Usefulness of Transthoracic and Transoesophageal Echocardiography in Recognition and Management of Cardiovascular Injuries After Blunt Chest Trauma

Home , Aortography, Echocardiography

Heart 1996;75:301-306 301 Usefulness of transthoracic and transoesophageal echocardiography in recognition and management of cardiovascular injuries after blunt chest trauma

Fabio Chirillo, Oscar Totis, Antonio Cavarzerani, Andrea Bruni, Antonio Famia, Mario Sarpellon, Paolo Ius, Carlo Valfre, Paolo Stritoni

Abstract About 10% of victims of major trauma sustain Objective-To assess the diagnostic poten- cardiac or aortic lesions.' Although the vast tial of transthoracic and transoesophageal majority of these individuals die at the scene,2 echocardiography for the detection of an increasing number of patients with previ- traumatic cardiovascular injuries in ously lethal injuries arrive at a treatment facility patients suffering from severe blunt chest because of improvements in prehospital care, trauma. more aggressive resuscitation in the field, and Design-Prospective study over a three rapid transportation to appropriate units.3 The year period. identification of cardiovascular injury in these Setting-A regional cardiothoracic centre. patients is often difficult because of severe Patients-134 consecutive patients (94 associated lesions which are more apparent M/40 F; mean age 38 (SD 14) years) suffer- and require prompt treatment.4 The diagnos- ing from severe blunt chest trauma (injury tic value of routine techniques for the detec- severity score 33 5 (18.2)). Most patients tion of traumatic cardiovascular injuries is (89%) were victims of motor vehicle acci- limited56 and the accuracy of more refined dents. techniques, such as computed tomography Evaluation-All patients underwent trans- and magnetic resonance imaging, has been thoracic and transoesophageal echocar- questioned.7 9 diography within 8 h of admission. Aortography is still considered to be the Aortography was performed in the first 20 gold standard technique for the detection of patients and in a further five equivocal traumatic aortic rupture,'01' but it cannot be cases. always performed on an emergency basis, has Results-Transthoracic echocardiography the complications involved in an invasive tech- provided suboptimal images in 83 patients, nique,'2 does not identify myocardial or valvar detecting three aortic ruptures, 28 pericar- lesions, and has a low diagnostic yield and very dial effusions (one cardiac tamponade), 35 high cost when performed on all trauma left pleural effusions, and 15 myocardial patients. contusions. Transoesophageal echocardio- Transthoracic echocardiography has been graphy was feasible in 131 patients and used in previous studies'3-'5 on patients suffer- detected 14 aortic ruptures (13 at the ing from chest trauma: it proved helpful in isthmus), 40 pericardial effusions, 51 detecting myocardial contusion, whereas no left pleural effusions, 34 periaortic data have been reported on identification of haematomas, 45 myocardial contusions, aortic rupture. right atrial laceration in one patient with Transoesophageal echocardiography over- cardiac tamponade, one tricuspid valve comes most of the technical limitations of rupture, and one severe mitral regurgita- transthoracic echocardiography and provides heart and almost Department of tion caused by annular disruption. For the excellent visualisation of the Cardiology, Regional detection of aortic rupture trans- the entire thoracic aorta in nearly all patients.'6 17 Hospital, Treviso, Italy oesophageal echocardiography showed The aim of this study was to assess the diagnos- F Chirillo and 98% tic potential of transthoracic and trans- O Totis 93% sensitivity, 98% specificity, A Cavarzerani accuracy. Time to surgery was signifi- oesophageal echocardiography for the detection A Bruni candy shorter (30 (12) v 71 (21) min; P < oftraumatic cardiovascular lesions and to evalu- P Stritoni 0.05) for patients operated on only on the ate the impact of transoesophageal echocardio- Department of basis of transoesophageal echocardiographic findings on the management of patients Anaesthesiology and Intensive Care, graphic findings. following severe blunt chest trauma. Regional Hospital, Conclusions-Transthoracic echocardiog- Treviso, Italy raphy has low diagnostic yield in severe A Famia blunt chest trauma, while trans- Methods M Sarpellon oesophageal echocardiography provides Between September 1991 and September Department of Cardiovascular accurate diagnosis in a short time at the 1994 134 consecutive patients (94 men, 40 