https://orcid.org/0000-0002-5661-1324
Figures
Abstract
Importance
Patients with Gram-negative bloodstream infections (GN-BSI) are classified as non-immunocompromised (n-IC) or immunocompromised (IC). However, immunosuppressive condition should not be considered univocally.
Objective
To investigate epidemiological characteristics, management and outcome of GN-BSI in IC and non-IC patients.
Methods
Retrospective single-center study of hospitalized patients with GN-BSI conducted over a 7-year period. Patients with GN-BSI were divided in: solid organ transplant (SOT) recipients, patients with hematologic malignancy (HM), patients with metastatic solid cancer (mSC), and non-major IC patients (nm-IC).
Results
3544 patients analysed: 76.7% nm-IC, 6.5% SOT, 8.0% HM and 8.8% mSC. SOT and HM patients were younger (SOT: 56.6 ± 13.1 years; HM: 56.4 ± 14.5; nm-IC: 72.4 ± 16.1; mSC: 68.6 ± 13.1, p < 0.001) and had lower CCI value (SOT: 4.5 ± 2.4; HM: 4.1 ± 2.1; nm-IC: 5.5 ± 2.6; mSC: 9.7 ± 2.5, p < 0.001). Urinary tract infection was the most common source of BSI in nm-IC (nm-IC: 50.1%, HM:15%; SOT: 33.3%; mSC: 25.9%, p < 0.001), intra-abdominal infection was the more frequent source among SOT and mSC (SOT:42.3%; mSC: 49.3%, nm-IC: 27.8%, HM:29%; p < 0.001). Primary BSI was the first cause of GN-BSI in HM (HM: 62.1%; SOT: 18.5%; nm-IC: 17.2%; mSC: 10.6%, p < 0.001). The lowest rate of death was observed in SOT and the highest in mSC (SOT 8.2%; nm-IC 13.4%; HM 14.9%; mSC 19.9%, p < 0.001). Relapse rate was highest in SOT (SOT: 18.8%; HM: 11.8%; NMIC: 7.2%; aST: 7.1%, p < 0.001). Follow-up bloodcultures were associated with a lower mortality only among NMIC (HR = 0.317, 95% CI 0.178–0.563, p < 0.001) and aST (HR = 0.198, 95% CI 0.058–0.673, p = 0.010). The role of treatment duration on relapse was not evident in any group, conversely receiving at least 7 days of treatment was associated with a lower risk of 90-day mortality in SOT and HM patients.
Citation: Toschi A, Pascale R, Gibertoni D, Pasquali R, Grechi A, Grassi I, et al. (2025) Epidemiological characteristics and management of Gram-negative bacteraemia in different immunocompromised hosts: Observational single-center study. PLoS One 20(7): e0327535. https://doi.org/10.1371/journal.pone.0327535
Editor: Ennio Polilli, Pescara General Hospital, ITALY
Received: December 4, 2024; Accepted: June 17, 2025; Published: July 7, 2025
Copyright: © 2025 Toschi et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: The work reported in this publication was funded by the Italian Ministry of Health, RC-2025-2796965 project.
Competing interests: The authors have declared that no competing interests exist.
Introduction
In the last 15 years Gram negative bloodstream infection (GN-BSI) has emerged as a main infectious complication, especially among hospitalized patients and those with chronic underlying conditions, in which it is associated with high rates of morbidity and mortality [1]. The standardization of diagnostic and therapeutic management of GN-BSI is challenging [2–3] due to the heterogeneity of GN-BSI in terms of host characteristics, causative agents and susceptibility patterns. As regards host characteristics, patients are mainly classified as non-immunocompromised (n-IC) or immunocompromised (IC), including in this latter group a wide spectrum of conditions with different degree and duration of immunosuppression. However, immunosuppressive condition should not be considered univocally, due to the different pathogenetic mechanisms underlying the alteration of the immune system. Immunosuppression due to anti-rejection therapy in patients undergoing solid organ transplant is quite different from that caused by an altered function of the immune system typical of haematological patient or determined by chemotherapy in a patient with a solid neoplasm. Consequently, IC patients with GN-BSI may have different aetiologies, therapeutic needs and outcomes depending on the underlying condition.
