Although the actual incidence of blunt aortic rupture is unknown, autopsy series have documented aortic rupture in 12 to 23% of deaths from blunt trauma.5,61,62 In the United States, the incidence of thoracic aortic injury among motor vehicle collision victims has been approximated by the vehicular crash database to be 1.5%.63 Seventy to eighty percent of these injuries occur in 36- to 40-year-old men.4,64,65 A majority of patients exsanguinate at the scene, but successful resuscitation of the few who survive long enough to be seen in the emergency department depends on accurate diagnosis and intervention in the acute setting. For instance, 75% of patients presenting to the hospital with an aortic rupture after blunt trauma are initially hemodynamically stable,4 but up to 50% die before definitive surgery.64,66
Traumatic aortic disruptions typically occur in motor vehicle drivers, passengers, or pedestrians hit by vehicles.4,5,61 Alcohol or other substance abuse is involved in greater than 40% of these motor vehicle accidents.61 Patients ejected from the vehicle are twice as likely to sustain traumatic aortic injury, and seatbelt use can decrease this risk fourfold.61 Overall, seatbelts have been demonstrated to be more effective than airbags at preventing blunt aortic injury,67 and aortic rupture of both the ascending and descending aorta has been attributed to the deployment of an air bag.68,69 Falls from significant height, crush injuries, and airplane accidents have also caused aortic rupture.4,5,62,70,71
Pathogenesis of Blunt Aortic Injury
The pathogenesis of aortic transection remains controversial and no integrated understanding of these forces has been achieved. The "whiplash" theory proposes that a combination of traction, torsion, shear, and bursting forces interact owing to differential deceleration of tissues within the mediastinum, thereby causing adequate stress to rupture the aorta at the isthmus.61,72–76 Aortic mobility is limited by the ligamentum arteriosum, the left main stem bronchus, and the paired intercostal arteries. Investigations have shown that displacement of the aorta in a longitudinal direction may cause a tear at the isthmus.75 Alternatively, Crass and Associates have argued that the differential forces of deceleration, torsion, and hydrostatics alone have inadequate magnitude in vehicular accidents to result in aortic tearing and have proposed the "osseous pinch" mechanism based on quantifiable thoracic compression.77,78 With this mechanism, anterior thoracic osseous structures (manubrium, first rib, and clavicular heads) rotate posteriorly and inferiorly and may impact the vertebral column, pinching the aortic isthmus and proximal descending thoracic aorta.77 This pinch is theorized to cause shearing of the aorta and some clinical data do support the osseous pinch mechanism.79 Overall, diversity among the direction and magnitude of forces generated by a blunt traumatic event in patients suffering descending thoracic aortic rupture have prohibited identification of a single pathogenetic mechanism.
Traumatic aortic disruptions occur most commonly at the aortic isthmus (Fig. 56-4), and an autopsy series identified 54% at this site, with 8% in the ascending aorta, 2% in the arch, and 11% in the distal descending aorta.62 In those who survive, the periadventitial tissues around the isthmus appear to provide protection against free rupture and allow time for transfer to a hospital; consequently, surgical case series have reported 84 to 97% of aortic ruptures occurring at the isthmus.3,4,80–83 The strength of the aortic wall is in its adventitial layer, and despite the increased incidence of transection at the aortic isthmus, there is no evidence to suggest that the adventitia in this area is deficient.84 Additionally, the structure of the aortic wall surrounding the transection has not demonstrated any defect and atherosclerotic disease does not play a role in this injury.5,61,62 The transverse transection caused by blunt aortic trauma typically involves all three layers of the wall and the edges may be separated by several centimeters.5,62 Occasionally, noncircumferential or partial aortic wall disruptions have been described, particularly posteriorly, and under these circumstances intramural hematomas and focal dissections may occur.5,62,85
Anatomical diagram of the thoracic aorta and its major branches.
