention cannot be underestimated. However, the distribution of pathogenic bacteria in BSIs of patients in different regions varies greatly, which makes it difficult for early intervention of BSIs.

Gram-positive bacteria plays a predominant role as pathogens in certain haematology centres in Europe and America (Mikulska et al., 2014). However, the situation is markedly different in haematological wards in China, where Gram-negative bacteria are more prevalent (Deng et al., 2012). This disparity not only involves the types of pathogens but also extends to the establishment of medical infrastructure and the formulation of prevention and control strategies. Furthermore, regional variations also impact the antimicrobial susceptibility of specific pathogens (Deng et al., 2012; Mikulska et al., 2014). Resistance patterns of the same pathogenic bacteria may differ among different regions underscoring the need to consider the local microbial environment and resistance profiles when devising treatment plans and preventive strategies. This consideration is particularly crucial in the treatment of haematologic patients, where drug selection and administration must be tailored to the patient’s condition, the type of pathogenic bacteria, and the local susceptibility data. This integration ensures the optimal therapeutic outcomes and prognostic results (Deng et al., 2012; Mikulska et al., 2014). Therefore, as the etiological differences in different regions, there are some difficulties in empirical determination of medication for patients with BSIs and of prophylactic medication in patients at high risk of BSIs.

This study conducted a retrospective analysis of pathogen distribution, drug resistance, the risk factors associated with mortality within 30 days of BSI, and long-term prognosis of AML patients with BSI in a subset of 56 AML patients with BSIs, including 18 cases that resulted in death. This study provides a robust basis for clinically determining treatment approaches for AML patients with concurrent BSI and for improving patient prognosis in China.

BSI is a serious systemic infectious disease caused by various pathogenic microorganisms that invade the blood circulation and release toxins and metabolites (Jang-yong et al., 2020). It is one of the common complications of AML patients and is associated with poor prognosis (Collin et al., 2001; Nørgaard et al., 2006). Leukopenia, especially the decrease in neutrophils, can be caused by patients with AML themselves and chemotherapy, which leads to immunodeficiency, and the incidence of BSIs is significantly increased. Previous exposure to antibiotics or inappropriate empirical antibiotic treatment contributes to an increase in the incidence of BSI, drug resistance and mortality. This trend was corroborated by Italian studies conducted by Gudiol C et al. and Trecarichi EM et al., wherein the incidence of MDR escalated from 13.7% in 2011 to 31.5% in 2023, accompanied by an increase in mortality rates from 24% to 34.4% (Gudiol et al., 2011; Trecarichi et al., 2023).

The distribution of types and drug resistance rates of pathogenic bacteria of BSIs varies in different regions and times. Studies have shown that Gram-positive bacteria were the most common cause of BSIs in cancer patients before 2000, and the infection rate of Gram-positive bacteria reached 76% in 2000 (Wisplinghoff et al., 2003; Ortega et al., 2005). Trecarichi et al. found that the pathogenic bacteria of BSIs in patients with HMs changed from Gram-positive bacteria to Gram-negative bacteria from 2006 to 2016 (Trecarichi and Tumbarello, 2014; Trecarichi et al., 2015). A study on BSIs in 357 AML patients in Finland from 2003 to 2011 showed that most of the pathogens were Gram-positive bacteria (65.7%), with coagulase-negative staphylococci accounting for the highest proportion at 24.7% among Gram-positive bacteria (Kolonen et al., 2017). An Italian study on adult haematological malignancies from 2009 to 2012 revealed that Gram-negative bacteria constituted 52.8% of infections (Trecarichi et al., 2015).