Indeed, among solid organ transplant recipients, GN-BSI account for 6% to 45% of all BSI, with reported mortality rates ranging from 3% to 52% [4]. Similarly, in haematological neutropenic patients, bacteraemia caused by GN rods occurs in 25% to 74% of cases, with Escherichia coli and Pseudomonas aeruginosa being the primary etiological agents. Mortality rates in these patients are highly variable, ranging from 6% to 40% [5]. Finally, in patients with solid tumors, GN aetiology account for over 60% of BSI, with a mortality rate of approximately 20%. Escherichia coli is the most frequently reported microorganism (47%) [6].
However, current literature typically focuses on a specific IC condition, sometimes comparing these patients to the general population [4–6]. We deem that a comparison of epidemiological characteristics and outcome of patients with GN-BSI according with the underlying IC status may provide useful information about the potential need of a different approach to the diagnostic and therapeutic management.
Therefore, the aim of this study is to investigate bacteremia characteristics, aetiology distribution, management, and outcome in different types of IC patients with GN-BSI. In addition, the impact of follow-up bloodcultures (FUBCs) on all-cause 30-day mortality and duration of antibiotic treatment on 90-day relapse were assessed overall and in each population.
Methods
Study design and setting
Single-center retrospective observational study of adult patients with GN-BSI hospitalized at IRCCS Azienda Ospedaliero-Universitaria di Bologna, a 1450-bed tertiary-care hospital in Bologna, Italy, over a 7-year period (January 1st 2013 to December 31st 2019). Follow-up for mortality/relapse was 90 days after the index BCs (time 0 for all outcomes).
During hospitalization, patient management was determined by attending physicians and was not dictated by local protocol. Data were anonymously collected from 5th May 2023–12th June 2023 using hospital records and inserted in a dedicated REDCap electronic case report form (eCRF) hosted by IRCCS Azienda Ospedaliero-Universitaria di Bologna [7]. The study was conducted according to the declaration of Helsinki and Good Clinical Practice guidelines and was approved by the local Ethics Committee (n° 894/2021/Oss/AOUBo).
Population
All adult (≥ 18 years) patients hospitalized and diagnosed with GN-BSI during the study period were screened for inclusion using local microbiology registries. The inclusion criteria were as follows: hospitalized patients aged ≥ 18 years; diagnosis of GN-BSI, defined as one or more positive BCs obtained for ruling out an infection. The exclusion criteria were: patients receiving palliative care; unavailability of study data, particularly missing information regarding the type of pathogen, management, and outcomes. Patients with multiple episodes of infection were considered only once, at the time of first episode (index BCs).
Patients were classified into IC patients and non-major immunocompromised (nm-IC). IC patients were defined as patients with a reduced immune function and belonging to one of these three groups according to their underlying condition: solid organ transplant (SOT) recipients, patients with hematologic malignancies (HM) and patients with metastatic solid cancer (mSC) as recorded by Charlson comorbidity index [8] (full definitions of the different CI conditions are provided in S1 File). Conversely, non-major immunocompromised (nm-IC) patients were defined as patients without the previous conditions, including both patients with no underlying immunocompromising condition and patients with minor immunocompromising conditions (e.g., advanced age, chronic renal insufficiency, liver cirrhosis).
If patients had more than one major immunocompromising condition, the most clinically relevant one at the time of index BSI was considered, according to authors’ judgment.
Variables and definitions
Primary endpoint was all-cause 30-day mortality. The secondary endpoint was infection relapse within 90 days, defined by a positive BC for the same pathogen of the index BC after treatment course was completed with clinical cure. Clinical cure was defined as the resolution of all signs and symptoms of infection according to vital signs, evolution of SOFA score and laboratory data, assessed at 7 days from index blood cultures BCs. FUBCs were defined as BCs drawn between 48 hours and 7 days after index BCs. Results of FUBCs were classified as positive for the same pathogen; positive for a different pathogen from that of index BCs; and negative. Duration of active antibiotic therapy was defined as the number of consecutive days during which the patient received an appropriate antibiotic regimen. A detailed description of variables and definitions is reported in S1 File.
