Bacterial bloodstream infections in pediatric oncology patients in a tertiary medical center: pathogen distribution, antimicrobial nonsusceptibility, and outcomes Authors: Menucha Jurkowicz, Eugene Leibovitz, Hila Rosental, Nathan Keller, Gilad Sherman, Efrat Dresner, Guy Katzenellenbogen, Elina Gelman, Michal Paret, Reut Myaskovetzky-Yitshack, Hana Golan, Assaf Barg, Amos Toren, Michal Stein Published: 2026 ABSTRACT The purpose of this study was to analyze bacterial bloodstream infections (BSIs) occurring in hospitalized pediatric hemato-oncology patients during 2012 to 2023. Data on 775 bacterial BSI episodes occurring in 456 patients (292, 63.04%, hematological malignancies and 164, 35.96%, solid tumors) were retrospectively collected and analyzed. In total, 362 (46.71%) bacterial-BSIs had profound neutropenia; 849 pathogens were isolated; gram-negative (GN) (68.78%, 584/849) were more common than gram-positive (GP) pathogens (30.86%, 262/849), P < 0.001. The most frequent GN isolates were Klebsiella spp. (20.38% of all isolates), Escherichia coli (19.20%), and Pseudomonas aeruginosa (12.01%); the most common GP isolates were Streptococcus viridans (9.19%) Enterococcus spp. (6.36%), and Staphylococcus aureus (6.12%, 21.15% MRSA). An increase in GN rates was observed in 2018–2023 versus 2012–2017 (74.16% vs. 63.22%, P = 0.001). Overall nonsusceptibility rates of Klebsiella spp. and E. coli were 80.13% and 90.07% for piperacillin, 16.76% and 9.82% for amikacin, 69.19% and 50.31% for ceftriaxone, 34.68% and 23.92% for piperacillin/tazobactam, and 8.72% and 1.85% for meropenem. Nonsusceptibility rates of Acinetobacter baumannii were 57.14% for both piperacillin/tazobactam and meropenem. The appropriateness of empiric antibiotic therapy (EAT) with piperacillin/tazobactam was lower than piperacillin/tazobactam + amikacin (62.61% vs. 90.45%, P < 0.0001). PICU admission was required in 230 (29.68%); 7-day and 30-day mortality rates were 6.36% and 14.69%, respectively.Conclusions:We report an increase in GN-BSIs and in nonsusceptibility rates of Klebsiella spp. and E. coli to several first/second-line antibiotics used in empiric antibiotic therapy. In our setting, an EAT protocol including piperacillin/tazobactam and amikacin covered most pathogens and should be considered in stable children with suspicion of BSI. What is Known:• Bloodstream infections and antibiotic resistance rates are increasing in pediatric hemato-oncology patients.What is New:• We found a dominance of gram-negative pathogens and moderate to high resistance rates of Klebsiella spp. and Escherichia coli for piperacillin/tazobactam, ciprofloxacin, and ceftriaxone.• Empiric antibiotic therapy was appropriate for 77.07% of evaluable pathogens.• The appropriateness of empiric antibiotic therapy with piperacillin/tazobactam alone was lower than that of piperacillin/tazobactam plus amikacin. Introduction Bloodstream infections (BSIs) occur in up to 50% of pediatric cancer patients [1–6] and the characteristics of BSI acquisition and mortality can be cancer-specific [3–5, 7, 8]. In particular, BSIs are more frequent in patients with acute myeloid leukemia (AML) and during induction therapy, while BSI characteristics differ between hematologic malignancies and solid tumors [9–11]. The choice of empiric antibiotic treatment (EAT) for suspected BSI must consider the patient’s clinical status including previous colonization with relevant pathogens, predominant local pathogens, and regional antimicrobial susceptibility patterns [3, 4]. In recent decades, the epidemiology of BSIs in hemato-oncology patients has changed considerably, with a significant decrease in gram-positive (GP) pathogens and an increase in gram-negative (GN) bacteria [12–15]. More recently, a progressive rise in multidrug-resistant (MDR) strains has been reported [16–18]. In Europe, a 2014 survey from 39 hematology centers reported infection incidence rates of 15–24% for extended-spectrum β-lactamase-producing (ESBL) Enterobacteriaceae, 5–14% for aminoglycoside-resistant GN bacteria, and 5–14% for carbapenem-resistant Pseudomonas aeruginosa [19]. In a multicenter cohort study including 811 GN BSI episodes in Italian patients, Trecarichi et al. [20] reported a significant reduction in use of fluoroquinolone