Predictive Value of Thromboelastography (TEG) for Neurological Presentation and Outcome After Aneurysmal Subarachnoid Hemorrhage using Machine Learning
Predictive Value of Thromboelastography (TEG) for Neurological Presentation and Outcome After Aneurysmal Subarachnoid Hemorrhage Using Machine Learning
Introduction: Thromboelastography (TEG) is a routine, blood test that measures the coagulative properties of a blood sample. Previous studies have demonstrated its efficacy in predicting coagulation profiles in patients with aneurysmal subarachnoid hemorrhage (aSAH), however, its impact on the neurological presentation and outcome remain understudied.
Methods: A retrospective analysis of all patients with an aSAH who presented to a large quaternary center from 7/1/20 to 9/30/22 was performed. All patients with a TEG with and without platelet mapping performed on initial hospital presentation wre included. Outcomes measured were Hunt-Hess grade (HH), modified Fisher Grade (mFG), Glasgow Coma Scale (GCS), and discharge modified Rankin Scale (D-mRS). A probability-boosted multivariate linear regression analysis was performed to predict outcomes using values in TEG. For D-mRS, a sensitivity analysis was conducted to identify the added accuracy of TEG to clinical parameters.
Results: During the study period 139 aSAH patients met inclusion criteria. The mean age was was 60.93 (SD = 13.39). TEG was found to accurately predict HH (R = 0.56, p < 0.001), mFG (R = 0.44, p = 0.002), GCS (0.636, p < 0.001). PLM6s AA and ADP Inhibition were found to be most correlative with outcomes. In the prediction of discharge mRS, the TEG and clinical model were the most predictive (R = 0.748, p < 0.001) compared to the TEG-only and clinical-only models, with a 59.35% and 61.67% reduction in sum-of-squared error (SSE).
Conclusion : In this machine learning analysis, the utility of TEG in predicting clinical presentation-based parameters, such as HH, mFG, and GCS, was determined. TEG was found to add value in predicting discharge mRS. This data presents a powerful argument for the increased utilization of TEG in the management of aSAH patients to better monitor their progression and outcome. Further studies in the utility of TEG, specifically in long-term outcomes are warranted.