Statistical analysis
Patients’ characteristics were described as absolute and relative frequencies. Continuous variables were summarized with mean and standard deviation if normally distributed or with median and interquartile range (IQR) if non-normally distributed. Comparison of patients’ characteristics across the four subpopulations of IC and nm-IC patients was performed using ANOVA, Kruskal-Wallis test or chi square test, as appropriate. The post-hoc comparison for statistically significant tests was carried out by Sidàk post hoc test (after ANOVA), Holm post-hoc test (after Kruskal-Wallis test) applying alpha-level adjustment for family-wise error rate. After chi-square test, the categories having a Pearson residual greater > 2 (proportion greater than expected) and <−2 (proportion lower than expected) were identified as those contributing to the test’s statistical significance [9]. Survival analyses to identify predictors of 30-day mortality and 90-day relapse since the date of BSI were carried out using Royston-Parmar parametric survival models [10], because the proportionality of hazards assumption between patients with and without FUBC was not met. Several model parametrizations (exponential, log-logistic, probit) with 1–5 splines were checked, and among them the one providing the lower Bayesian Information Criterion (BIC) value was chosen. To remove immortal time bias, patients with FUBC entered the analysis at the date of FUBC (performed at a median time of 4 days after BSI). Moreover, patients who died within 5 days from BSI and without FUBC were removed from this analysis. When 30-day mortality was the outcome, FUBC performed within 7 days was set as the main exposure and as time-dependent variable. When 90-day relapse was the outcome, death was included as competing risk, survival time started at the end of antibiotic therapy course, and duration of therapy (< 7 days or ≥7 days) was the main exposure. The cumulative incidence of death and relapse was obtained according to the method described in Hinchliffe et al. [11], which entails including the two causes as both main effects and time-dependent effects, and creating an indicator for the interaction of each competing risk and levels of the exposure. In this way, a stratified model is obtained with separate baseline survival estimated for relapse and death. The multivariable models were performed first on the overall population and subsequently on each subpopulation separately, and included as potential confounders variables selected for their clinical interest. Confounders expressed as continuous variables (age, Charlson Comorbidity Index, SOFA score) were centred around their mean. Stata v.18.0 was used for all analyses; specifically, the stpm2 [12] and stpm2cif [11] packages were used for multivariable survival analysis.
Results
In the study period, 4497 patients were hospitalized with diagnosis of GN-BSI, among which 3544 were included in the study. Of these, 2719 (76.7%) were nm-IC, 232 (6.5%) SOT, 282 (8.0%) HM and 311 (8.8%) mSC (Fig 1).
Legend: Flow chart illustrating patient inclusion, exclusion criteria, and final stratification for analysis. GN-BSI = Gram negative BSI.
As shown in Table 1, male gender was more frequent in the SOT group (SOT: 72.4% vs nm-IC: 53.7%; HM: 61.7%; mSC: 54.3%, p < 0.001). SOT and HM patients on average were younger (SOT: 56.6 ± 13.1 years vs HM: 56.4 ± 14.5; nm-IC: 72.4 ± 16.1; mSC: 68.6 ± 13.1, p < 0.001) and had lower CCI value than patients with mSC and nm-IC (SOT: 4.5 ± 2.4 vs HM: 4.1 ± 2.1; nm-IC: 5.5 ± 2.6; mSC: 9.7 ± 2.5, p < 0.001). CPE rectal colonization was observed more frequently in SOT and mSC patients (SOT: 33.6%; mSC: 20.6%; nm-IC: 13.0%; HM: 7.1%, p < 0.001). Septic shock at the time of GN-BSI onset was less frequently observed in HM (4.0%, p = 0.001). Specific characteristics of SOT, HM and mSC are summarized in S1–S3 Tables. Among secondary BSI, urinary tract infection (UTI) was the most common source of BSI in nm-IC, while intra-abdominal infection (IAI) was the more frequent source among SOT and mSC patients. Primary BSI was the first cause of GN-BSI in HM (HM: 62.1%; SOT: 18.5%; nm-IC: 17.2%; mSC: 10.6%, p < 0.001). Regarding aetiology, Enterobacterales were more common than non-fermentative Gram-negative (NFGN) rods. Among Enterobacterales, Escherichia coli was the leading pathogen (61.9%), followed by Klebsiella spp. (27.6%). NFGN rods were more frequently isolated in HM than other populations (18.8%, p < 0.001) and Pseudomonas spp. was the most common microorganism. SOT group had significant higher rate of resistance strains compared to all groups, except for FQR higher in HM group (72.7%, p < 0.001).