The natural history of a traumatic aortic transection has been increasingly investigated as the concept of medical therapy and delayed operative repair has evolved. Previously reported mortality rates drawn from autopsy reports have claimed that 86% of patients with traumatic aortic disruption die at the scene, whereas only 11% survive greater than 6 hours.5 Considering those patients who reach definitive medical care, recent surgical series report mortality rates from 0 to 50% depending on the size of the series.3,4,64–66,70,86–89
Clinical Presentation and Initial Evaluation
The leading cause of death in patients with aortic injury who make it to the hospital remains exsanguinating aortic rupture, which occurs in at least 20% of patients.4 Among patients with blunt aortic injury who are hemodynamically stable on arrival to the emergency room, 4% die in the hospital of aortic rupture before surgical repair.4 Therefore, an organized, efficient, and effective evaluation of these patients is necessary to prevent unnecessary loss of life.
A multitrauma patient should be evaluated according to standard ATLS protocols regardless of whether aortic disruption is suspected. The primary and secondary survey, routine radiographs, and hemodynamic stabilization must be completed before the team can begin investigating specific injuries. The first step in diagnosing a blunt traumatic aortic injury is identifying the at-risk patient. Motor vehicle collisions, falls from height, explosions, and crush injuries have the impact and deceleration forces required to cause aortic transection, therefore these patients should undergo imaging directed at ruling out this potentially fatal injury.2,70,90,91
Ninety-five percent of patients with aortic disruption have associated injuries, and Table 56-3 lists the frequency of associated injuries from data accrued from approximately 50 trauma centers across the United States and Canada (American Association for the Surgery of Trauma [AAST] trial).4 Closed head injuries have been identified in 51% of patients, 46% have had associated rib fractures, 38% have had pulmonary contusions, 20 to 35% have had orthopedic injuries, and only 4% have had concomitant cardiac contusion.4 The mean injury severity score in the AAST trial was 42.1, implying that current advancements in emergency medical resuscitation in the field have provided more patients the opportunity to reach the hospital and receive aggressive definitive care.
Table 56-3 Associated (Surgical) Injuries in Hospitalized Patients with Traumatic Aortic Disruption ||Download (.pdf)
Table 56-3 Associated (Surgical) Injuries in Hospitalized Patients with Traumatic Aortic Disruption
|Schmidt et al154 n = 8O, %||Hilgenberg et al157 n = 51, %||Duhaylongsod et al75 n =108, %||Sturm et al125 n = 37, %||Fabian et al4 n = 274, %||Crestanello et al144 n = 72, %|
|Central nervous system||25||39||34||27||51||49|
| Diaphragm||13|| 2||12|| 7||—|| 8|
| Heart||10||10||18||—|| 4|| 3|
| Rib/clavicle fractures||40||39||55||35||46||—|
| Kidney|| 9||12||11|| 5||—|
| Bowel||10||—||15|| 3|| 7|
| Other abdominal||11||—|| 9|| 8||14|
| Spine|| 5||10||20||—||12||31|
| Maxillofacial|| 5||10||20||—||13||18|
There are often clues evident in the initial evaluation of a trauma patient that can suggest aortic disruption (Table 56-4). As with blunt injury to the arch vessels, patients sustaining descending thoracic aortic rupture are often hemodynamically stable on presentation at the emergency department. Although patients may complain of dyspnea or back pain and display differential blood pressures in the upper versus lower extremities, specific signs or symptoms of aortic rupture have been identified in less than 50% of cases.92–95 Complete de-gloving injury may result in intussusception of the aortic media into the descending thoracic aorta, with resultant "pseudo-coarctation" variable amounts of distal perfusion compromise (Fig. 56-5). In the majority of trauma patients, a supine chest radiograph is obtained as part of the initial evaluation, and the constellation of grossly widened mediastinum, hemothorax, and transient hemodynamic instability on arrival appear to be predictive of early in-hospital death from blunt thoracic aortic injury.96
Table 56-4 Clues that Suggest Aortic Disruption ||Download (.pdf)
Table 56-4 Clues that Suggest Aortic Disruption
Motor vehicle crash >50 km/h
Motor vehicle crash into fixed barrier
No seat belt
Ejection from vehicle
Broken steering wheel
Motorcycle or airplane crash
Pedestrian hit by motor vehicle
Falls greater than 3 meters
Crush or cave-in injuries
Loss of consciousness
Hemodynamic shock (systolic blood pressure <90 mm Hg)
Fracture of sternum, first rib, clavicle, scapula, or multiple ribs
Steering wheel imprint on chest
Unequal extremity blood pressures
Paraplegia or paraparesis
Pseudocoarctation of the aorta caused by circumferential transection of the aorta with the wall flap telescoping into the distal aorta. Blood flow between the proximal and distal aortic segments is maintained by the surrounding adventitia alone.