prophylaxis and a significant increase in susceptibility rates to ciprofloxacin. Antibiotic resistance adversely affects survival among patients with hematological malignancies and hematopoietic stem cell transplant (HSCT) recipients [8, 15, 18–21]. O’Connor et al. reported a 30-day mortality rate of 3% among pediatric cancer patients with BSIs [22], while Willis et al. [23] reported a mortality rate of 6.5% in pediatric oncologic/HSCT patients with BISs compared with 0.7% in patients without BISs. The high incidence of BSIs in hemato-oncology patients and the emergence of antibiotic resistance have led to an increased use of broad-spectrum antibiotics, including carbapenems, either as monotherapy or as combination therapy [3, 4, 8]. Knowledge of the local epidemiology of BSIs is pivotal in guiding the correct selection of antibiotic therapy in pediatric cancer patients. Local epidemiology should be incorporated into institutional guidelines and standard operating procedures that guide EAT at the time of initial presentation. In this study, we aimed to evaluate the epidemiology, microbiology, antimicrobial susceptibility patterns, EAT management, and outcomes of bacterial BSIs in pediatric hemato-oncology patients during a 12-year period (2012–2023). Patients and methods We retrospectively analyzed bacterial BSI episodes diagnosed between 2012 and 2023 in oncologic children aged < 18 years at Sheba Medical Center (a 2000-bed, tertiary, university-affiliated hospital in Israel). The study was approved by the Sheba Medical Center Institutional Review Board (SMC-D-0915–23). The type of underlying disease was classified as follows: (1) hematologic malignancies including acute lymphoblastic leukemia (ALL), AML, Hodgkin’s disease, non-Hodgkin lymphoma, and other leukemias and (2) solid tumors including neuroblastoma, bone or soft tissue sarcoma, central nervous system tumor, and other solid tumors. Inclusion criteria were a diagnosis of BSI (due to GP and GN pathogens or Mycobacteria spp.) in hemato-oncology patients treated with chemotherapy or allogeneic HSCT. In cases of suspected infection, two sets of blood samples had been drawn and processed in the microbiology laboratories. Regarding coagulase-negative staphylococci, Micrococcus species, and other skin contaminants, at least 2 consecutive sets of positive blood cultures were required to confirm a BSI [24]. Clinical management In each episode, demographic, clinical, microbiology, management, and outcome data were retrieved using MDClone ADAMS platform, a self-service query tool providing comprehensive patient-level data of wide-ranging variables in a defined timeframe around an index event (mdclone.com). The first-line dedicated EAT protocol for hemato-oncology patients admitted from the community with suspected BSI, including fever and neutropenia, was piperacillin plus amikacin or piperacillin/tazobactam from 2012 to 2015 and piperacillin/tazobactam plus amikacin for the period 2016–2023. The second-line EAT dedicated protocol was meropenem, prescribed in patients who developed signs of breakthrough infection during treatment with first-line protocol or in patients who appeared to be in severe sepsis. The patients received antineoplastic therapy according to international protocols and recommendations, and the only antibiotic prophylaxis used was anti-Pneumocystis jiroveci prophylaxis with trimethoprim-sulfamethoxazole (twice daily, thrice weekly) which was routinely administered during the study period to all children with malignancies and after engraftment post-HSCT. Definitions Fever was defined as a single-temperature measurement > 38.0°C at admission. Neutropenia was defined as ≤ 0.5 × 1099/L, and profound neutropenia was defined as ≤ 0.1 × 109/L. BSI was defined as the detection of true pathogen bacteria in 1 or more blood cultures and polymicrobial BSIs were defined when ≥ 2 true pathogens were detected in a single blood culture. For pathogen susceptibilities analyzed, the “intermediate” susceptibility category was considered nonsusceptible for clinical purposes. The antibiotic therapy administered was considered appropriate when the isolated pathogen was susceptible to at least one of the antibiotics used empirically in the first 48–72 h. Mortality was assessed