OBJECTIVE: To explore the clinical application value of metagenomic next-generation sequencing (mNGS) in pathogen detection of bloodstream infection secondary to hematologic diseases in children. METHODS: 42 children with bloodstream infections secondary to hematologic diseases admitted to the Children's Hematology and Tumor Center of the Affiliated Hospital of Guangdong Medical University from November 2021 to May 2023 were included in the study, and their clinical data, results of peripheral blood mNGS and traditional blood culture, pathogen distribution characteristics, and diagnostic efficacy of mNGS were retrospectively analyzed. RESULTS: Among the 42 children included, there were 2 cases (4.8%) of aplastic anemia (AA), 27 cases (64.3%) of acute lymphoblastic leukemia (ALL), 7 cases (16.7%) of acute myeloid leukemia (AML), 1 case (2.4%) of chronic myeloid leukemia (CML), 2 cases (4.8%) of hemophagocytic lymphohistiocytosis (HLH), 2 cases (4.8%) of non-Hodgkin lymphoma (NHL), and 1 case (2.4%) of Wiskott-Aldrich syndrome (WAS). In mNGS testing, pathogens were detected in 31 peripheral blood samples, with a positive rate of 73.8% (31/42), significantly higher than the pathogen positive rate of 16.7% (7/42) detected by traditional blood culture, and the difference was statistically significant (P < 0.05). Among the pathogen-positive cases detected by mNGS, 23 cases (74.2%) were positive for bacteria, 12 cases (38.7%) were positive for viruses, and 9 cases (29.0%) were positive for fungi. 32.2% (10/31) of the pathogen-positive samples detected by mNGS were mixed pathogens, which could not be effectively detected by traditional blood culture. CONCLUSION Peripheral blood mNGS has advantages in the detection of pathogens of bloodstream infection secondary to hematologic diseases, with a higher detection rate of pathogen positivity than traditional blood cultures. It can detect viruses, rare pathogens and mixed pathogens, and has good clinical application value.
Child ; Child, Preschool ; Female ; Humans ; Male ; Hematologic Diseases/immunology* ; High-Throughput Nucleotide Sequencing ; Metagenomics ; Retrospective Studies ; Sepsis/microbiology*

OBJECTIVE: To analyze the clinical characteristics, microbiological analysis, and drug resistance patterns of intensive care unit (ICU) bloodstream infection. METHODS: A prospective cohort study method was employed to collect clinical data from patients suspected of bloodstream infection (BSI) during their stay in ICUs across 67 hospitals in 16 provinces and cities nationwide, from July 1, 2021, to December 31, 2022. Electronic data collection technology was used to gather general information on ICU patients, including gender, age, length of hospital stay, as well as diagnostic results, laboratory tests, imaging studies, microbiological results (including smear, culture results, and pathogen high-throughput testing), and prognosis. Patients were divided into a BSI group and a non-BSI group based on the presence or absence of BSI; further, patients with BSI were categorized into a drug-resistant group and a non-drug-resistant group based on the presence or absence of drug resistance. Differences in the aforementioned indicators between groups were analyzed and compared; variables with P < 0.10 in the univariate analysis were included in a multivariate Logistic regression analysis to identify risk factors for mortality and drug resistance in ICU patients with BSI. RESULTS: A total of 2 962 ICU patients suspected of BSI participated in the study, including 790 in the BSI group and 2 172 in the non-BSI group. Patients in the BSI group were mainly from East China and Southwest China, with significantly higher age and mortality rates than those in the non-BSI group. Among ICU patients with BSI, Staphylococcus had the highest detection rate (8.10%), followed by Klebsiella pneumoniae (7.47%); there were 169 cases in the drug-resistant group and 621 cases in the non-drug-resistant group; 666 cases survived, and 124 cases died (mortality was 15.70%). There were statistically significant differences between the death group and the survival group in terms of age, regional distribution, and bloodstream infections caused by Gram negative (G^-) bacilli, Enterococcus faecium, Aspergillus, and Klebsiella pneumoniae; multivariate Logistic regression analysis showed that age [odds ratio (OR) = 1.01, 95% confidence interval (95%CI) was 1.00-1.03], regional distribution (OR = 4.07, 95%CI was 1.02-1.34), Enterococcus faecium infection (OR = 3.64, 95%CI was 1.16-11.45), and Klebsiella pneumoniae infection (OR = 2.64,95%CI was 1.45-4.80) were independent risk factors for death in ICU patients with BSI (all P < 0.05). There were statistically significant differences between the drug-resistant group and the non-drug-resistant group in terms of age and bloodstream infections caused by Gram positive (G⁺) cocci and G^- bacilli; multivariate Logistic regression analysis showed that age (OR = 1.01,95%CI was 1.00-1.03), G^- bacilli infection (OR = 2.18, 95%CI was 1.33-3.59), Escherichia coli infection (OR = 0.28,95%CI was 0.09-0.84), and Enterococcus faecium infection (OR = 3.35, 95%CI was 1.06-10.58) were independent risk factors for drug resistance in ICU patients with BSI (all P < 0.05). CONCLUSIONS Bloodstream infections may increase the mortality of ICU patients. Older age, regional distribution, Enterococcus faecium infection and Klebsiella pneumoniae infection can increase the mortality rate of ICU patients with BSI; bloodstream infections caused by G^- bacilli are prone to drug resistance, but have no significant impact on the mortality of ICU patients with BSI.
