Skip to main content

Circulating endothelial signatures correlate with worse outcomes in COVID-19, respiratory failure and ARDS.

Critical care (London, England)2025-10-15PubMed
Total: 77.0Innovation: 8Impact: 0Rigor: 0Citation: 0

Summary

Using unsupervised deconvolution of blood transcriptomes, the authors quantified circulating endothelial cell signatures that were higher at baseline in non-survivors and independently associated with 28-day mortality. Elevated signatures also tracked worse respiratory trajectories, supporting a non-invasive biomarker of endothelial injury relevant to ARDS pathobiology.

Key Findings

  • Baseline ECS% was higher in non-survivors vs survivors in adults with COVID-19 (2.9% vs 2.7%, n=932, p<0.001).
  • Each 1% increase in baseline ECS% independently increased mortality risk (adjusted OR 1.36, 95% CI 1.03–1.79).
  • In ventilated pediatric patients, day 0 ECS% was higher in non-survivors (2.8% vs 2.6%, n=244, p<0.05).
  • Higher ECS% associated with worse respiratory trajectories across oxygen-requirement categories (p<0.001).

Clinical Implications

ECS% could inform early risk stratification and trial enrichment for endothelial-targeted therapies in ARDS and severe COVID-19, enabling non-invasive monitoring of vascular injury.

Why It Matters

Introduces a scalable transcriptomic metric of endothelial damage with prognostic value across pediatric and adult cohorts, bridging mechanism and risk stratification in ARDS-related respiratory failure.

Limitations

  • Observational design limits causal inference and residual confounding may persist.
  • Deconvolution-based ECS% depends on reference signatures and may vary across platforms and sample processing.

Future Directions

Prospective validation with predefined thresholds, integration into clinical decision-support, and biomarker-driven trials testing endothelial-targeted therapies in ARDS.

Study Information

Study Type
Cohort
Research Domain
Prognosis
Evidence Level
III - Prospective/retrospective cohort analyses showing prognostic associations
Study Design
OTHER