Multiplane transesophageal echocardiography (TEE) and helical computed tomography (CT) of the chest have been validated separately against aortography for the diagnosis of acute traumatic aortic injuries (ATAI). However, their respective diagnostic accuracy in identifying blunt traumatic cardiovascular lesions has not been compared.
During a 3-yr period, 110 consecutive patients with severe blunt chest trauma (age: 41 +/- 17 yr; injury severity score: 34 +/- 14) prospectively underwent TEE and chest CT as part of their initial evaluation. Results of both imaging methods were interpreted independently by experienced investigators and subsequently compared. All cases of subadventitial acute traumatic aortic injury were surgically confirmed.
Seventeen patients had vascular injury and 11 had cardiac lesions. TEE and CT identified all subadventitial disruptions involving the aortic isthmus (n = 10) or the ascending aorta (n = 1) that necessitated surgical repair. In contrast, CT only depicted one disruption of the innominate artery. TEE detected injuries involving the intimal or medial layer, or both, of the aortic isthmus in four patients with apparently normal CT results who underwent successful conservative treatment. All cardiac injuries but two were identified only by TEE.
In patients with severe blunt chest trauma, TEE and CT have similar diagnostic accuracy for the identification of surgical acute traumatic aortic injuy. TEE also allows the diagnosis of associated cardiac injuries and is more sensitive than CT for the identification of intimal or medial lesions of the thoracic aorta.
BLUNT traumatic cardiovascular injuries after severe chest trauma must be diagnosed and repaired rapidly to improve the prognosis of patients who survive long enough to reach the hospital. 1A recent multicenter trial 2that included 274 patients with acute thoracic aortic injuries (ATAIs) reported an overall mortality rate of 31%, with 63% of deaths attributable to traumatic aortic disruptions. In addition, the frequency of cardiac injuries is presumably underscored in patients with severe blunt chest trauma. 3
Traditionally, angiography is the accepted reference imaging method for the diagnosis of ATAI. 4Recently, transesophageal echocardiography (TEE) 5–13and contrast-enhanced computed tomography (CT) of the chest 14–17have been proposed as alternative noninvasive imaging methods for the routine evaluation of patients at high risk of ATAI. Recent development of multiplane high-frequency TEE probes and helical CT have overcome previous limitations of these imaging techniques and potentially improved their diagnostic capability. Although independently validated against angiography, the results of these imaging methods have not been prospectively compared in patients with suspected traumatic cardiovascular injuries.
Accordingly, we conducted a prospective study to determine the diagnostic accuracy of multiplane TEE and helical CT for the detection of traumatic cardiovascular injuries in patients with severe blunt chest trauma.
During a 3-yr-period, consecutive patients at high risk of blunt traumatic cardiovascular injury who were referred to our institution underwent multiplane TEE and helical CT of the chest as part of their initial evaluation. Imaging methods were performed in random order, according to the availability of the medical staff, within 24 h of admission. If multiplane TEE or helical CT was contraindicated, 5,18only the alternative imaging method was performed. Angiography was performed when (1) traumatic injury to aortic branches was suspected, 5(2) multiplane TEE and helical CT both were inconclusive or yielded discrepant results, or (3) a large hemomediastinum without apparent ATAI was identified. 19Informed consent was not necessary to obtain in this clinical setting; the study protocol was approved by our Institutional Review Board (Dupuytren University Hospital, Limoges, France).
High risk for blunt traumatic cardiovascular injuries was defined by the presence of at least one of the following criteria 5,18: (1) history of violent deceleration trauma (head-on collision or fall from a height > 3 m); (2) unrestrained patient ejection from the vehicle or death in the accident; (3) pedestrian or bicyclist having been struck by an automobile; (4) chest trauma necessitating mechanical ventilation; (5) presence of acute coarctation syndrome or an external sign of direct chest injury, (6) associated traumatic injury reflecting violent impact to the body ( e.g. , diaphragmatic rupture, mesenteric tear); (7) unexplained hypotension or shock; and (8) enlarged mediastinum as seen on admission chest radiograph. Because ATAI and cardiac injury may occur even if a chest radiograph is unremarkable, 9,17patients with normal mediastinal contours were also enrolled in the study if one or more of the aforementioned criteria were present.