Adult ; Humans ; Bacteremia/microbiology* ; China/epidemiology* ; Cohort Studies ; Cross Infection/microbiology* ; Drug Resistance, Bacterial ; Intensive Care Units/statistics & numerical data* ; Prospective Studies ; Risk Factors ; Sepsis/microbiology*

OBJECTIVE: To investigate the potential mechanisms of sepsis pathogenesis in intensive care unit (ICU), with a specific focus on the role of skin microbiota, and to evaluate the causal relationships between skin microbiota and ICU sepsis using Mendelian randomization (MR). METHODS: A two-sample MR analysis was performed using skin microbiota genome-wide association study (GWAS) summary data from German population cohorts as exposures, combined with ICU sepsis susceptibility and 28-day mortality GWAS summary data from the IEU OpenGWAS database as outcomes. The primary causal effect estimates were generated using the inverse variance weighted (IVW) method, supplemented by validation through MR-Egger and weighted median approaches. Heterogeneity and pleiotropy tests, along with sensitivity analyses, were conducted to evaluate the robustness of the results. RESULTS: Regarding risk of ICU sepsis, IVW analysis showed that order Pseudomonadales [odds ratio (OR) = 0.93, 95% confidence interval (95%CI) was 0.88-0.98], family Flavobacteriaceae (OR = 0.93, 95%CI was 0.90-0.96), and genus Acinetobacter (OR = 0.96, 95%CI was 0.93-0.99) were significantly negatively correlated with the risk of ICU sepsis (all P < 0.05). There was a significant positive correlation between the risk of ICU sepsis and the presence of β-Proteobacteria (OR = 1.05, 95%CI was 1.00-1.11) and Actinobacteria (OR = 1.05, 95%CI was 1.00-1.11), both P < 0.05. Regarding 28-day mortality of ICU sepsis, IVW analysis showed that phylum Bacteroidetes (OR = 0.92, 95%CI was 0.86-0.99), family Streptococcaceae (OR = 0.92, 95%CI was 0.85-0.98), family Flavobacteriaceae (OR = 0.90, 95%CI was 0.83-0.97), genus Streptococcus (OR = 0.92, 95%CI was 0.86-0.99), ASV016 [Enhydrobacter] (OR = 0.92, 95%CI was 0.87-0.98), and ASV042 [Acinetobacter] (OR = 0.92, 95%CI was 0.88-0.97) were significantly negatively correlated with the 28-day mortality of ICU sepsis (all P < 0.05); family Moraxellaceae (OR = 1.09, 95%CI was 1.00-1.18) and ASV008 [Staphylococcus] (OR = 1.08, 95%CI was 1.03-1.14) was significantly positively correlated with the 28-day mortality of ICU sepsis (both P < 0.05). Sensitivity analysis and MR-PRESSO showed no heterogeneity, pleiotropy, or horizontal pleiotropy between skin microbiota and ICU sepsis risk and 28-day mortality rate. Analysis of confounding factors showed that single nucleotide polymorphisms (SNPs) associated with relevant skin bacteria could independently and causally affect the risk of ICU sepsis or ICU sepsis related mortality rate, independent of other confounding factors. The Steiger test results indicated that the established causal relationship was not due to reverse causality. CONCLUSIONS Skin microbiota composition may influence both sepsis susceptibility and 28-day mortality in ICU settings. Family Flavobacteriaceae demonstrated protective effects against sepsis onset and mortality. These findings provide new perspectives for early detection and management strategies.
