ORIGINAL ARTICLES: GENERAL THORACIC

GENERAL THORACIC SURGERY:

The Annals of Thoracic Surgery CME Program is located online at http://cme.ctsnetjournals.org. To take the CME activity related to this article, you must have either an STS member or an individual nonmember subscription to the journal.

Triage and Outcome of Patients with Mediastinal Penetrating Trauma

Joshua H. Burack, MD, Emad Kandil, MD, Ahmed Sawas, BS, Patricia A. ONeill, MD, Salvatore J. A. Sclafani, MD, Robert C. Lowery, MD, and Michael E. Zenilman, MD
Departments of Surgery and Radiology, Kings County Hospital Center, and State University of New YorkDownstate, Brooklyn, New York

Background. A retrospective study was conducted detailing an experience with echocardiography and contrast-enhanced helical computed tomographic angiographic (CTA) scans in the evaluation of stable patients with mediastinal penetrating trauma (MPT). Methods. Unstable patients underwent emergent operative intervention, and stable patients underwent chest roentgenogram, transthoracic echocardiography (TTE), and CTA. Further testing (angiogram, bronchoscopy, esophagoscopy, esophagogram) was done only if one of these studies revealed evidence of a trajectory in the vicinity of major vasculature or viscera. Results. Between 1997 and 2003, 207 patients had MPT. Seventy-two (35%) were unstable (45 gun shot wounds, 27 stab wounds) and 19 died in the emergency department. Fifty-three had emergent intervention and 32 survived. Work-up was done on 135 stable patients (65%) consisting of 46 gunshot wounds and 89 stab wounds, of

which 5 had a positive TTE result and underwent a repair of a cardiac injury. CTA evaluation was normal in almost 80% of patients, who subsequently did not require further evaluation or treatment. In the stable patients, endoscopy or esophagography conrmed one tracheal injury and no esophageal injury. In the entire group, 10 patients (7%) had occult injury, and there were no deaths or missed injuries. Conclusions. In cases of MPT, unstable patients require surgery, and in stable patients, TTE and chest CTA are effective screening tools. Patients with a negative TTE and CTA results can be observed and may not require further testing or endoscopy, whereas patients with positive TTE or CTA results require further assessment to exclude occult injury. (Ann Thorac Surg 2007;83:377 82) 2007 by The Society of Thoracic Surgeons

pproximately 150,000 people die each year in the United States as a result of trauma, and 25% of the deaths can be directly related to thoracic injury [1]. Almost all patients with thoracic trauma are treated conservatively with a successful outcome. Most patients survive as a result of a prompt resuscitation, efcient diagnostic testing, and simple therapeutic maneuvers. In a contemporary multicenter experience, urgent operative treatment was required in only 0.5% of blunt and 2.8% of penetrating thoracic injuries [2, 3]. In contrast are those patients with penetrating injury to the mediastinum, a group of patients who have a substantial operative intervention rate and mortality. Historically, a gunshot injury traversing the mediastinum precipitates hemodynamic instability related to a thoracic

Accepted for publication May 18, 2006. Presented at the Forty-second Annual Meeting of The Society of Thoracic Surgeons, Chicago, IL, Jan 30 Feb 1, 2006. Address correspondence to Dr Burack, Department of Surgery, Box 40, State University of New YorkDownstate, 450 Clarkson Ave, Brooklyn, NY 11203; e-mail: jburack@downstate.edu.

vascular injury in approximately 50% of patients and has an operative mortality rate of 20% to 40% [4 6]. Certainly, the lethal nature of the injury is related to the high concentration of major vascular and visceral structures that reside within the mediastinum, and in many instances, it is implausible that a transmediastinal penetrating injury has avoided a major structure. Indeed, another distinct subgroup of patients present with stable vital signs and mediastinal penetrating trauma (MPT), who typically require diagnostic evaluation, followed by observation alone. The essential clinical task is to efciently triage the unstable patients to the operating room and correctly diagnose and operate on only those stable patients who harbor potentially lethal injuries. The evolution of imaging techniques in ultrasound and computed tomography (CT) have been assets in the rapid and accurate diagnosis of injury and have gradually replaced the cumbersome invasive evaluation of mediastinal injury by pericardial drainage, endoscopy, and routine angiography [4 9]. The purpose of the present study was to assess the reliability and the efcacy of transthoracic echocardiography (TTE) and CT angiography (CTA)
0003-4975/07/$32.00 doi:10.1016/j.athoracsur.2006.05.107

