A Machine Learning Approach for the Prediction of Traumatic Brain Injury Induced Coagulopathy

Abstract

Background: Traumatic brain injury-induced coagulopathy (TBI-IC), is a disease with poor prognosis and increased mortality rate.Objectives: Our study aimed to identify predictors as well as develop machine learning (ML) models to predict the risk of coagulopathy in this population.Methods: ML models were developed and validated based on two public databases named Medical Information Mart for Intensive Care (MIMIC)-IV and the eICU Collaborative Research Database (eICU-CRD). Candidate predictors, including demographics, family history, comorbidities, vital signs, laboratory findings, injury type, therapy strategy and scoring system were included. Models were compared on area under the curve (AUC), accuracy, sensitivity, specificity, positive and negative predictive values, and decision curve analysis (DCA) curve.Results: Of 999 patients in MIMIC-IV included in the final cohort, a total of 493 (49.35%) patients developed coagulopathy following TBI. Recursive feature elimination (RFE) selected 15 variables, including international normalized ratio (INR), prothrombin time (PT), sepsis related organ failure assessment (SOFA), activated partial thromboplastin time (APTT), platelet (PLT), hematocrit (HCT), red blood cell (RBC), hemoglobin (HGB), blood urea nitrogen (BUN), red blood cell volume distribution width (RDW), creatinine (CRE), congestive heart failure, myocardial infarction, sodium, and blood transfusion. The external validation in eICU-CRD demonstrated that adapting boosting (Ada) model had the highest AUC of 0.924 (95% CI: 0.902–0.943). Furthermore, in the DCA curve, the Ada model and the extreme Gradient Boosting (XGB) model had relatively higher net benefits (ie, the correct classification of coagulopathy considering a trade-off between false- negatives and false-positives)—over other models across a range of threshold probability values.Conclusions: The ML models, as indicated by our study, can be used to predict the incidence of TBI-IC in the intensive care unit (ICU).

A metabolomic and proteomic analysis of pathologic hypercoagulability in traumatic brain injury patients after dura violation

“…Beyond fibrinolysis, ornithine has been described as a marker of endotheliopathy, competing with arginine as a substrate with nitric oxide synthase; the end result in the setting of TBI is that ornithine induces oxidative stress and decreases blood-brain barrier integrity 48 . Creatinine, another important metabolite in the arginine and proline pathway, significantly increased after dura violation in our population; in a previous study aimed to identify predictors of coagulopathy in TBI and develop machine learning (ML) models to predict the risk of coagulopathy, Yang et al found that creatinine was a core component in contributing to prediction of coagulopathy in TBI patients 1 …”

“…48 Creatinine, another important metabolite in the arginine and proline pathway, significantly increased after dura violation in our population; in a previous study aimed to identify predictors of coagulopathy in TBI and develop machine learning (ML) models to predict the risk of coagulopathy, Yang et al found that creatinine was a core component in contributing to prediction of coagulopathy in TBI patients. 1 To complement the metabolomic data which characterizes a shift toward platelet hyperactivity and fibrinolysis shutdown, our proteomic data demonstrated key increases in proteins related to platelet activation, degranulation, and aggregation which correlated to thrombelastographic measurements of platelet activity. Examining the mechanisms and natural history of TBI has been recently empowered by proteomic data, which has further characterized the neuroinflammatory response to TBI.…”

“…The coagulopathy of traumatic brain injury (TBI) remains poorly understood. While coagulopathy occurs in nearly half of patients who incur a head injury, obstacles remain in predicting these patients and understanding the derangements in hemostatic capacity after TBI, in particular if also accompanied by multiple trauma or hemorrhagic shock 1 . Contradictory descriptions in the literature highlight the distinction between systemic and local coagulation environment, with characterization of hypocoagulopathy and progression of intracranial hemorrhage to hypercoagulability due to systemic immune response, inflammation, and platelet hyperactivity 2–4 .…”

“…While coagulopathy occurs in nearly half of patients who incur a head injury, obstacles remain in predicting these patients and understanding the derangements in hemostatic capacity after TBI, in particular if also accompanied by multiple trauma or hemorrhagic shock. 1 Contradictory descriptions in the literature highlight the distinction between systemic and local coagulation environment, with characterization of hypocoagulopathy and progression of intracranial hemorrhage to hypercoagulability due to systemic immune response, inflammation, and platelet hyperactivity. [2][3][4] This latter pathologic hypercoagulability has been linked to higher rates of macrothrombotic morbidity including deep venous thrombosis and microthrombotic morbidity like acute lung and kidney injury.…”

Age, admission platelet count, and mortality in severe isolated traumatic brain injury: A retrospective cohort study

“…[18][19][20][21] Trauma-induced platelet dysfunction is well described. 7,22,23 In addition, although more recent work has tended to focus on the documentation and assessment of trauma-associated platelet dysfunction, plateletassociated disturbances-in function and numbersappear to be common and specific to TBI [24][25][26][27][28][29][30][31][32][33][34] but overall are frustratingly poorly correlated with clinical observations of bleeding. 35 However, despite the marked excess risk for TBI and TBI mortality with aging, 3,4 few studies have attempted to isolate the potential independent associations of age with platelet dysfunction in TBI, tending instead to bin such effects with medication and/or comorbidities and/or to bin TBI with other severe trauma.…”