Surgery, Regional bedside, is inexpensive, mimally inva- women) aged 15-78 years (mean 38 (SD 14) Hospital, Treviso, Italy sive, and does not interfere with other years) admitted to our hospital following P Ius diagnostic or therapeutic procedures. severe blunt chest trauma were prospectively C Valfre and Correspondence to: evaluated for presence of cardiac injury Dr F Chirillo, Cardiologia, (Heart 1996;75:301-306) aortic rupture (table 1). Patients were entered Ospedale Regionale, 31 100 into the study who were found to have Treviso, Italy. con- Accepted for publication Keywords: blunt chest trauma; traumatic cardiovascu- haemothorax, pneumothorax, pulmonary 22 August 1995 lar injuries; echocardiography tusion, fracture of the first two or multiple 302 Chirillo, Totis, Cavarzerani, Bruni, Farnia, Sarpellon, et al

Table 1 Clinical characteristics of 134 patients with a 100 cm gastroscope and attached to a Sonos blunt chest trauma evaluatedfor cardiovascular injury 1000 Hewlett-Packard ultrasound system. Sex M/F 94/40 Introduction of the probe was accomplished Age (years) 38 (SD 14) Mechanism of injury, n (%) after the administration of intravenous Drivers 54 (40) droperidole or morphine, when necessary, for Occupants 33 (25) sedation. Pedestrians struck by car 14 (10-5) (Motor)cyclists 18 (13-5) Cross sectional echocardiography identified Fall from height 12 (9) Animal kicks 3 (2) myocardial contusion in presence of hypokine- Injury severity score 33-5 (SD 18-2) sis, akinesis, or dyskinesis of any segment of Associated injuries, n (%) 86 (64) Cranial 38 (28) the left ventricle or right ventricular free wall.20 Abdominal 41 (30 5) Patients with myocardial contusion underwent a Pelvic 16 (2) Major extremity fractures 78 (58) control transthoracic echocardiogram before Emergency surgical procedures, n 145 discharge. Pericardial and pleural effusion were Neurosurgical 15 (11) Abdominal 35 (26) diagnosed when an echo-free space was seen Urological 7 (5) between the visceral and parietal pericardial Orthopaedic 55 (41) Thoracic (non-cardiovascular) 19 (14) and pleural echoes in all imaging planes. Cardiovascular 15 (11) Periaortic haematoma was diagnosed when a Chest film suggestive for aortic rupture, n (%) 65 (48 5) bright echo density was seen around the aorta. Death, n (%) 19 (14) Colour Doppler interrogation of the septa and valves identified intracardiac shunts or valvar regurgitation according to standard criteria.2' 22 Aortic rupture was diagnosed when two or ribs, or to have had a significant field evidence more of the following echographic criteria were of high energy transfer such as falls of more met: (1) abrupt and discrete change in aortic than 4 metres and automobile accidents with diameter (that is, normal diameter proximal structural intrusion, extrication difficulties, and distal to the site of rupture which exhibited passenger ejection, or death at the scene. widened lumen); (2) presence of one or more Motor vehicle accidents were the cause of consistent linear echoes indicative of transec- injury in most patients. The injury severity tion flaps dividing the aortic lumen into two or score'8 19 for the entire group was 33-5 (SD more lumina; (3) focal but complete intimal 18-2). Most patients (64%) experienced asso- and medial disruption with formation of a ciated cranial, abdominal and pelvic injury, pseudoaneurysmatic cavity; (4) colour Doppler often requiring emergency surgery. Ninety identification of a pseudocoarctation pattern seven patients arrived at our institution (that is, flow acceleration through the injured directly from the accident scene, and 37 were aortic segment); (5) periaortic haematoma. transferred from other hospitals. There were Findings on echocardiography were com- 19 deaths: five from head trauma, six from pared with aortography in 25 patients, surgery haemorragic shock, five from multisystem in 15 patients, and necropsy in 19 cases. organ failure, and three from aortic rupture. In Patients with transoesophageal echocardiogra- all but the latter three patients necropsy ruled phy negative for aortic rupture and repeat out major cardiac injury and aortic rupture. chest x ray negative for widened mediastinum Eighty patients were evaluated while endo- were considered free from aortic injury. tracheally intubated; 35 had chest tubes, 21 abdominal tubes, and five had subcutaneous STATISTICAL ANALYSIS emphysema. Continuous values are expressed as means Supine anteroposterior chest plain x ray film (SD). The sensitivity (true positives divided by was obtained in all patients at admission: true positives