within 7 and 30 days following the bacterial BSI diagnosis; 7-day mortality was considered as associated with the bacterial BSI episode. Microbiology laboratory procedures The blood cultures were processed in the Bactec system (Becton Dickinson) from 2012 to 2021, and BactAlert (BioMérieux) from 2022 to 2023. Bacteria were identified according to routine laboratory procedures, using matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) (Bruker Daltonics). Antimicrobial susceptibility testing (AST) was performed using the disk diffusion method according to CLSI guidelines [25]. Statistical analysis Patient demographics and clinical characteristics, as well as microbiological data, were reported as counts and percentages for categorical variables, and as median with interquartile range (IQR) for continuous variables. Categorical values and proportions were assessed using the chi-square test or Fisher’s exact test. For all analyses, two-sided p-values < 0.05 were considered statistically significant. All statistical analyses were performed using R Statistical Software (version 4.5.2; R Foundation for Statistical Computing, Vienna, Austria). Results Demographics and epidemiology In total, 775 bacterial BSI episodes were identified in 456 patients, with a median age of 5.51 years (Table 1). There were 273 (59.87%), 101 (22.15%) and 47 (10.31%) patients with 1, 2, or 3 bacterial BSI episodes. There were 292 (63.04%) patients with hematologic malignancies (mainly acute lymphoblastic leukemia and acute myeloid leukemia) and 164 (35.96%) with solid tumors (mainly neuroblastoma and primary brain/central nervous system (CNS) tumors). Table 1Patient characteristics (456 children with 775 bacterial BSI episodes during 2012–2023)CharacteristicPatients with bacterial-BSIGenderMale (n, %)249 (54.61)Female (n, %)207 (45.39)Age groups at BSI episode: < 1 years (n, %)54 (6.97)0–3 years (n, %)279 (36.00)4–7 years (n, %)211 (27.23)8–12 years (n, %)128 (16.52)13–17 years (n, %)157 (20.26)Age at BSI episode (years, median [IQR])5.51 [2.63–10.83]Comorbidities (n, %)54 (11.84)Central nervous system (n, %)17 (31.48)Infectious (n, %)11 (20.37)Respiratory (n, %)6 (11.11)Hematology (n, %)5 (9.26)Cardiovascular (n, %)5 (9.26)Other (n, %)10 (18.52)No. of patients with BSI episodes (n)456Patients with 1 episode (n, %)273 (59.87)Patients with 2 episodes (n, %)101 (22.15)Patients with 3 episodes (n, %)47 (10.31)Patients with > 3 episodes (n, %)35 (7.67)Hematologic malignancies (n, %)292 (64.03)Acute lymphoblastic leukemia (n, %)160 (54.79)Acute myeloid leukemia (n, %)75 (25.68)Non-Hodgkin’s lymphoma (n, %)20 (6.85)Hodgkin’s lymphoma (n, %)13 (4.45)Burkitt lymphoma (n, %)13 (4.45)Myelodysplastic syndrome (n, %)5 (1.71)Myelofibrosis (n, %)1 (0.34)Other (n, %)5 (1.71)Solid tumors (n, %)164 (35.96)Neuroblastoma (n, %)45 (27.44)Primary brain/CNS tumors (n, %)35 (21.34)Rhabdomyosarcoma and nonrhabdomyosarcomasoft tissue sarcomas (n, %)30 (18.29)Wilm’s tumor (n, %)15 (9.15)Ewing sarcoma (n, %)13 (7.93)Retinoblastoma (n, %)10 (6.10)Osteosarcoma (n, %)3 (1.83)Hepatoblastoma (n, %)3 (1.83)Other (n, %)10 (6.10)Anti-PCP prophylaxis (TMP/SMX) (n, %)456 (100.0)Hematopoietic stem cell transplant (n, %)36 (7.89) Table 1Patient characteristics (456 children with 775 bacterial BSI episodes during 2012–2023)CharacteristicPatients with bacterial-BSIGenderMale (n, %)249 (54.61)Female (n, %)207 (45.39)Age groups at BSI episode: < 1 years (n, %)54 (6.97)0–3 years (n, %)279 (36.00)4–7 years (n, %)211 (27.23)8–12 years (n, %)128 (16.52)13–17 years (n, %)157 (20.26)Age at BSI episode (years, median [IQR])5.51 [2.63–10.83]Comorbidities (n, %)54 (11.84)Central nervous system (n, %)17 (31.48)Infectious (n, %)11 (20.37)Respiratory (n, %)6 (11.11)Hematology (n, %)5 (9.26)Cardiovascular (n, %)5 (9.26)Other (n, %)10 (18.52)No. of patients with BSI episodes (n)456Patients with 1 episode (n, %)273 (59.87)Patients with 2 episodes (n, %)101 (22.15)Patients with 3 episodes (n, %)47 (10.31)Patients with > 3 episodes (n, %)35 (7.67)Hematologic malignancies (n, %)292 (64.03)Acute lymphoblastic leukemia (n, %)160 (54.79)Acute myeloid leukemia (n, %)75 (25.68)Non-Hodgkin’s lymphoma (n, %)20 (6.85)Hodgkin’s lymphoma (n, %)13 (4.45)Burkitt lymphoma (n, %)13 (4.45)Myelodysplastic syndrome (n, %)5 (1.71)Myelofibrosis (n, %)1 (0.34)Other (n, %)5 (1.71)Solid tumors (n, %)164 (35.96)Neuroblastoma (n, %)45 (27.44)Primary brain/CNS tumors (n, %)35 (21.34)Rhabdomyosarcoma and nonrhabdomyosarcomasoft tissue sarcomas (n, %)30 (18.29)Wilm’s tumor (n, %)15 (9.15)Ewing sarcoma (n, %)13 (7.93)Retinoblastoma (n, %)10 (6.10)Osteosarcoma (n, %)3 (1.83)Hepatoblastoma (n, %)3 (1.83)Other (n, %)10 (6.10)Anti-PCP prophylaxis (TMP/SMX) (n, %)456 (100.0)Hematopoietic stem cell transplant (n, %)36 (7.89) Patient characteristics (456 children with 775 bacterial BSI episodes during 2012–2023) Rhabdomyosarcoma and nonrhabdomyosarcoma soft tissue sarcomas (n, %) Background disease history/comorbidities were reported in 44 (9.65%) of the patients, the majority represented by central nervous system conditions (38.64% of all comorbidities, of them 5 patients with ataxia-telangiectasis, 4 facial nerve paralysis, and 2 epilepsy cases); infectious conditions (25%, 2 zoster, 2 meningitis, and 2 sinusitis cases); respiratory conditions (13.64%, 3 bronchiolitis and 2 pneumonia); and hematological nononcologic conditions (11.36%, 3 thrombocytopenia and 2 Pearson disease). Three (0.6%) patients suffered from Down’s syndrome. Overall, 297 (38.32%) episodes had a fever ≥ 38.0 °C at admission and 151 (19.48%) had a fever ≥ 39.0 °C (Table 2). Neutropenia ≤ 0.5 × 109/L was recorded in 446 (57.54%) episodes and profound neutropenia (absolute neutrophil count (ANC) ≤ 0.1 × 109/L) in 362 (46.71%) of the episodes. Fever ≥ 38.0 °C and neutropenia ≤ 0.5 × 109/L occurred in 381 (49.16%) of all episodes, with a median patient age of 6.49 (IQR 3.55–12.42). Fever ≥ 38.0 °C without neutropenia ≤ 0.5 × 109/L was recorded in 229 (29.55%) episodes, with a median patient age of 4.87 (IQR 2.43–8.48). Neutropenia ≤ 0.5 × 109/L without fever ≥ 38.0 °C occurred in 72 (9.29%) episodes with a median patient age of 6.40 (IQR 2.48–13.58). Table 2Laboratory, imaging, and outcome (456 patients with 775 bacterial-BSI episodes)CharacteristicPatients with bacterial-BSIFever at time of culture:Temperature ≥ 38.0˚C (n, %)610 (78.71)Fever ≥ 38.5˚C (n, %)297 (38.32)Fever ≥ 39.0˚C (n, %)151 (19.48)Fever (≥ 38.0˚C) without neutropenia (ANC* ≤ 0.5 × 109/L) (n, %)229 (29.55)Neutropenia (ANC* ≤ 0.5 × 109/L) (n, %)446 (57.55)Profound neutropenia (ANC* ≤ 0.1 × 109/L) (n, %)362 (46.71)Neutropenia (ANC* ≤ 0.5 × 109/L) without fever(≥ 38.0˚C):72 (9.29)Fever (≥ 38.0˚C) and neutropenia (ANC* ≤ 0.5 × 109/L) at time of culture (n, %):381 (49.16)Chest X-rays at time of culture**Normal (n, %)225 (61.98)New alveolar infiltrates (n, %)78 (21.49)Pleural effusion (n, %)33 (9.09)New diffuse interstitial infiltrates (n, %)32 (8.82)Septic shock (n, %)81 (10.45)PICU admission*** (n, %)230 (29.68)Mechanical ventilation*** (n, %)55 (7.10)Death7-day all-cause mortality (n, %)29 (6.36)30-day all-cause mortality (n, %)67 (14.69)*ANC refers to absolute neutrophil count**Information was available in 363 bacterial-BSI episodes***During first 14 days from positive culture Table 2Laboratory, imaging, and outcome (456 patients with 775 bacterial-BSI episodes)CharacteristicPatients with bacterial-BSIFever at time of culture:Temperature ≥ 38.0˚C (n, %)610 (78.71)Fever ≥ 38.5˚C (n, %)297 (38.32)Fever ≥ 39.0˚C (n, %)151 (19.48)Fever (≥ 38.0˚C) without neutropenia (ANC* ≤ 0.5 × 109/L) (n, %)229 (29.55)Neutropenia (ANC* ≤ 0.5 × 109/L) (n, %)446 (57.55)Profound neutropenia (ANC* ≤ 0.1 × 109/L) (n, %)362 (46.71)Neutropenia (ANC* ≤ 0.5 × 109/L) without fever(≥ 38.0˚C):72 (9.29)Fever (≥ 38.0˚C) and neutropenia (ANC* ≤ 0.5 × 109/L) at time of culture (n, %):381 (49.16)Chest X-rays at time of culture**Normal (n, %)225 (61.98)New alveolar infiltrates (n, %)78 (21.49)Pleural effusion (n, %)33 (9.09)New diffuse interstitial infiltrates (n, %)32 (8.82)Septic shock (n, %)81 (10.45)PICU admission*** (n, %)230 (29.68)Mechanical ventilation*** (n, %)55 (7.10)Death7-day all-cause mortality (n, %)29 (6.36)30-day all-cause mortality (n, %)67 (14.69)*ANC refers to absolute neutrophil count**Information was available in 363 bacterial-BSI episodes***During first 14 days from positive culture Laboratory, imaging, and outcome (456 patients with 775 bacterial-BSI episodes) *ANC refers to absolute neutrophil count **Information was available in 363 