Multiplane Transesophageal Echocardiography
Transesophageal echocardiography was performed as previously described 8using a SONOS 5500 imaging system (Hewlett-Packard, Andover, MA) equipped with a 7-MHz multiplane esophageal transducer. Blood pressure, electrocardiography, and arterial oxygen saturation were closely monitored. In patients undergoing spontaneous ventilation, the esophageal probe was inserted after pharyngeal anesthesia and mild sedation (midazolam 0.05–0.1 mg/kg). Patients undergoing ventilation received an intravenous injection of midazolam (0.15 mg/kg) and vecuronium bromide (0.1 mg/kg) as necessary. The nasogastric tube was removed. Particular attention was directed toward the evaluation of the different anatomic segments of the thoracic aorta, using two-dimensional echocardiography (transverse and longitudinal planes) and color Doppler. A complete examination of the heart was also routinely performed. TEE was recorded on videotapes for further interpretation by an investigator experienced in echocardiography and who was unaware of the medical history of the patient and the results of helical CT.
The TEE diagnosis of ATAI relied on the presence of (1) a (dilated) aortic isthmus with abnormal contour, (2) an acute false aneurysm formation or an intraluminal medial flap associated with a subadventitial disruption, or both, 8(4) echocardiographic signs consistent with a traumatic aortic dissection 8; (5) a mobile image appended to the aortic wall consistent with a intimal tear or a mural thrombus 8; and (6) a crescentic or circumferential thickening of the aortic wall (with or without echolucent areas and displaced intimal boundary), reflecting the presence of an intramural hematoma. 20,21The presence of a traumatic hemomediastinum was defined by an increased distance (> 3 mm) between the esophageal probe and the anteromedial wall of the aortic isthmus or the presence of blood between the posterolateral aortic wall and the left visceral pleura. 22A left-sided hemothorax was identified as the presence of fluid between the left lung and the thoracic wall. 23
In addition, the following TEE findings were recorded: presence of a regional wall motion abnormality without transmural myocardial infarction detected on electrocardiogram and consistent with a myocardial contusion, 24acute valvular regurgitation attributable to a traumatic injury, hemopericardium or mediastinal hematoma contiguous to the heart (eventually associated with a tamponade physiology).
Helical Chest Computed Tomography
If indicated, an unenhanced CT scan of the head and spine was performed before chest CT. Axial scanning was performed using a HiSpeed Advantage helical scanner (General Electric, Milwaukee, WI) and using the Advantage Windows 1.2 software (General Electric) for CT angiography reconstruction. In each patient, the following protocol was followed. The patient’s arms were raised overhead when possible, and a digital localization scout view of the chest was obtained during voluntary apnea (or ventilator pause). Nonionic contrast media (300 mg/ml iodine; Xenetix®, Guerbet, France) was injected through an arm vein (2.5 ml/kg) using a power injector at 2.0–2.5 ml/s. Continuous 5-mm helical scans were obtained from the mandibular angle to the diaphragm (1:1.5 pitch) at 120 kv and 290 mA. A retrospective two-dimensional reconstruction was performed every 3.0 mm. Images were recorded on 35.0 × 42.5-cm (14 × 17-in) radiographic films for subsequent analysis. Duration of helical acquisitions and computer reconstruction usually were 25–30 s and 12–15 min, respectively. Patients were typically immobilized in the radiographic suite for approximately 20 min.
All CT scans were independently interpreted by an experienced radiologist who was unaware of the medical history and had no access to the results of the multiplane TEE. Direct CT findings of ATAI included 18: (1) polypoid (clot) or linear (medial flap) intraluminal areas of low attenuation; (2) false aneurysm formation; (3) irregularity of the aortic wall or contour; (4) pseudocoarctation, (5) intramural hematoma; or (6) aortic dissection. In addition, signs on the CT scan associated with hemomediastinum, left hemothorax, hemopericardium, and mediastinal hematoma contiguous to the heart were also recorded.