2007 by The Society of Thoracic Surgeons Published by Elsevier Inc

GENERAL THORACIC

378

BURACK ET AL MEDIASTINAL PENETRATING TRAUMA

Ann Thorac Surg 2007;83:377 82

GENERAL THORACIC

Table 1. Unstable Hemodynamic State

1. Traumatic cardiac arrest or near arrest and an EDT 2. Cardiac tamponade 3. Persistent ATLS class III shock despite uid resuscitation (blood loss 15002000 mL, pulse rate 120, blood pressure decreased) 4. Chest Tube output 1500 mL of blood on insertion 5. Chest Tube output 500 mL/hour for the initial hour 6. Massive hemothorax after chest tube drainage
EDT Emergency department thoracotomy; trauma life support [1]. ATLS Advanced

in the diagnostic workup of patients with MPT and to determine whether the use of these diagnostic tests diminished the reliance on invasive investigative procedures. To record the lethal nature of all types of penetrating mediastinal trauma, injuries from stab wounds and gunshot wounds were both examined. Furthermore, a precise scheme for the classication of MPT was developed to clarify and organize the clinical presentation and outcome of these injuries.

Patients and Methods

Kings County Hospital Center (KCHC) is a New York State designated level 1 trauma center and, with approximately 1500 major trauma admissions yearly, is one of the busiest trauma centers in metropolitan New York City. All patients presenting with MPT to KCHC between April 1997 and April 2003 were included in the study, which was approved by the Institutional Review Board of the University Hospital of Brooklyn and the Kings County Hospital. Patients were identied using the New York State Trauma Registry. Under the direction of a dedicated, trauma surgeon, a team of surgical residents evaluated each patient with penetrating trauma. An unstable hemodynamic state was
Fig 2. Wound classication scheme for penetrating mediastinal trauma. Number of patients in each category listed in parenthesis. M1 (dot-dash arrow) base of the neck into mediastinum or pleura (dot-dash arrow); M2 (dotted arrow) one pleural cavity and mediastinal violation (central hematoma, visceral or spinal cord injury, metallic fragments in the mediastinum); M3 (solid arrow) parasternal injury within the nipple line or 4 cm from the sternum; M4 (dashed arrow) two pleural cavities and mediastinal traverse.

determined by the supervising trauma surgeon within the rst hour of admission and was dened by using standard bedside criteria (Table 1). All other patients were deemed stable. Dependent on vital signs, unstable patients underwent an abbreviated evaluation, followed by surgical exploration in the operating room. Some patients, without signs of life on admission or shortly thereafter, were pronounced dead in the emergency department (ED). Other patients, in extremis or traumatic cardiac arrest, underwent an ED thoracotomy (EDT). Stable patients were managed prospectively by preexisting management algorithms (Fig 1) [8]. The data

Table 2. Mechanism of Initial Clinical Presentation

Total SW GSW Fig 1. Management algorithm for a patient with penetrating mediastinal trauma. *Performed as needed. (CTA computed tomographic angiogram scan.)
a

Stable n (%) 89 (77) 46 (51)

2

Unstable n (%) 27 (23) 45 (49)a

Death n (%) 8 (7) 38 (42)a

116 91

.001 based on

analysis. SW stab wound.

GSW

gunshot wound;

Table 3. Injury Location at Initial Clinical Presentation

Total n GSW (%) M1 M2 M3 M4
GSW

Stable n (%) 10 (63) 26 (76) 92 (67) 7 (35)

M1M4

Unstable n (%) 6 (37) 8 (24) 45 (33) 13 (65)

see Fig 2.