Regarding management, FU-BCs were performed in 22.5% of patients and were more frequent in SOT and HM patients than in the other groups (SOT: 29.7%; HM: 29.4%; mSC: 26.7%; nm-IC: 20.6%, p < 0.001). Source control was performed in 24.5% of GN-BSI. Empiric therapy was appropriate in 71.3% of patients, more frequently in HM group (p < 0.001). Patients received a median treatment course of 11 days (IQR 7–15), with SOT patients receiving longer treatment course compared to other populations (SOT:13 [10–18] days; HM: 12 [7–16] days; mSC:12 [8–16] days; nm-IC: 11 [7–15] days, p < 0.001).
All-cause 30-day mortality was significantly lower in SOT patients, and higher in mSC (SOT 8.2%; nm-IC 13.4%; HM 14.9%; mSC 19.9%, p < 0.001). Kaplan Meier survival analysis curves are shown in S1 Fig. The best fitting model for 30-days mortality in the overall population was a probit with 4 splines (Table 2). Performance of FUBC was associated with a lower risk of death (HR: 0.345, 95%IC: 0.218–0.546, p < 0.001). Factors associated with a higher mortality were age (HR: 1.007, 95% IC: 1.002–1.012, p = 0.005), Charlson Comorbidity Index (CCI) (HR: 1.070, 95% IC: 1.043–1.098, p < 0.001), SOFA score (HR: 1.113, 95% IC: 1.087–1.139, p < 0.001), septic shock (HR: 1.463, 95% IC: 1.200–1.785, p < 0.001), CR profile (HR: 1.387, 95% IC:1.172–1.643, p < 0.001), NFGN etiology (HR: 1.252, 95%IC: 1.034–1.515, p = 0.022); conversely, UTI as source of infection (HR: 0.680, 95% IC:0.562–0.823, p < 0.001) and source control execution (HR: 0.789, 95% IC: 0.666–0.934, p = 0.006) were protective factors. Within subpopulations, FUBC confirmed its association with a lower mortality risk among nm-IC (HR = 0.317, 95% CI 0.178–0.563, p < 0.001), mSC (HR = 0.198, 95% CI 0.058–0.673, p = 0.010) and, not reaching statistical significance, SOT (HR = 0.342, 95% CI 0.092–1.272, p = 0.110), while in the HM subpopulation it was unrelated with mortality (HR = 1.024, 95% CI: 0.248–4.228, p = 0.974) (S4–S7 Tables). The non-proportionality of FUBC related hazard over time was particularly evident in mSC and nm-IC patients (S2 Fig).
Rate of 90-day relapse was overall 8.3% and was higher in SOT and HM (SOT: 18.8%; HM: 11.8%; nmIC: 7.2%; mSC: 7.1%, p < 0.001). At multivariable analysis, CR (HR = 2.021, 95% CI:1.597–2.557,p < 0.001) and NFR (HR = 1.541, 95% CI:1.166–2.038, p = 0.002) were the strongest risk factors for relapse, while treatment duration was not significantly associated with increased relapse risk (Table 3). However, as shown in Fig 2, the risk of death was higher than the risk of relapse in patients with less than 7 days of therapy. Among subpopulations, SOT patients and HM patients had the higher risk of relapse regardless of treatment duration; mSC patients had an extremely higher risk of death than relapse regardless of treatment duration, while SOT and HM patients displayed a reduced risk of death when therapy lasted at least 7 days, data as shown in S8–S11 Tables and Fig 3.
