Scandinavian Journal of Surgery 91: 41–45, 2002

Penetrating injuries of the chest

41

PENETRATING INJURIES OF THE CHEST: INDICATIONS FOR OPERATION D. Demetriades, G. C. Velmahos Department of Surgery, Division of Trauma and Critical Care, University of Southern California, Los Angeles California, U.S.A.

ABSTRACT

Penetrating injuries to the chest present a frequent and challenging problem. The majority of these injuries can be managed non-operatively. The selection of patients for operation or observation can be made by clinical examination and appropriate investigations. The trauma ultrasound has become a valuable first-line tool to rule out pericardial tamponade. Spiral computed tomography of the chest is increasingly used to evaluate transmediastinal gunshot wounds and direct, if needed, further organ-specific tests, such as esophagography, aortography, or bronchoscopy. Minimally invasive techniques have found sound application in the thoracoscopic evacuation of undrained hemothorax and the laparoscopic evaluation of diaphragmatic trauma. In the operative arena, lung-sparing techniques with the use of staplers, like wedge resection and tractotomy, have allowed easier, faster, and effective control of bleeding without sacrificing unnecessarily normal pulmonary parenchyma. Knowledge of the new advancements in the field of thoracic trauma will allow surgeons to provide expert care and improved outcomes. Key words: Thoracic trauma; pulmonary injury; trauma ultrasound; thoracic computed tomography; thoracoscopy; hemothorax; cardiac trauma; transmediastinal injury

INTRODUCTION Penetrating chest trauma is a common injury in most American inner-city trauma centers. At the Los Angeles County and University of Southern California Medical Center, which is the largest trauma center in the county, with about 7000 trauma admissions per year, penetrating chest trauma accounts for about 7 % of all trauma admissions or about 16 % of penetrating trauma admissions. The vast majority of penetrating chest trauma can safely be managed by simCorrespondence: D. Demetriades, M.D. Division of Trauma and Critical Care LAC + USC Medical Center, Rm 1105 1200 North State Street Los Angeles, California 90033 U.S.A. Email: [email protected]

ple thoracostomy tube drainage. In a prospective study of 543 patients with stab wounds to the chest only 14 % required an emergency thoracotomy or sternotomy (1). Similarly, only 15–20 % of gunshot wounds (GSW) to the chest require a thoracotomy. However, the selection of patients who need a thoracotomy is not always easy and there are many controversies regarding the indications for operation or observation. In this manuscript we will discuss these issues and present our experience and practices. INITIAL EVALUATION The extent and means of the initial evaluation in the emergency room (ER) depends on the clinical condition of the victim. The patients could be divided in three groups. First, patients with cardiac arrest or imminent cardiac arrest require an immediate ER resuscitative thoracotomy without any investigations.

42

D. Demetriades, G. C. Velmahos

anoxia, other organs may be saved and donated after a successful ER thoracotomy. Lastly, an ER thoracotomy is an excellent drill for the medical and nursing staff, and future patients may benefit from their experience. In these cases the procedure should be performed under controlled conditions in order to avoid accidental injuries to the staff. Blood transfusions should be avoided once the irreversibility of the cardiac arrest is established. The overall survival rate varies a lot and depends on the mechanism of injury (GSW’s versus stab wounds), the pre-hospital time, and the indications for resuscitative thoracotomy. Velmahos et al (2) in a retrospective study of 670 cases with penetrating trauma reported an overall survival of 8.3 % for stab wounds and 4.4 % for GSW’s. Fig. 1. Multiple gunshot wounds to the chest, abdomen, and neck. A trauma ultrasound and plain films may be valuable in determining the source of hypotension.

Second, patients with severe hypotension, but not in immediate danger of cardiac arrest, require an emergent operation in the operating room, often without any specific investigations. A trauma ultrasound may be useful, provided the ultrasound machine is immediately available in the ER and the investigation can be performed by the ER physicians or the surgeons without any delay. This is particularly helpful in cases in which the source of hypotension is not clear because there are multiple penetrating injuries involving many body areas. For the same reasons, in cases of multiple GSW’s, chest and abdominal plain radiographs may be helpful, provided they can be performed without any delay (Fig. 1). A third group of patients presents in a fairly stable condition and many of these patients may benefit from more complex and time-consuming investigations for identification of contained vascular injuries, aerodigestive and diaphragmatic injuries and other less life-threatening conditions. EMERGENCY ROOM THORACOTOMY An ER thoracotomy is indicated in all penetrating injuries of the chest with imminent cardiac arrest or cardiac arrest which occurred en route or in the ER. There is considerable controversy regarding the indications for resuscitative thoracotomy in patients who lost vital signs in the field. The authors strongly believe that in an urban environment with short prehospital times a resuscitative thoracotomy should be performed liberally and no time should be wasted trying to establish the time of loss of vital signs. The documentation of the time of loss of vital signs may be unreliable taking into account the stressful and suboptimal environment under which the paramedics work. We have saved patients with “documented” absence of vital signs for more than 10 minutes. A second reason for a policy of liberal thoracotomies is the possibility of organ donation. Even if, the brain is irreversibly damaged due to prolonged