Humans ; Sepsis/mortality* ; Intensive Care Units ; Mendelian Randomization Analysis ; Microbiota ; Skin/microbiology* ; Genome-Wide Association Study ; Risk Factors ; Skin Microbiome

OBJECTIVE: To explore the clinical features of patients with necrotizing soft tissue infection (NSTI) and the related factors for sepsis, so as to provide a basis for early intervention and improvement of patients' prognosis. METHODS: A retrospective case series study was conducted to analyze the clinical data of NSTI patients admitted to the department of burns and wound repair surgery of Guangdong Provincial People's Hospital from October 2021 to December 2024. Demographic information, underlying diseases, infection characteristics, laboratory test results and etiological findings at admission, treatment status, occurrence of complications (including sepsis) and prognosis were collected. Univariate and multivariate Logistic regression analyses were used to identify the associated factors for sepsis in NSTI patients. Receiver operator characteristic curves (ROC curves) were plotted to evaluate the predictive value of individual and combined factors for sepsis. RESULTS: A total of 159 NSTI patients were enrolled, mainly middle-aged and elderly males. Most patients had comorbidities, including diabetes mellitus (110 cases, 69.2%) and hypertension (67 cases, 42.1%). The main infection site was the lower extremities (104 cases, 65.4%). Common symptoms included redness (96 cases, 60.4%), swelling (129 cases, 81.1%), local heat (60 cases, 37.7%), pain (100 cases, 62.9%), and skin ulceration or necrosis (9 cases, 5.7%). Imaging findings included soft tissue swelling (66 cases, 57.9%), gas accumulation (41 cases, 36.0%), and abnormal signal/density shadows (50 cases, 43.9%). Staphylococcus aureus was the main pathogenic bacterium [12.0% (31/259)], and drug-resistant Escherichia coli had the highest detection rate among drug-resistant bacteria [35.1% (13/37)]. Regarding debridement and repair, most patients (80 cases, 50.3%) underwent debridement ≥ 72 hours after admission, while only 10.1% (16 cases) received debridement within 6 hours. Most patients underwent multiple debridements, with 2 times of debridements being the most common (68 cases, 42.8%), and the maximum times of debridements reached 6. The largest number of patients received secondary suture (44 cases, 27.7%). In terms of complications, sepsis was the most common (66 cases, 41.51%), followed by acute kidney injury, respiratory failure requiring mechanical ventilation, and multiple organ dysfunction syndrome (MODS), while disseminated intravascular coagulation (DIC) was the least common. During the follow-up period, 9 patients (5.66%) were readmitted within 90 days, and 11 patients died, with a mortality rate of 6.92%. Univariate analysis showed that diabetes, coronary heart disease, gout, body temperature, heart rate, C-reactive protein, platelet count, total bilirubin, albumin, creatinine, out-of-hospital treatment, and out-of-hospital use of antimicrobial agents were significantly associated with sepsis in NSTI patients (all P < 0.05). Multivariate Logistic regression analysis showed that coronary heart disease [odds ratio (OR) = 30.085, 95% confidence interval (95%CI) was 2.105-956.935], C-reactive protein (OR = 1.026, 95%CI was 1.009-1.054), and total bilirubin (OR = 1.436, 95%CI was 1.188-1.948) were independent associated factors for sepsis in NSTI patients (all P < 0.05). ROC curve analysis revealed that the combination of the three predictors yielded the highest AUC for predicting sepsis in NSTI patients compared to any individual predictor [area under the curve (AUC) = 0.799 (95%CI was 0.721-0.878)]. CONCLUSIONS The clinical features of NSTI patients show certain regularity. Coronary heart disease, C-reactive protein, and total bilirubin are independent associated factors for sepsis in NSTI patients.