bacterial-BSI episodes ***During first 14 days from positive culture Chest X-rays were performed at admission in 363 episodes; a new alveolar infiltrate and a new diffuse interstitial infiltrate were diagnosed in 78/363 (21.49%) and 32/363 (8.82%) of examinations. A viral co-infection was diagnosed in 42 episodes (cytomegalovirus (CMV), Epstein–Barr virus (EBV), respiratory syncytial virus (RSV), varicella, and rhinovirus in 14, 7, 7, 7, and 6 patients, respectively). Blood culture results A total of 849 pathogens were identified in 775 bacterial BSI episodes (Table 3). There were 584 (68.78%) GN pathogens, 262 (30.86%) GP pathogens, and 3 (0.4%) Mycobacterium spp. Monomicrobial bacterial BSI accounted for 706 (91.10%) episodes, while polymicrobial bacterial BSI accounted for 69 episodes (8.90%). Table 3Microbiological data (849 pathogens recovered in 775 bacterial BSI episodes)Total pathogensPart of polymicrobial BSIGram-negative bacteria (n, % of all pathogens)584 (68.78)*Klebsiella spp.173 (20.38)35 (20.23)**Escherichia coli163 (19.20)31 (19.02)Pseudomonas aeruginosa102 (12.01)21 (20.59)Enterobacter spp.53 (6.24)8 (15.09)Acinetobacter baumannii14 (1.65)3 (21.43)Citrobacter spp.14 (1.65)7 (50.0)Stenotrophomonas maltophilia13 (1.53)5 (38.46)Salmonella non-typhi spp.13 (1.53)1 (7.69)Serratia spp.7 (0.82)3 (42.86)Proteus mirabilis6 (0.71)1 (16.67)Moraxella catarrhalis5 (0.59)0 (0.0)Bacteroides spp.4 (0.47)0 (0.0)Other17 (2.00)2 (11.76)Gram-positive bacteria (n, %)262 (30.86)Streptococcus viridans spp.78 (9.19)4 (5.13)Enterococcus spp.54 (6.36)14 (25.93)Staphylococcus aureus52 (6.12)5 (9.62)Coagulase-negative Staphylococcus44 (5.18)0 (0.0)Streptococcus pneumoniae17 (2.00)1 (5.88)Abiotrophia defectiva5 (0.59)1 (20.00)Clostridium spp.4 (0.47)0 (0.0)Streptococcus pyogenes3 (0.35)0 (0.0)Bacillus cereus3 (0.36)0 (0.0)Streptococcus dysgalactiae2 (0.24)1 (50.00)Mycobacterium spp.3 (0.4)0 (0.0)*Percentage of total number of isolated pathogens**Percentage of total isolates of each pathogen Table 3Microbiological data (849 pathogens recovered in 775 bacterial BSI episodes)Total pathogensPart of polymicrobial BSIGram-negative bacteria (n, % of all pathogens)584 (68.78)*Klebsiella spp.173 (20.38)35 (20.23)**Escherichia coli163 (19.20)31 (19.02)Pseudomonas aeruginosa102 (12.01)21 (20.59)Enterobacter spp.53 (6.24)8 (15.09)Acinetobacter baumannii14 (1.65)3 (21.43)Citrobacter spp.14 (1.65)7 (50.0)Stenotrophomonas maltophilia13 (1.53)5 (38.46)Salmonella non-typhi spp.13 (1.53)1 (7.69)Serratia spp.7 (0.82)3 (42.86)Proteus mirabilis6 (0.71)1 (16.67)Moraxella catarrhalis5 (0.59)0 (0.0)Bacteroides spp.4 (0.47)0 (0.0)Other17 (2.00)2 (11.76)Gram-positive bacteria (n, %)262 (30.86)Streptococcus viridans spp.78 (9.19)4 (5.13)Enterococcus spp.54 (6.36)14 (25.93)Staphylococcus aureus52 (6.12)5 (9.62)Coagulase-negative Staphylococcus44 (5.18)0 (0.0)Streptococcus pneumoniae17 (2.00)1 (5.88)Abiotrophia defectiva5 (0.59)1 (20.00)Clostridium spp.4 (0.47)0 (0.0)Streptococcus pyogenes3 (0.35)0 (0.0)Bacillus cereus3 (0.36)0 (0.0)Streptococcus dysgalactiae2 (0.24)1 (50.00)Mycobacterium spp.3 (0.4)0 (0.0)*Percentage of total number of isolated pathogens**Percentage of total isolates of each pathogen Microbiological data (849 pathogens recovered in 775 bacterial BSI episodes) *Percentage of total number of isolated pathogens **Percentage of total isolates of each pathogen The most frequent GN pathogens were Klebsiella spp. (20.38% of all pathogens), followed by Escherichia coli (19.20%) and Pseudomonas aeruginosa (12.10%). The Enterobacteriaceae group represented 50.53% of all pathogens and 73.46% of all GN pathogens. Among GN pathogens, Klebsiella spp., E. coli, and P. aeruginosa represented 29.62%, 27.91%, and 17.47% of all isolates. The most common GP pathogens were Streptococcus viridans spp. (9.19% of all isolates), Enterococcus spp. (6.36%), and Staphylococcus aureus (6.12%). Among GP pathogens, S. viridans spp., Enterococcus spp., and S. aureus represented 29.77%, 20.61%, and 19.85 of all isolates, respectively. Overall, 397 pathogens were isolated during the first 6 years of the study (2012–2017) and 449 (a 13.1% increase) during the last 6 years of the study (2018–2023). An increase of 17.30% in the percentages of GN pathogens was observed between the two