Daily Sepsis Research Analysis
Three impactful studies advanced sepsis science today: a multicenter cohort using deep learning showed that antibiotic timing can be tailored by objective risk strata; a clinically relevant mouse model revealed that delayed, continuous antithrombin III infusion improves survival by mitigating liver-specific thrombin-driven injury; and a Mendelian randomization plus single-cell analysis identified CD39+ CD8 T cells and androsterone sulfate as mediators of sepsis risk and mortality.
Summary
Three impactful studies advanced sepsis science today: a multicenter cohort using deep learning showed that antibiotic timing can be tailored by objective risk strata; a clinically relevant mouse model revealed that delayed, continuous antithrombin III infusion improves survival by mitigating liver-specific thrombin-driven injury; and a Mendelian randomization plus single-cell analysis identified CD39+ CD8 T cells and androsterone sulfate as mediators of sepsis risk and mortality.
Research Themes
- Risk-stratified antibiotic timing using deep learning
- Coagulation–endothelium axis and organ-targeted therapy in sepsis
- Immunometabolic mechanisms (CD39+ T cells and steroid metabolites) driving sepsis
Selected Articles
1. Mortality and antibiotic timing in deep learning-derived surviving sepsis campaign risk groups: a multicenter study.
In a 34,087-patient multicenter cohort, deep learning models objectively stratified sepsis risk at triage. Patients with possible sepsis unlikely to develop shock had similar mortality whether antibiotics were given within 1–3 hours or later, whereas patients with probable sepsis had significantly lower mortality with antibiotics within 1 hour.
Impact: This study operationalizes SSC risk groups with objective models and links them to outcome by antibiotic timing, informing nuanced, risk-tailored antimicrobial stewardship.
Clinical Implications: Adopt risk-stratified antibiotic timing: maintain 1-hour targets for probable sepsis, but consider more lenient timing for low-risk patients to enable diagnostic clarification and antimicrobial stewardship. Prospective interventional validation is needed before changing protocols.
Key Findings
- Two prospectively applied deep learning models stratified patients by sepsis probability and shock risk at triage across two health systems.
- Group mortality and median time to antibiotics: (1) shock likely/probable sepsis 23.2%, 1.7 h; (2) shock likely/possible sepsis 17.7%, 3.0 h; (3) shock unlikely/probable sepsis 5.0%, 2.8 h; (4) shock unlikely/possible sepsis 1.9%, 4.6 h.
- In probable sepsis, antibiotics within 1 hour were associated with lower mortality; in possible sepsis unlikely to develop shock, mortality was similar regardless of 1–3 hour windows.
Methodological Strengths
- Large multicenter cohort with external validation and prospectively applied risk models.
- Objective, model-based risk stratification aligned with SSC categories, enabling pragmatic comparison by antibiotic timing.
Limitations
- Observational design limits causal inference; unmeasured confounding in antibiotic timing is possible.
- Generalizability may be limited to health systems studied; details of model transportability and clinician behavior were not fully explored.
Future Directions: Prospective, randomized or quasi-experimental trials testing risk-stratified antibiotic timing and real-time model deployment to confirm safety and benefits for stewardship.