INDICATIONS FOR OPERATING ROOM THORACOTOMY Patients with severe hypotension require an emergency thoracotomy without any delays. However, in a significant number of patients the indications for operation or observation are not clear and their management remains a controversial issue. These issues include mild shock, blood loss in the thoracostomy tube, persistent air leaks, transmediastinal GSW’s, and thoracoabdominal injuries. Although many authors believe that any shock on admission following penetrating chest trauma should be an indication for operation, our experience has shown that mild shock is not an absolute indication for thoracotomy. If the hypotension can easily be corrected with intravenous fluids, the patient should be further evaluated by echocardiogram, X-rays, ECG, and in the appropriate cases with CT scan. In a prospective study of 156 hypotensive patients following penetrating chest trauma, we found that only half required an operation (1). The amount of blood loss in the thoracostomy tube has been extensively cited as a reliable indication for thoracotomy. Various arbitrary figures have been reported as critical points for operation: 150–200 ml per hour for 3 or more hours, initial blood loss in the thoracostomy drain of more than 800–100 ml. We believe that these figures are often unreliable indicators of the severity of cardiovascular trauma. A clotted large hemothorax or a cardiac tamponade may not be associated with significant blood loss in the chest drain. On the other hand, peripheral lung lacerations or an intercostal venous injury may be associated with continuous slow bleeding but in the vast majority of cases this bleeding is self-controlled. Our practice is to perform an emergency operation if the initial blood loss is 1,000–1,500 ml. In the rest of the cases the decision to operate or not should be made in conjunction with the hemodynamic condition, the rate of bleeding, and the findings of appropriate investigations. Major air leaks following penetrating chest trauma have been considered by many surgeons as strong indications for operation. In our experience almost all of these leaks heal spontaneously, and it

Penetrating injuries of the chest

Fig. 2A. Mediastinal gunshot wound with a bullet trajectory near the aorta. This patient needs an aortography.

is very uncommon that a surgical intervention is required. In persistent leaks following mediastinal trauma, especially GSW’s, a bronchoscopy is advisable to evaluate the trachea and the major bronchi. In mechanically ventilated patients major air leaks may interfere with effective ventilation. These cases can successfully be managed with high frequency percussive ventilation which maintains a low peak inspiratory pressure, allowing for spontaneous closure of the leak. Transmediastinal GSW’s are associated with a high incidence of severe injuries to vital intrathoracic structures. Most patients with significant cardiovascular injuries are dead on admission or present in severe shock, and an immediate thoracotomy is required. Patients who are hemodynamically stable on admission often do not require surgical intervention. After appropriate investigations about 60 % of these patients can be managed non-operatively (4). The usual evaluation of these injuries includes angiography, esophagography, and endoscopy. At our center we have been using spiral CT scanning to select patients who might benefit from further invasive investigations, such as angiography or endoscopy. The CT scan should be performed with intravenous contrast and its purpose is to delineate the missile trajectory. Trajectories in close proximity to major vessels or the aerodigestive tract require further evaluation by means of angiography and endoscopy or contrastswallow studies. In patients with missile trajectories away from vital structures no further investigations are required (Fig 2A, 2B). In a prospective study of 24 hemodynamically stable patients with mediastinal GSW’s, spiral CT scanning eliminated the need for angiography in 75 % of them. Similarly, on the basis of the CT scan esophageal studies were avoided in 75 % of these patients (5). Patients with penetrating thoracoabdominal injuries, especially on the left side, pose special diagnostic and therapeutic problems. Although physical examination is reliable in identifying significant intraabdominal injuries, diaphragmatic injuries may be asymptomatic. Failure to recognize and repair these

43

Fig. 2B. Mediastinal gunshot wound with trajectory away from the major vessels and the aerodigestive tract. No angiography or esophageal studies are required.