Legend: Comparison of the cumulative risks of death and relapse among patients stratified by the duration of therapy. Patients receiving less than 7 days of therapy exhibited a higher risk of death compared to relapse.
Legend: Cumulative incidence functions for death and relapse for patients with SOT, HM, and mSC, stratified by treatment duration (<7 vs ≥ 7 days). SOT and HM patients exhibited a higher risk of relapse regardless of treatment length, while mSC patients showed a consistently higher risk of death. A ≥ 7-day therapy was associated with reduced mortality in SOT and HM subgroups. nm-IC = non major immunocompromised condition; SOT = solid organ transplantation; HM = haematological malignancies; mSC = metastatic solid cancer.
Discussion
We analysed a large cohort of patients with GN BSI focusing on three major groups of IC hosts as SOT recipients, patients with haematological malignancy (HM) and those with metastatic solid cancer (mSC). We observed peculiar differences regarding clinical and microbiological characteristics, and outcome between these populations and non-major immunocompromised study population. Specifically, we observed the highest 30-day mortality rates in the mSC and the lowest in the SOT group. Conversely, the 90-day relapse rate was higher in the SOT group compared to all other populations. Furthermore, we observed that FUBC execution was associated with a lower risk of death in overall population, while its impact was uncertain in SOT and absent in HM. Finally, treatment duration was not associated with a higher risk of relapse, but SOT and HM patients displayed a reduced risk of death when therapy lasted at least 7 days.
Lower mortality rates in SOT recipients with sepsis and/or bacteraemia compared with non-SOT patients have been reported suggesting a protective role of immunosuppressive therapy and/or a more aggressive management in such high-risk setting [3–15]. Conversely, Eichenberger and colleagues recently reported that there was no difference in mortality between SOT and immunocompetent patients with GN-BSI [16]. However, none of these studies provided a detailed description of comparator groups, as in our study. Here, we confirmed that mortality is lower in SOT than in the other groups, but it is worth remarking younger age and lower rate of comorbidities observed in SOT, mainly when compared with mSC and nm-IC patients. Conversely, SOT had the highest risk of relapse. Unfortunately, studies assessing the association between the net state of immunosuppression and outcome of BSI in SOT recipients are not available yet. Otherwise, as for the management, FU-BCs were almost protective against 30-day mortality as well as ≥7 days of treatment duration was associated with reduced 90-day mortality risk.
In patients with HM, a higher rate of fluoroquinolone (FQ) resistance and non-fermenting pathogens were observed compared with other groups. The first is a common place in centres still using FQ prophylaxis [17,18]. In this setting both mortality and relapse rates were higher compared with the other groups, although only for relapse difference was statistically significant. Compared to our cohort, lower mortality rates were reported in a large multicenter Spanish study, ranging from 11% in allogeneic to 5% in autologous HSCT patients [19]. However, the authors reported the outcomes of HSCT patients with BSI due to both Gram-negative and Gram-positive bacteria. In our population, higher SOFA and BSI due to a non-fermenting GNB were associated with increased risk of death, appropriate empiric therapy was protective, while FU-BC was not associated with improved outcome in HM patients (S6 Table). Further, at least 7 days of treatment was not associated with lower relpase but with a reduced 90-day mortality risk. Similarly, Puerta-Alcalde and collegues [20], reported P. aeruginosa aetiology and inappropriate empiric antibiotic therapy for GN rods as independent risk factors for mortality in HSCT patients with BSI. In addition, a recent study assessing impact of treatment duration on 30-day recurrence or mortality showed that 7–11 days was not inferior to 12–21 days in 434 haematological patients with P. aeruginosa bacteraemia [21]. In our cohort, the median treatment duration of GN-BSI in HM patients was 12 (IQR 7–16) days, also in patients with BSI due to P. aeruginosa [12 (7–16) days].
Patients with mSC showed the highest mortality rate (19.9%). This finding is in line with previous evidence, although not all specifically focus on GN etiology, with mortality rates ranging from 13% to 36% [18,22]. The highest mortality rate in our population could be explained also by older age and high rate of comorbidities, while relapse was not more frequent than in other groups. Both FU-BCs and source control were associated with improved 30-day outcome, while treatment duration was not associated with 90-day relapse or mortality risk.
