Introduction

Bacteremia remains a significant concern among pediatric patients diagnosed with Acute Lymphoblastic Leukemia (ALL) and Acute Myeloid Leukemia (AML). These patients are highly susceptible to infections due to their immunosuppression resulting from the disease itself or from intensified chemotherapy regimens [1]. Despite advancements in supportive care, bacteremia continues to be a major cause of morbidity and mortality [2]. Understanding the distribution of bacteremia across different pediatric age groups is crucial for developing individualized targeted prevention and treatment strategies.

The general pediatric population is heterogeneous in nature, with variations in infection susceptibility, clinical manifestations, and pathogen distribution across age groups. Infants, for instance, might have different vulnerabilities due to immature immune systems and less developed mucosal barriers compared to older children and adolescents, who may have distinct patterns of exposure and immune responses [3]. Some studies have reported age-dependent variation in bacterial isolates, with Escherichia coli becoming increasingly prevalent after 5 years of age and peaking among children aged ≥10 years, while Klebsiella pneumoniae has been observed most frequently in the 5–9 years age group [4].

Variations in leukemia subtypes (ALL vs. AML), treatment intensity and treatment phase could further influence the risk of bacteremia [5]. Existing literature reports overall incidence rates or focuses on specific treatment phases, such as induction or consolidation; few have assessed the incidence of bacteremia throughout the entire treatment course or phases with less intensive chemotherapy such as maintenance and palliative care [6–9].

While some studies have reported age-stratified data, they are limited and do not comprehensively assess differences across the full pediatric age spectrum or throughout leukemia treatment. Additionally, data linking age-specific pathogen distribution with antibiotic susceptibility, particularly across key classes such as β-lactams, aminoglycosides and glycopeptides, remain limited. Accordingly, this study aimed to investigate the distribution of bacterial pathogens and their susceptibility profile across different age groups of pediatric patients with acute leukemia throughout the entire course of treatment.

Materials and methods

This was a single-center retrospective study conducted at King Hussein Cancer Center (KHCC) located in Amman, Jordan. KHCC is a leading hospital in the Middle East region that offers pediatric oncology services, including advanced treatment for pediatric patients with leukemia, lymphoma, and solid tumors. The study was approved by the KHCC institutional review board (24 KHCC 113), with a waiver of consent given the retrospective nature of the study.

We included all pediatric patients with ALL and AML who were diagnosed with bacteremia upon admission or during their hospital stay between 01/01/2015–31/12/2024. Data collection covered this entire period; however, data were accessed and extracted for this research study between 01/07/2024 and 31/12/2024. Patients under the bone marrow transplant service and those who initiated their leukemia treatment prior to being transferred to KHCC were excluded from the study. The electronic medical records were used to identify patients with bacteremia, defined as the presence of viable bacterial microorganisms in the bloodstream evidenced by one or more positive blood cultures with associated clinical symptoms such as documented fever (≥38°C), chills/shivering, tachycardia, hypotension, mucositis [10].

Patients were categorized into three groups based on age: infants (≤1 year), children (>1 to <12 years), and adolescents (≥12 to ≤18 years). The treatment course was defined as the period beginning at initial diagnosis and extending through all planned phases of treatment. For patients who experienced disease relapse or transitioned to palliative care, these phases were also considered part of the treatment course. The following data were collected for each bacteremia episode: gender, type and subtype of acute leukemia, disease risk and treatment phase of acute leukemia as per KHCC guidelines (risk group classified as high, standard, or low; phases as induction, consolidation, maintenance, relapse, or palliative care) [11–15], remission status, central nervous system (CNS) involvement, presence and type of indwelling central venous catheter (CVC), degree of neutropenia based on absolute neutrophils counts (ANC), presenting signs and symptoms and the causative bacterial pathogen along with its susceptibility and resistance testing.

At our center, newly diagnosed pediatric patients with ALL are classified and treated according to the St. Jude Total XV and Total XIII protocols [11,12]. For relapsed ALL, management follows either the UKALLR3 or the St. Jude ALL-R16 protocol, depending on physician preference and patient clinical status [13,14]. Newly diagnosed pediatric patients with AML are treated according to the St. Jude AML02 trial [15]. Infection prophylaxis included oral and perianal care (sodium bicarbonate mouthwash, mycostatin or miconazole oral gel, povidone solution, and zinc oxide cream), along with trimethoprim/sulfamethoxazole administered twice daily on weekends for Pneumocystis jirovecii prophylaxis. Patients with contraindications to trimethoprim/sulfamethoxazole received pentamidine or dapsone. In addition, ciprofloxacin and vancomycin prophylaxis were administered to AML patients post-chemotherapy until count recovery, defined as an ANC > 1000 cells/µL [15].