Death n (%) 2 (13) 5 (15) 28 (20) 12 (60)

16 (31) 34 (50) 137 (36) 20 (100)

gunshot wound;

were collected retrospectively by medical records review and database analysis. All mediastinal wounds were classied according to the location of the wounds or trajectory of the knife or missile, or both (Fig 2). To avoid confounding data from complex thoracoabdominal injury, patients were excluded from the review if the trajectory of the wound entered the abdominal cavity, if a diagnostic peritoneal lavage was positive for injury, or if the patient underwent a concomitant laparotomy. Knife impalement injuries were also excluded because these patients underwent operative removal, regardless of the hemodynamic state or associated injury. In each stable patient, an initial work-up included a portable chest roentgenogram, TTE, and a CTA. The ultrasound scan was performed in the emergency department and was read by a cardiology fellow, in conjunction with the trauma team. A TTE result was deemed positive only if any appreciable pericardial uid was seen on the subcostal view. The CTA was performed in a remote suite with a single-detector CT scanner (model CTI, GE Medical, Milwaukee, WI), with the administration of intravenous contrast timed to visualize the vasculature and oral contrast to visualize the esophagus and with 10-mm imaging slices. The study was read by a staff radiologist and the trauma team. In the radiology suite, patients had continuous monitoring of oxygen saturation, electrocardiogram, and arterial blood pressure by noninvasive or catheter technique. For the purposes of this study, a CTA was considered positive if the tract of the injury was in proximity to vital mediastinal vascular or visceral structures or if a substantial collection of air or blood was in proximity to the tract. Indications for endoscopy were determined by individual clinical circumstance, and in particular, if a pneumomediastinum or a posterior mediastinal tract was identied, patients underwent additional endoscopic evaluation. Additional patients, those with a hematoma or a missile tract near major vasculature, were triaged to formal angiography. Patients were generally observed either on the regular ward or in the surgical intensive care unit for 48 hours. Routine follow-up was done at an interval of 5 to 7 days in the outpatient clinic.

years and 183 (88%) were men. The mechanism of injury was almost evenly distributed: 116 patients (56%) had stab wounds and 91 patients (44%) had gunshot wounds (Table 2). The specic location of the injury was most commonly parasternal (M3) in 137 patients (66%; Table 3). The diagnostic algorithm was used with 135 patients (65%) who presented with stable vital signs. Half of the patients (49%) with gunshot wounds were unstable and required emergent surgery, and less than one quarter of patients (23%) with stab wounds also required immediate surgery. All of the mediastinal traverse injuries (M4) were caused by gunshot wounds, and this trajectory had the highest rate of instability and subsequent operative intervention. Injuries from a cephalad direction (M1), which were predominately stab wounds, were responsible for the second highest incidence of instability and subsequent operative intervention. Mediastinal traverse injuries (M4) had the highest observed mortality rate (60%), which was at least three times higher than all other injury trajectories. The mortality rate was 42% for gunshot wounds and 7% for stab wounds. The presence of a gunshot wound, based on 2 analysis, was associated with signicant risk of both instability and death. In the entire population of 207 patients, 72 patients (35%) were unstable. Within the cohort of unstable patients, 19 patients (26%) with conrmed MPT died shortly after admission to the ED, without any intervention. An EDT was done in 13 patients (18%), with 2 survivors (2.8%), both of whom had sustained a cardiac injury. Excluding the patients who died shortly after presentation, 16 patients had cardiac injuries, and 11 of these patients were unstable; 12 patients in this group survived. Forty patients underwent formal operative exploration, either by thoracotomy or sternotomy, with pulmonary and great vessel injury being the most common injuries.

Results
During a 60-month interval, 207 patients were retrospectively identied with a MPT. The average age was 28.5

Fig 3. Evaluation of the stable patients: results of the diagnostic algorithm. Number of patients in each category listed in parenthesis and percentages reect the group of stable patients only. (CTA computed tomographic angiogram scan.)

