Risk factors for patient outcomes following penetrating traumatic brain injury

Traumatic brain injury annually affects around 69 million individuals worldwide. Penetrating traumatic brain injuries (pTBI) have a high morbidity and mortality rate. There has been an increase in the incidence of pTBI and an increase in the rate of operative intervention. However, there’s still limited data on how injury details and treatment can impact patient outcomes. This project will explore the factors associated with poor outcomes following pTBI using the United States National Trauma Data Bank (NTDB). Data from the NTDB for patients with a pTBI between 2017 and 2021 were collected. Logistic regression was completed on data from 2017 and 2018 to develop a multivariable model. This model was tested internally and externally. The multivariable regression model for in-hospital complications included the emergency department Glasgow Coma Scale (GCS), penetration greater than 2cm, external ventricular drain (EVD) placement, Injury Severity Score (ISS), and transfusion. This model achieved an accuracy of 89.5% and 87.1% in internal and external testing cohorts, respectively. The model for length of stay included ISS, transfusion, and EVD placement. The complications model included EVD placement, transfusion, and ISS. The accuracy for the length of stay model was around 71%, while it was around 93% for the complication model. As the prevalence of pTBI remains high in the US, it is vital to understand the factors associated with poor prognosis. The severity of the overall injury, patient age, and procedures were the most predictive parameters. Interestingly, the location of fractures and associated complications was not as important in this model.