Abstract
Abstract
Thoracic trauma is a major contributor to morbidity in severely injured patients and is frequently associated with prolonged mechanical ventilation, tracheostomy, and extended intensive care unit (ICU) treatment. Early identification of patients at risk for unfavorable respiratory-support trajectories remains challenging. Small extracellular vesicles (sEVs) are rapidly released in response to cellular activation and injury and may provide biological information complementary to established clinical parameters. In our study we analyzed CD9⁺CD14⁺ and CD9⁺CD61⁺ sEV subpopulations in 45 patients with thoracic trauma. sEV levels were quantified within 24 h after admission and evaluated in relation to clinically relevant respiratory-support trajectories, including the predefined primary respiratory-support endpoints of tracheostomy and prolonged mechanical ventilation (> 5 days). Prolonged ICU stay (> 5 days) was analyzed as a supportive endpoint reflecting overall intensive-care burden. Predictive performance was assessed using repeated cross-validated receiver operating characteristic (ROC) analysis and compared with predefined clinical injury, laboratory, and oxygenation-based models. Residual analyses were performed to determine whether EV-associated information extended beyond established thoracic injury severity scores. CD9⁺CD14⁺ sEVs represented the dominant EV signal and were significantly elevated in patients requiring tracheostomy (3.84-fold,
p
= 0.009), prolonged mechanical ventilation > 5 days (4.88-fold,
p
= 0.003), and prolonged ICU treatment > 5 days (5.62-fold,
p
= 0.003). Within this exploratory cohort, CD9⁺CD14⁺ sEVs demonstrated performance comparable to the clinical injury model and consistently performed significantly better than laboratory- and oxygenation-based approaches across multiple respiratory-support trajectory endpoints. Sensitivity analyses using stricter thresholds (> 10 days) yielded similar findings. Residual analyses further demonstrated that patients with higher-than-expected CD9⁺CD14⁺ levels relative to AIS Chest and TTSS exhibited increased rates of tracheostomy and prolonged mechanical ventilation, suggesting that EV-associated information is not fully explained by anatomical injury severity alone. We concluded that early CD9⁺CD14⁺ sEV levels are associated with unfavorable respiratory-support trajectories following thoracic trauma. While established injury scores primarily describe the extent of anatomical damage, EV-associated biomarkers may provide complementary information regarding the biological response to injury. These findings support further investigation of EV-based biomarkers as tools for early risk stratification in thoracic trauma and warrant validation in larger independent cohorts.