Gram-negative bacteria have been the main cause of infection in related studies in China, of which Escherichia coli is the most common (Chen et al., 2010; Li et al., 2022). Escherichia coli typically colonizes in the gastrointestinal tract. During the treatment of AML, cytotoxic chemotherapy alters the integrity of the gastrointestinal mucosa and impairs the immune system, increasing the risk of invasive infections caused by the translocation of colonising bacteria (Montassier et al., 2015; Garcia-Vidal et al., 2018). Similarly, gram-negative bacteria are the main pathogens in our centre. However, in our study, Klebsiella pneumoniae took precedence among Gram-negative bacteria. Some patients with potential prior infection, having received β-lactam cephalosporins and quinolones antibiotics for infection treatment, exhibited Klebsiella pneumoniae in the screening. Previous antimicrobial therapy, particularly β-lactams and cephalosporins, can also elevate the risk of multidrug resistance, leading to a high incidence of Klebsiella pneumoniae in the region (Garcia-Vidal et al., 2018). This emphasizes the need for cautious use of broad-spectrum antibiotics in our centre. In the study by Chen et al., the results are similar to those of this study, with Klebsiella pneumoniae being the predominant gram-negative bacterium (Esbenshade et al., 2015; Chen et al., 2019). In our centre, the predominant pathogen spectrum comprises Gram-negative bacteria. Standard empirical treatment typically involves broad-spectrum β-lactam antibiotics, while enterococci exhibit complete resistance to β-lactam antibiotics (Kristich et al., 2014; Gagetti et al., 2019). Consequently, the incidence of enterococci has exhibited a progressive increase from 2013 to 2021 in our centre. Enterococci has become predominant Gram-positive bacteria from 2019 to 2021 in our centre, likely attributed to the inadequacy of empirical antimicrobial therapy to address enterococcal infections. In cases where a Gram-positive bacterial infection is suspected, the addition of vancomycin to empirical treatment regimens may be considered.

In other studies, Gram-positive bacteria are mainly represented by Streptococcus viridans and Staphylococcus epidermidis (Chen et al., 2019; Carvalho et al., 2020), which is totally different from the situation in our centre. The clinical medication principle involves tailoring treatments based on blood culture and susceptibility results to ensure efficacy while minimising the development of drug-resistant pathogens. However, blood pathogen cultures and susceptibility testing can be time-consuming. Clinical treatment for BSI often relies on empirical approaches. As antimicrobial drug usage increases, pathogen resistance grows, making empirical anti-infection treatment more challenging. Thus, understanding the microbial antimicrobial spectra and resistance profiles in different healthcare centres is crucial for guiding clinical empirical drug administration, improving susceptibility testing, and reducing overall risk.

In our centre, the resistance rate of Klebsiella pneumoniae to cephalosporins and quinolones was high, and the resistance rate to aminoglycosides was low. The resistance rates of Escherichia coli to β-lactam/β-lactamase inhibitors, sulfonamides and cephalosporins were high, their resistance rate to carbapenems was low.

Therefore, aminoglycosides and carbapenems can be recommended for AML patients with BSI suspected of being infected with Gram-negative bacteria, which consistent with the findings of Trecarichi et al (Trecarichi et al., 2015; Åttman et al., 2021). Although the predominance of Gram-positive bacteria in our study differs from other studies, consistent with our results, most of them exhibited high sensitivity to vancomycin (Chen et al., 2019; Li et al., 2022). Therefore, in our centre, empirical treatment with vancomycin may be considered for AML patients suspected of Gram-positive bacterial infection prior to obtaining definitive pathogen and drug sensitivity results.

Studies conducted by Gudiol C et al. have highlighted the consequences of the inappropriate use of antibiotics, demonstrating an associated increase in drug resistance (Gudiol et al., 2011; Trecarichi et al., 2023). Comparing drug resistance rates from 2013 to 2021 in our centre revealed a gradual upward trend, attributed to the use of carbapenem in empirical anti-infective therapy. Specifically, the incidence of ESBL and MDR increased during the period of 2016–2018 compared to 2013–2015, potentially linked to the irrational use of antibiotics. Kollef MH et al. found that prolonged antibiotic exposure was a risk factor for MDR (Munoz-Price et al., 2016; Kollef et al., 2021). Interestingly, our study observed a decreasing trend in MDR incidence during 2019–2021 compared to 2016–2018, potentially attributed to reduce antibiotic exposure duration and decreased use of empirical antibacterial drugs in our centre.