Patients with non-major immunocompromised condition showed a trend toward lower mortality and relapse rates compared with the other groups. FU-BCs were strongly protective, while there was no difference between shorter and longer treatment duration in terms of 90-day relapse or mortality risk. The role of FU-BC is highly debated as randomized controlled trial (RCT) is not available yet [23]. Despite this, two recent metanalysis showed a favourable impact of FU-BC on outcome of patients with GN-BSI [24–25], a recent large population-based study showed no difference in mortality in patients receiving or not FUBC (10.1% vs 8.9%, respectively) [26]. As for treatment duration, the findings from this group confirm what already observed in both observational and RCTs showing non-inferiority of shorter treatment duration in patients with GN-BSI in terms of relapse and/or mortality [27–29]. However, the main limitations of such studies were low representation of immunocompromised patients. In this retrospective analysis on three main groups of immunocompromised patients, surprisingly we did not observe an impact of treatment duration on 90-day relapse risk but a reduced risk of 90-day mortality in SOT and HM patients treated at least 7 days. Further multicentre studies are needed to confirm these findings.
There are several limitations in this study. We analysed a large cohort of patients, but the single-center design could limit the generalizability of our results. We focused on three specific types of IC condition, for which the increased risk of infectious complications is well-defined, considering all the other patients as not having a major IC status. However, patients included in this group could have been affected by other types and degrees of IC. We tried to take into account this occurrence by adjusting multivariable analysis for the Charlson index score. As for solid cancer, we considered only patients with a metastatic solid cancer for their higher fragility, but also patients with localized solid tumor could present with several degrees of immunosuppression. Finally, despite the adjustment for potential confounding variables, the use of FU-BC as time dependent variable, and considering patients since the end of therapy for the analysis of treatment duration, selection and immortal time bias could not be fully eliminated.
In conclusion, our data suggest that the characteristics and outcome of GN-BSI is different between the three specific IC categories of SOT recipients, patients with HM and those with metastatic solid cancer. Management of GN-BSI should be targeted according with the type of IC patient. Large interventional studies are needed to investigate the impact of specific procedures, as FU-BC, as well as the optimal treatment duration in each IC group.
Supporting information
S2 Table. Hematologic malignancy patients characteristics.
https://doi.org/10.1371/journal.pone.0327535.s003
(DOCX)
S3 Table. Characteristics of patients with metastatic solid tumor.
https://doi.org/10.1371/journal.pone.0327535.s004
(DOCX)
S4 Table. Multivariable survival analysis of 30-day mortality in nm-IC population.
https://doi.org/10.1371/journal.pone.0327535.s005
(DOCX)
S5 Table. Multivariable survival analysis of 30-day mortality in SOT population.
https://doi.org/10.1371/journal.pone.0327535.s006
(DOCX)
S6 Table. Multivariable survival analysis of 30-day mortality in HM population.
https://doi.org/10.1371/journal.pone.0327535.s007
(DOCX)
S7 Table. Multivariable survival analysis of 30-day mortality in mSC population.
https://doi.org/10.1371/journal.pone.0327535.s008
(DOCX)
S8 Table. Multivariable survival analysis of 90-day relapse or death in nm-IC population.
https://doi.org/10.1371/journal.pone.0327535.s009
(DOCX)
S9 Table. Multivariable survival analysis of 90-day relapse or death in SOT population.
https://doi.org/10.1371/journal.pone.0327535.s010
(DOCX)
S10 Table. Multivariable survival analysis of 90-day relapse or death in HM population.
https://doi.org/10.1371/journal.pone.0327535.s011
(DOCX)
S11 Table. Multivariable survival analysis of 90-day relapse or death in mSC population.
https://doi.org/10.1371/journal.pone.0327535.s012
(DOCX)
S1 Fig. 30-day overall survival by IC subpopulation.