GENERAL THORACIC

Ann Thorac Surg 2007;83:377 82

BURACK ET AL MEDIASTINAL PENETRATING TRAUMA

379

380

BURACK ET AL MEDIASTINAL PENETRATING TRAUMA

Ann Thorac Surg 2007;83:377 82

GENERAL THORACIC

Table 4. Occult Injury in Stable Patients with MPT

Patient 1 2 3 4 5 Injury Injury Injury Injury Angiographic Findings (thrombosis) to 4th Intercostal artery to the vertebral artery at the thoracic inlet to the Internal Mammary artery to the left Subclavian artery Treatment Observation Coil embolization Coil embolization Sternotomy/thoracotomy and interposition graft of the subclavian artery Endovascular stent graft, thoracotomy, and tracheal resection

Inominate artery pseudoaneurysm and thoracic tracheal injury

There were no negative explorations and no apparent missed injuries. Intraoperative endoscopy was liberally applied in patients with pneumomediastinum or with an injury to the posterior mediastinum. In the unstable group of patients, there were two injuries to the esophagus, which were successfully repaired, and none to the trachea. In the entire group of unstable patients, 32 survived for an overall survival rate of 44%. The survival rate increased to 75% in those patients who survived to transfer to the operating room. The results of the stable patients, who were triaged by diagnostic algorithm, were recorded (Fig 3.) After the initial echocardiogram, 5 patients (3.7%) were transferred to the operating room for successful repair of a cardiac injury. The remaining stable patients underwent a CTA evaluation, which was unremarkable in almost 80%. None of the patients became unstable during the course of the diagnostic work-up or required emergent transfer to the operating room. The 23 patients (21%) with a positive CTA result were referred for formal angiography, which, again, was unremarkable in approximately 80%. Angiography did conrm ve occult vascular injuries, of which three were managed with percutaneous therapies and two required operative intervention (Table 4). Bronchoscopy, esophagoscopy, and esophagogram were selectively performed in 21 patients and identied one thoracic tracheal injury and no esophageal injuries. Overall, in the stable patients with MPT, there was no evidence of a missed injury, either during hospitalization or at clinic follow-up. In addition, no patients died and occult injury was detected in only 10 patients, which represents 7% of the entire stable cohort.

Comment
In general, patients with penetrating mediastinal trauma triage themselves between operative intervention or evaluation and observation. The decision to pursue operative exploration is relatively straightforward given the initial clinical presentation, and essential clues about the severity of injury can be obtained by evaluation of the timing and degree of posttraumatic shock. In the rst hour after hospital admission, thoracic vascular and neurologic trauma are the most common causes of death. In a large contemporaneous series of urban trauma patients, the three signicant risk factors for early death were a major thoracic

injury, a major neurologic injury, and admission systolic blood pressure of less than 90 mm Hg [1, 10]. In the current series, one third of the patients presented in such a fashion, with both shock and a penetrating thoracic injury. Interestingly, one third of the unstable patients had been injured by a stab woundnot a gunshot wound emphasizing the lethal nature of any penetrating injury to the mediastinum In the operating room, the culprit was invariably an injury to the heart, the great vessels, or the pulmonary hilum. The overall mortality rate in the entire unstable group was 55%, with an 85% mortality rate in the moribund patients who underwent an EDT. The operative mortality was 25% for patients who survived long enough to transfer to the operating room. This is similar to the results in several small series of transmediastinal gunshot wounds, where the operative mortality for unstable patients was 20% to 36% [4 6]. Selective management of transmediastinal gunshot wounds was rst proposed in 1981 by Richardson and colleagues [4] and has become the standard for the current management of stable patients. In a recent prospective study of transmediastinal gunshot wounds, all ED patients with an admission systolic blood pressure of more than 100 mm Hg and no obvious bleeding safely underwent a complete diagnostic evaluation [5]. No patients in the current series decompensated during the diagnostic phase and no deaths were recorded in the entire group of stable patients. Additionally, after the diagnostic angiogram, several patients with vascular injury underwent percutaneous interventions with coil embolization or placement of an endovascular stent [11]. Only 7% of stable patients in the current study required intervention for injury. In several recent small clinical series of transmediastinal gunshot wounds, the surgical exploration rate in stable patients was 8% to 31% [6 8, 12]. Uniform survival in a stable patient after a transmediastinal gunshot wound has been repeatedly observed [7, 8]. Other reports have noted an occasional death in a patient who becomes unstable in angiography and in patients with a failed esophageal repair or those with late multiorgan failure and sepsis [6, 12]. Injuries to the trachea and the esophagus are rare: only three cases were observed in the entire cohort, for an incidence rate of 1.4%. In several other series of mediastinal gunshot wounds, the rate of injury to the