Data Analysis and Definitions
Acute thoracic aortic injuries were classified as subadventitial 8or superficial, 5,8involving the intimal layer ( e.g. , intimal tear, mural thrombus) or the medial layer ( e.g. , intramural hematoma) of the thoracic aorta. This distinction is mainly based on distinct management of subadventitial (surgical) and more superficial injuries (conservative treatment) of the aortic isthmus. 5,8,20
Diagnosis of cardiovascular injury using either multiplane TEE or spiral CT and confirmed by angiography, surgery, or necropsy was considered as a true-positive result. When multiplane TEE and helical CT yielded discrepant results, the following protocol was followed. In the presence of a suspected subadventitial ATAI, angiography was performed. When a superficial ATAI was suspected, a serial follow-up using the same noninvasive diagnostic method was performed. An abnormal aortogram or diagnosis confirmed during this follow-up period was interpreted as a false-negative result of the alternative noninvasive diagnostic technique. Conversely, a normal aortogram or the identification of radiologic 18or echocardiographic 25findings associated with imaging artifacts corresponded to a false-positive result of the imaging method erroneously interpreted as abnormal. Finally, a true-negative result was defined by the absence of cardiovascular injury shown by multiplane TEE and helical CT in patients with unremarkable courses. Accordingly, the follow-up of all patients was obtained by calling the respective attending physicians after patient discharge from the hospital.
Descriptive statistics were applied to patient characteristics and diagnostic findings during multiplane TEE and helical CT. Results are expressed as mean ± standard deviation or percentage. The diagnostic capability of multiplane TEE and helical CT of the chest for the identification of ATAI was assessed by calculating their respective sensitivity, specificity, and positive and negative predictive values, with corresponding 95% confidence intervals.
During the study period, 285 patients were admitted to our Intensive Care Unit because of severe blunt chest trauma. Among them, 110 (mean age: 41 ± 17 yr; 88 men and 22 women; injury severity score: 34 ± 14) were considered to be at high risk of traumatic cardiovascular injury (fig. 1). A single noninvasive imaging method was performed for 15 patients (14%). Multiplane TEE was contraindicated in three patients (3%) because of severe facial trauma, unstable cervical spine injuries, or pneumoperitoneum, and the esophageal probe could not be inserted in another patient. Helical CT was not performed in 10 patients (9%) who were hemodynamically unstable and could not be transported to the radiology suite (fig. 1). All TEE studies were judged as technically adequate for interpretation, whereas helical CT was not deemed proper for analysis in one patient. Finally, 95 patients underwent both procedures (fig. 1). Angiography was performed in 11 patients (10%) for suspected injury to aortic branches (n = 4), inconclusive studies or discrepant results (n = 5), or large traumatic hemomediastinum without evidence of associated ATAI (n = 2). Twelve patients (11%) died during their stay in the intensive care unit. One died rapidly in the emergency room of exsanguination secondary to traumatic aortic disruption diagnosed using multiplane TEE.
A traumatic arterial injury was diagnosed in 17 patients (15.5%). Among them, 16 had ATAI, and the one other patient presented with a laceration of the innominate artery. With the exception of one subadventitial disruption of the ascending aorta, all ATAI involved the aortic isthmus. ATAI included subadventitial disruptions (n = 11), traumatic dissection (n = 1), and superficial aortic lesions (intimal tear [n = 1], mobile thrombi, [n = 2], intramural hematoma [n = 1]). Patients with superficial ATAI underwent conservative treatment. With the exception of two patients who died before surgery, the other patients with ATAI underwent surgical repair (fig. 1).
A traumatic cardiac lesion was diagnosed in 11 patients (10%): uncomplicated hemopericardium (n = 1), large anterior mediastinal hematoma contiguous to the heart (n = 3), acute severe aortic regurgitation as a result of valvular disruption (n = 1), thrombus entrapped in a patent foramen ovale (n = 1), and severe myocardial contusions (right ventricle, n = 2; left ventricle, n = 3). The patient with the aortic valve disruption also had a subadventitial disruption of the aortic isthmus. 26Surgery was performed in 3 of 10 patients with hemodynamic compromise. No evidence of traumatic cardiovascular lesion was found in the other 83 patients who did not undergo additional work-up (fig. 1). Among them, all patients who were discharged from the hospital were alive, with no symptoms attributable to a potential cardiovascular condition during follow-up (17 ± 11 months).