IC = immunocompromised condition; nm-IC = non major immunocompromised condition; SOT = solid organ transplantation; HM = haematological malignancies; mSC = metastatic solid cancer.
https://doi.org/10.1371/journal.pone.0327535.s013
(JPG)
S2 Fig. Model-estimated survival function regarding 30-day mortality for FUBC execution vs not execution.
FUBC = follow up bloodcultures; nm-IC = non major immunocompromised condition; SOT = solid organ transplantation; HM = haematological malignancies; mSC = metastatic solid cancer.
https://doi.org/10.1371/journal.pone.0327535.s014
(JPG)
References
- 1. Kern WV, Rieg S. Burden of bacterial bloodstream infection-a brief update on epidemiology and significance of multidrug-resistant pathogens. Clin Microbiol Infect. 2020;26(2):151–7. pmid:31712069
- 2. Diallo K, Thilly N, Luc A, Beraud G, Ergonul Ö, Giannella M, et al. Management of bloodstream infections by infection specialists: an international ESCMID cross-sectional survey. Int J Antimicrob Agents. 2018;51(5):794–8. pmid:29309899
- 3. Heil EL, Bork JT, Abbo LM, Barlam TF, Cosgrove SE, Davis A, et al. Optimizing the management of uncomplicated gram-negative bloodstream infections: consensus guidance using a modified delphi process. Open Forum Infect Dis. 2021;8(10):ofab434. pmid:34738022
- 4. Kritikos A, Manuel O. Bloodstream infections after solid-organ transplantation. Virulence. 2016;7(3):329–40. pmid:26766415
- 5. Gustinetti G, Mikulska M. Bloodstream infections in neutropenic cancer patients: a practical update. Virulence. 2016;7(3):280–97. pmid:27002635
- 6. Amanati A, Sajedianfard S, Khajeh S, Ghasempour S, Mehrangiz S, Nematolahi S, et al. Bloodstream infections in adult patients with malignancy, epidemiology, microbiology, and risk factors associated with mortality and multi-drug resistance. BMC Infect Dis. 2021;21(1):636. pmid:34215207
- 7. Harris PA, Taylor R, Minor BL, Elliott V, Fernandez M, O’Neal L. The REDCap consortium: building an international community of software platform partners. J Biomed Inform. 2019;95:103208.
- 8. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373–83. pmid:3558716
- 9.
Siegel S, Castellan NJ Jr. Nonparametric statistics for the behavioral sciences. 2nd ed. New York: McGrawHill; 1988. https://doi.org/10.1177/014662168901300212
- 10.
Royston P, Lambert PC. Flexible parametric survival analysis using Stata: beyond the Cox model. First ed. College Station, Texas: StataCorp LP; 2011. p 347.
- 11. Hinchliffe SR, Lambert PC. Extending the flexible parametric survival model for competing risks. Stata J Promot Commun Stat Stata. 2013;13(2):344–55.
- 12. Lambert PC, Royston P. Further development of flexible parametric models for survival analysis. Stata J Promot Commun Stat Stata. 2009;9(2):265–90.