Humans ; Retrospective Studies ; Male ; Sepsis ; Soft Tissue Infections/microbiology* ; Female ; Middle Aged ; Aged ; Adult ; Prognosis ; Risk Factors ; Necrosis ; Logistic Models ; Fasciitis, Necrotizing

OBJECTIVE: To investigate the effects of liriodendrin on the intestinal flora and the ferroptosis signaling pathway in renal tissue of rats with sepsis-induced acute kidney injury (AKI). METHODS: Thirty male Sprague-Dawley (SD) rats were randomly divided into sham operation group (Sham group), sepsis model induced by cecal ligation and puncture group (CLP group), and liriodendrin intervention group (CLP+LIR group), with 10 rats in each group. The CLP+LIR group was given 0.2 mL of 100 mg/kg liriodendrin by gavage 2 hours before modeling; Sham group and CLP group were given the same volume of normal saline by gavage. The samples were collected after anesthesia 24 hours after modeling. The pathological changes of renal tissue were observed by hematoxylin-eosin (HE) staining. The levels of inflammatory factors such as tumor necrosis factor-α (TNF-α), interleukins (IL-1β, IL-6) were detected by enzyme linked immunosorbent assay (ELISA). The levels of renal function indicators such as creatinine (Cr), and urea nitrogen (UREA) in peripheral blood, and the content of malondialdehyde (MDA) and Fe²⁺ in renal tissue were detected. Western blotting was used to detect the expressions of nuclear factor E2-related factor 2 (Nrf2), glutathione peroxidase 4 (GPX4) and heme oxygenase-1 (HO-1) in renal tissues. The changes of intestinal flora were detected by 16S rDNA high-throughput sequencing. RESULTS: Compared with the Sham group, the CLP group showed significantly enlarged glomeruli, noticeable renal interstitial edema, disorganized kidney tissue, and significantly increased pathological scores. The contents of TNF-α, IL-1β, IL-6, Cr, and UREA in peripheral blood and the levels of MDA and Fe²⁺ in renal tissue were significantly increased. The protein expressions of Nrf2, GPX4, and HO-1 in renal tissue were significantly down-regulated. The species richness of intestinal flora decreased significantly, and the relative abundances of pathogenic bacteria such as Morganella, Citrobacter, Proteus, Klebsiella, Shigella, Aggregatibacter, and Enterococcus increased significantly, while the relative abundances of beneficial bacteria such as Butyricimonas, Veillonella, Prevotella, Lactobacillus, Bifidobacterium, and Ruminococcus decreased significantly. Compared with the CLP group, CLP+LIR group could significantly reduce the pathological damage of renal tissue, the pathological score significantly decreased (1.80±0.84 vs. 4.20±1.30, P < 0.05), and improve the composition of intestinal flora, reduce the relative abundances of pathogenic bacteria such as Proteus, Klebsiella, Shigella, Aggregatibacter, and Enterococcus, and significantly increase the relative abundances of Lactobacillus, Bifidobacterium, and Ruminococcus, significantly reduce the contents of TNF-α, IL-1β, IL-6, Cr, and UREA in peripheral blood and the levels of MDA and Fe²⁺ in renal tissue [blood TNF-α (ng/L): 191.31±7.23 vs. 254.90±47.89, blood IL-1β (ng/L): 11.15±4.04 vs. 23.06±1.67, blood IL-6 (ng/L): 163.20±17.83 vs. 267.69±20.92, blood Cr (μmol/L): 24.14±4.25 vs. 41.17±5.43, blood UREA (mmol/L): 4.59±0.90 vs. 8.01±1.07, renal MDA (μmol/g): 9.67±0.46 vs. 16.05±0.88, renal Fe²⁺ (mg/g): 0.71±0.07 vs. 0.93±0.04, all P < 0.05], and increase the protein expressions of Nrf2, GPX4, and HO-1 (Nrf2/GAPDH: 1.21±0.01 vs. 0.39±0.01, GPX4/GAPDH: 0.74±0.04 vs. 0.48±0.04, HO-1/GAPDH: 0.91±0.01 vs. 0.41±0.02, all P < 0.05). CONCLUSIONS Liriodendrin has an obvious protective effect on sepsis-induced AKI. The mechanism may involve regulating the intestinal flora, increasing the activation of the Nrf2/HO-1/GPX4 signaling pathway in renal tissue, and reducing ferroptosis.