Microbiology results and susceptibility patterns of pathogens towards antibiotics were based on local testing performed at KHCC in accordance with Clinical and Laboratory Standards Institute (CLSI) guidelines. Bacterial isolates were classified as non-susceptible if they showed resistant or intermediate susceptibility [16]. Clinical analyses were performed at the bacteremia episode level, while microbiological analyses were conducted at the bacterial isolate level. Unless otherwise stated, patient-level characteristics are counted at the episode level, and patients with multiple episodes may contribute more than once. Blood cultures that were considered contaminants were excluded. A contaminant was defined as a single positive culture without accompanying clinical signs or symptoms of infection (fever, chills/shivering, tachycardia, hypotension, or mucositis). Polymicrobial cultures obtained within the same clinical infectious episode were counted as one bacteremia episode but multiple bacterial isolates; cultures yielding a different organism after clinical resolution or after a documented negative blood culture were counted as a new bacteremia episode. Duplicate cultures yielding the same organism during the same admission were considered a single bacteremia episode and isolate unless separated by a documented negative blood culture, in which case they were considered a new bacteremia episode and isolate.

Categorical data were presented as counts and percentages while continuous data were presented as median and interquartile range (IQR).

Results

Over the study period, 308 bacteremia episodes for 160 patients were identified. Of these, 291 (94%) episodes occurred in ALL patients and 17 (6%) in AML patients. The median age was 5 years (IQR 3–11 years) with children, adolescents, and infants representing 78%, 21%, and 1%, respectively. Males accounted for 194 (63%) of the episodes. Across all age groups, bacteremia was most observed during induction, consolidation and maintenance phases combined; while fewer episodes were seen during relapse and palliative care. Table 1 outlines the baseline demographics and clinical characteristics of bacteremia episodes in pediatric patients with acute leukemia.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Baseline demographics and clinical characteristics of bacteremia episodes in pediatric patients with acute leukemia.

https://doi.org/10.1371/journal.pone.0358779.t001

A total of 330 bacterial isolates were identified, including 312 isolates from patients with ALL (284 B-cell ALL and 28 T-cell ALL) and 18 isolates from patients with AML. Among patients with B-cell ALL, bacterial isolates were most frequent during maintenance (26%), induction (25%), and consolidation (24%). Gram-positive organisms predominated across all treatment phases, particularly among children, whereas a higher proportion of Gram-negative organisms was observed among adolescents during relapse and the palliative care phases. In T-cell ALL, isolates were most common during consolidation (57%). The overall pathogen distribution was broadly comparable to that observed in B-cell ALL, with Staphylococcus spp. representing the most frequently isolated organisms. A relatively higher proportion of Micrococcus spp. was noted among both children and adolescents. Table 2 presents the distribution of bacterial isolates by age group and treatment phase in pediatric patients with acute leukemia.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Distribution of bacterial isolates by age group and treatment phase in pediatric patients with acute leukemia.

https://doi.org/10.1371/journal.pone.0358779.t002

The distribution of bacterial isolates was summarized across leukemia risk categories and age groups. The largest number of bacterial isolates was observed among high-risk patients with ALL and AML across the age groups. Staphylococcus spp. were the most frequently observed organisms in most subgroups. Gram-negative organisms, including Escherichia coli, Pseudomonas spp., and Klebsiella spp., were also observed among children and adolescents in the standard- and high-risk groups. Table 3 presents the distribution of bacterial isolates by age group and leukemia risk category.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Distribution of bacterial isolates by age group and leukemia risk group.

https://doi.org/10.1371/journal.pone.0358779.t003

Among all 330 bacterial isolates, antimicrobial resistance was confined to Escherichia coli and Klebsiella spp. exhibiting Extended-Spectrum β-lactamase (ESBL) and Carbapenem-Resistant Enterobacterales (CRE) phenotypes. ESBL-producing isolates showed reduced susceptibility to beta-lactams, ciprofloxacin, and trimethoprim–sulfamethoxazole. Among CRE-producing Klebsiella spp. isolates, resistance was observed to ertapenem, while susceptibility to imipenem and meropenem was retained. Detailed susceptibility percentages for all antibiotics are presented in Table 4.

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Antimicrobial susceptibility patterns of Escherichia coli and Klebsiella spp. exhibiting ESBL and CRE phenotypes.

https://doi.org/10.1371/journal.pone.0358779.t004

Discussion

This study aimed to describe the bacterial profile and antibiotic susceptibility of bloodstream infections among infants, children, and adolescents with acute leukemia across all treatment phases. Our findings summarize descriptively age- and treatment-phase related differences in pathogen distribution, with potential implications for empiric antimicrobial selection and infection-prevention strategies in pediatric oncology.