BACKGROUND: The current Surviving Sepsis Campaign (SSC) guidelines provide recommendations on timing of administering antibiotics in sepsis patients based on probability of sepsis and presence of shock. However, there have been minimal efforts to stratify patients objectively into these groups and describe patient outcomes as a function of antibiotic timing recommendations based on risk stratification using this approach. METHODS: We conducted an observational cohort study using prospectively applied patient data from two large health systems using patient encounters between 2016 and 2024. At the time of clinical suspicion of sepsis, two deep learning (DL) models were used to stratify patients objectively into groups analogous to the SSC risk groups, based on a patient's likelihood of having sepsis and likelihood of developing shock. These risk groups were: (1) shock likely to develop and sepsis probable, (2) shock likely to develop and sepsis possible, (3) shock unlikely to develop and sepsis probable, and (4) shock unlikely to develop and sepsis possible. The primary outcome was short-term mortality, a composite of in-hospital mortality and transition to hospice care, across each risk group. RESULTS: We identified 34,087 adult patients with potential sepsis. At the development site, risk group mortality rates (%) and median time to antibiotics [IQR] were as follows: (1) 23.2%, 1.7 [1.0-3.1] hours; (2) 17.7%, 3.0 [1.7-6.2] hours; (3) 5.0%, 2.8 [1.5-5.1] hours; and (4) 1.9%, 4.6 [2.7-8.0] hours. Results from the validation site were similar. Mortality rates were similar for patients with possible sepsis unlikely to develop shock regardless of antibiotic administration within 1, 3 or more hours from triage. For patients with probable sepsis at the development site, regardless of risk of shock, mortality was significantly lower if antibiotics were administered within the first hour from triage. CONCLUSIONS: Our data suggest that patients who are at low risk of developing shock and possible sepsis had similar rates of mortality in the 1-hour vs. > 1-hour and 3-hour vs. > 3-hour time to antibiotic administration groups. Thus, a more lenient time to antibiotic administration could allow for more detailed evaluations and judicious administration of antibiotics without impacting patient mortality. Patients with probable sepsis had lower mortality if antibiotics were administered within 1 h from triage, regardless of risk of shock. Additional prospective studies are required to validate these findings and guide optimal antibiotic timing in patients with suspected sepsis.
2. Mechanisms of Sepsis From a Metabolic Immunology Perspective: A Bidirectional Mendelian Randomization and Single-Cell Sequencing Study of CD39 + Cells.
Using bidirectional Mendelian randomization and single-cell transcriptomics, the authors implicate CD3+CD39+CD8+ T cells as causal contributors to sepsis incidence and 28-day mortality. Androsterone sulfate partially mediated this risk, underscoring immunometabolic crosstalk as a targetable pathway.
Impact: Integrating human genetics with single-cell data links a specific immune subset to sepsis outcomes and identifies a steroid metabolite mediator, opening avenues for CD39–adenosine axis or metabolic interventions.
Clinical Implications: CD39+ CD8 T-cell abundance and related metabolite signatures could inform risk stratification and therapeutic targeting (e.g., CD39/adenosine pathway, androgen metabolism). Interventional studies are required before clinical adoption.
Key Findings
- CD39 expression was upregulated in immune cells from sepsis patients in single-cell analyses.
- Bidirectional two-sample MR implicated CD3+CD39+CD8+ T cells as risk factors for sepsis incidence (OR 1.053, P=0.008) and 28-day mortality (OR 1.108, P=0.037).
- Seventy-three metabolites correlated with these cells; androsterone sulfate mediated 4.97% of sepsis risk (P=0.026), supporting immunometabolic mediation.
Methodological Strengths
- Bidirectional Mendelian randomization across large GWAS with multiple immune phenotypes and metabolites.
- Orthogonal validation via single-cell transcriptomics (clustering, differential expression, pathway enrichment).
Limitations
- MR assumptions (relevance, independence, exclusion restriction) may be violated; effect sizes are modest.
- Single-cell validation relies on one dataset; mediation by androsterone sulfate explains a small proportion of risk.
Future Directions: Test pharmacologic modulation of the CD39–adenosine pathway and steroid metabolism in preclinical models and early-phase trials; expand multi-omics datasets and longitudinal validation.
BACKGROUND: Interactions between immune phenotypes and metabolites in sepsis pathogenesis remain poorly defined. We integrated Mendelian randomization (MR) and single-cell transcriptomics to investigate metabolic mediation in immune-sepsis associations. METHODS: Bidirectional two-sample MR analyzed sepsis genome-wide association studies (11,643 cases), 1,400 metabolites, and 731 immune phenotypes. Single-cell analysis of GSE167363 (sepsis vs. controls) included clustering, differential expression and pathway enrichment. RESULTS: CD39 expression was upregulated in sepsis immune cells. MR-identified CD3 + CD39 + CD8 + T cells as risk factors for sepsis incidence (odds ratio = 1.053, P = 0.008) and 28-day mortality (odds ratio = 1.108, P = 0.037). These cells correlated with 73 metabolites, notably androsterone sulfate, which mediated 4.97% of sepsis risk ( P = 0.026). CONCLUSION: CD39 + CD8 + T cells drive sepsis progression through metabolic intermediates like androsterone sulfate, highlighting immunometabolic crosstalk as a therapeutic target.
