https://orcid.org/0000-0002-3305-0077
Figures
Abstract
Purpose
Corticosteroids are suggested for selected patients with septic shock, but their association with outcomes has not been specifically investigated in patients with hospital-acquired bloodstream infections (HA-BSI). We explored the association between hydrocortisone use and day-28 mortality in this population.
Methods
Data from the prospective EUROBACT-2 study were used to assess the association between hydrocortisone use and day-28 mortality in ICU patients with HA-BSI and sepsis. Patients receiving hydrocortisone were compared with patients not receiving hydrocortisone. A 1:1 propensity score matching procedure was performed using nearest-neighbour matching without replacement and a caliper of 0.2 standard deviation of the logit propensity score, followed by multivariable logistic regression with cluster-robust standard errors by matched pair.
Results
Among 2,401 analyzed patients, 608 (25%) received hydrocortisone during the HA-BSI episode. The final propensity score included age, sex, Charlson comorbidity score, ICU admission diagnosis, SOFA score without cardiovascular component, septic shock, country of inclusion, lactate at HA-BSI diagnosis, and ICU-acquired versus non-ICU-acquired HA-BSI. Matching yielded 521 pairs (1,042 patients). In the multivariable model, hydrocortisone use was associated with increased day-28 mortality in the overall matched cohort (OR 1.448; 95% CI 1.096–1.914; p = 0.009). This association was observed in patients with sepsis without shock (OR 1.697; 95% CI 1.051–2.740; p = 0.031) and in the female subgroup (OR 1.718; 95% CI 1.055–2.796; p = 0.030), but not in patients with septic shock or in those receiving norepinephrine-equivalent dose ≥0.25 µg/kg/min.
Conclusion
Hydrocortisone use was not associated with improved day-28 survival in ICU patients with HA-BSI. In the matched analysis, hydrocortisone use was associated with higher day-28 mortality in the overall cohort, whereas no significant association was observed in patients with septic shock or high-dose norepinephrine. These findings are observational and hypothesis-generating and should not be interpreted causally.
Citation: Duclos G, Boucekine M, Tabah A, Ruckly S, Buetti N, Timsit JF, et al. (2026) Association between hydrocortisone use and outcomes in sepsis patients with hospital-acquired bloodstream infections: An ancillary analysis of the EUROBACT-2 cohort. PLoS One 21(9): e0356721. https://doi.org/10.1371/journal.pone.0356721
Editor: Iddya Karunasagar, Nitte University, INDIA
Received: January 6, 2026; Accepted: August 5, 2026; Published: September 15, 2026
Copyright: © 2026 Duclos 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: Data available upon reasonable requests to Claire Dupuis (data protection officer) of the Outcomerea network. It will also require the scientific acceptance of the steering committee of Eurobact2 trial and the scientific committee of the Outcome rea network. Please contact cdup83@gmail.com or outcomerea@outcomerea.org.
Funding: The author(s) received no specific funding for this work.
Competing interests: Competing interests: GD received fees for lectures from AOP Pharma. MB has no disclosure to declare AT has no disclosure to declare SR has no disclosure to declare NB has no disclosure to declare JFT received grants from Merck and Pfizer for research unit, consulting fees from Advanz, Pfizer, Menarini and Merck serves as lecturer for Mundipharma and Biomerieux. ML serves as lecturer for Grifols, Shinogi and consultant for AOP Pharma, Viatris, Previa.
Introduction
The Surviving Sepsis Campaign (SSC) guidelines suggest intravenous hydrocortisone, usually 200 mg daily, for adults with septic shock and an ongoing vasopressor requirement. The 2021 SSC guidelines refer to a norepinephrine or epinephrine dose ≥0.25 µg/kg/min for at least 4 hours after initiation as a clinically relevant threshold for vasopressor requirement [1]. Although corticosteroids shorten shock duration in septic shock, their effect on mortality remains uncertain. Furthermore, major randomized trials of corticosteroids in septic shock or severe infections included heterogeneous populations, and microbiological documentation was incomplete in a substantial proportion of patients [2–6]
The EUROBACT-2 international cohort prospectively enrolled 2,600 adult patients with HA-BSI managed in 333 ICUs across 52 countries. Because inclusion required a positive blood culture, all patients had microbiologically documented infection [7]. HA-BSI represents a late-onset, hospital-acquired and microbiologically documented infectious context that differs from many randomized corticosteroid trials. We used this database to explore the association between hydrocortisone use and outcomes in critically ill patients with HA-BSI, while accounting for measured severity, centre/country effects, and potential confounding by indication.