plus false negatives), specificity mediastinal widening, a blurred aortic knob, (true negatives divided by true negatives plus apical capping of the lung, rightward displace- false positives), positive predictive value (true ment of the trachea, and depression of the left positives divided by true positives plus false main bronchus were considered findings sug- positives), negative predictive value (true nega- gestive of aortic rupture. Chest x ray was tives divided by true negatives plus false nega- repeated in the erect position before discharge tives) and accuracy (true positives plus true from the hospital. negatives divided by total number of examina- Transthoracic and transoesophageal echo- tions performed) of transoesophageal echocar- cardiography were performed in all patients diography for aortic rupture were calculated. within 8 h of admission. Thirty seven echocar- A P value < 0Q05 was considered statistically diographic studies were done while patients significant. were undergoing other diagnostic or therapeutic procedures: peritoneal lavage (7), drainage of pneumothorax or haemothorax (10), reduc- Results tion of extremity fractures (8), or packing of Transthoracic echocardiography provided abdominal lesions (12). Echocardiographic suboptimal images of the heart and the aorta studies were carried on in the operating room in 83 patients. Reasons were: mechanical ven- (20), intensive care unit (85), x ray depart- tilation (58), obesity (8), subcutaneous ment (8), and emergency room (20). emphysema (5), chest or abdominal tubes (7). Studies were performed by expert cardiolo- In five patients transthoracic studies were not gists using 2 5 MHz and 3.5 MHz trans- performed in order to avoid contamination of thoracic probes and a transoesophageal 5 sterile instrumentation used for other ongoing MHz single plane probe mounted at the end of diagnostic and therapeutic procedures. Usefulness oftransthoracic and transoesophageal echocardiography in recognition and management ofcardiovascular injuries after blunt chest trauma 303

Table 2 Findings on transthoracic and transoesophageal Oesophageal intubation for transoeso- echocardiography in 134 patients sufferingfrom blunt chest phageal echocardiography was impossible in trauma three patients (one had tracheostomy after TTE, n (%) TOE, n (%) laryngectomy and two had fracture of the cer- Feasibility 51 (38) 131 (98) vical spine with initial neurological deficit). Aortic rupture 3 (6) 14 (10-5) detected Myocardial contusion 15 (21-5) 45 (34) Transoesophageal echocardiography Valvar lesions 0 2 (1-5) 40 pericardial effusions, 51 left pleural effu- Periaortic haematoma 5 (10) 34 (26) undetected Pericardial effusion 28 (55) 40 (30 5) sions (five were large and by chest Left pleural effusion 35 (69) 51 (39) x ray, and were drained after transoesophageal 45 myo- TTE, transthoracic echocardiography; TOE, transoesophageal study), 34 periaortic haematomas, echocardiography. cardial contusions (eight patients had left ventricular ejection fraction < 35%). Trans- oesophageal echocardiography identified one Findings on transthoracic echocardiography tricuspid valve rupture, laceration of the right are shown in table 2. In two patients with aor- atrial lateral wall in one patient with cardiac tic rupture at the isthmus a pseudocoarctation tamponade (fig 1), and one severe mitral pattern was noted. There were four false posi- regurgitation caused by disruption of the ante- tives on transthoracic echocardiography for rior anulus (fig 2). These three patients were the identification of aortic rupture. successfully operated on in the emergency department. Incidental findings on transoesophageal echocardiography were: one cor triatriatum, one mitral stenosis, and eight aneurysms of the atrial septum (four with mild left to right shunts). Sixty five patients had suspected rupture of the aorta on chest x ray. Confirmatory aortography was performed in the first 20 cases (to complete the learning curve), and in another five cases with equivocal transoesophageal echocardiographic findings. Transoesophageal echocardiography disclosed 14 aortic ruptures (13 at the isthmus) (fig 3); there were two false positives (due to reverberations in the ascending aorta) and one false negative (sensitivity 93%, specificity 98%, positive predictive value 87%, negative predictive value 99%, accuracy 98%). This latter patient had pneumomediastinum caused by laceration of the left main- stem bronchus following a fall from a 5 m height. Transoesophageal study provided poor visualisation of the entire thoracic aorta due