3. Continuous antithrombin III infusion in a clinically relevant sepsis model.
In a polymicrobial CLP mouse model with clinically relevant delayed treatment, continuous antithrombin III infusion improved 7-day survival compared with both saline and bolus AT. Protection was liver-specific, reducing thrombin-driven vascular leak and inflammation without altering bacterial burden.
Impact: Demonstrates that dosing schedule and organ context determine AT efficacy, potentially explaining prior clinical failures and guiding future trial design toward continuous infusion and liver-focused endpoints.
Clinical Implications: Suggests evaluating continuous AT infusion in abdominal sepsis, potentially combined with agents protecting non-hepatic organs, with pharmacokinetic/pharmacodynamic guidance. Translation requires careful patient selection and organ-specific endpoints.
Key Findings
- Delayed (~6 h), continuous AT infusion improved 7-day survival vs saline (65% vs 29%, p=0.018) and vs single bolus AT (65% vs 19%, p=0.003).
- Protective effects were confined to liver: reduced vascular leakage and inflammatory cytokines; kidney and lung were not protected.
- AT did not change organ bacterial counts; liver had the highest bacterial and thrombin accumulation at 48 h, implicating a thrombin-driven injury axis.
Methodological Strengths
- Clinically relevant polymicrobial CLP model with treatment delayed to mimic real-world initiation.
- Direct comparison of continuous infusion vs bolus dosing using osmotic minipumps ensures steady exposure.
Limitations
- Mouse model findings may not translate directly to humans; organ-specific effects could differ clinically.
- Did not test multiple dosing rates or other infection sources; combination strategies were not evaluated experimentally.
Future Directions: Conduct PK/PD-guided dose-ranging trials of continuous AT in abdominal sepsis with liver-focused biomarkers and explore rational combinations for extrahepatic protection.
BACKGROUND: Bolus antithrombin-III (AT) improved sepsis/organ dysfunction and survival in lipopolysaccharide/monomicrobial infusion pretreatment animal models; however, AT failed in clinical trials. Because insults and drug administration schedules differed between pre-clinical and clinical settings, we re-examined AT using a clinically relevant polymicrobial insult (cecal ligation and puncture, CLP) and a new method to continuously infuse AT after animals became ill. METHODS: Mice were catheterized with saline-filled osmotic minipumps. During CLP surgery we inserted AT- or saline-containing minipumps. We created and validated a ~ 6 h delay between sepsis induction and treatment. We compared delayed, continuous AT infusion with a conventional bolus AT injection using survival studies and 48 h studies. RESULTS: 6 h delayed, continuous AT infusion significantly improved 7d survival vs saline infusion (65% vs 29%, n = 21, p = 0.018) and vs a single injection of AT (65% vs. 19%, n = 21, p = 0.003). Delayed, continuous AT attenuated liver but not kidney or lung injury. Vascular leakage and inflammatory cytokines were suppressed only in liver. The highest accumulation of bacteria and thrombin at 48 h was in liver. AT did not change organ bacterial counts. CONCLUSIONS: Delayed, continuous AT infusion improved 7d survival after CLP compared to single bolus AT injection or continuous vehicle. Liver may be critical in abdominal sepsis because of bacterial accumulation and subsequent thrombin generation. AT may be protective due to attenuation of thrombin-induced vascular leakage, inflammation, and liver injury during CLP sepsis. Because other organs were unprotected, AT may be combined with drugs protecting different organs.