Risk scoring models fail to predict pulmonary embolism in trauma patients

Abstract

BackgroundWe aimed to identify risk factors and risk scoring models to help identify post-traumatic pulmonary embolisms (PE). MethodsWe performed a retrospective review (2014–2019) of all adult trauma patients admitted to our Level I trauma center that received a CT pulmonary angiogram (CTPA) for a suspected PE. A systematic literature search found eleven risk scoring models, all of which were applied to these patients. Scores of patients with and without PE were compared. ResultsOf the 235 trauma patients that received CTPA, 31 (13%) showed a PE. No risk scoring model had both a sensitivity and specificity above 90%. The Wells Score had the highest area under the curve (0.65). After logistic regression, no risk scoring model variables were independently associated with PE. ConclusionsIn trauma patients with clinically suspected PE, clinical variables and current risk scoring models do not adequately differentiate patients with and without PE.