Diagnostic Capabilities of Imaging Methods
Diagnostic accuracy of multiplane TEE and helical CT for the identification of ATAI is detailed in table 1. Both imaging methods were reliable for the identification of surgical ATAI, whereas multiplane TEE was more sensitive for the detection of traumatic lesions involving the intimal or medial layers of the thoracic aorta (table 2). All subadventitial disruptions of the aortic isthmus or confined to the ascending aorta, and the single traumatic dissection of the descending thoracic aorta, were accurately detected by both imaging procedures. Multiplane TEE missed the laceration of the innominate artery that was depicted by helical CT scan (figs. 2A and B) and confirmed during surgery. Helical CT failed to detect intimal injuries in three patients, whereas multiplane TEE clearly depicted intimal tear or mural thrombi of the aortic isthmus (figs. 2C–F). TEE follow-up showed progressive healing of these aortic changes. Similarly, multiplane TEE identified an intramural hematoma confined to the aortic isthmus in a patient whose CT scan was unremarkable and who died of massive pulmonary embolism (table 2).
Multiplane TEE allowed rapid diagnosis of cardiac lesions in 11 patients (fig. 3), whereas helical CT only depicted the presence of an uncomplicated hemopericardium and retrosternal hematoma. Both imaging methods identified a traumatic hemomediastinum in a similar proportion of patients, whereas multiplane TEE tended to be more sensitive for the detection of mild hemothorax (table 2).
In the 15 patients assessed using a single imaging method, multiplane TEE identified two subadventitial disruptions of the aortic isthmus and four traumatic cardiac injuries that resulted in hemodynamic compromise.
This study is the first to prospectively compare the diagnostic accuracy of multiplane TEE and helical CT for the detection of ATAI and cardiac injuries in patients with severe blunt chest trauma. Both imaging techniques were reliable for the identification of subadventitial traumatic injuries to the aortic isthmus or the ascending aorta, which potentially can lead to lethal adventitial rupture. Helical CT also allowed the diagnosis of one supraaortic vascular lesion, whereas multiplane TEE seemed to be more sensitive for the detection of superficial aortic injuries (i.e. , involving the aortic intimal or medial layers of the thoracic aorta) and blunt cardiac trauma (table 2).
The 25% incidence of traumatic cardiovascular injuries in our high-risk study population (27 of 110 patients) emphasizes the critical importance of the choice of the first-line imaging method routinely used for the assessment of patients with severe blunt chest trauma. Aortography has long been recognized as the only reliable screening imaging method for the assessment of patients with suspected ATAI. 4,27Recently, alternative noninvasive imaging methods, such as multiplane TEE 6–8,12,13and the new-generation CT scanning, 14–16have been validated in this setting. Their numerous advantages over angiography 4,5allow a more liberal use, potentially decreasing the number of undiagnosed traumatic cardiovascular injuries. 9,11,12Accordingly, indications of angiography become frequently restricted to contraindicated or inconclusive noninvasive tests, and suspected injuries to aortic branches, as in the current series.
Diagnostic Accuracy of Multiplane Transesophageal Echocardiography and Helical Computed Tomography
Both imaging methods had a similar diagnostic accuracy for the identification of subadventitial disruptions of the aortic isthmus that necessitated surgery (table 2). Compared with aortography, TEE sensitivity and specificity have been reported to range from 57–100% and 87–100%, respectively. 6–8,12,13The higher diagnostic accuracy observed in our series (table 1) is presumably attributed to the routine use of multiplane TEE in the setting of severe blunt chest trauma by experienced operators. This probably overcame substantial limitations related to either technical characteristics of monoplane esophageal probes 6,7or an inadequate learning curve. 28The current study also confirmed the reliability of helical CT for the diagnosis of subadventitial ATAI. In this clinical setting, 14–17this imaging method had a 95–100% sensitivity and approximately an 82–100% specificity. In contrast, authors 29have emphasized the substantial limitations of conventional CT for the detection of traumatic injuries to the thoracic aorta or its branches.
In the current series, multiplane TEE missed a laceration of the innominate artery, which was correctly identified by helical CT (fig. 2). Although multiplane TEE allows the visualization of the initial portion of supraaortic arteries, 30aortic branch lesions are usually out of the scope of vision of the esophageal probe. 31The diagnosis, using helical CT, of traumatic injuries of supraaortic arteries has only been sporadically reported. 14,17,32However, distal arterial injuries 14or intracranial vascular lesions 32may not be detected by this imaging method. Although duplex ultrasound may be an alternative noninvasive approach, 33selective angiography is necessary when traumatic lesions of aortic branches are suspected. 8,14
Multiplane TEE identified a superficial ATAI that was missed by helical CT in four patients (table 2). Using reconstruction algorithms of helical CT scans, localized aortic lacerations and small pseudoaneurysm formations may be obscured by volume averaging with the normal aortic wall or adjacent aortic lumen. 17,18In these cases, retrospective reconstruction of additional axial CT images may be useful to detect smaller aortic lesions. 18The higher spatial resolution of high-frequency esophageal probes (∼2 mm) accounts for the greater sensitivity of multiplane TEE for the diagnosis of traumatic intimal tears, mural thrombi, or small aortic wall lesions (fig. 2).