study periods (251/397, 63.22% vs. 333/449, 74.16%, P = 0.001), accompanied by a decrease of 29.74% in the percentages of GP pathogens (146/397, 36.78% vs. 116/449, 25.84%, P = 0.001) (Fig. 1). An increase in Enterobacter spp. rates (15/251, 5.98%, vs. 38/333, 11.41%, P = 0.024) among all GN pathogens was observed in the second study period. A decrease in the percentage of E. coli (82/251, 32.67%, vs. 81/333, 24.32%, P = 0.026) among all GN pathogens was observed in the second study period. No differences between the two study periods were recorded in the percentages of Klebsiella spp., P. aeruginosa, and A. baumannii. Regarding GP pathogens, no statistically significant differences were observed between the two study periods (Fig. 1).Fig. 1Distribution of A gram-negative and B gram-positive pathogens during the two study periods (2012–2017 and 2018–2023).*p ≤ 0.05 and.**p ≤ 0.01 Distribution of A gram-negative and B gram-positive pathogens during the two study periods (2012–2017 and 2018–2023).*p ≤ 0.05 and.**p ≤ 0.01 Antibiotic nonsusceptibility The nonsusceptibility to antibiotics was analyzed for 706/849 (83.16%) pathogens, including the most common five GN (Klebsiella spp., E. coli, P. aeruginosa., Enterobacter spp., and A. baumannii; total 505 pathogens) and four GP (S. viridans, Enterococcus spp., S. aureus, and S. pneumoniae; total 201 pathogens) pathogens. Overall nonsusceptibility rates of Klebsiella spp. and E. coli were 80.13% and 90.07% for piperacillin, 16.76% and 9.82% for amikacin, 58.23% and 50.92% for ciprofloxacin, 69.19% and 50.31% for ceftriaxone, 34.68% and 23.92% for piperacillin/tazobactam, and 8.72% and 1.85% for meropenem. ESBLs were identified in most evaluable Klebsiella spp. and of E. coli isolates (98.10% and 100.00%, respectively). Nonsusceptibility rates of P. aeruginosa. to piperacillin, amikacin, ciprofloxacin, piperacillin/tazobactam and meropenem were 8.91%, 2.97%, 5.00, 8.91%, and 5.00%, respectively. The nonsusceptibility rates of Enterobacter spp. to piperacillin/tazobactam and meropenem were 19.23% and 5.66%, respectively, while the respective rates of nonsusceptibility of A. baumanii were 57.14% and 57.14%, respectively. Among GP isolates, 4.20% (11/262, 21.15% of all S. aureus isolates) were MRSA; 67.95%, 29.49%, and 59.21% S. viridans isolates were nonsusceptible to penicillin, ceftriaxone, and erythromycin. Ten S. pneumoniae isolates were resistant to penicillin and all seventeen isolates were sensitive to ceftriaxone. Of 31 E. faecium isolates, 4 (12.90%) were vancomycin-resistant. Antibiotic nonsusceptibility dynamics (2012–2017 vs. 2018–2023) Between the two study periods, an increase was observed in the nonsusceptibility rates of K. pneumoniae to ciprofloxacin (48.57 to 65.00%, P = 0.03) and meropenem (1.43 to 13.73%, P = 0.001), and in the nonsusceptibility rates of E. coli to ciprofloxacin (41.46 to 60.49%, P = 0.02); the E. coli isolates showed an increasing trend for nonsusceptibility to piperacillin/tazobactam (18.29 to 29.63%, P = 0.09) and meropenem (0.00 to 3.70%, P = 0.08) (Fig. 2). Among GP pathogens, no changes in MRSA rates were observed during the two periods studied while a decrease was reported in the nonsusceptibility of S. viridans to penicillin (77.78% vs. 54.55%, P = 0.05) and ceftriaxone (42.22% vs. 12.12%, P = 0.04).Fig. 2Nonsusceptibility of Gram-negative pathogens during the two study periods (2012–2017 and 2018–2023): A Klebsiella spp., B Escherichia coli, C Pseudomonas aeruginosa, and D Enterobacter spp. *p ≤ 0.05 and.**p ≤ 0 Nonsusceptibility of Gram-negative pathogens during the two study periods (2012–2017 and 2018–2023): A Klebsiella spp., B Escherichia coli, C Pseudomonas aeruginosa, and D Enterobacter spp. *p ≤ 0.05 and.