Methods
From the EUROBACT-2 database, we included adult ICU patients with HA-BSI, defined as a positive blood culture sample taken more than 48 hours after hospital admission. HA-BSI had to be managed in the ICU, either because it was acquired in the ICU or because the patient was transferred to the ICU for management of the infection [7].
All eligible patients with HA-BSI from the EUROBACT-2 cohort were considered. Hydrocortisone administration during the HA-BSI episode defined the exposed group, whereas patients not receiving hydrocortisone served as controls. Patients with chronic corticosteroid use of more than 20 mg per day for at least four weeks and patients receiving a corticosteroid other than hydrocortisone during the ICU stay were excluded according to the EUROBACT-2 study definitions.
Ethics statement
Initial ethical approval as a low-risk research project with waiver of individual consent was granted by the Human Research Ethics Committee of the Royal Brisbane & Women’s Hospital, Queensland, Australia (LNR/2019/QRBW/48376). Each study site then obtained ethical and governance approvals according to national and/or local regulations.
Primary goal
Our primary objective was to assess the association between hydrocortisone use during the HA-BSI episode and day-28 mortality in ICU patients with HA-BSI and sepsis.
Secondary goals
Secondary objectives were to assess the association between hydrocortisone use and day-28 mortality in predefined subgroups: patients with HA-BSI-related septic shock, patients with HA-BSI-related sepsis without shock, female patients, and patients receiving norepinephrine-equivalent dose ≥0.25 µg/kg/min at HA-BSI diagnosis. Exploratory secondary outcomes included vasopressor-free days at day 28, mechanical ventilation-free days at day 28, and SOFA variation between HA-BSI diagnosis and day 7.
Patient characteristics
- Demographic data, main diagnosis at admission to the ICU, and comorbidities were collected. The Charlson score was used as a continuous variable.
- Illness severity was assessed using the Simplified Acute Physiology Score II (SAPS II) and the Sequential Organ Failure Assessment (SOFA) score. The SOFA score without the cardiovascular component was calculated. Sepsis and septic shock were defined according to Sepsis-3 criteria on the day of blood culture sampling [1].
- The maximum dose of norepinephrine equivalent was calculated in µg/kg/min. For patients who acquired HA-BSI in the ICU, the highest dose administered on the day of HA-BSI diagnosis, which was defined as the day of blood culture sampling, was recorded. For those who acquired HA-BSI prior to ICU admission, the maximum dose administered on the first day of ICU stay was recorded.
- The maximum dose of norepinephrine equivalent was calculated by adding a dose of norepinephrine and epinephrine using a 1:1 dose equivalence [8,9].
- A norepinephrine-equivalent dose threshold of ≥0.25 µg/kg/min was used to define high-dose norepinephrine requirement. This threshold was considered a clinically relevant vasopressor-dose threshold in line with SSC guidance, rather than a formal definition of septic shock [1].
Characteristics of HA-BSI
- The sources of HA-BSI were categorized into six groups (catheter-related, respiratory, intra-abdominal, urinary, primary HA-BSI, and others) determined by the senior physician in charge. For patients with more than one possible source of HA-BSI, the sources were classified according to the most probable locus of infection.
- The time for adequate administration of empirical therapy was defined as the time in hours between HA-BSI and the administration of at least one antibiotic agent effective on the responsible pathogen, according to its antibiogram and infection source. The time of reference was when blood culture sampling was undertaken.