to the presence of air in the mediastinum. The Figure 1 Transoesophageal basalfour chamber view taken in a patient with cardiac tamponadefollowing a horse kick. (Left) The arrow indicates a smallfissuring on the right study was repeated after left pneumectomy and atrial wall. There is pericardial effusion and the right ventricle looks compressed. (Right) detected rupture of the distal ascending aorta. A portion ofthe torn right atrial wall (arrow) is seen floating within the right atrium just Time to surgery was significantly longer for above the tricuspid valve (la, left atrium; Iv, left ventricle; ra, right atrium; rv, right ventricle). patients undergoing aortography (71 (21) min v 30 (12) min; P < 0-05). One patient died during confirmatory aortography and another died during transfer to the operative room soon after transoesophageal echocardiographic evaluation. Twelve patients underwent reparation of the aortic rupture (fig 4) in the emergency department. There was one operative death from disseminated intravascular coagulation; 11 patients were discharged in good general condition. Comparison with surgical and necropsy findings showed that transoesophageal echocardiography was fully diagnostic in 11 cases. It could not identify one case of avulsion of the innominate artery concomi- tant to injury of the ascending aorta, one of avulsion of the left subclavian artery concomi- tant to aortic transection at the isthmus, and one of retrograde extension of the isthmic transection to the distal arch. In this patient surgeons had serious difficulties in approaching the arch rupture through a left thoracotomy, but operation was carried out successfully. Figure 2 Transoesophageal basal two chamber view demonstrating a large tear in the anterior mitral annulus (arrows). In the right panel colourflow imaging shows massive Repeat echo showed improvement in left paravalvar regurgitation (ao, aorta; la, left atrium). ventricular systolic function in all patients with 304 Chirdlo, Totis, Cavarzerani, Bruni, Farnia, Sarpellon, et al

such as enlargement of the thoracic aorta secondary to hypertension or atherosclerosis and thoracic neoplasms. In our experience an initial emergency room chest radiograph was obtainable in most patients only in the supine position because of the severity of trauma. Most of the abnormalities observed on the initial radiograph disappeared on the erect chest x ray obtained before discharge from the hospital. Transthoracic echocardiography provided suboptimal images in 62% of patients and the bad quality of the images obtained was the main cause of the low sensitivity of transthoracic echocardiography, especially when aortic rupture in the region of the isthmus (poorly visualised by the suprasternal notch) is searched for. In the light of our experience we do not recommend transthoracic echocardiography as a Figure 3 Traumatic aortic rupture at the isthmus. Transoesophageal scanning shows the routine technique in the evaluation of blunt presence of an echolucent structure presenting a "valve-like" movement (left is a diastolic frame, right a systolicframe). This aspect corresponded to the adventitia and chest trauma, either for the identification of periadventitial tissues invaginated into the aortic lumen through the traumatic tear. aortic injury or for myocardial contusion. The use of this technique would consume time and increase the cost of management of trauma patients with low clinical yield. The only 134 rather specific sign for isthmic transection on CHEST TRAUMA transthoracic echocardiography might be the presence of a pseudocoarctation pattern (peak 17 117 transisthmic gradient > 20 mm Hg). CARDIOVASCULAR INJURIES NO CARDIOVASCULAR Transoesophageal echocardiography offers INJURIES several advantages over other imaging methods currently used in evaluating patients with 5 2 16 101 chest trauma. Firstly, it is at most a semi-inva- MERGENCY SURGERY DEATH DE LV sive procedure which has been shown to be AWAITING SURGERY D A safe when used by experienced personnel. Potential complications may occur in patients 141 with traumatic oesophageal and gastric ALIVE DIED lesions, since the probe is not provided with a fibreoptic capability. Excessive neck and head Figure 4 Outcome of 134 patients with severe blunt chest trauma. movements to facilitate oesophageal intubation might be harmful in patients who may have injuries to the cervical spine. Patients myocardial contusion and normalisation of all with suspected aortic injury should be heavily segmental abnormalities in 38 patients. sedated to avoid any increase