Frequent exposure to initially inappropriate empirical antibiotic treatment may play a pivotal role in the risk of MDR-related death, hospital bacterial colonisation or infection and viral reactivation are also closely related to MDR morbidity and mortality (Hotchkiss et al., 2013; Zilberberg et al., 2014; van Vught et al., 2016). Compared with non-MDR infection, patients with MDR bacterial infection have an increased risk of mortality (Cornaglia et al., 2011; Vardakas et al., 2013). Our data also revealed an increased mortality rate within 30 days attributable to MDR infections, despite a declining trend in MDR incidence. However, due to the limited sample size, further data collection is warranted for comprehensive research. Among the 8 patients with MDR infection from 2019 to 2021, only one patient had previous colonisation of MDR strain (Enterobacter asburiae), which differed from the MDR strain (Escherichia coli) causing BSI. This patient received antifungal treatment combined with cephalosporin, with subsequent adjustment of antibiotic therapy based on drug sensitivity results. Prior to the identification of the causative pathogen in bloodstream infections, 6 out of the remaining 7 patients received empirical antifungal therapy, primarily in combination with β-lactam, carbapenem, or quinolone antibiotics, until the pathogens of BSI were identified.

Septic shock, denotes a complex syndrome of septicemia induced by microbial agents and their toxic byproducts, resulting in shock (Shankar-Hari et al., 2016). The use of chemotherapeutic drugs leads to neutropenia, which leads to severe immunodeficiency and increases the probability of septic shock (Na et al., 2022).The incidence of septic shock in patients with HMs ranged from 20.7% to 31% (Royo-Cebrecos et al., 2022; Wang et al., 2023; Zhang et al., 2023). In this study, the incidence of septic shock was 37.3%, with 50% of cases resulting in death within 30 days after BSI. The elevated incidence of septic shock in our centre may be associated with neutropenia occurrence among patients, with over half of them experiencing neutropenia lasting over 20 days (Na et al., 2022; Islas-Muñoz et al., 2024).

In addition, coagulation disorders associated with haematological malignancies can also increase mortality during septic shock (Lemiale et al., 2023). This study found that septic shock or the use of vasoactive drugs is an independent risk factor for the death of patients with AML after BSI. Royo-Cebrecs et al’ s study is similar to our findings (Chen et al., 2020; Royo-Cebrecos et al., 2022).

Albumin plays a crucial role in mitigating inflammation, maintaining vascular endothelial integrity, and regulating acid-base balance. It also acts to inhibit the progression of inflammation and damage to microcirculation and tissues (Ferrer et al., 2018; Hariri et al., 2018). In the context of infection, capillary permeability increases, causing albumin to seep from the plasma into the interstitial space, resulting in hypoalbuminemia (Gradel et al., 2018). In this study, lower levels of albumin were found to be indicative of adverse outcomes within 30 days following BSI. Notably, research conducted by Benjamin M. Greenberg and others has demonstrated a linear association between albumin levels and the prognosis of bacteremic patients (Greenberg et al., 2005; Magnussen et al., 2016). Diminished albumin levels in patients can lead to immunosuppression following BSI, contributing to poorer patient outcomes (McSorley et al., 2017). Furthermore, low albumin levels in patients with potential infectious foci can lead to infection recurrence (Leal et al., 2018; Abdel-Razik et al., 2020). At the same time, reduced albumin levels are associated with a decreased effectiveness of antibiotic treatment (Ibrahim et al., 2000; Holmes et al., 2018). Low albumin levels serve as a risk factor for treatment failure in Gram-negative bacteremia (Hashiguchi et al., 2020).

Agranulocytosis refers to neutropenia, bone marrow suppression caused by the use of chemotherapeutic drugs or by haematological tumours hinders the generation of an effective immune response, and patients with reduced immunity are prone to infection. The longer the agranulocytosis, the higher the incidence of infection, and the mortality rate of blood stream infection in patients with agranulocytosis and fever in patients with HMs can reach 42% (Zuber et al., 2012; Wohlfarth et al., 2014). The longer the duration of agranulocytosis, the worse the prognosis of HMs patients with infection (Darmon et al., 2002; Li et al., 2018). This study also confirmed this result. However, this study suggests that the duration of neutropenia is not an independent risk factor for death within 30 days after BSI. Tang’s study is consistent with our findings (Tang et al., 2018).This may be due to the early application of empirical antibacterial drugs to timely control the infection caused by agranulocytosis, which reduced the incidence of serious complications and mortality (Falcone et al., 2020; Zhang et al., 2020). This correlation requires additional data for further study and verification.

