Daily Sepsis Research Analysis
Three papers stand out today: an interpretable machine-learning framework (GroupFasterRisk) that matches black-box ICU mortality prediction while remaining sparse and transparent; a macrophage-targeted, mannose-modified exosomal miRNA therapy that alleviates sepsis-related acute lung injury in vivo; and a meta-analysis showing no outcome difference between continuous versus intermittent hydrocortisone in septic shock, informing pragmatic steroid protocols.
Summary
Three papers stand out today: an interpretable machine-learning framework (GroupFasterRisk) that matches black-box ICU mortality prediction while remaining sparse and transparent; a macrophage-targeted, mannose-modified exosomal miRNA therapy that alleviates sepsis-related acute lung injury in vivo; and a meta-analysis showing no outcome difference between continuous versus intermittent hydrocortisone in septic shock, informing pragmatic steroid protocols.
Research Themes
- Interpretable AI for critical care prognosis
- Targeted nanomedicine/miRNA therapy for sepsis-induced organ injury
- Optimization of corticosteroid administration strategies in septic shock
Selected Articles
1. Fast and interpretable mortality risk scores for critical care patients.
GroupFasterRisk yields sparse, interpretable ICU mortality risk scores that outperform OASIS/SAPS II and match APACHE IV/IVa with a fraction of parameters. In sepsis, AMI, heart failure, and AKI subgroups, it surpassed OASIS and SOFA, and its variable selection improved performance of other ML models.
Impact: It provides a practical path to deploy transparent, high-performing prediction tools acceptable to clinicians and regulators, addressing a key barrier to AI adoption in critical care and sepsis.
Clinical Implications: Hospitals can adopt interpretable risk scores with competitive accuracy to guide triage, resource allocation, and sepsis pathways, while enabling clinician oversight and auditing. Variable sets identified by GroupFasterRisk may also refine existing scoring systems.
Key Findings
- GroupFasterRisk outperformed OASIS and SAPS II and matched APACHE IV/IVa using at most one-third of parameters.
- In sepsis/septicemia, AMI, heart failure, and AKI subgroups, GroupFasterRisk beat OASIS and SOFA.
- Variable selection by GroupFasterRisk improved performance of other mortality ML models.
- The algorithm enforces sparsity, group structure, and monotonicity, and yields multiple equally good models.
Methodological Strengths
- Evaluation on the largest public ICU datasets (MIMIC III and eICU) with comprehensive benchmarks.
- Model design incorporates sparsity, group structure, and monotonicity, enhancing interpretability and domain alignment.
Limitations
- Retrospective validation; no prospective impact evaluation on patient outcomes.
- Exact sample sizes and external multi-center clinical deployment are not reported.
Future Directions: Prospective multi-center trials to assess clinical impact and fairness, EHR integration, and tailored sepsis-specific score deployment with clinician-in-the-loop refinement.
OBJECTIVE: Prediction of mortality in intensive care unit (ICU) patients typically relies on black box models (that are unacceptable for use in hospitals) or hand-tuned interpretable models (that might lead to the loss in performance). We aim to bridge the gap between these 2 categories by building on modern interpretable machine learning (ML) techniques to design interpretable mortality risk scores that are as accurate as black boxes.
MATERIAL AND METHODS: We developed a new algorithm, GroupFasterRisk, which has several important benefits: it uses both hard and soft direct sparsity regularization, it incorporates group sparsity to allow more cohesive models, it allows for monotonicity constraint to include domain knowledge, and it produces many equally good models, which allows domain experts to choose among them. For evaluation, we leveraged the largest existing public ICU monitoring datasets (MIMIC III and eICU).
RESULTS: Models produced by GroupFasterRisk outperformed OASIS and SAPS II scores and performed similarly to APACHE IV/IVa while using at most a third of the parameters. For patients with sepsis/septicemia, acute myocardial infarction, heart failure, and acute kidney failure, GroupFasterRisk models outperformed OASIS and SOFA. Finally, different mortality prediction ML approaches performed better based on variables selected by GroupFasterRisk as compared to OASIS variables.
DISCUSSION: GroupFasterRisk's models performed better than risk scores currently used in hospitals, and on par with black box ML models, while being orders of magnitude sparser. Because GroupFasterRisk produces a variety of risk scores, it allows design flexibility-the key enabler of practical model creation.
CONCLUSION: GroupFasterRisk is a fast, accessible, and flexible procedure that allows learning a diverse set of sparse risk scores for mortality prediction.
2. Mannose-modified exosomes loaded with MiR-23b-3p target alveolar macrophages to alleviate acute lung injury in Sepsis.
miR-23b is downregulated in macrophages during sepsis-related ALI; intratracheal miR-23b mimics alleviate injury by suppressing M1 activation via the Lpar1–NF-κB pathway. Mannose-modified MSC-derived exosomes enable targeted delivery to macrophages, offering a low-immunogenic platform for pulmonary miRNA therapy.
Impact: Introduces a macrophage-targeted exosomal miRNA therapy with mechanistic validation, opening a translational path for treating sepsis-induced lung injury.
Clinical Implications: While preclinical, the platform suggests a route to reduce inflammatory lung damage in sepsis by precisely delivering anti-inflammatory miRNA to alveolar macrophages, potentially lowering doses and systemic toxicity.
