A Single center descriptive retrospective cohort study was conducted in patients who underwent allo-HSCT from January 2016 to December 2021. Only human leukocyte antigen (HLA) MSD allo-HSCT are performed in the center of this study. Patients with at least one positive blood culture for significant bacteria within 100 days after allo-HSCT were included in the bacteremia cohort and all the others were included in the non-bacteremia cohort. Clinical and biological data were collected from the patient's medical records.

All patients had double central venous catheter (CVC) insertion before starting the conditioning regimen.

Screening for MDR bacteria was performed with weekly deep rectal swabs before the start of conditioning and until hospital discharge to detect extended-spectrum beta-lactamase (ESBL), carbapenemase producers and vancomycin-resistant enterococci (VRE). First-line antibiotic treatment consisted of piperacillin-tazobactam and amikacin in the absence of colonization by MDR GN bacteria. Second-line therapy consisted of antifungal therapy (caspofungin or voriconazole). Glycopeptide was used empirically as a second-line therapy in patients with evidence of CVC or skin infections, mucositis Grade IV, and colonization by staphylococcus or streptococcus resistant to piperacillin-tazobactam.

Nasal and sputum swabs were performed weekly for Aspergillus sp. as screening tests for invasive aspergillosis. Patients were monitored once a week from engraftment to Day +100 for cytomegalovirus (CMV) using the pp 65 antigenemia test or by quantitative polymerase chain reaction (PCR). All patients received prophylactic fluconazole (400 mg/day in adults and 6 mg/kg/day i.v. in children) from Day -7 to Day +75 and acyclovir as prophylaxis against the herpes simplex virus (HSV) from Day +1 to Day +180. No antibiotic prophylaxis was used.

All patients were transfused with phenotyped, leukodepleted, and irradiated red blood cells, and with leukodepleted and irradiated single donor platelets or random donor platelets. Patients received granulocyte colony-stimulating factor (G-CSF -5 mg/kg/day) from Day +7 until the neutrophil count reached 1 × 109/L on three consecutive days. Engraftment was defined as the first of three consecutive days with an absolute neutrophil count higher than 0.5 × 109/L.6 All patients were treated in laminar airflow isolation rooms with high-efficiency particulate air filtration and received gut decontamination with oral colistin, gentamicin and fungizone.

A blood culture was performed in the presence of systemic symptoms of infection (fever, hypotension or malaise). Surveillance blood cultures were carried out in asymptomatic patients during treatment with steroids to detect occult bacteremia. Blood samples for aerobic and anaerobic growth conditions, from central lines and concomitant blood samples from a peripheral vein for blood cultures were routinely performed. C-reactive protein (CRP) was monitored every other day.

Bacteremia was defined as the isolation of bacteria from at least one blood culture. For coryneforms or coagulase negative staphylococci (CoNS), two consecutive positive blood cultures were required. Early bacteremia was considered as occurring between Day 0 and Day +100 after allo-HSCT. Bacteremia was considered catheter-related when the causative microorganism isolated in the blood was also recovered from a catheter hub or tip, or a positive culture of a blood sample from CVC was found at least two hours earlier than a positive result of a peripheral blood culture.7 A repeat bacteremic episode was considered when the same microorganism was isolated after an interval of more than seven days from the earlier episode with adequate treatment or when a different microorganism was isolated after an interval of more than three days from the earlier episode.8 Polymicrobial bacteremia was defined as the isolation of more than one microorganism from the same blood sample or from two consecutive blood samples within 24 hours.9

Blood samples were analyzed according to the “Référentiel en Microbiologie Médicale”.10 Bacterial identification was based on morphologic, cultural and biochemical characteristics (Api systems, BioMérieux®). Antimicrobial susceptibility testing was performed by the diffusion method on agar medium according to the Comité de l'Antibiogramme de la Société Française de Microbiologie (CAS-FM) standards.11 The minimal inhibitory concentrations (MIC) of colistin for ESBL producing Enterobacteriales (ESBL-E), MDR P. aeruginosa and MDR A. baumannii were performed using the microdilution method (Biocentric®). The MIC of glycopeptides for methicillin-resistant S. aureus (MRSA) and VRE were determined by the microdilution method (Biocentric®) and E-test (BioMérieux®), respectively. ESBL identification was determined by the double disk synergy test. MDR bacteremia was defined as the isolation in the blood of MDR bacteria (ESBL-E, P. aeruginosa or A. baumannii resistant to at least three families of antibiotics - β-lactam, aminoglycoside, fluoroquinolone, colistin, - MRSA and VRE).

Bacteremia was initially treated by empiric broad-spectrum intravenous antibiotics. Antibiotics were modified according to the susceptibility of the isolated bacteria and response to treatment. Initial empirical treatment was defined as appropriate if an antibiotic prescribed within 24 hours matched the in vitro susceptibility of a pathogen deemed to be the likely cause of infection.12 Antibiotic treatment was continued for at least 14 days, although it was continued until the resolution of systemic symptoms in patients with signs of complicated bacteremia, such as the presence of secondary deep infections, persistent bacteremia, and implantable devices.