**p ≤ 0 Empiric antibiotic treatment (EAT) EAT appropriateness was analysed in 689/849 (81.15%) evaluable pathogens, including the most common five GN (Klebsiella spp., E. coli, P. aeruginosa., Enterobacter spp., and A. baumannii; total 500 pathogens) and four GP (S. viridans, Enterococcus spp., S. aureus, and S. pneumoniae; 189 pathogens) pathogens. No EAT was administered in 15 episodes. The most frequent choices of EAT were piperacillin/tazobactam plus amikacin (199/689 isolates, 28.88%), piperacillin/tazobactam (115, 16.69%), piperacillin plus amikacin (90, 13.06%), and meropenem (77, 11.18.2%). Vancomycin was used for 60 isolates (8.71%, as single therapy for 3 isolates and as part of double antibiotic therapy for 41). EAT appropriateness The EAT was appropriate for 531/689 (77.07%) evaluable pathogens, of them for 408/500 (81.6%) GN and 124/189 (65.61%) GP pathogens. The appropriateness of therapy was 75.59%, 82.1%, 93.0%, 88.68%, and 53.85% for Klebsiella spp., E. coli, P. aeruginosa, Enterobacter spp., and A. baumannii, respectively. The most frequently used EAT (piperacillin/tazobactam plus amikacin) was found appropriate in 180/199 (90.45%) treatments. The appropriateness of EAT with piperacillin/tazobactam alone was significantly lower (72/115, 62.61%), P < 0.0001. Patient outcomes PICU admissions were required in 230 (29.68%, 252 pathogens) episodes, and mechanical ventilation was reported in 55 (7.10%) episodes. The four most common recovered pathogens in patients admitted to PICU were: E. coli (61/252, 24.21%), K. pneumoniae (59/252, 23.41%), P. aeruginosa (28/252, 11.11%), and Enterococcus spp. (18/252, 7.14%, 9 E. faecium). Septic shock occurred in 81 (10.45%) BSI episodes (monobacterial in 73 episodes and polybacterial in 8 episodes). Five of these polymicrobial BSI episodes were caused by Klebsiella spp. plus E. coli. From a total of 88 pathogens (77 GN, 11 GP) isolated in patients with septic shock, the most common were K. pneumoniae and E. coli (28/88, 31.82% and 27/88, 30.68%) isolates, followed by P. aeruginosa (8, 9.09%) and Enterobacter spp. (7, 7.95%). Overall, 10.81% of the GN isolates responsible for septic shock (8/74 checked) were nonsusceptible to meropenem. The respective nonsusceptibility rates to piperacillin/tazobactam, amikacin, piperacillin/tazobactam plus amikacin, and meropenem plus amikacin were 24.32% (18/74), 17.57% (13/74), 13.51% (10/74), and 8.1% (6/74), respectively. Death occurring during the first 7 days was recorded in 29/456 (6.36%, 32 pathogens) of patients. The three most common recovered pathogens in fatalities within 7 days were K. pneumoniae (10/32, 31.25%), Enterococcus spp. (4/32, 12.5%), and P. aeruginosa (4/32, 12.5%). The 30-day mortality rate was 14.69% (67/456). Discussion Recent information on the microbiological characteristics, antibiotic resistance patterns, therapeutic protocols, and outcomes in pediatric patients hospitalized in hematology-oncology units in Israel is limited. In Jerusalem, Vinker-Shuster et al. [26] characterized 85 BSI episodes caused by gram-negative–resistant pathogens in 270 hemato-oncologic and post-HSCT children admitted during 2010–2014. Overall, 10% of the pathogens were carbapenem-resistant. Carbapenem-resistant vs. carbapenem-sensitive bacteremia was associated with longer duration of bacteremia, higher risk for PICU admission, and higher mortality rates. Arad-Cohen et al. [27] reported on GN bacteremias among 62 children with de novo acute myelocytic leukemia; 77.4% had 1–8 bacteremia episodes. Of 238 chemotherapy courses, 98 (41.2%) developed bacteremia episodes: 66 (67.3%) isolates were GN rods (mostly E. coli and Klebsiella spp.), mortality was high (9.7%), and most isolates were sensitive to vancomycin plus meropenem (94.75%). Our retrospective study enrolled a considerable number of children with cancer and bacterial BSIs over a 12-year period. We found that 21.29% episodes were afebrile at admission and profound neutropenia was reported in 46.71% of the episodes. The proportion of neutropenic children in this study was comparable with other studies [13, 14, 20–22, 26, 28]. In addition, a new alveolar infiltrate and a new diffuse interstitial infiltrate were diagnosed in 21.49% and 8.82% of patients in whom X-rays examinations were performed at admission, emphasizing the involvement of respiratory infections in the etiology of BSIs in hemato-oncologic patients. GN pathogens represented 68.78% of all pathogens, with a significant increase to 77.26% during the last 6 years of the study. These findings are in accordance with recent reports emphasizing the return of GNs predominance in most BSIs in children with cancer [3, 4, 14, 18, 19]. We showed that most common GN pathogens were represented by Klebsiella spp., E. coli, and P. aeruginosa, and the