in blood pres- Repeat x ray chest film did not show any sure which may precipitate aortic rupture.26 significant enlargement of the mediastinum in Notwithstanding these potential risks, morbid- patients with transoesophageal echocardiogra- ity and mortality of transoesophageal studies phy negative for aortic rupture. were nil, as in previous studies.27 28 Secondly, transoesophageal echocardiography can be performed quickly (since it does Discussion not require the assembly of an on-call radiol- The evaluation ofpatients suffering from blunt ogy team), at the bedside (avoiding trans- chest trauma for the detection ofcardiovascular portation of patients, which it is often difficult injury is challenging because most cardiovas- for the presence of mechanical ventilation, cular lesions occur in a milieu of multisystem orthopaedic devices, etc, and may be danger- trauma and often go undetected. In these ous in such unstable patients), and does not patients the diagnostic value of routine tech- interfere with other diagnostic, therapeutic, niques is limited because the electrocardio- and resuscitative procedures. graphic changes are non-specific and a chest Thirdly, transoesophageal echocardiogra- film cannot be always performed in the upright phy is accurate for the detection of aortic rup- position, resulting in uncertainties over medi- ture and provides diagnostic information astinal dimensions and structure.23-25 The about possible myocardial and cardiac valvar portable chest radiograph is frequently com- injury as well as about presence and magni- promised by limited exposure capability, expi- tude of pericardiil and left pleural effusion. ratory views, patient motion, difficult patient Many different criteria for the diagnosis of positioning, and magnification and distortion traumatic aortic rupture using transoeso- of the mediastinum in the supine position. phageal echocardiography have been sug- Moreover actual widening of the mediastinum gested.28-32 The inclusion of periaortic may result from chronic anatomical changes haematoma among our diagnostic criteria for Usefulness oftransthoracic and transoesophageal echocardiography in recognition and management ofcardiovascular injuries after blunt chest trauma 305

aortic rupture may appear questionable. In patients underwent both transoesophageal fact Sandor33 reported that blood in the medi- echocardiography and aortography with a astinum was secondary to aortic rupture in mean 105 (56) min interval before achieving only 12-5% of trauma patients and to non- definite diagnosis. After the death of one aortic bleeding in 87-5%. Other sources of patient during confirmatory aortography, we periaortic haematoma include tearing of medi- decided that was ethically correct to send astinal veins, rib fractures, sternal fractures, patients with transoesophageal echocardio- injury to brachiocephalic vessels, and caudal graphic evidence of aortic rupture to operation dissection of blood from cervical injuries.34 without confirmatory aortography. Aorto- Moreover Woodring et all5 have reported nine graphy was performed subsequently in only cases of mediastinal widening associated with five patients with suspected involvement of the fractures of the cervical, thoracic, and lumbar ascending aorta and arch vessels. This reduces spine. Such lesions were found in five patients the ability of the present study to assess the in the present series. Periaortic haematoma diagnostic accuracy of transoesophageal echo- therefore has poor specificity as a sign of aortic cardiography in the diagnosis of traumatic aor- rupture following blunt chest trauma. tic rupture, because this aim could only be However, it should be emphasised that in the properly explored by a comparison with aor- present study it was not included among the tography, so far considered the gold standard diagnostic criteria on its own, and it presented technique. This was carried out in the study 100% sensitivity. This may partly be depen- by Smith et al.28 Their reported sensitivity dent on the absence of patients with intimal (100%) and positive predictive value (99%) injury36 (that is, subclinical traumatic aortic were higher than in the present study. This injury which may occur without extravasation may be because all the ruptures were localised of blood from the aorta). at the isthmus (clearly visualised by trans- Limitations of transoesophageal echocar- oesophageal imaging using single plane tech- diography include: absent visualisation of the nology) and because no injuries of the arch distal ascending aorta and proximal arch vessels