Key Findings
- miR-23b expression is reduced in macrophages within ALI tissue.
- Intratracheal miR-23b mimics alleviate ALI by suppressing M1 macrophage activation via the Lpar1–NF-κB pathway.
- Mannose-modified MSC-derived exosomes enable targeted delivery of miR-23b to macrophages, reducing immunogenicity concerns.
- The targeted exosomal miRNA strategy mitigated sepsis-induced lung injury in vivo.
Methodological Strengths
- Mechanistic linkage to the Lpar1–NF-κB pathway with functional rescue using miRNA mimics.
- Targeted delivery using mannose-modified MSC-derived exosomes demonstrated in vivo.
Limitations
- Preclinical animal study without human safety, dosing, or pharmacokinetic data.
- Comparative efficacy versus standard anti-inflammatory or anti-cytokine therapies not assessed.
Future Directions: Define safety, biodistribution, and dosing in larger animals; optimize exosome manufacturing; and conduct early-phase clinical trials in sepsis-related ALI.
The anti-inflammatory role of miR-23b-3p (miR-23b) is known in autoimmune diseases like multiple sclerosis, systemic lupus erythematosus, and rheumatoid arthritis. However, its role in sepsis-related acute lung injury (ALI) and its effect on macrophages in ALI remain unexplored. This investigation aimed to evaluate miR-23b's therapeutic potential in macrophages in the context of ALI. The study found reduced miR-23b expression in macrophages within ALI tissue. Intratracheal delivery of miR-23b mimics alleviated ALI by partially inhibiting M1 macrophage activation through the Lpar1-NF-κB pathway. Effective delivery systems are crucial for prolonging miR-23b activity in the lungs, reducing dosage, and minimizing side effects by specifically targeting macrophages. However, current vector systems for nucleic acid delivery, including viral, lipid-based, polymer-based, and peptide-based vectors, face limitations due to eliciting immune responses. Exosomes have garnered significant attention as a leading gene delivery system due to the safety, effectivity and low immunogenicity. We further isolated exosomes from bone marrow-derived mesenchymal stem cells, modified exosomes with mannosylated ligands to enhance the targeted delivery of miR-23b to macrophage. This approach represents a promising novel therapeutic strategy for treating sepsis-induced ALI.
3. Influence of hydrocortisone infusion method on the clinical outcome of patients with septic shock: A systematic review and meta-analysis.
Across 7 studies (n=554), intermittent bolus and continuous hydrocortisone infusion yielded no significant differences in short-term mortality or key secondary outcomes in septic shock. Findings support choosing administration method based on logistics and safety rather than expected efficacy.
Impact: Provides synthesis with trial registration that clarifies an active clinical question, enabling protocol standardization and reducing unwarranted practice variation.
Clinical Implications: Clinicians can select bolus or continuous hydrocortisone based on workflow, monitoring, and adverse-effect profile, focusing attention on timely initiation and appropriate total dosing in septic shock.
Key Findings
- No significant difference in short-term mortality between intermittent bolus and continuous hydrocortisone infusion.
- No differences in ICU/hospital length of stay, vasopressor-free days, hyperglycemia, hypernatremia, or ICU-acquired weakness.
- PROSPERO-registered analysis integrating RCTs and cohort studies (n=554).
Methodological Strengths
- Systematic review and meta-analysis with PROSPERO registration and predefined outcomes.
- Inclusion of both RCTs and cohorts with standardized effect estimates (OR, MD) and 95% CIs.
Limitations
- Modest total sample size and potential heterogeneity across studies.
- Administration protocols and co-interventions may vary, limiting precision of pooled estimates.
Future Directions: Large, adequately powered RCTs with standardized protocols to compare administration methods and evaluate patient-centered outcomes, adverse events, and resource utilization.
BACKGROUND: The effect of the modality of hydrocortisone administration on clinical outcomes in patients with septic shock remains uncertain. This systematic review and meta-analysis evaluate the impact of intermittent bolus and continuous infusion of hydrocortisone on these outcomes.
METHODS: We searched the PubMed, Embase databases, and Cochrane Library for randomized controlled trials (RCTs) and cohort studies published from inception to January 1, 2023. We included studies involving adult patients with septic shock. All authors reported our primary outcome of short-term mortality and clearly compared the clinically relevant secondary outcomes (ICU length of stay, hospital length of stay, vasopressor-free days, hyperglycemia, hypernatremia, and ICU-acquired weakness [ICUAW]) of intermittent bolus and continuous infusion of hydrocortisone. Results were expressed as odds ratio (OR) and mean difference (MD) with accompanying 95% confidence interval (CI). The PROSPERO registration number is CRD42023392160.
RESULTS: Seven studies, including 554 patients, were included. The primary outcome of this meta-analysis showed no statistically significant difference in the short-term mortality between intermittent bolus and continuous infusion groups (OR=1.21, 95% CI: 0.84 to 1.73;
CONCLUSIONS: This meta-analysis indicated no significant difference in short-term mortality between intermittent bolus or continuous hydrocortisone infusion in patients with septic shock. Additionally, the hydrocortisone infusion method was not associated with ICU length of stay, hospital length of stay, vasopressor-free days, hyperglycemia, hypernatremia, or ICUAW.