trachea and the esophagus was quite variable, ranging from 0% to 33% [4 8, 12]. Regardless of the variability of the incidence, the disastrous consequences of a missed hollow viscus injury will continue to mandate the use of traditional bronchoscopy, esophagoscopy, or contrast esophagography in those patients where there is a clinical suspicion related to proximity of the injury or pneumomediastinum. Flexible endoscopy, particularly in the agitated or intubated patient or contrast esophagraphy in the awake and cooperative patient can accurately diagnose esophageal injury [13]. Flexible bronchoscopy similarly provides direct visualization and is highly accurate in the diagnosis of airway injury [14]. In the current report, 15% of the unstable patients had a penetrating cardiac injury, and patients explored in the ED had a survival of 18%, and patients who were able to transfer to the operating room had a survival of 71%. In stable patients, a signicant echocardiographic pericardial effusion was detected in 4%, which led to intervention and successful repair in each instance. An echocardiogram is advantageous owing to the portability of the ultrasound equipment and the rapidity of the exam. The diagnosis of a traumatic pericardial effusion can be made by the visualization of an echolucent region between the heart and pericardium, and right ventricular diastolic collapse will conrm tamponade [9]. If a skilled sonographer is present, ultrasound imaging appears to be as accurate as the traditional subxiphoid pericardial window in making the diagnosis of effusion, with an accuracy, sensitivity, and specicity that exceeds 95% [9, 15]. A combination of aggressive operative intervention in the unstable patient and ultrasound evaluation of the stable patient provided an overall survival of 40% in the 30 patients with known cardiac injury, which is similar to the results in a large recent series of penetrating cardiac trauma [16]. The diagnostic evaluation of the stable patient in the current study rested on two primary diagnostic testsCT scans and ultrasound imagingwith the limited, selective use of angiography, endoscopy, and esophagography. The current diagnostic algorithm has replaced the traditional invasive evaluation because it is safe, efcient, and cost-effective [8]. The ever-increasing accuracy of imaging techniques has expedited the transition from reliance on invasive diagnostic tests to rapid noninvasive evaluation. Pericardial drainage and endoscopy are being replaced by trauma-specic roentgenograms and ultrasound protocols. The current series was based on the results from a single-detector CT scan, which has been eclipsed by the current multidetector CT scans that provide exquisite anatomic detail, three-dimensional reconstructions, and potentially, vascular imaging similar to a traditional angiogram. In a recent series of stable patients with mediastinal gunshot wounds, 67% of patients had a satisfactory imaging of the missile tract with a multidetector CT scan, thus avoiding all other diagnostic testing, including echocardiography [7]. In the future, the evaluation of many stable penetrating