were reported. An interesting finding because of the interposition of the air filled tra- which emerges from comparative studies on chea and left main bronchus; poor visualisa- transoesophageal echocardiography and aor- tion of the epiaortic vessels and low resolution tography2836 is that this latter gold standard in presence of pneumomediastinum and pneu- technique also has false negatives (incomplete mopericardium.16 Biplane'7 and omniplane aortic rupture, thrombosis of the pouch, or transoesophageal probes37 can provide better technical deficiencies such as poor opacifica- visualisation of the ascending aorta, but do not tion or inadequate number of views) and false allow detailed analysis of the brachiocephalic positives (ductal diverticula or atheromatous vessels and cannot eliminate interference from aorta). air in the airways or in the mediastinum. In the evaluation of aortic injury trans- It is therefore clear that avulsion of the bra- oesophageal echocardiography should be chiocephalic vessels is not amenable to trans- considered not as an alternative, but as a com- oesophageal echocardiographic diagnosis, as plement to angiography. Transoesophageal are other conditions such as rupture of the dis- echocardiography might be a means of selecting tal ascending aorta and proximal arch, injury patients who must proceed directly to surgery of the abdominal aorta, or avulsion of thoracic (unequivocal evidence of aortic rupture), and abdominal aortic collaterals. who must undergo aortography (periaortic Previous studies30-32 dealing with smaller haematoma and otherwise unexplained medi- series ofpatients have reported different values astinal widening), and who may benefit from of accuracy for detection of aortic rupture by aortography (patients with negative trans- transoesophageal echocardiography. oesophageal echocardiography). This latter In our series one false negative and two false policy will pick up any false negatives and also positives occurred at the beginning of our the patients with ascending aorta, arch, and experience. Later on in the study both cardiol- great vessel pathology. ogists and surgeons became confident with Which method is to be preferred in sus- transoesophageal echocardiographic images pected cardiovascular injury depends on the and seven patients were operated upon with- human and material resources available within out aortography. This underscores the accu- each institution. We believe that trans- racy of the technique in evaluating aortic oesophageal echocardiography, when avail- rupture and the need of a learning curve, able, should be considered first in cases of even for experienced echocardiographers. suspected traumatic cardiovascular lesions Experience is needed to select the views, to because of its safety, speed, accuracy, and con- optimise partial and total gain (to avoid sono- venience. Time from admission to diagnosis graphic artefacts) and in the interpretation of and surgery should be as short as possible in images.29 If operators are not sufficiently patients suffering cardiovascular injuries. In skilled we recommend sending patients for our institution transoesophageal echocardiog- angiography in the presence of dubious find- raphy has become the preferred technique to ings. This will undoubtedly increase initial achieve this objective with expert operators on cost, but will serve to complete the learning call 24 hours a day. curve and ultimately reduce cost and increase diagnostic yielding in a near future.'8 1 Editorial. Traumatic injury of the heart. Lancet 1990; 336:1287-9. Time to surgery is crucial in acute aortic 2 LoCicero J, Mattox KL. Epidemiology of chest trauma. pathology. In the study by Smith et al28 all Surg Clin North Am 1989;69:15-19. 306 Chirillo, Totis, Cavarzerani, Bruni, Farnia, Sarpellon, et al

3 Krasna MJ, Flancbaum MJ. Blunt cardiac trauma: clinical Trauma 1988;28:69-77. manifestations and management. Semin Thorac 20 Tenzer ML. The spectrum of myocardial contusion. 7 Cardiovasc Surg 1992;4:195-202. Trauma 1985;25:620. 4 Parmely LF, Mattingly TW, Manion WC, Jahnke EJ. Non- 21 Perry GJ, Helmcke F, Nanda NC, Byard C, Soto B. penetrating traumatic injury of the aorta. Circulation Evaluation of aortic insufficiency by Doppler color flow 1958;17:1086-101. imaging. J Am Coll Cardiol 1987;9:952-9. 5 Akins CW, Buckley MJ, Daggen W, McIlduff JB, Austen 22 Mohr-Kahaly S, Erbel R, Zenker G, et al. Semiquantitative WG. Acute traumatic disruption of the thoracic aorta: a grading of mitral regurgitation by color Doppler echocar- ten-year experience. Ann Thorac Surg 1981;31:305-9. diography. Inty Cardiol 1989;23:223-30. 