Fig. 3. Left diaphragmatic hernia many years after a stab wound to the left lower chest. Laparoscopy at the acute stage would have prevented this complication.

injuries may result in late diaphragmatic hernias, which are often associated with significant morbidity and mortality (Fig. 3). The incidence of diaphragmatic violation in thoracoabdominal injuries is very high. In a prospective study from our center, 106 patients with left thoracoabdominal injuries underwent a laparotomy in the presence of peritoneal signs or a laparoscopy in otherwise asymptomatic patients. The overall incidence of diaphragmatic injuries was 43 % (63 % in GSW’s and 30 % in stab wounds). In 31 % of proven diaphragmatic injuries there were no signs of peritonitis and in 40 % the chest film was normal (6). In another prospective study, of 110 asymptomatic patients with penetrating trauma to the left thoracoabdominal area, routine laparoscopy revealed diaphragmatic injury in 24 % of cases (26 % for stab wounds and 13 % for GSW’s) (7). Right diaphragmatic injuries rarely progress to diaphragmatic hernias because of the protective anatomical presence of the liver. However, we have seen right diaphragmatic hernias in anterior thoracoabdominal injuries. It is now part of our standard protocols to evaluate lapa-

44

D. Demetriades, G. C. Velmahos

hope that they will be more effective than the first drain. Such efforts are usually futile, increase the risk of empyema, and the delay may make the definitive evacuation more difficult. Our protocol for suspected retained hemothorax is CT scan evaluation within 48 hours of admission and thoracoscopic evacuation within 3–6 days of admission (9). COMBINED PENETRATING CHEST AND ABDOMINAL INJURIES

Fig. 4. Chest X-ray shows a large right hemothorax. CT scan shows a large intrapulmonary hematoma with no significant hemothorax. Planning therapeutic interventions on chest films alone can be dangerous.

roscopically all patients with left thoracoabdominal or anterior right thoracoabdominal injuries who are asymptomatic. We advocate observation for at least 6–8 hours before laparoscopic evaluation in order to exclude any associated intra-abdominal injuries which have not manifested clinically during the early hours of admission. In cases with a retained hemothorax, thoracoscopy may be preferable to a laparoscopy because the clotted hemothorax may be evacuated at the same session during which the diaphragm is evaluated. RESIDUAL HEMOTHORAX Undrained residual hemothorax may become infected or cause fibrosis with lung entrapment and respiratory compromise. Persistent opacifications on chest films following thoracostomy tube insertion may be due to a clotted hemothorax, lung contusion, or atelectasis. The radiological differential diagnosis of these intrapulmonary conditions from a residual hemothorax is often difficult or even impossible (Fig. 4). In a prospective study of 58 patients with suspected residual hemothorax on chest X-rays and with CT scan or thoracoscopy used as gold standards, the radiologist’s and surgeon’s interpretation was wrong in about half the cases (8). It is therefore of major practical importance to evaluate all suspected hemothoraces by CT scan before any therapeutic decision is made. In confirmed large residual hemothorax early evacuation by thoracoscopy or a small anterolateral thoracotomy should be performed within 3–6 days of injury. Longer delays often result in inflammation and organization of the clot, making a subsequent evacuation more difficult, especially if decortication becomes necessary. In a series of 703 patients with traumatic hemothorax only 12 cases (1.7 %) required surgical evacuation of retained hemothorax (8). A common mistake in the management of a persistent large hemothorax is the insertion of a second or even a third thoracostomy tube with the

Multiple penetrating injuries involving both, the chest and abdomen, are not unusual in urban trauma centers. In the presence of hypotension it might be a difficult task to identify the source of the problem and plan the operation. If the clinical examination or a reliable trauma ultrasound clearly identify the source of the hypotension in the chest or the abdomen, the suspect cavity should be explored first. If the chest is opened first, the bleeding is controlled and the cavity is isolated with gauze packs before the abdomen is explored. Care should be taken to avoid contamination of the mediastinum or the pleural cavity from any gastrointestinal spillage. If the patient is hemodynamically stable, the chest cavity should be closed before the abdomen is opened. All patients with penetrating abdominal injuries undergoing anesthesia for another operation should be explored, irrespective of signs or symptoms, because the abdomen cannot reliably be observed perioperatively. If the abdominal cavity is explored first, any bleeding or intestinal spillage are controlled by packing or clamping and a transdiaphragmatic window is performed to rule out cardiac tamponade. If there is no cardiac injury or any evidence of major intrapleural bleeding, as shown by chest X-rays or by the thoracostomy tube output, the patient is spared the thoracotomy. SPECIFIC PENETRATING CHEST INJURIES The diagnosis and management of penetrating cardiac trauma have been previously described in this journal (10) and will not be discussed in this issue. Penetrating lung injuries rarely need operative repair. Due to the low-pressure vascular system and the rich concentration of tissue thromboplastin in the lungs, any bleeding is usually self-controlled. However, in the presence of severe bleeding, usually from a large lung vessel, a thoracotomy is necessary. In some cases with extensive parenchymal damage or hilar injuries, a lobectomy may be necessary. In this situation en-masse stapled pneumonectomy is preferable over an anatomical lobectomy with isolation of the hilar structures, because it is easier, faster, and associated with less blood loss. In situations with significant bleeding or major air leak from a missile or knife tract, a tractotomy with direct control of the vascular or bronchial injury is the preferred technique (11). Suturing of the entry and exit of the tract is ill-advised because it creates an ideal environment