Although less frequently reported, 34blunt traumatic cardiac injuries may be life-threatening or lead to severe complications when unrecognized (fig. 3). 11,12In the current study, helical CT detected only two of the blunt traumatic cardiac injuries diagnosed using multiplane TEE (table 2). Current helical CT scanners are not suited to provide accurate examination of the heart. Newly developed, ultrafast CT provides valuable information on global and regional cardiac function, 35,36but these scanners are not widely available and require extensive postprocessing.
Although patients at high risk of blunt traumatic cardiovascular injury can be objectively identified, 5,18the choice of the optimal screening imaging method is controversial. 4,37Because of similar diagnostic accuracy, multiplane TEE and helical CT can interchangeably be used (e.g ., contraindication) for the assessment of hemodynamically stable patients with suspected subadventitial ATAI (fig. 4). In these cases, chest radiography at admission typically shows an enlarged mediastinum, with associated radiographic findings reflecting the presence of a hemomediastinum or acute false aneurysm formation. 9,18Associated chest injuries noted on the chest radiograph at admission (e.g. , flail chest wall, severe pulmonary contusion) or the suspicion of associated extrathoracic lesions (e.g. , severe head injury, intraabdominal hemorrhage) should lead to use of helical CT as the screening imaging method. 14When chest radiography at admission shows an unremarkable mediastinum in a patient at high risk of traumatic cardiovascular injury (e.g. , violent deceleration), 9a multiplane TEE should be used because of its greater sensitivity for the detection of superficial ATAI and blunt cardiac trauma (fig. 4).
In hemodynamically unstable patients presenting with multisystem trauma, transportation to the radiology department is unsafe and may result in undue delay. 38,39In these cases, multiplane TEE can be safely performed at the bedside during ongoing short-term care and seems ideally suited to rapidly rule out ATAI and to detect potential associated blunt cardiac injuries (fig. 4). 11,12,26Finally, patients with suspected traumatic injury to aortic branches should still be referred for selective angiography, 4,8,31but helical CT shows promise in identifying supraaortic lesions (table 2). Similarly, angiography should be performed if results of examinations are inconclusive, 5,14,17especially when a large hemomediastinum is observed without apparent ATAI or associated vertebrate fractures. 17,19
The main limitation of this prospective study was the absence of systematic angiography, traditionally referred to as the reference imaging technique for the diagnosis of ATAI. Importantly, TEE and helical CT were previously validated against aortography. 6–8,13–16Accordingly, we only sought to compare the respective diagnostic capabilities of these noninvasive imaging methods in patients at high risk of cardiovascular injury. We considered routine angiography to be unethical for the following reasons. First, this additional procedure would have considerably increased the patient evaluation time and further delayed potential surgical repair. 7Second, angiography may also lead to false-negative, 8,9,27and false-positive results. 18Third, it has been argued that clinical follow-up findings–in contrast to negative angiography–may not constitute adequate proof that ATAI is not present. 15However, on the basis of the natural history of untreated ATAI, 40the absence of a complicated course during follow-up can be considered as strong evidence that combined negative multiplane TEE and helical CT were reliable in excluding a traumatic injury to the thoracic aorta. 15Finally, the observed traumatic cardiac injuries (n = 11) would not have been visualized using aortography. 26
In the current study, multiplane TEE and helical CT showed a similar diagnostic accuracy for the identification of subadventitial ATAI. Multiplane TEE provided rapid diagnosis of superficial ATAI and blunt cardiac injuries and proved of particular value in patients with hemodynamic compromise. Based on the results of the current study, a novel, noninvasive approach can be proposed for initial evaluation of patients at high risk of blunt traumatic cardiovascular injuries.
The authors thank Professor R.M. Lang (Director, Noninvasive Cardiac Imaging Laboratories, University of Chicago, Chicago, Illinois) for reviewing the manuscript and Dr J.L. Pascaud (Radiology Department, Dupuytren Hospital, Limoges, France) for his valuable technical help.