Attributable mortality was defined as death with persistent signs or symptoms of sepsis, or persistent positive blood cultures, or persistence of a focus of infection with no other reason for death.13

Proportions were compared using a Fisher exact test or chi-square test, and continuous variables were compared using a Student t-test or Mann Whitney U test, as appropriate. Acute GvHD was censored at day 100 after allo-HSCT. Overall survival (OS) was defined as the time between allo-HSCT and death from any cause or the last follow-up. Survivor curves were estimated using the Kaplan-Meier method and were compared using the log-rank test. All factors with p-values < 0.2 by univariate analysis were included in multivariable analysis using a Cox regression proportional hazard model with a p-value of less than 0.05 being considered statistically significant.

Between January 2016 and December 2021, 245 patients received allo-HSCT from MSD. Patient characteristics are detailed in Table 1.

A total of 40 episodes of early bacteremia occurred in 36 of 245 transplanted patients. Overall, 14.6% of transplanted patient were infected with 1.11 episodes per infected patient. Two patients presented two episodes of bacteremia and one patient presented three episodes. One episode of bacteremia was poly-microbial (two different GN bacteria). Fifteen episodes (37.5%) were due to GP and 25 episodes (62.5%) were due to GN bacteria. Forty-one microorganisms were isolated in the 40 episodes of bacteremia. The GP/GN ratio was 0.16. Twenty-nine episodes of bacteremia (72.5%) occurred during the first 30 days after allo-HSCT with the median time from allo-HSCT to bacteremia being nine days (range: 0-90 days). Episodes of bacteremia were catheter-related in 20% of the cases.

Coagulase negative staphylococci, Klebsiella and Pseudomonas species were the most common isolated microorganisms in blood cultures. The distribution of the microorganisms isolated in the blood cultures is detailed in Table 2.

Coagulase negative staphylococci were methicillin-resistant in 75% of cases (6/8) and all GP bacilli were penicillin-resistant (4/4). The only enterococcus faecium isolated was resistant to vancomycin. GN bacilli were MDR in 61.5% of cases (16/26). Of the 15 Enterobacteriaceae isolated, nine were identified as ESBL producers and one case of Klebsiella pneumonia was identified as carbapenemase-producing Enterobacteriaceae. Of the 11 non-fermenting GN bacilli, six were identified as MDR (Pseudomonas sp. and Aeromonas (5/10) and Acinetobacter baumannii (1/1)).

The clinical and laboratory characteristics of patients at the time of bacteremic episodes are summarized in Table 3 (all bacteremic episodes were counted in patients who had more than one).

Concurrent infectious focuses were present in 18 patients (45%), including pneumonia (n = 6), enterocolitis (n = 4) and catheter-related bloodstream infections (n = 8) due to CoNS (n = 4), Pseudomonas aeruginosa (n = 1), Raoultella terrigenia (n = 1), Serratia marcescens (n = 1), and Corynebacterium jeikeium (n = 1). Prior rectal colonizations by MDR bacteria were detected in 14 patients (35%) within the month prior to bacteremia. Klebsiella pneumoniae was the most common microorganism isolated (seven patients) followed by Escherichia coli (four patients) and Enterococcus faecium (two patients). Of the 14 patients previously colonized by MDR bacteria, the bacteria detected in blood cultures were the same as the colonizing bacteria in six patients. Three patients had septic shock due to E. coli, Raoultella terrigenia, and P. aeruginosa. Corticosteroid therapy was prescribed in 50% of patients for GvHD.

At the time of the bacteremic episode, 30 patients (75%) had an elevated CRP (>5mg/L). Procalcitonin and lactate were not routinely assessed during the study period.

Median time between the onset of antibiotic therapy and the first positive blood culture was six days. Sixteen episodes (40%) occurred during antibiotic therapy prescribed for fever of unknown origin. The episode of bacteremia was a new infectious episode in these patients and the cultures were collected at the same time that the antibiotics were started for each infectious episode. Antibiotics were appropriate in 57.5% of the cases of bacteremia (23/40). Initial antibiotic therapy was modified in 50% of cases (n = 20): inappropriate (n = 16) and used for synergistic effect (n = 4). Two patients had methicillin-resistant CoNS bacteremia with a favorable outcome with piperacillin-tazobactam. Teicoplanin was the most frequently prescribed antibiotic (n = 21) followed by imipenem (n = 19) and piperacillin-tazobactam (n = 13). The median duration of treatment for bacteremia was 14 days (range: 9-26 days). One patient died after nine days following bacteremia due to septic shock. The CVC was removed in eight patients (20%), four of whom had catheter-related bacteremia. The CVC was removed within an average of five days of bacteremia (range: 1-8 days) and within three days (range: 1-4 days) in patients with appropriate antibiotic therapy. Catheter removal and antibiotic therapy contributed to infection control in seven bacteremic episodes (17.5%).

In multivariate analysis, the use of bone marrow as graft source and CMV reactivation were significantly associated with early bacteremia after allo-HSCT in the present study (Table 4).

After a median follow-up of 24 months (range: 5 days-83 months), the overall survival was 80.4% in patients with early bacteremia and 78.7% in patients without early bacteremia (p-value = 0.8).

Of the 245 transplanted patients, 53 (21.6%) patients died (26 from relapse and 27 from non-relapse mortality. The non-relapse mortality was 11.1% (4/36) in patients with bacteremia and 11% (23/209) in patients without bacteremia (p-value = 0.72). Attributable mortality was 2.8% (1/36); caused by an MDR-Raoultella terrigena bacteremia despite appropriate antibiotic therapy and CVC removal.