Enterobacteriaceae spp. group represented 73.46% of all GN pathogens. The most common GP pathogens were represented by S. viridans, Enterococcus spp., and S. aureus. We found an increase in Enterobacter spp. rates isolated in the second study period while Klebsiella spp. P. aeruginosa, A. baumannii, and S. aureus recovery rates were stable during the study period. Among GN pathogens, resistance to carbapenems in infections caused by K. pneumoniae and Acinetobacter spp. has recently been reported as one of the major emerging causes of severe and fatal infections in hemato-oncology patients [29–35]. We reported high (> 55%) nonsusceptibility rates of Klebsiella spp. and E. coli to piperacillin, ceftriaxone, and ciprofloxacin and moderately high nonsusceptibility rates (34.68% and 23.92%) to piperacillin/tazobactam. ESBLs were identified in most Klebsiella spp. and of E. coli isolates. We also showed relatively high nonsusceptibility rates (57.14% and 57.14%, respectively) to piperacillin/tazobactam and meropenem among A. baumannii isolates. Among the GP pathogens, 21.15% of S. aureus isolates were MRSA and the nonsusceptibility of S. viridans of penicillin, ceftriaxone, and erythromycin was high. These findings are of concern and are similar to data previously reported in the medical literature, although major differences were observed between different countries [13–17, 27–29]. The most frequent EAT used were piperacillin/tazobactam plus amikacin, piperacillin/tazobactam alone, piperacillin plus amikacin, and meropenem. Recent guidelines for the management of fever and neutropenia in pediatric cancer patients recommend monotherapy with an antipseudomonal penicillin, such as piperacillin-tazobactam, or a fourth-generation cephalosporin or a carbapenem, as EAT. No difference in treatment outcome has been described between the suggested antibiotics [3, 4, 19, 20, 28]. However, we observed relatively high nonsusceptibility rates to piperacillin/tazobactam among the main GN pathogens and therefore cannot recommend, in our setting, use of this drug alone as the first-line EAT. We consider that our antibiotic nonsusceptibility findings are worrisome, particularly in respect to ceftazidime and piperacillin/tazobactam nonsusceptibility rates vs. Klebsiella spp. and E. coli, which may result in an increased risk of treatment failure [13, 14, 17, 19, 20, 28, 29, 31]. During the last years of the study, an EAT was used with a β-lactam plus β-lactamase inhibitor combination (piperacillin/tazobactam) plus amikacin, for clinically stable patients. The first-line EAT combination (piperacillin/tazobactam plus amikacin) was found to cover the majority (90.45%) of pathogens causing bacterial BSIs at our medical center. Carbapenems, such as meropenem, were usually reserved for treatment of more resistant pathogens, clinically unstable patients, and as EAT in sites with a high prevalence of resistant GN [3, 4, 20, 28]. For clinically unstable patients, even when at low risk of infection with resistant pathogens, patients with septic shock, those with previous bacterial BSIs or previously colonized with multidrug-resistant pathogens or as a second line of therapy after failure of first-line therapies, we recommended the use of an empiric treatment with piperacillin/tazobactam, meropenem, or meropenem plus amikacin, with the possible addition of vancomycin. This study has some limitations. First, we report a single-center experience, which limits the generalizability of our findings. To balance this limitation, we assessed a large cohort of 838 BSIs through 12 years in children with a wide spectrum of malignancies. Second, the retrospective study design might have missed some clinical, laboratory, and treatment characteristics and may have increased the risk of bias in some cases. In conclusion, we reported an overall dominance of GN-BSI isolates during the study period with an increase in GN-BSIs rates during the last 6 years of the study. We reported increased antimicrobial resistance rates of Klebsiella spp. and E. coli against piperacillin/tazobactam and of A. baumannii against meropenem. In our setting, use of an EAT protocol including piperacillin/tazobactam plus amikacin was found to cover the majority of pathogens isolated in the study patients.