- Source control was defined as the performance of the measures required to treat the origin of the HA-BSI (e.g., percutaneous, surgical drainage, and catheter removal). Source control was classified into three categories according to the treating clinician: 1) source control measures were not indicated, 2) source control measures were indicated and appropriately performed, and 3) source control measures were indicated but were neither performed nor partially performed.
- HA-BSI was considered acquired after ICU admission when the first positive blood culture was obtained more than 48 hours after ICU admission.
Characteristics of hydrocortisone use
For each patient, hydrocortisone administration and number of treatment days were recorded. However, the exact date and time of hydrocortisone initiation relative to blood culture sampling or HA-BSI diagnosis were not available. Each centre reported its usual protocol regarding hydrocortisone indications, dosage, and association with fludrocortisone, but details on the exact daily dose were not reported at the patient level.
Centre characteristics
Country and income categories were determined according to United Nations M49 standards. Clinical pharmacists’ consultations were categorized as follows in the original study [7]: “available during business hours,” “available on request 24/7,” “available as part of permanent ICU staff,” “available on scheduled meeting,” and “never consulted.” For this study, this outcome was assessed as “any consultation with a clinical pharmacist,” as opposed to no consultation with a clinical pharmacist.
Statistical analysis
Baseline characteristics are presented for the full enrolled sample and for each subgroup. The Wilcoxon rank-sum test and the chi-square test were used to test for differences for continuous and categorical variables, respectively. A Fisher’s exact test was used for categorical variables with individual cell counts below five.
For all calculations, the time of reference was the moment of the first positive blood culture sampling. Vital status on day 28 after the first positive blood culture sampling was collected. Exposure to invasive therapies (mechanical ventilation, vasopressors, renal replacement therapy, and extra-corporeal membrane oxygenation therapy) was also noted. To prevent exposure bias, exposure-free days were calculated for survivors on day 28. All patients who died during the period had an exposure-free day from invasive therapies, calculated as equal to zero.
Directed acyclic graphs (DAGs) were used to identify significant variables to include in propensity score matching and the final multivariate model. DAGs, or causal diagrams, are graphical representations of causal structures that help to understand and illustrate potential biases, including biases arising from confounding, selection, and misclassification [10] (S1 Fig).
A propensity score model estimating the probability of receiving hydrocortisone was developed using logistic regression and variables selected from the DAGs and clinically relevant baseline characteristics [11]. Matching was performed using 1:1 nearest-neighbour matching without replacement and a caliper width of 0.20 standard deviation of the logit of the propensity score [11,12]. Patients outside the region of common support were not included in the matched analysis.
The final propensity score included age, sex, Charlson comorbidity score, diagnosis at ICU admission, SOFA score without cardiovascular component at HA-BSI diagnosis, septic shock, country of inclusion, lactate at HA-BSI diagnosis, and ICU-acquired versus non-ICU-acquired HA-BSI. Matching quality was assessed using propensity score distribution plots and standardized mean differences, with SMD < 0.1 considered indicative of good balance.
The primary multivariable logistic regression model assessed the association between hydrocortisone exposure and day-28 mortality in the matched sample. Covariates included respiratory source of infection, time to adequate antimicrobial therapy, source control category, clinical pharmacist never consulted, centre use of fludrocortisone, and norepinephrine-equivalent dose ≥0.25 µg/kg/min at HA-BSI diagnosis. Hydrocortisone duration was removed from the primary model because it occurred after treatment initiation and could be influenced by subsequent clinical evolution and survival.
Predefined subgroup analyses were performed in patients with septic shock at HA-BSI diagnosis, patients with sepsis without shock, female patients, and patients receiving norepinephrine-equivalent dose ≥0.25 µg/kg/min at HA-BSI diagnosis. The formal hydrocortisone × sex interaction was also assessed to avoid overinterpretation of sex-specific subgroup findings.