trauma victims may be reduced to a solitary evaluation with rapid multidetector CT scan [17]. It is well established that a gunshot wound remains the most signicant risk factor for both hemodynamic instability and death when compared with all types of penetrating trauma [16]. This observation was conrmed in the current series. Stab wound injury to the mediastinum is also responsible for substantial operative intervention and mortality rates, however, and the study of MPT should include both injury mechanisms, which will reect the current trend in clinical trauma research to classify injury by either blunt or penetrating mechanism [2, 3, 10]. Various denitions of mediastinal penetrating trauma have been proposed, but the several common anatomic trajectories have yet to be classied [4, 5, 7, 8]. None of the previous denitions have described a penetrating injury that traverses the base of the neck, the thoracic inlet, and ultimately enters the mediastinum or pleural space (M1). Although not as lethal as any of the other subtypes of mediastinal wounds in the current series, this trajectory was responsible for a relatively high incidence of hemodynamic instability and operative intervention. Vascular injury in the region of the thoracic inlet, with rapid hemorrhage into the free pleural cavity, is the probable mechanism. Further prospective observations on the individual behavior of the various subtypes of injury may help with clinical decision-making and triage in the future. In summary, accurate triage of the unstable patient, along with a classication system of injury trajectory, combined with modern diagnostic evaluation of the stable patient who may be harboring an occult injury, will continue to improve the management and outcomes of patients with penetrating mediastinal trauma.

References
1. American College of Surgeons Subcommittee on Trauma. Advanced trauma life support program for doctors, 7th ed. Chicago, IL: American College of Surgeons; 2004. 2. Karmy-Jones R, Jurkovich GJ, Shatz DV, et al. Management of traumatic lung injury: a Western Trauma Association multicenter review. J Trauma 2001;51:1049 53. 3. Karmy-Jones R, Jurkovich GJ, Nathens AB, et al. Timing of urgent thoracotomy for hemorrhage after trauma. Arch Surg 2001;136:5137. 4. Richardson JD, Flint LM, Snow NJ, Gray LA, Trinkle JK. Management of transmediastinal gunshot wounds. Surgery 1981;90:671 6. 5. Renz BM, Cava RA, Feliciano DV, Rozycki GS. Transmediastnal gunshot wounds: a prospective study. J Trauma 2000; 48:416 22. 6. Degiannis E, Benn CA, Leandros E, Goosen J, Boffard K, Saadia R. Transmediastinal gunshot injuries. Surgery 2000; 128:54 8. 7. Hampeter DE, Demetriades D, Asenio JA, et al. Helical computed tomographic scan in the evaluation of mediastinal gunshot wounds. J Trauma 2000;49:689 95. 8. Stassen NA, Lukan JK, Spain DA, Millar FB, Carrillo E, Richardson JD. Reevaluation of diagnostic procedures for transmediastinal gunshot wounds. J Trauma 2002;53:635 8. 9. Rozycki GS, Feliciano DV, Ochsner MG, et al. The role of ultrasound in patients with possible penetrating cardiac wounds: a prospective mutlticenter study. J Trauma 1999;46:54352.

GENERAL THORACIC

Ann Thorac Surg 2007;83:377 82

BURACK ET AL MEDIASTINAL PENETRATING TRAUMA

381

382

BURACK ET AL MEDIASTINAL PENETRATING TRAUMA

Ann Thorac Surg 2007;83:377 82

GENERAL THORACIC

10. Demetriades D, Murray J, Charalambides K, et al. Trauma fatalities: time and location of hospital deaths. J Am Coll Surg 2004;198:20 6. 11. Blattman SB, Landis GS, Knight M, Panetta TF, Sclafani SJA, Burack JH. Combined endovascular and open repair of a penetrating innominate artery and tracheal injury. Ann Thorac Surg 2002;74:2379. 12. Nagy KK, Robert RR, Smith RF, et al. Transmediastinal gunshot wounds: are stable patients really stable? World J Surg 2002;26:124750. 13. Flowers JL, Graham SM, Ugarte MA, et al. Flexible endoscopy for the diagnosis of esophageal trauma. J Trauma 1996; 40:261 6.