In the typical hemodynamically stable blunt trauma patient, head and abdominopelvic CT should be conducted to identify closed-head or intra-abdominal injury. Those with an abnormal chest x-ray or a traumatic mechanism consistent with aortic injury should undergo helical CT scan of the chest with intravenous contrast at this time. Operative management of intracranial space-occupying lesions and intra-abdominal hemorrhage takes priority over nonbleeding aortic injuries. Hemodynamically unstable patients with signs of exsanguinating hemorrhage should go directly to the operating room for control of hemorrhage, and TEE may be used to evaluate for aortic injury. Aortography may be useful when helical CT or TEE fail to definitively identify or adequately characterize an aortic injury. Thoracoscopy has also been used to evaluate traumatic hemothoraces, however with experienced practitioners intraoperative TEE has superb sensitivity and specificity for aortic transection.97
Anti-impulse therapy with beta-blockers should be initiated in patients proceeding directly to the operating room for aortic repair, as well as in those selected for delayed repair, to reduce blood pressure and thereby reduce aortic wall stress.90,91,98 In-hospital aortic rupture rates have been reduced through aggressive beta-blockade without adversely affecting the outcome of associated injuries.90
During the evaluation of a blunt trauma patient, an anteroposterior chest radiograph is routinely obtained and ought to be examined for one of the 15 signs that have been associated with aortic rupture (Table 56-5).99 Widening of the mediastinum to a width exceeding 25% of the total chest width, obliteration of the aortic knob, apical pleural capping, and fractures of the sternum, scapula, clavicle, or first rib are some of the most common findings (Fig. 56-6). None of these signs have demonstrated sufficient sensitivity or specificity to effectively rule out aortic injury, however, and up to 40% of patients with aortic rupture have had chest x-ray findings interpreted as normal.70,90,93,100–103 When abnormalities are identified, however, they can aide the practitioner in determining which patients require aggressive imaging to definitively rule out an aortic injury.
Table 56-5 Chest X-Ray Findings Associated with Blunt Aortic Disruption ||Download (.pdf)
Table 56-5 Chest X-Ray Findings Associated with Blunt Aortic Disruption
|Widened mediastinum (>8.0 cm)|
|Mediastinum-to-chest width ratio >0.25|
|Tracheal shift to the patient's right|
|Blurred aortic contour|
|Irregularity or loss of the aortic knob|
|Left apical cap|
|Depression of the left main bronchus|
|Opacification of the aortopulmonary window|
|Right deviation of the nasogastric tube|
|Wide paraspinal lines|
|First rib fracture|
|Any other rib fracture|
|Thoracic spine fracture|
Chest radiograph of an 18-year-old male victim of a motor vehicle collision showing a widened mediastinum.
