Quantification of traumatic hemomediastinum using transesophageal echocardiography: impact on patient management

STUDY OBJECTIVES: To determine whether the quantitative evaluation of hemomediastinum using transesophageal echocardiography (TEE) is predictive of the presence of a traumatic disruption of the thoracic aorta (TDA) or its branches in patients who have sustained severe blunt chest trauma. DESIGN: Retrospective study. SETTING: ICU of a tertiary referral teaching hospital. PATIENTS: Forty-one patients sustaining severe blunt chest trauma (32 men, nine women; mean age, 43+/-16 years; mean Injury Severity Score, 39+/-22) who underwent a TEE study were divided into two groups, patients with (group TDA+, n=15) or without (group TDA-, n=26) major vascular injury diagnosed using an alternative method such as aortography, surgery, or necropsy. The control group included 41 age- and sex-matched patients with an unremarkable TEE study performed to rule out an intracardiac source of emboli. INTERVENTIONS: The presence of hemomediastinum was quantitatively assessed by measuring the distances between the esophageal scope and anteromedial aortic wall (distance 1), and between the posterolateral aortic wall and left visceral pleura (distance 2) at the level of the aortic isthmus. An observer who was unaware of both medical history and final diagnosis measured the distances. MEASUREMENTS AND RESULTS: In group TDA+, TEE demonstrated aortic injuries in 13 patients, revealed an isolated hemomediastinum in one patient (ruptured intercostal arteries), and was unremarkable in the remaining patient, who sustained a disrupted right subclavian artery. No associated major vessel injuries were diagnosed in the group TDA- (normal aortograms). When compared to the control group, mean distances were greater in patients with chest trauma (distance 1=5.5+/-4.4 mm vs 2.7+/-0.8 mm, p=0.001; distance 2=3.8+/-5.0 mm vs 1.2+/-0.3 mm, p=0.02). The corresponding distances were even greater in group TDA+ when compared with group TDA- (distance 1=8.6+/-5.9 mm vs 3.7+/-1.5 mm, and distance 2=7.1+/-7.0 mm vs 2.0+/-1.7; for both differences, p<0.01). A threshold value of 5.5 mm for distance 1 or 6.6 mm for distance 2 had a sensitivity of 80%, a specificity of 92%, a positive and negative predictive value of 86% and 89%, respectively, for the diagnosis of underlying major vascular injury. CONCLUSIONS: TEE allows quantitative assessment of traumatic hemomediastinum. The presence of a large hemomediastinum requires further evaluation by aortography, even if the thoracic aorta appears normal during the TEE examination, in order to rule out an underlying major vascular injury which may be outside the field of view of the echocardiographer.