Daily Sepsis Research Analysis
Three sepsis studies stood out today: a massive machine learning study developed and externally validated a non-laboratory ED screening tool (qSepsis) that outperformed SIRS, qSOFA, and MEWS; a mechanistic study uncovered Shh/Gli1-driven Dio3 reactivation as a driver of muscle metabolic dysfunction in sepsis; and a pre-registered statistical analysis plan for the ANDROMEDA-SHOCK-2 RCT sets a transparent, hierarchical win-ratio approach for phenotype-based, capillary refill time-targeted resuscit
Summary
Three sepsis studies stood out today: a massive machine learning study developed and externally validated a non-laboratory ED screening tool (qSepsis) that outperformed SIRS, qSOFA, and MEWS; a mechanistic study uncovered Shh/Gli1-driven Dio3 reactivation as a driver of muscle metabolic dysfunction in sepsis; and a pre-registered statistical analysis plan for the ANDROMEDA-SHOCK-2 RCT sets a transparent, hierarchical win-ratio approach for phenotype-based, capillary refill time-targeted resuscitation.
Research Themes
- AI-enabled, non-laboratory sepsis screening in the ED
- Thyroid hormone inactivation (Dio3) and Shh/Gli1 signaling in sepsis muscle metabolism
- Transparent trial methodology for phenotype-guided, capillary refill time-targeted resuscitation
Selected Articles
1. Development and validation of a screening tool for sepsis without laboratory results in the emergency department: a machine learning study.
Using non-laboratory ED data, the qSepsis machine-learning model (logistic regression) achieved AUROC 0.862 internally and 0.766 in external validation, outperforming SIRS, qSOFA, and MEWS. Precision-recall analysis (AUPRC 0.213) indicates better case finding than comparators but with low PPV that may increase false alarms in practice.
Impact: Demonstrates scalable, lab-free sepsis screening that generalizes across continents and outperforms widely used bedside scores, addressing a critical time bottleneck in the ED.
Clinical Implications: qSepsis can aid rapid triage when labs are delayed or unavailable, especially in prehospital or high-throughput ED settings. It should augment, not replace, clinician judgment, and be paired with confirmatory testing due to low PPV; prospective implementation studies are needed.
Key Findings
- Logistic regression qSepsis achieved AUROC 0.862 (95% CI 0.855–0.869) internally and 0.766 (0.758–0.774) externally.
- qSepsis outperformed SIRS (AUROC 0.704), qSOFA (0.579), and MEWS (0.600) in external validation.
- AUPRC was 0.213 for qSepsis versus 0.071 (SIRS), 0.096 (qSOFA), and 0.083 (MEWS), but with low PPV indicating potential false alarms.
Methodological Strengths
- Extremely large derivation cohort (n=414,864) and external validation (n=200,089) across countries
- Direct benchmarking against SIRS, qSOFA, and MEWS; multiple ML algorithms compared
Limitations
- Retrospective design with potential labeling bias tied to Sepsis-3 definitions
- Low PPV may increase false alarms; no prospective impact or clinical workflow evaluation
Future Directions: Prospective, multi-site implementation trials to assess clinical impact, alarm burden, and equity; model calibration across diverse EDs; integration with prehospital triage systems.
2. Type 3 deiodinase activation mediated by the Shh/Gli1 axis promotes sepsis-induced metabolic dysregulation in skeletal muscles.
In septic rats and human tissues, Dio3 was upregulated early, with rT3 associating with organ dysfunction. Skeletal-muscle–targeted Dio3 inhibition restored thyroid hormone responsiveness, preserved GLUT4 function and muscle mass, and maintained protein turnover balance. Dio3 reactivation was transcriptionally driven by Shh/Gli1 induced by STAT3, and Shh inhibition improved systemic TH actions.
Impact: Reveals a previously underappreciated mechanism linking Shh/Gli1 signaling to tissue thyroid hormone inactivation and sepsis-associated muscle wasting, highlighting Dio3/Shh as therapeutic targets.
Clinical Implications: While preclinical, targeting Dio3 or upstream Shh/Gli1 signaling could mitigate septic muscle catabolism and anabolic resistance; biomarker-guided approaches using rT3 may stratify patients for future trials.
Key Findings
- Early sepsis upregulated Dio3 in skeletal muscle and lung, with rT3 strongly associated with organ function.
- Muscle-targeted Dio3 inhibition restored tissue TH actions, preserved GLUT4 function, prevented fast-to-slow fiber shift, and maintained protein synthesis–proteolysis balance, preserving muscle mass.
- Dio3 reactivation was transcriptionally driven by Shh/Gli1 induced by STAT3; Shh inhibition (cyclopamine) improved systemic TH responsiveness.
Methodological Strengths
- Multi-system approach with CLP rat model and human biopsy samples
- Mechanistic validation via RNA-seq and ChIP-qPCR alongside functional metabolic readouts
Limitations
- Preclinical study; translational applicability and safety of Dio3/Shh targeting remain untested in humans
- Human tissue sampling details and sample sizes are not specified in the abstract
Future Directions: Phase 0/1 studies to assess target engagement and safety of Dio3/Shh modulation; exploration of rT3 as a stratification biomarker; integration with nutrition and rehabilitation strategies in sepsis.
3. Statistical analysis plan for hemodynamic phenotype-based, capillary refill time-targeted resuscitation in early septic shock: the ANDROMEDA-SHOCK-2 randomized clinical trial.
This pre-registered SAP for a multicenter RCT specifies a hierarchical primary endpoint analyzed via stratified win ratio, with detailed subgroup and sensitivity analyses for capillary refill time-targeted, phenotype-based resuscitation versus standard care in early septic shock. Publishing the SAP before database lock aims to minimize analysis bias and enhance transparency.
Impact: Sets a rigorous, transparent analytic framework (stratified win ratio) for a pivotal resuscitation RCT, enabling credible evaluation of a phenotype-guided, CRT-targeted strategy likely to influence practice if effective.
Clinical Implications: While outcomes are pending, the SAP increases confidence that trial results will be interpretable and less prone to bias; if positive, CRT-targeted, phenotype-based resuscitation may be adopted and studied in implementation.
Key Findings
- Hierarchical primary outcome will be analyzed using a stratified win ratio to compare CRT-targeted, phenotype-based resuscitation versus standard care.
- Prespecified secondary/tertiary outcomes, subgroup analyses, and sensitivity analyses are detailed to minimize analytic flexibility.
- Comprehensive presentation plan includes mock tables, baseline characteristics, and treatment effect reporting prior to database lock.
Methodological Strengths
- Pre-registered statistical analysis plan with hierarchical win ratio for clinically meaningful composite assessment
- International, multicenter RCT design with prespecified subgroup and sensitivity analyses
Limitations
- No clinical outcomes reported yet; impact depends on forthcoming trial results
- Sample size and operational nuances not detailed in the abstract
Future Directions: Execute the RCT per SAP, then assess generalizability and implementation of CRT-targeted, phenotype-based resuscitation; consider patient-centered outcomes within a hierarchical framework.