Sensitivity analysis
Sensitivity analyses were conducted to assess the robustness of the findings. First, patients admitted to the ICU for COVID-19 were excluded. Second, time to adequate antimicrobial therapy and source control were removed from the multivariable model because these variables may occur after baseline and may be influenced by clinical evolution. Third, patients with coagulase-negative staphylococcal bloodstream infection were excluded from the already matched cohort without rematching.
Exploratory secondary outcomes were analyzed according to hydrocortisone exposure. Vasopressor-free days and mechanical ventilation-free days at day 28 were analyzed, as was the change in SOFA score between HA-BSI diagnosis and day 7. For delta SOFA, negative values indicated improvement and positive values indicated worsening between HA-BSI diagnosis and day 7.
Multivariable analyses were performed in the matched cohort and predefined subgroups using the same adjustment strategy. Logistic regression models were used for day-28 mortality and linear regression models for exploratory continuous secondary outcomes. Cluster-robust standard errors by matched pair were used. Results were expressed as odds ratios (ORs) with 95% confidence intervals (CIs) for mortality and as adjusted mean differences (β) with 95% CIs for secondary outcomes. Statistical analyses were performed using R (R Core Team (2020), R: A language and environment for statistical computing) and Python for additional figure and table generation.
Results
Patient characteristics
The EUROBACT-2 cohort included 2,600 patients, of whom 162 were excluded because they received chronic corticosteroid treatment (Fig 1). Among the 2,438 remaining patients, 37 were excluded because they received corticosteroids other than hydrocortisone during the ICU stay. Among the 2,401 analyzed patients, 608 (25%) received hydrocortisone during the HA-BSI episode (Fig 1). Patient characteristics are reported in Table 1, with additional details in S1–S4 Tables. The mean age was 61 ± 16 years, and 1,543 patients (64%) were male. At ICU admission, the mean SAPS II and Charlson score were 49 ± 17 and 2.27 ± 2.57, respectively. The mean interval between hospital admission and ICU admission was 8.5 ± 22.6 days, and the mean time between hospital admission and HA-BSI diagnosis was 18 ± 29 days. HA-BSI was diagnosed before ICU admission in 500 (21%) patients, during the first 48 hours of ICU admission in 210 (8.8%) patients, and more than 48 hours after ICU admission in 1,685 (70%) patients.
Characteristics of corticosteroid use according to centre
Centre-level hydrocortisone practices are described in S5 Table. In 26 (8%) of 330 centres, hydrocortisone was never prescribed. In the other centres, usual indications for hydrocortisone included sepsis, septic shock, and septic shock with high-dose vasopressors. Among the 608 patients receiving hydrocortisone, most received a centre-reported regimen corresponding to 200 mg/day by intermittent bolus or continuous infusion. Hydrocortisone was usually administered until shock resolution or for a fixed 7-day course, according to centre practice.
Outcome
Propensity score matching using the final propensity score model yielded 521 pairs, corresponding to 1,042 matched patients (Fig 1). The effect of matching on variables included in the propensity score, variables included in the multivariable model, centre/country/income variables, and additional clinically relevant variables is shown in S2–S4 Figs. A descriptive comparison of matched and unmatched patients is provided in S7 Table, supporting the interpretation that the matched analysis applies to the region of common support.
Day-28 mortality was 36% (n = 875) in the unmatched cohort. In the matched cohort, day-28 mortality was 50% (260/521) in the hydrocortisone group and 42% (220/521) in the control group. In the multivariable analysis (N = 935), hydrocortisone use was associated with day-28 mortality in the overall matched cohort (OR 1.448; 95% CI 1.096–1.914; p = 0.009) (Fig 2 and Table 2).
Subgroup analyses
Among the matched patients, 598 had septic shock at HA-BSI diagnosis. Day-28 mortality in this subgroup was 53% (318/598). Mortality was 56% (168/300) in the hydrocortisone group and 50% (150/298) in the control group. In the multivariable analysis (N = 546), hydrocortisone use was not significantly associated with day-28 mortality in patients with septic shock (OR 1.392; 95% CI 0.952–2.036; p = 0.088) (Fig 2 and Table 2).
