Subacute Trends in Metabolic and Hematological Markers Days After Severe Traumatic Brain Injury Predict Long-Term Outcomes

Shawn R Eagle; Maxwell Wang; Regan M Shanahan; Anna L Slingerland; Shovan Bhatia; Michael R Kann; Tyler Augi; Ava Puccio; David O Okonkwo

doi:10.1227/neu.0000000000003863

Subacute Trends in Metabolic and Hematological Markers Days After Severe Traumatic Brain Injury Predict Long-Term Outcomes

Neurosurgery. 2025 Nov 25. doi: 10.1227/neu.0000000000003863. Online ahead of print.

Authors

Shawn R Eagle¹, Maxwell Wang², Regan M Shanahan¹, Anna L Slingerland², Shovan Bhatia^{1

2}, Michael R Kann¹, Tyler Augi², Ava Puccio¹, David O Okonkwo¹

Affiliations

¹ Department of Neurological Surgery, University of Pittsburgh, Pittsburgh , Pennsylvania , USA.
² Department of Medicine, University of Pittsburgh, Pittsburgh , Pennsylvania , USA.

PMID: 41288377
DOI: 10.1227/neu.0000000000003863

Abstract

Background and objectives: Prognostic models for severe traumatic brain injury (TBI) patients do not perform well enough to be clinically useful, possibly because included predictors for the most well-established models are from hospital presentation only. We sought to investigate fluctuations in metabolic and hematological markers over the subacute time period (ie, up to 14 days postinjury) and associate these changes with functional outcome at 6 months postinjury.

Methods: This cohort study included prospectively collected patients with severe TBI treated at a single level I trauma center (n = 315). Measured metrics included basic vitals (eg, heart rate), arterial blood gases, basic metabolic panel (eg, glucose, sodium), blood counts (eg, hemoglobin, platelets), and endocrine panels (eg, cortisol) taken periodically from admission to 14 days postinjury. Functional outcome was defined as Glasgow Outcome Scale-Extended score at 6 months. Random forest regression trees were used to associate these markers to outcome. Model performance was compared to using International Mission on Prognosis and Analysis of Clinical Trials metrics collected on hospital admission.

Results: The International Mission on Prognosis and Analysis of Clinical Trials model predicted 6-month Glasgow Outcome Scale-Extended with a coefficient of determination (R 2 ; ie, variance accounted for) of approximately 19%. Incorporation of subacute biomarker trends improved model performance by 8% (R 2 = 27%). Fluctuations in glucose, sodium, and platelets over the first 2 weeks postadmission formed reliable patterns across participants and correlated significantly with eventual outcomes (correlation = -0.3 to -0.4 for sodium and glucose, correlation = +0.2 to +0.3 for platelets). In a subset of participants, fluctuations in glucose and platelets were associated to cortisol levels taken within 48 hours of admission.

Conclusion: Fluctuations in several hematological and metabolic markers over the days to weeks after severe TBI form consistent patterns that can be linked to endocrinological disruptions and 6-month functional outcome.

Keywords: Endocrine; IMPACT; Metabolic; Prognostication; TBI.