6 Plotnick GD, Hamilton S, Lee YC. The cardiologist and 23 Woodring JH, Dillon ML. Radiographic manifestations of the trauma patient: noninvasive testing. Semin Thorac mediastinal hemorrhage from blunt chest trauma. Ann Cardiovasc Surg 1992;4:168-76. Thorac Surg 1984;37:171-8. 7 Nelson JB, Bresticker MA, Nahrwood DL. Computed 24 Mirvis SE, Bidwell JK, Buddenmeyer EU, et al. Value of tomography in the initial evaluation ofpatients with blunt chest radiography in excluding traumatic aortic rupture. trauma. Trauma 1992;33:722-7. Radiology 1987;163:487-93. 8 Miller FB, Richardson JD, Thomas HA, Cryer HM, 25 Bumey RE, Gundry SR, MacKenzie JR, Whitehouse WM, Willing SJ. Role of CT in diagnosis of major arterial Wu SC. Chest roentgenograms in diagnosis of traumatic injury after blunt thoracic trauma. Surgery 1989;106: aortic rupture: observer variation in interpretation. Chest 596-603. 1984;85:605-9. 9 Mirvis SE, Templeton PA. Imaging of thoracic trauma. 26 Warren RL, Akins CW, Conn AKT, Hilgenberg AD, Semin Thorac Cardiovasc Surg 1992;4:177-86. McCabe CJ. Acute traumatic disruption of the thoracic 10 Mirvis SE, Bidwell JK, Buddenmeyer EU, Diaconis JN, aorta: emergency department management. Ann Emerg Pais SO, Whitley JE. Imaging diagnosis of traumatic aor- Med 1992;21:391-6. tic rupture: a review and experience at a major trauma 27 Daniel WG, Erbel R, Kasper W, et al. Safety of trans- center. Invest Radiol 1987;22: 187-96. esophageal echocardiography. A multicenter survey of 11 Gundry SR, Williams S, Burney RE, MacKenzie JR, Cho 10,419 examinations. Circulation 1991;83:817-21. KJ. Indications for aortography: radiography after blunt 28 Smith MD, Cassidy JM, Souther S, et al. Transesophageal chest trauma: a reassessment of the radiographic findings echocardiography in the diagnosis of traumatic rupture of associated with traumatic rupture of the aorta. Invest the aorta. NEnglJMed 1995;332:356-62. Radiol 1983;18:230-7. 29 Brooks SW, Young JC, Cmolik BL, et al. The use of trans- 12 LaBerge JM, Jeffrey RB. Aortic lacerations: fatal complica- esophageal echocardiography in the evaluation of chest tions of thoracic aortography. Radiology 1987;165:36-9. trauma. J Trauma 1992;32:761-8. 13 Miller FA, Seward JB, Gersh BJ, Tajik AJ, Mucha P. Two- 30 Shapiro MJ, Yanofsky SD, Trapp J, et al. Cardiovascular dimensional echocardiographic findings in cardiac evaluation in blunt chest trauma using transesophageal trauma. Am Cardiol 1982;50:1022-7. echocardiography. J Trauma 199 1;31:835-40. 14 Helling TS, Duke P, Beggs CW, Crouse W. A prospective 31 de Belder A, Thomas M, Marrinan M. Traumatic rupture of evaluation of 68 patients suffering blunt chest trauma for the thoracic aorta diagnosed by transesophageal echocar- evidence of cardiac injury. Y Trauma 1989;29:961-5. diography. BrHeartJ 1993;70:393-4. 15 Bu'Lock FA, Prothero A, Shaw C, et al. Cardiac involve- 32 Dacosta A, Billard JL, Zeni F, et al. Rupture traumatique ment in seatbelt-related and direct sternal trauma: a de l'isthme aortique diagnostiqee par l'echographie tran- prospective study and management implications. Eur soesophagienne. Arch Mal Coeur 1994;87:101-4. Hearty 1994;15:1621-7. 33 Sandor F. Incidence and significance of traumatic medi- 16 Seward JB, Khandheria BK, Oh JK, et al. Transesophageal astinal hematoma. Thorax 1967;22:43-62. echocardiography: technique, anatomic correlations, 34 Rich NM. Injuries of the intrathoracic branches of the aor- implementations, and clinical applications. Mayo Clin tic arch. In: Rich NM, ed. Vascular trauma. Philadelphia: Proc 1988;63:649-80. WB Saunders, 1978:287. 17 Seward JB, Khandheria BK, Edwards WD, et al. Biplanar 35 Woodring JH, Lee C, Jenkins K. Spinal fracture in blunt transesophageal echocardiography: anatomic correla- chest trauma. J Trauma 1988;28:789-93. tions, image orientation, and clinical applications. Mayo 36 Davis GA, Sauerisen S, Chandrasekaran K, Karalis DG, Clin Proc 1990;65:1193-213. Ross J, Mintz GS. Subclinical traumatic aortic injury 18 Baker SP, O'Neill B, Haddon W, Long WB. The Injury diagnosed by transesophageal echocardiography. Am Severity Score: a method for describing patients with HeartJ 1992;123:534-6. multiple injuries and evaluating emergency care. Y 37 Khandheria BK, Oh JK. TEE: state of the art and future Trauma 1974;14:187-96. directions. Am_J Cardiol 1992;69:61-75H. 19 Copes WS, Champion HR, Sacco WJ, Lawnick MM, 38 Vlahakes GJ, Warren RL. Traumatic rupture of the aorta. Keast SL, Bain LW. The Injury Severity Score revisited. Y NEnglJ Med 1995;332:389-90.