Penetrating injuries of the chest

for air embolism. Injuries to the thoracic great vessels occur in about 5 % of gunshot wounds and 2 % of stab wounds to the chest (12). Most of the victims reach the hospital dead or in severe shock. The overall mortality of thoracic aortic injuries is higher than 90 % and in subclavian vascular injuries about 65 % (12, 13, 14). Many of these victims require an emergency room thoracotomy (about 80 % of aortic injuries and 23 % of subclavian vascular injuries) and the survival is very poor. Injury to the thoracic duct may occur after penetrating trauma in the left supraclavicular area or the mediastinum. If recognized intraoperatively by the presence of milky fluid, ligation of the injured duct should be performed. However, very often the diagnosis is made a few days after the injury by the presence of a chylothorax or chyle in a drain. Although the appearance of the chyle is usually milky, in many cases, especially in parenterally fed patients, it may be clear. The diagnosis is confirmed by the high protein content (> 3 gm %), high total fat content (> 0.4 gm %), high triglyceride levels (> 200 mg %) and marked lymphocytic predominance. Lymphangiography confirms the size and site of thoracic duct injury. Almost all patients with traumatic thoracic duct leaks heal nonoperatively with drainage and low-fat diet or total parenteral nutrition. Surgical intervention is very rarely necessary and should be considered only in patients with persistent major leaks continuing for more than two weeks with no signs of improvement. In these situations open or thoracoscopic ligation of the duct should be considered. REFERENCES 01. Demetriades D, Rabinowitz B, Markides N: Indications for thoracotomy in stab injuries of the chest: a prospective study of 543 patients. Br J Surg 1986;73:880–890 02. Velmahos GC, Degiannis E, Souter I, Alwood AC, Saadia R: Outcome of a strict policy on emergency department thora-

45

cotomies. Arch Surg 1995;130:774–777 03. Asensio JA, Berne JD, Chan L, Murray J, Falabella A, Gomez H, Chahwan S, Velmahos G, Cornwell EE, Belzberg H, Shoemaker W, Berne TV: One hundred five penetrating cardiac injuries: a 2-year prospective evaluation. J Trauma, 1998;44:1073–1082 04. Renz BM, Cara RA, Feliciano DV, Rozycky GS: Transmediastinal gunshot wounds: A prospective study. J Trauma 2000; 48:416–422 05. Hanpeter DE, Demetriades D, Asensio JA, Berne TV, Velmahos G, Murray J: Helical computed tomographic scan in the evaluation of mediastinal gunshot wounds. J Trauma 2000;49: 689–695 06. Murray JA, Demetriades D, Cornwell EE III, Asensio JA, Velmahos G, Belzberg H, Berne TV: Penetrating left thoracoabdominal trauma: The incidence and clinical presentation of diaphragmatic injuries. J Trauma 1997;43:624–626 07. Murray JA, Demetriades D, Asensio JA, Cornwell EE III, Velmahos G, Belzberg H, Berne TV: Occult injuries to the diaphragm: prospective evaluation of laparoscopy in penetrating injuries to the left lower chest. J Am Coll Surg 1998;187: 626–630 08. Velmahos GC, Demetriades D, Chan L, Tatevossian R, Cornwell EE III, Yassa N, Murray JA, Asensio JA, Berne TV: Predicting the need for thoracoscopic evacuation of residual traumatic hemothorax: chest radiography is insufficient. J Trauma 1999;46:65–70 09. Velmahos GC, Demetriades D: Early thoracoscopy for the evacuation of undrained hemothorax. Eur J Surg 1999;165: 924–929 10. Asensio JA, Gambaro E, Forno W, Steinberg D, Tsai KJ, Rowe V, Navarro Nuño I, Leppäniemi A, Demetriades D: Penetrating cardiac injuries: A complex challenge. Ann Chir Gyn 2000;89:155–166 11. Velmahos GC, Baker C, Demetriades D, Goodman J, Murray JA, Asensio JA: Lung-sparing surgery after penetrating trauma using tractotomy, partial lobectomy and pneumonorrhaphy. Arch Surg 1999;134:186–9 12. Demetriades D: Penetrating injuries to the thoracic great vessels. J Card Surg 1997;12:173–180 13. Demetriades D, Theodorou D, Murray J, Asensio JA, Cornwell EE, Velmahos G, Belzberg H, Berne TV: Mortality and prognostic factors in penetrating injuries of the aorta. J Trauma 1996;40:761–763 14. Demetriades D, Chahwan S, Gomez H, Peng R, Velmahos G, Murray J, Asensio J, Bongard F: Penetrating injuries of the subclavian and axillary vessels. J Am Coll Surg 1999;188:290– 295

Received: December 20, 2000