Approximately 1% of blunt trauma patients have a thoracic aortic injury identified by helical computed tomography (CT).90,100 Since its introduction in the early 1990s, CT has become the screening tool of choice at most medical institutions to detect traumatic aortic rupture because of its availability, speed, and ease of interpretation. Additionally, sensitivities and negative predictive values nearing 100% have been reported for volumetric helical or spiral CT.70,90,103–106 Following injection of nonionic contrast media, 3- to 5-mm thick images of the chest can be obtained in approximately 90 seconds.106 Normal aorta portrays homogeneous enhancement, whereas filling defects, contrast extravasation, intimal flaps, periaortic hematoma, pseudoaneurysm, and mural thrombi may suggest the presence of an aortic injury (Fig. 56-7).106 False-positive studies can occur, however, and a ductus diverticulum remnant is an example of a nontraumatic aortic abnormality that may be mistaken for vessel rupture.106 If CT identifies a luminal or mural aortic irregularity in the absence of periaortic hematoma or vice versa, the diagnosis of aortic rupture should be questioned and pursued with additional imaging. Moreover, the enhanced resolution of CT imaging has allowed identification of minimal aortic injuries, such as small intimal flaps with minimal or no mediastinal changes, that may be safely managed with anti-impulse therapy.90,106,107
Helical computed tomography scan of the chest in a 30-year-old man after a high-speed motor vehicle collision demonstrating an intimal flap in the proximal descending thoracic aorta and periaortic hematoma.
Transesophageal echocardiography (TEE) has become a valuable tool in cardiothoracic surgery because of its ability to image the entire descending thoracic aorta along with portions of the ascending aorta arch and its portability. In unstable blunt trauma patients requiring laparotomy, TEE can be used to evaluate the descending aorta for evidence of rupture, such as a mural flap or a thickened vessel wall concerning for mural thrombus. Multiplanar TEE probes permit acquisition of cross-sectional images at different angles along a single rotational axis. The typical 5- or 7-MHz transducer permits adequate resolution of structures as small as 1 to 2 mm. Doppler mapping of turbulent blood flow near a vessel wall abnormality may be suggestive of blunt aortic disruption, and time-resolved imaging allows evaluation of the movement of anatomical structures, thereby enhancing the ability to define the physiologic consequences of such abnormalities. Chronic atheromatous disease of the aorta can complicate obtaining and interpreting TEE images; therefore, observation of multiple related signs of injury, such as mural flap with a surrounding mediastinal hematoma, is more reliable.
A disadvantage of TEE, and potential inhibitor to its widespread use as a screening tool for aortic injury, is its operator-dependent nature, with sensitivities as low as 63% documented for this modality.108 A prospective comparison of imaging techniques for diagnosis of blunt aortic trauma reported, however, sensitivity and specificity of 93 and 100% for TEE compared with 73 and 100% for helical CT.105 TEE is more invasive than helical CT, but overall the associated risk is low. Contraindications include concomitant injury to the cervical spine, oropharynx, esophagus, or maxillofacial structures.105
The role of aortography in evaluating blunt thoracic injuries was firmly established prior to the advent of noninvasive, sophisticated imaging techniques, and it may still be considered the gold standard. In experienced hands its sensitivity and specificity both approach 100%.109 Intra-arterial digital subtraction is most often used because it allows rapid generation of images (Fig. 56-8). In the past, intravenous digital subtraction was used as well. After injecting intravenous contrast, time-delayed images of the arch and descending aorta were obtained, and although this technique greatly decreased the duration of the procedure, it was abandoned because the diagnostic accuracy for aortic disruption was less than 70%.110 With the availability and speed of helical CT, aortography is now rarely used for diagnosis, but routinely used for endovascular stent-graft placement. This intervention requires a highly trained team of endovascular specialists and can be time consuming; therefore, trauma patients with additional life- or limb-threatening injuries should be otherwise stabilized before entering the endovascular suite. Rates of exsanguination and death of up to 10% have previously been reported during diagnostic aortography, but this incidence has decreased significantly as endovascular proficiency has improved.95,110,111 In fact, complication rates attributed directly to aortography are low, but patients may suffer contrast reactions, contrast nephropathy, groin hematomas, or femoral artery pseudoaneurysms. False-positive studies are usually attributed to atheromata or ductal diverticula.
Digital subtraction arteriogram of an acute traumatic aortic disruption near the isthmus.
Magnetic Resonance Angiography
Vascular structures are well imaged by magnetic resonance angiography (MRA), particularly the thoracic aorta, and its utility in the diagnosis and follow-up of complex aortic disease, including aortic dissections and aneurysms, is firmly established.110,112,113 The time required to capture images inhibits the utility of MRA in the acute evaluation of a trauma patient; however, it may be effective in post-therapeutic surveillance of traumatic thoracic injuries.