Sepsis without shock was documented in 444 matched patients. Day-28 mortality in this subgroup was 36% (162/444). Mortality was 42% (92/221) in the hydrocortisone group and 31% (70/223) in the control group. This association persisted in the multivariable analysis (N = 389) (OR 1.697; 95% CI 1.051–2.740; p = 0.031) (Fig 2 and Table 2).
The matched cohort included 368 female patients, with day-28 mortality of 48% (177/368). Mortality was 54% (97/181) in the hydrocortisone group and 43% (80/187) in the control group. In the multivariable analysis (N = 329), hydrocortisone use was associated with day-28 mortality in the female subgroup (OR 1.718; 95% CI 1.055–2.796; p = 0.030) (Fig 2 and Table 2). However, the formal hydrocortisone × sex interaction was not statistically significant; therefore, this subgroup finding should be considered exploratory.
In contrast, hydrocortisone use was not associated with day-28 mortality among patients receiving norepinephrine-equivalent dose ≥0.25 µg/kg/min at HA-BSI diagnosis (S8 Table).
Sensitivity analysis
A sensitivity analysis excluding patients admitted to the ICU for COVID-19 included 813 patients in the complete-case multivariable model. In this analysis, hydrocortisone use remained associated with increased day-28 mortality in the overall matched cohort (OR 1.445; 95% CI 1.072–1.947; p = 0.016), whereas no significant association was observed among patients with septic shock or high-dose norepinephrine (Fig 3).
Additional sensitivity analyses are presented in Figs 4 and 5 and in the Supplementary Materials. When time to adequate antimicrobial therapy and source control were removed from the multivariable model, the association between hydrocortisone use and day-28 mortality persisted in the overall matched cohort, in patients with sepsis without shock, and female subgroup, but not in patients with septic shock or high-dose norepinephrine. After exclusion of patients with coagulase-negative staphylococcal bloodstream infection from the already matched cohort, the overall association also persisted, while no significant association was observed in patients with septic shock or high-dose norepinephrine.
Secondary outcomes
Exploratory secondary outcomes are presented in Table 3. Hydrocortisone exposure was associated with fewer vasopressor-free days and fewer mechanical ventilation-free days at day 28, and with a less favorable SOFA trajectory between HA-BSI diagnosis and day 7. These secondary analyses were exploratory and should be interpreted cautiously because of the observational design and the post-baseline nature of these outcomes.
Discussion
To the best of our knowledge, this is the first study to assess the association between hydrocortisone use and day-28 mortality in ICU patients with sepsis and HA-BSI. In the propensity-matched cohort, hydrocortisone use was associated with increased day-28 mortality in the overall population. This association persisted after excluding patients admitted for COVID-19, after excluded potential confounding factor regarding the quality of post exposure treatment and after excluding patients with coagulase-negative staphylococcal bloodstream infection. However, no significant association was observed among patients with septic shock or among those receiving norepinephrine-equivalent dose ≥0.25 µg/kg/min. These findings suggest that the observed signal was mainly driven by patients outside the subgroup most consistent with current guideline-based hydrocortisone indications.
A strength of this study is that it assesses real-life hydrocortisone use in a large international cohort of patients with microbiologically documented HA-BSI. Unlike several corticosteroid trials in sepsis or severe infections, inclusion in EUROBACT-2 required a positive blood culture, ensuring microbiological documentation of the infectious episode [6,7]. This setting allowed us to explore hydrocortisone use in a specific hospital-acquired proved infection context that differs from many randomized trials conducted in broader septic shock or respiratory infection populations where the microbiological documentation vary between 55 and 72% [2–4,6].