14. Cassada DC, Munyikwa MP, Moniz MP, Dieter RA, Schuchman GE, Enderson BL. Acute injuries of the trachea and major bronchi: importance of early diagnosis. Ann Thorac Surg 2000:69:15637. 15. Jimenez E, Martin M, Krukenkamp I, Barrett J. Subxiphoid pericardotomy versus echocardiography: a prospective analysis of the diagnosis of occult penetrating cardiac injury. Surgery 1990;108:676 80. 16. Asensio JA, Berne JD, Demetriades D, et al. One hundred ve penetrating cardiac injuries: a 2-year prospective evaluation. J Trauma 1998;44:1073 82. 17. Mirvis SE. Diagnostic imaging of acute thoracic injury. Semin Ultrasound CT MRI 2004;25:156 79.

DISCUSSION
DR TIMOTHY L. VAN NATTA (Iowa City, IA): It was a very nice presentation. Any time you can simplify the algorithm for something like this, its most helpful. My question actually has two parts. First, who does the echocardiogram in the emergency department? Is it the emergency room physician, the surgeon, or is it cardiology? If its cardiology, then are you depending on the fellow or the staff to come in and do it? If you are depending on cardiology, would not the FAST exam, the ultrasound done by the surgeon, tell you right away that there is pericardial uid, so you can eliminate that in the stable patient, and then go on and get your helical CT. If that showed some trajectory close to the heart, maybe you could work backwards, if the patient remains stable, and then get a transthoracic or transesophageal echocardiogram at that point. Thank you. DR BURACK: To answer your rst question, in this series the echoes were done by the cardiology fellows, who were for some reason always available to come in the middle of the night. I guess they like this sort of thing. It was done in conjunction with the trauma attending surgeon. I think that in the literature, the FAST exam is the established evaluation that is being done for pericardial effusion in the setting of trauma, and actually were making a transition now from cardiology fellows to the FAST exam being done by the surgeons. Also, your comments about the helical CT, in the future it may be just a helical CT, particularly the new 64-slice CTs, which can document a pericardial effusion probably better than an ultrasound exam, and in our next prospective series we may eliminate the echocardiogram and it just may be a CAT scan.

INVITED COMMENTARY
Burack and associates [1] have analyzed 20 patients with mediastinal traverse injuries from a cohort of 207 patients with penetrating wounds to the mediastinum, with entry from the front, cephalad, or side. The 20 mediastinal traverse patients were those at greatest risk for tracheal, bronchial, or esophageal injury, and historically have posed one of the great challenges for surgeons determining which such patients should receive additional endoscopic or contrast studies. All the unstable patients in this group (35%) either died in the emergency center or had operative intervention, where two esophageal injuries and no tracheal injuries were discovered at direct exploration. Among the stable patients (65%), 7 had mediastinal traverse, and 1 was found to have a tracheal injury. None were found to have esophageal injury at endoscopy or esophagoscopy. Among this group of 7 patients, the analysis of aerodigestive injuries is most helpful, as the unstable patients had direct evaluation at the time of surgery. Although the stable cohort available for such an analysis is very small, it appears that an algorithm based on the computed tomography (CT) determined trajectory in penetrating mediastinal and mediastinal traverse injuries can be benecial. When CT demonstrates the trajectory in stable patients is not specically in the area of the trachea or the esophagus, bronchoscopy, esophagoscopy, or esophagography are apparently not required. Esophageal injury can be caused by caustic ingestion, iatrogenic perforation, blunt trauma, foreign body perforation, and mediastinal penetrating trauma. It is in the area of mediastinal traverse that the clinician has historically had the most difculty determining the exact diagnostic modality and planning for an operation. This article, with its small numbers of mediastinal traverse patients provides some guidance. Kenneth Mattox, MD Department of Surgery Baylor College of Medicine One Baylor Plaza Houston, TX 77030 e-mail: kmattox2@aol.com

Reference
1. Burack JH, Kandil E, Sawas A, et al. Triage and outcome of patients with mediastinal penetrating trauma. Ann Thorac Surg 2007;83:377 82.

2007 by The Society of Thoracic Surgeons Published by Elsevier Inc

0003-4975/07/$32.00 doi:10.1016/j.athoracsur.2006.10.061