Stenotrophomonas maltophilia is an opportunistic pathogen that often occurs in hospitals and in immunosuppressed patients (Looney et al., 2009; Adegoke et al., 2017). The infection rate of this bacterium increased gradually because of the low immune function and the wide use of broad-spectrum antibiotics (Said et al., 2024). TMP-SMX is the first choice of treatment for Stenotrophomonas maltophilia infection, but TMP-SMX treatment may cause myelosuppression, hyperkalaemia and hypersensitivity syndrome, and Stenotrophomonas maltophilia infection is relatively rare. Hence, TMP-SMX is not generally used for empirical treatment of BSIs (Brooke, 2012; Chang et al., 2015). In this study, seven patients with Stenotrophomonas maltophilia infection were included, and no drug sensitivity test was performed. Five patients died within 30 days after BSI with Stenotrophomonas maltophilia, with a mortality rate of 71.4%. Sumida et al. also reported similar findings that Stenotrophomonas maltophilia infection was associated with adverse outcomes in patients (Cho et al., 2014; Sumida et al., 2015). A history of recurrent infections and prior antibiotic use are risk factors for Stenotrophomonas maltophilia infection (Apisarnthanarak et al., 2003; Bao et al., 2020). In our study, all the patients with Stenotrophomonas maltophilia infection had a history of repeated infection before BSI, but no multiple infection between Stenotrophomonas maltophilia and other bacteria were found during BSI (Apisarnthanarak et al., 2003; Bao et al., 2020). Among seven patients with Stenotrophomonas maltophilia infection, 5 patients who died within 30 days after BSI had a history of empirical use of carbapenem antibiotics before the pathogen was identified. Stenotrophomonas maltophilia exhibits resistance to some empirically used clinical treatments, particularly β-lactam antibiotics (including carbapenems) and aminoglycosides. As a result, routine empiric therapy often fails to achieve timely and effective control of the Stenotrophomonas maltophilia infection before blood culture results become available (Gales et al., 2019; Mojica et al., 2019).

Nonetheless, these results must be interpreted with caution and a number of limitations should be borne in mind. This study is a single-centre study in Zhejiang Province People’s Hospital, with limited sample size and certain regional characteristics. Thus, patients from other places were still need to be included. In addition, some factors that have been associated with the prognosis of patients with post-infection AML in other studies were not relevant in this study and may require confirmation with a larger sample size.

In this study, 56 AML patients complicated by BSIs were primarily infected with Gram-negative bacteria. Aminoglycosides and carbapenems are recommended for empirical treatment based on the pathogenic spectrum and drug resistance status, pending definitive drug sensitivity results. Serum albumin levels and the presence of septic shock emerged as independent risk factors for mortality within 30 days among AML patients with BSI. Furthermore, prolonged agranulocytosis exceeding 20 days and septic shock were associated with heightened mortality rates in terms of long-term prognosis for AML patients with BSI. In addition, Stenotrophomonas maltophilia infection may be associated with poor prognosis, with TMP-SMX serving as a potential treatment option to manage infection and enhance patient outcomes.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The study was approved by the research involving humans were approved by the ethics committee of Zhejiang people's Hospital (QT2023130). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required from the participants or the participants’ legal guardians/ next of kin in accordance with the national legislation and institutional requirements.

HW: Software, Validation, Writing – original draft, Writing – review & editing. ML: Writing – original draft, Data curation, Writing – review & editing. CS: Data curation, Writing – review & editing, Formal analysis. FS: Writing – review & editing, Data curation. XS: Writing – review & editing. QH: Data curation, Writing – review & editing. YW: Writing – review & editing. YC: Writing – review & editing. XT: Writing – review & editing.