Across all age groups, Staphylococcus spp. were the predominant pathogen, consistent with previously published cohorts of pediatric patients with acute leukemia [6,9]. This predominance is expected given the widespread use of CVC and chemotherapy induced mucositis, which is causing a shift toward Gram-positive coccal bacteremia [17]. However, a higher proportion of Gram-negative pathogens was observed among adolescents, including Escherichia coli, Klebsiella spp., and Pseudomonas spp. These findings may be relevant when considering empiric antimicrobial selection in older pediatric age groups, particularly in patients with recurrent admissions, bacteremia episodes, or prolonged neutropenia.

The distribution of pathogens showed descriptive differences across treatment phases. Bacteremia was most frequent during induction and consolidation, which represents the most intensive parts of leukemia therapy. During these phases, patients —particularly those in high-risk groups—receive intensive chemotherapy, have CVCs inserted, and experience treatment-related complications such as profound myelosuppression, febrile neutropenia, and mucositis, often necessitating prolonged hospitalization and increasing the risk of infections from endogenous flora and nosocomial sources. These findings reinforce the importance of close infection monitoring, timely empiric broad-spectrum antimicrobial therapy, and supportive care measures during early intensive treatment phases. Although maintenance therapy is less intensive, it still accounted for a substantial number of episodes because of its prolonged duration of two to three years and the limited neutrophil recovery following steroid pulses [18]. These findings are consistent with previous studies reporting that most bacteremia infections occur during induction and maintenance [6,9].

In contrast, fewer bacteremia episodes were observed during relapse and palliative care. Relatively, more Gram-negative pathogens were identified, particularly among adolescents. However, these observations should be interpreted cautiously given a smaller number of episodes in these treatment phases. Differences in prior antibiotic exposure – with increasing multidrug resistance –, cumulative chemotherapy intensity, and the combined effects of age and treatment intensity may have contributed to the observed distribution of pathogens [19].

In AML, the overall number of bacteremia episodes was lower, which may be partly attributable to the smaller patient population at our center compared with ALL. In addition, the routine use of prophylactic vancomycin and ciprofloxacin when the ANC declines may potentially influence pathogen distribution and infection frequency; however, this study was not designed to evaluate the effect of prophylactic antimicrobial strategies [8,20].

The antimicrobial agents included in susceptibility testing were selected based on institutional microbiology protocols, antibiograms and commonly used empiric and targeted therapies for febrile neutropenia and bloodstream infections in pediatric oncology patients. Attention was given to β-lactams, carbapenems, aminoglycosides, and glycopeptides due to their frequent use in empiric management of high-risk febrile neutropenia. In this study, most antibiotic resistance was observed among ESBL- and CRE-producing Escherichia coli and Klebsiella spp., consistent with global trends in multidrug-resistant Enterobacterales [21]. A study by Hassouneh et al. also reported that the highest prevalence of multidrug-resistant pathogens was observed in Escherichia coli (45%) and Klebsiella pneumoniae (21%) [22]. Surveillance studies from Europe, Asia, and the Middle East have reported a rising prevalence of ESBL- and CRE-producing Escherichia coli and Klebsiella spp. in both hospital and community settings, driven by broad antibiotic selection and the dissemination of mobile resistance genes. This phenomenon is particularly pronounced in pediatric and immune-suppressed populations, such as children with leukemia, who are frequently exposed to broad-spectrum antibiotics [23].

This study has several limitations. As a retrospective single-center analysis, findings may not be generalizable to other settings with different microbiological profiles or infection-control practices. Contamination was defined solely based on the absence of clinical signs or symptoms, which may have led to misclassification. Clinical outcomes including mortality, and inferential statistical analyses were not examined but represent key areas for future research. Multiple bacteremia episodes from the same patient may also have contributed to repeated observations. Furthermore, the infants, adolescents and AML subgroups were limited in number, restricting the interpretation and generalizability of subgroup-specific findings.

Despite these limitations, this study addresses an important gap by providing an age-stratified assessment of bacteremia patterns across the full course of leukemia treatment and may help inform future studies evaluating age-tailored empiric antimicrobial strategies in pediatric oncology. Future multicenter studies with larger AML, infants, and adolescents cohorts are needed to better characterize age-specific pathogen distribution and resistance patterns in pediatric patients with acute leukemia. Prospective studies evaluating clinical outcomes, recurrent bacteremia episodes, antimicrobial prophylaxis strategies, and the impact of empiric antibiotic selection on resistance trends would further strengthen infection-management approaches in this population.

Conclusions

Descriptive differences in bacteremia episodes and bacterial isolates were observed across age groups, treatment phases, and risk categories. Larger multicenter studies are needed to determine whether these patterns should inform age- or phase-specific empiric therapy strategies.

Acknowledgments

The authors have no acknowledgments to declare.