Our findings should be interpreted in the context of the heterogeneous literature on corticosteroids in sepsis. Recent guideline updates support hydrocortisone use in selected patients with septic shock, mainly to accelerate shock resolution, but the effect on survival remains less certain [1,13]. In our cohort, hydrocortisone was not associated with increased mortality among patients with septic shock or high-dose norepinephrine. By contrast, the association with mortality was observed mainly in patients without shock and in the low-dose norepinephrine subgroup. These observations are consistent with guidelines and the hypothesis that corticosteroid effects may vary according to sepsis phenotype and severity [13,14]. In a large, emulated target trial framework based on 4,000 patients and stratified by a machine learning method, Rajendran et al. found that the use of corticosteroids in critically ill patients with infection was associated with an increase in organ failure, notably among those having a mild severe infection and those showing rapid improvement [14]. Consequently, the association explored by this work align with the existing data that suggests corticosteroids may be deleterious in patients with less severe sepsis.
The association observed in female patients should be interpreted cautiously. Although sex-related differences in sepsis and immune response have been described [15], the deleterious hydrocortisone & sex association found in the female subgroup result should be considered exploratory and hypothesis-generating rather than evidence of a sex-specific treatment effect.
This study has several limitations. First, hydrocortisone use was not randomized, and the rationale for administration was not recorded; residual confounding by indication may persist despite propensity score matching and multivariable adjustment. Second, the exact date and time of hydrocortisone initiation relative to blood culture sampling or HA-BSI diagnosis were not available, and time-dependent bias cannot be excluded. Third, daily patient-level hydrocortisone dosing was not available, and centre-level practices differed regarding dose, duration, and association with fludrocortisone [16]. Fourth, exact same-centre or same-country matching was not feasible without major loss of sample size, and residual centre/country effects may persist. Fifth, propensity score matching restricted the analysis to patients within the region of common support; therefore, findings may not apply to the entire EUROBACT-2 HA-BSI population. Finally, HA-BSI patients remained heterogeneous in terms of severity, source of infection, microbiology, immune status, and inflammatory phenotype, which may influence corticosteroid response [16,17].
Conclusion
Hydrocortisone use was associated with increased day-28 mortality in the overall propensity-matched cohort of ICU patients with HA-BSI. This association persisted in several sensitivity analyses but was not observed among patients with septic shock or high-dose norepinephrine. These findings should be interpreted cautiously as observational and hypothesis-generating, and they do not establish a causal harmful effect of hydrocortisone.
Supporting information
S2 Table. Bloodstream infection information and physiological conditions at the time of blood culture sampling.
https://doi.org/10.1371/journal.pone.0356721.s002
(DOCX)
S3 Table. Bacteriological characteristics of bloodstream infections.
https://doi.org/10.1371/journal.pone.0356721.s003
(DOCX)
S6 Table. Descriptive analysis of centres using fludrocortisone.
https://doi.org/10.1371/journal.pone.0356721.s006
(DOCX)
S7 Table. Characteristics of matched and unmatched patients after propensity score matching.
https://doi.org/10.1371/journal.pone.0356721.s007
(DOCX)
S8 Table. Multivariable analysis in the high-dose norepinephrine subgroup.
https://doi.org/10.1371/journal.pone.0356721.s008
(DOCX)
S2 Fig. Balance for propensity score variables and patient-level multivariable covariates.
https://doi.org/10.1371/journal.pone.0356721.s010
(TIF)
S3 Fig. Balance for centre, country, and income variables.
https://doi.org/10.1371/journal.pone.0356721.s011
(TIF)
S4 Fig. Balance for relevant variables not included in the propensity score or multivariable model.
https://doi.org/10.1371/journal.pone.0356721.s012
(TIF)
Acknowledgments
The authors are indebted to professor Djillali Annane (service de Médecine Intensive Réanimation, Hôpital Raymond Poincaré, Garches, France) and professor Karim Asehnoune (Departement d’anesthésie et de réanimation, Hôtel-Dieu, Université de Nantes, Nantes, France) for their fruitful comments and reviewing of this manuscript. The authors are also indebted to national coordinators of the Eurobact-2 study: The Eurobact 2 study group: National coordinators, scientific committee, and participating intensive care units. List of EUROBACT 2 investigators can be found in supplemental.
