Prognostic value of admission ROTEM in trauma: enhancing 30-day all-cause mortality prediction using machine learning

Area of Science:

• Trauma resuscitation and critical care medicine.
• Coagulation diagnostics and monitoring.
• Biomarker discovery in emergency medicine.

Background:

• Hemorrhage is a primary cause of trauma mortality.
• Early coagulation markers are underutilized for predicting long-term trauma outcomes.
• Rotational thromboelastometry (ROTEM) may offer prognostic value.

Purpose of the Study:

• To evaluate the predictive capability of admission ROTEM for 30-day all-cause mortality in trauma patients.
• To compare ROTEM's predictive performance against established trauma scores.
• To determine if integrating ROTEM improves existing trauma scoring systems.

Main Methods:

• Retrospective cohort study of 1,498 adult trauma patients.
• Rotational thromboelastometry (ROTEM) tests (EXTEM, INTEM, FIBTEM) performed on admission.
• Machine learning models developed using ROTEM alone, conventional trauma scores, and combinations for mortality prediction.

Main Results:

• ROTEM alone demonstrated strong predictive performance for 30-day mortality (AUROC 0.80), comparable to RTS and NISS.
• Integration of ROTEM significantly enhanced the specificity of all tested trauma scores.
• Predictive variables included clotting time, clot firmness time, and fibrinolysis indices; performance was lower in female patients.

Conclusions:

• Admission ROTEM is a valuable tool for predicting 30-day trauma mortality.
• Combining ROTEM with existing trauma scores substantially improves predictive accuracy.
• Viscoelastic data from ROTEM provides crucial prognostic information for trauma outcome assessment.