Animals ; Acute Kidney Injury/microbiology* ; Rats, Sprague-Dawley ; Sepsis/complications* ; Male ; Ferroptosis/drug effects* ; Gastrointestinal Microbiome/drug effects* ; Rats ; Signal Transduction ; Kidney/metabolism* ; Tumor Necrosis Factor-alpha/metabolism*

Humans ; Young Adult ; Fusobacterium necrophorum ; Pharyngitis/microbiology* ; Streptococcal Infections ; Sepsis/diagnosis*

To understand the clinical characteristics of Staphylococcus aureus bloodstream infection and the main risk factors affecting clinical prognosis, providing a reference for clinical prevention and control of Staphylococcus aureus bloodstream infection. In this study, the clinical data of 152 patients with Staphylococcus aureus bloodstream infection admitted to Guangdong Provincial People's Hospital from January 2019 to December 2021 were retrospectively analyzed by reviewing the electronic medical record system, including underlying diseases, clinical characteristics, risk factors, and bacterial resistance. Statistical methods such as Chi-Squared Test and t Test were used to analyze the related risk factors that may affect the clinical characteristics and prognosis of patients with Staphylococcus aureus and methicillin-resistant Staphylococcus aureus (MRSA) bloodstream infection, then the variables with P<0.05 in univariate analysis were included in the multivariate logistic regression model to analyze the independent risk factors of poor prognosis. The results showed among 152 patients with Staphylococcus aureus bloodstream infection, 50 patients (32.89%) were infected with MRSA. In comparison, 102 patients (67.11%) were infected with methicillin-sensitive Staphylococcus aureus (MSSA). Except for rifampicin, the resistance rate of MRSA to commonly used antibiotics was all higher than that of MSSA, and the difference was statistically significant (Chi-square values were 8.272, 11.972, 4.998, 4.776, respectively;all P-values are less than 0.05). Strains resistant to vancomycin, linezolid, and quinupristin/dalfopristin were not found. In the MRSA group, indwelling catheter and drainage tube, carbapenems, and β-lactamase inhibitor treatment were significantly higher than the MSSA group. The difference was statistically significant (P<0.05). The incidence of poor prognosis of bloodstream infection in the MRSA group was higher than that in the MSSA group (34.00% vs 13.73%), and the difference was statistically significant (χ²=8.495, P<0.05). No independent risk factors associated with poor prognosis were found in the included patients with MRSA bloodstream infection.Multivariate Logistic regression model analysis showed that solid malignant tumors (OR=13.576, 95%CI: 3.352-54.977, P<0.05), mechanical ventilation (OR=7.468, 95%CI: 1.398-39.884, P<0.05) were the most important independent risk factors for poor prognosis in patients with Staphylococcus aureus bloodstream infection. In summary, the poor prognosis rate of MRSA bloodstream infection is higher than that of MSSA. The clinical evaluation of related risk factors should be strengthened, targeted prevention and control interventions should be taken to improve the prognosis of patients with Staphylococcus aureus bloodstream infection, and the use of antibiotics should be rational and standardized, to control bacterial infection and drug resistance effectively .
Humans ; Methicillin-Resistant Staphylococcus aureus ; Staphylococcus aureus ; Retrospective Studies ; Prognosis ; Staphylococcal Infections/microbiology* ; Anti-Bacterial Agents/pharmacology* ; Methicillin/therapeutic use* ; Sepsis

Blood stream infection (BSI),a blood-borne disease caused by microorganisms such as bacteria,fungi,and viruses,can lead to bacteremia,sepsis,and infectious shock,posing a serious threat to human life and health.Identifying the pathogen is central to the precise treatment of BSI.Traditional blood culture is the gold standard for pathogen identification,while it has limitations in clinical practice due to the long time consumption,production of false negative results,etc.Nanopore sequencing,as a new generation of sequencing technology,can rapidly detect pathogens,drug resistance genes,and virulence genes for the optimization of clinical treatment.This paper reviews the current status of nanopore sequencing technology in the diagnosis of BSI.
Humans ; Nanopore Sequencing ; Sepsis/diagnosis* ; Bacteremia/microbiology* ; Bacteria ; Blood Culture/methods*

OBJECTIVE: To investigate the risk factors of oral ulcers and bloodstream infection in patients with hematopoietic stem cell transplantation. METHODS: The clinical data of 401 hematopoietic stem cell transplant patients in the First Affiliated Hospital of Zhengzhou University from January 2020 to December 2021 were retrospective analyzed, and the risk factors of oral ulcers and bloodstream infection statistical and analyzed. RESULTS: Among the 401 patients, the incidence of oral ulcers was 61.3% (246/401), and the incidence of bloodstream infection was 9.0% (36/401). A total of 40 strains of pathogenic bacteria were isolated from 36 patients, including 26 strains of Gram negative strains (65%), 13 strains of Gram positive strains (32.5%), and 1 strain of fungi (2.5%). Single-factor analysis showed that oral hygiene was associated with the occurrence of bloodstream infection, and the Multi-factor analysis showed that age ≥14 years old, disease diagnosis of leukemia, and allogeneic hematopoietic stem cell transplantation were risk factors for oral ulcers. CONCLUSION The incidence of oral ulcers in patients with hematopoietic stem cell transplantation is high. The age ≥14 years, disease diagnosis of leukemia, and allogeneic hematopoietic stem cell transplantation were risk factors for oral ulcers in patients, and oral hygiene was associated with the occurrence of bloodstream infection.