- 13. Battle SE, Al-Hasan MN. Paradoxical outcomes of gram-negative bloodstream infection in solid organ transplant recipients. Transpl Infect Dis. 2022;24(6):e13964. pmid:36411497
- 14. Ackerman KS, Hoffman KL, Díaz I, Simmons W, Ballman KV, Kodiyanplakkal RP, et al. Effect of sepsis on death as modified by solid organ transplantation. Open Forum Infect Dis. 2023;10(4):ofad148. pmid:37056981
- 15. Mezochow A, Anesi JA. The intricate interplay of immunosuppression and outcomes following Gram-negative bloodstream infection in solid organ transplantation. Transpl Infect Dis. 2022;24(6):e13966. pmid:36411541
- 16. Eichenberger EM, Troy J, Ruffin F, Dagher M, Thaden JT, Ford ML, et al. Gram-negative bacteremia in solid organ transplant recipients: clinical characteristics and outcomes as compared to immunocompetent non-transplant recipients. Transpl Infect Dis. 2022;24(6):e13969. pmid:36411527
- 17. Averbuch D, Tridello G, Hoek J, Mikulska M, Akan H, Yanez San Segundo L, et al. Antimicrobial Resistance in gram-negative rods causing bacteremia in hematopoietic stem cell transplant recipients: intercontinental prospective study of the infectious diseases working party of the european bone marrow transplantation group. Clin Infect Dis. 2017;65(11):1819–28. pmid:29020364
- 18. Gudiol C, Bodro M, Simonetti A, Tubau F, González-Barca E, Cisnal M, et al. Changing aetiology, clinical features, antimicrobial resistance, and outcomes of bloodstream infection in neutropenic cancer patients. Clin Microbiol Infect. 2013;19(5):474–9. pmid:22524597
- 19. Moreno A, Cervera C, Gavaldá J, Rovira M, de la Cámara R, Jarque I, et al. Bloodstream infections among transplant recipients: results of a nationwide surveillance in Spain. Am J Transplant. 2007;7(11):2579–86. pmid:17868067
- 20. Puerta-Alcalde P, Chumbita M, Charry P, Castaño-Díez S, Cardozo C, Moreno-García E, et al. Risk factors for mortality in hematopoietic stem cell transplantation recipients with bloodstream infection: points to be addressed by future guidelines. Transplant Cell Ther. 2021;27(6):501.e1-501.e6. pmid:33891882
- 21. Feng X, Qian C, Fan Y, Li J, Wang J, Lin Q, et al. Is short-course antibiotic therapy suitable for pseudomonas aeruginosa bloodstream infections in onco-hematology patients with febrile neutropenia? results of a multi-institutional analysis. Clin Infect Dis. 2024;78(3):518–25. pmid:37795577
- 22. Velasco E, Byington R, Martins CAS, Schirmer M, Dias LMC, Gonçalves VMSC. Comparative study of clinical characteristics of neutropenic and non-neutropenic adult cancer patients with bloodstream infections. Eur J Clin Microbiol Infect Dis. 2006;25(1):1–7. pmid:16424972
- 23. Huttner BD, Sharland M, Huttner A. On culture and blood cultures. Clin Microbiol Infect. 2023;29(9):1100–2. pmid:37263416
- 24. Gatti M, Bonazzetti C, Tazza B, Pascale R, Miani B, Malosso M, et al. Impact on clinical outcome of follow-up blood cultures and risk factors for persistent bacteraemia in patients with gram-negative bloodstream infections: a systematic review with meta-analysis. Clin Microbiol Infect. 2023.
- 25. Thaden JT, Cantrell S, Dagher M, Tao Y, Ruffin F, Maskarinec SA, et al. Association of follow-up blood cultures with mortality in patients with gram-negative bloodstream infections: a systematic review and meta-analysis. JAMA Netw Open. 2022;5(9):e2232576. pmid:36136334
- 26. Ong SWX, Luo J, Fridman DJ, Lee SM, Johnstone J, Schwartz KL, et al. Follow-up blood cultures do not reduce mortality in hospitalized patients with Gram-negative bloodstream infection: a retrospective population-wide cohort study. Clin Microbiol Infect. 2024;30(7):890–8. pmid:38552794
- 27. Giannella M, Pascale R, Toschi A, Ferraro G, Graziano E, Furii F, et al. Treatment duration for Escherichia coli bloodstream infection and outcomes: retrospective single-centre study. Clin Microbiol Infect. 2018;24(10):1077–83. pmid:29371138
- 28. Yahav D, Franceschini E, Koppel F, Turjeman A, Babich T, Bitterman R, et al. Seven versus 14 days of antibiotic therapy for uncomplicated gram-negative bacteremia: a noninferiority randomized controlled trial. Clin Infect Dis. 2019;69(7):1091–8. pmid:30535100
- 29. Molina J, Montero-Mateos E, Praena-Segovia J, León-Jiménez E, Natera C, López-Cortés LE, et al. Seven-versus 14-day course of antibiotics for the treatment of bloodstream infections by Enterobacterales: a randomized, controlled trial. Clin Microbiol Infect. 2022;28(4):550–7. pmid:34508886