References
- 1. Evans L, Rhodes A, Alhazzani W, Antonelli M, Coopersmith CM, French C, et al. International guidelines for management of sepsis and septic shock 2021. Crit Care Med. 2021;49:e1063–e1143.
- 2. Dequin PF, Meziani F, Quenot JP, Kamel T, Ricard JD, Badie J, et al. Hydrocortisone in severe community-acquired pneumonia. N Engl J Med. 2023;388(23):1931–41.
- 3. Sprung CL, Annane D, Keh D, Moreno R, Singer M, Freivogel K, et al. Hydrocortisone therapy for patients with septic shock. N Engl J Med. 2008;358(2):111–24. pmid:18184957
- 4. Venkatesh B, Finfer S, Cohen J, Rajbhandari D, Arabi Y, Bellomo R, et al. Adjunctive glucocorticoid therapy in patients with septic shock. N Engl J Med. 2018;378:797–808.
- 5. Bode C, Weis S, Sauer A, Wendel-Garcia P, David S. Targeting the host response in sepsis: current approaches and future evidence. Crit Care. 2023;27(1):478. pmid:38057824
- 6. Annane D, Renault A, Brun-Buisson C, Megarbane B, Quenot J-P, Siami S, et al. Hydrocortisone plus fludrocortisone for adults with septic shock. N Engl J Med. 2018;378(9):809–18. pmid:29490185
- 7. Tabah A, Buetti N, Staiquly Q, Ruckly S, Akova M, Aslan AT, et al. Epidemiology and outcomes of hospital-acquired bloodstream infections in intensive care unit patients: the EUROBACT-2 international cohort study. Intensive Care Med. 2023;49(2):178–90. pmid:36764959
- 8. Goradia S, Sardaneh AA, Narayan SW, Penm J, Patanwala AE. Vasopressor dose equivalence: a scoping review and suggested formula. J Crit Care. 2021;61:233–40. pmid:33220576
- 9. Kotani Y, Di Gioia A, Landoni G, Belletti A, Khanna AK. An updated “norepinephrine equivalent” score in intensive care as a marker of shock severity. Crit Care. 2023;27(1):29. pmid:36670410
- 10. Digitale JC, Martin JN, Glymour MM. Tutorial on directed acyclic graphs. J Clin Epidemiol. 2022;142:264–7. pmid:34371103
- 11. Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res. 2011;46(3):399–424. pmid:21818162
- 12. Warisawa T, Cook CM, Ahmad Y, Howard JP, Seligman H, Rajkumar C, et al. Deferred versus performed revascularization for left main coronary disease with hemodynamic significance. Circ Cardiovasc Interv. 2023;16(6):e012700. pmid:37339234
- 13. Chaudhuri D, Nei AM, Rochwerg B, Balk RA, Asehnoune K, Cadena R, et al. Focused update: guidelines on use of corticosteroids in sepsis, acute respiratory distress syndrome, and community-acquired pneumonia. Crit Care Med. 2024;52(1):e219–33.
- 14. Rajendran S, Xu Z, Pan W, Zang C, Siempos I, Torres L, et al. Corticosteroids for infectious critical illness: a multicenter target trial emulation stratified by predicted organ dysfunction trajectory. 2024:2024.03.07.24303926.
- 15. Lakbar I, Einav S, Lalevée N, Martin-Loeches I, Pastene B, Leone M. Interactions between gender and sepsis-implications for the future. Microorganisms. 2023;11(3):746. pmid:36985319
- 16. Pirracchio R, Annane D, Waschka AK, Lamontagne F, Arabi YM, Bollaert PE, et al. Patient-level meta-analysis of low-dose hydrocortisone in adults with septic shock. NEJM Evid. 2023;2:EVIDoa2300034.
- 17. Smit JM, Van Der Zee PA, Stoof SCM, Van Genderen ME, Snijders D, Boersma WG, et al. Predicting benefit from adjuvant therapy with corticosteroids in community-acquired pneumonia: a data-driven analysis of randomised trials. Lancet Respir Med. 2025;13(3):221–33. pmid:39892408