Humans ; Adolescent ; Retrospective Studies ; Oral Ulcer/etiology* ; Bacteremia/microbiology* ; Hematopoietic Stem Cell Transplantation/adverse effects* ; Sepsis ; Risk Factors ; Leukemia

OBJECTIVE: To investigate the clinical features, distribution of pathogenic bacteria, and drug resistance of bloodstream infection in children with acute leukemia. METHODS: Clinical data of 93 blood culture-positive children with acute leukemia from January 2015 to December 2019 in Department of Pediatrics, The Second Hospital of Anhui Medical University were analyzed retrospectively. RESULTS: In these 93 cases, 78 cases were in the period of neutrophil deficiency. There were 54 Gram-negative bacteria (G^-) (58.1%) found through blood culture, and the top 4 strains were Escherichia coli (15.1%), Klebsiella pneumoniae (13.9%), Pseudomonas aeruginosa (6.5%), and Enterobacter cloacae (6.5%). There were 39 Gram-positive bacteria (G⁺) (41.9%) detected, and the top 4 strains were Staphylococcus epidermidis (10.8%), Streptococcus pneumoniae (6.5%), Staphylococcus hemolyticus (5.4%), and Staphylococcus human (5.4%). Among 74 strains of pathogenic bacteria from acute lymphoblastic leukemia (ALL) children, there were 29 strains of G⁺ bacteria (39.2%) and 45 strains of G^- bacteria (60.8%). While in 19 strains from acute myeloblastic leukemia (AML) patients, G^- bacteria accounted for 47.4% and G⁺ bacteria accounted for 52.6%. In 15 ALL children without neutropenia, G⁺ bacteria made up the majority of the strains (66.7%). In the 93 strains of pathogenic bacteria, 13 (13.9%) strains were multidrug-resistant. Among them, extended-spectrum β-lactamases accounted for 42.9%, carbapenemase-resistant enzyme Klebsiella pneumoniae 15.4%, and carbapenemase-resistant enzyme Enterobacter cloacae strains 33.3%, which were detected from G^- bacteria. While, 13.3% of methicillin-resistant coagulase-negative Staphylococci accounted for 13.3% detected from G⁺ bacteria, but linezolid, vancomycin, teicoplanin Staphylococcus and Enterococcus resistant were not found. The average procalcitonin (PCT) value of G^- bacteria infection was (11.02±20.282) ng/ml, while in G⁺ infection it was (1.81±4.911) ng/ml, the difference was statistically significant (P<0.05). The mean value of C-reactive protein (CRP) in G^- infection was (76.33±69.946) mg/L, and that in G⁺ infection was (38.34±57.951) mg/L. The prognosis of active treatment was good, and only one case died of septic shock complicated with disseminated intravascular coagulation (DIC) and gastrointestinal bleeding caused by carbapenemase-resistant enzyme enterobacteriaceae. CONCLUSION G^- is the major bacteria in acute leukemia children with bloodstream infection, but the distribution of ALL and AML strains is different. G^- bacteria dominates in ALL, while G⁺ bacteria and G^- bacteria are equally distributed in AML. Non-agranulocytosis accompanied by bloodstream infections is dominant by G⁺ bacteria. The mean value of PCT and CRP are significantly higher in G^- bacteria infection than in G⁺ bacteria.
Acute Disease ; Anti-Bacterial Agents/therapeutic use* ; Bacteremia/microbiology* ; Bacteria ; Child ; Drug Resistance, Bacterial ; Humans ; Leukemia, Myeloid, Acute/drug therapy* ; Microbial Sensitivity Tests ; Precursor Cell Lymphoblastic Leukemia-Lymphoma/drug therapy* ; Procalcitonin ; Retrospective Studies ; Sepsis/drug therapy*