Monthly ReportMar 1–31, 2025
Sepsis, March 2025 edition
We read 92 papers and selected 5.
Summary
March 2025 sepsis research converged on precision host biology, rapid diagnostics, immune-endothelial injury, and pragmatic infection prevention. Mechanistic studies identified endothelial ferroptosis, TLR4/redox-driven microbiome disruption, and phase-specific immune dysfunction as actionable pathways, while AI-based vasopressor optimization offered a clinically testable approach to septic shock management. Large clinical and global-health studies supported scalable prevention strategies, including intrapartum azithromycin, and reinforced the importance of antimicrobial stewardship and resistance surveillance. Overall, the month showed a sustained transition from descriptive sepsis phenotyping toward biomarker-guided, mechanism-based, and implementation-oriented care.
Selected Articles
1. Extracellular vesicle-packaged GBP2 from macrophages aggravates sepsis-induced acute lung injury by promoting ferroptosis in pulmonary vascular endothelial cells.
This preclinical study identified a macrophage extracellular-vesicle GBP2–OTUD5–GPX4 pathway that promotes endothelial ferroptosis and vascular barrier disruption in sepsis-associated lung injury. Findings were supported by patient EV samples, cellular experiments, cecal ligation and puncture models, and endothelial-specific Gpx4 knockout models. Plantainoside D disrupted the GBP2–OTUD5 interaction, reduced GPX4 ubiquitination, and attenuated lung injury in experimental systems.
Impact: The paper links extracellular-vesicle signaling to endothelial ferroptosis using multiple complementary models and identifies a druggable molecular interaction with a candidate small molecule.
Clinical Implications: EV-associated GBP2 could become a biomarker of endothelial injury, while GBP2, OTUD5, GPX4 ubiquitination, and Plantainoside D provide a translational pipeline for lung-protective therapy. Pharmacokinetic, toxicology, dose-response, and prospective biomarker studies are required.
Key Findings
- Macrophage-derived extracellular vesicles promoted endothelial ferroptosis and barrier disruption.
- GBP2 bound OTUD5 and promoted GPX4 ubiquitination.
- Plantainoside D disrupted the GBP2–OTUD5 interaction and reduced experimental lung injury.
2. Sublethal systemic LPS in mice enables gut-luminal pathogens to bloom through oxygen species-mediated microbiota inhibition.
In mice, physiologic-range systemic LPS rapidly caused 100- to 10,000-fold expansion of facultative pathogens, including Klebsiella, Escherichia coli, Enterococcus, and Salmonella, without overt enteropathy. The process depended on TLR4-driven increases in luminal reactive oxygen species, which inhibited normal fermentation and favored oxidative-respiration growth by facultative anaerobes.
Impact: The study offers a compelling host-side mechanism for pathogen blooms and secondary infections during systemic inflammation, identifying TLR4 and luminal redox regulation as preventive targets.
Clinical Implications: The findings support investigation of host-directed approaches, such as modulation of TLR4 signaling, preservation of microbial fermentation, or targeted luminal redox control. Translation to humans requires confirmation in critically ill patient cohorts and careful assessment of infection and immune risks.
Key Findings
- Systemic LPS caused 100- to 10,000-fold expansion of several facultative gut pathogens within 24 hours.
- The response was dependent on TLR4 signaling.
- Increased luminal reactive oxygen species inhibited fermentation and favored pathogen growth.
3. Optimal Vasopressin Initiation in Septic Shock: The OVISS Reinforcement Learning Study.
The OVISS study used multicenter electronic health-record data and reinforcement learning to derive and externally validate a policy favoring earlier vasopressin initiation at lower norepinephrine doses. In validation across 227 hospitals, treatment concordant with the policy was associated with lower in-hospital mortality, with an adjusted odds ratio of 0.81.
Impact: The study presents an externally validated, clinically interpretable AI-derived treatment policy that could alter vasopressor sequencing and be integrated into electronic decision support.
Clinical Implications: The results justify prospective pragmatic trials of earlier vasopressin initiation and cautious evaluation of electronic decision support. Because the evidence is based on observational data and causal-policy modeling, it should not yet be interpreted as proof that the strategy improves mortality.
Key Findings
- The reinforcement-learning policy recommended vasopressin in 87% of cases versus 31% under usual care.
- Policy-concordant initiation occurred earlier and at lower norepinephrine doses.
- Concordant treatment was associated with lower mortality across external validation datasets.
4. Innate Immune Activation Is a Strong Suppressor of CCL22 and Impedes Regulatory T Cell-Dendritic Cell Interaction.
This study showed that TLR, RIG-I-like receptor, and STING activation suppresses CCL22 production by dendritic cells and reduces regulatory T cell–dendritic cell clustering. The mechanism was reproduced in infection models and supported by lower serum CCL22 in patients with sepsis, suggesting a phase-specific immune biomarker and a framework for timing immunomodulatory therapy.
Impact: The paper provides a mechanistic explanation for impaired regulatory immune-cell interactions during sepsis and connects the pathway to a measurable human biomarker.
Clinical Implications: Serum CCL22 could potentially help identify immune phases and stratify patients for time-sensitive immunomodulation. However, longitudinal validation is needed because CCL22 levels may vary with infection source, disease severity, and treatment exposure.
Key Findings
- TLR, RIG-I-like receptor, and STING activation suppressed CCL22 expression.
- Reduced CCL22 impaired regulatory T cell–dendritic cell clustering.
- Sepsis patient sera showed decreased CCL22 concentrations.
5. Effectiveness of intrapartum azithromycin to prevent infections in planned vaginal births in low-income and middle-income countries: a post-hoc analysis of data from a multicentre, randomised, double-blind, placebo-controlled trial.
A post-hoc analysis of the A-PLUS randomized trial involving 29,278 women found that a single 2 g oral intrapartum dose of azithromycin reduced maternal infection from 5.6% to 4.0% (RR 0.71, 95% CI 0.64–0.79). No increase in neonatal infection or clear safety signal was detected, and complementary modeling suggested potential cost savings in low- and middle-income settings.
Impact: The study provides large-scale randomized evidence for a simple, scalable maternal infection-prevention intervention with immediate global-health and policy relevance.
Clinical Implications: Health systems in low- and middle-income countries may consider intrapartum azithromycin within context-specific implementation programs, but rollout should include antimicrobial-resistance surveillance, stewardship safeguards, and continued neonatal safety monitoring.
Key Findings
- Among 29,278 randomized women, maternal infection was 4.0% with azithromycin versus 5.6% with placebo.
- The relative risk of maternal infection was 0.71 with a 95% confidence interval of 0.64–0.79.
- No increase in neonatal infection or major safety signal was detected.