Delayed Repair versus Nonoperative Management
Uniformly, hemodynamically unstable blunt trauma patients should bypass diagnostic imaging and be taken to the operating room immediately. During a damage control exploratory laparotomy or thoracotomy, TEE may be conducted to diagnose contained aortic rupture.105 Operative repair of the aorta should not be attempted at this time, however, and this group of patients benefit from immediate transfer to the intensive care unit for further resuscitation. Once hemodynamic stability is achieved, anti-impulse therapy with beta-blockade should be initiated to minimize aortic wall stress.90
Hemodynamically stable patients diagnosed with blunt aortic injury, and lacking severe associated injuries requiring interventions, warrant immediate repair. Treatment of all non–life-threatening injuries should be delayed until after definitive aortic repair. Delayed management has demonstrated safety and effectiveness in carefully selected patients with severe associated injuries or comorbidities.64–66,86,114–118 Patients with thoracic, intraperitoneal, or retroperitoneal hemorrhage, or intracranial bleeding that causes mass effect should be managed with aggressive anti-impulse therapy to minimize the risk of aortic rupture while these injuries are addressed.118 The goal of anti-impulse therapy should be to maintain a systolic blood pressure less than 120 mm Hg and/or a mean arterial pressure less than 80 mm Hg.90 The aortic insult should also be monitored by TEE during surgical repair of concomitant injuries, and routinely imaged with CT during the delayed management period.115 The mortality rate of patients awaiting aortic repair has ranged from 30 to 50%, but the majority of deaths have not been related to the aortic injury.115,116,118 In fact, in one retrospective study of nonoperative management, 5 of 15 patients died during the index hospitalization from head trauma, and the remainder were doing well at 2.5 years of follow-up with stable aortic imaging.115 Additionally, a few small series have demonstrated survival rates of 67 to 72% in select, high-risk patients treated nonoperatively.117,118 In the AAST trial, those presenting in extremis or with evidence of free aortic rupture were excluded, and the mortality rate of patients with associated injuries that precluded initial aortic repair was 55%.4 Therefore, evidence supports operative delay or nonoperative management in select patients with blunt aortic injury who may be considered poor operative candidates.
Blunt thoracic aortic injury has traditionally undergone open repair with interposition graft placement, and this approach has proved to be safe, effective, and durable, thereby establishing it as the standard with which new repair strategies should be compared. Mortality after an open repair has been approximated at 20%, and the morbidity rate may be as high as 14%, largely attributable to the incidence of spinal cord ischemia.119 The popularity of endovascular stent-grafting (EVSG) of traumatic aortic disruptions has grown immensely in the current era because of expected decreases in operating room time, complication rate, morbidity, and mortality, but no Level I data exist to confirm these trends.120 This approach is particularly attractive in multitrauma patients, and several single institution series have reported good short-term outcomes.65,87,88,121–125 Many anatomical details must be considered when pursuing EVSG for traumatic aortic rupture in a young, otherwise healthy patient with a normal-caliber thoracic aorta. The landing zones may not coincide with those expected when the commercially available stent-graft was designed for the treatment of aneurysm disease; therefore, successful placement of the stent-graft relies on individual surgeon ingenuity and long-term durability of these repairs remains unknown. EVSG techniques continue to evolve, however, and are not uniformly applicable; therefore, surgeons treating thoracic aortic disruption must be comfortable with conventional open repair techniques.
The major controversy regarding open operative repair of blunt aortic trauma involves spinal cord protection. Some still report safety and efficacy with a "clamp-and-sew" technique, whereas the majority of surgeons have successfully reduced the historical 10% paraplegia rate through some form of lower body perfusion to minimize spinal cord and visceral organ ischemia.3,4,80,81,83,102,126,127
Arterial Supply to the Spinal Cord
Blood supply to the spinal cord relies on three longitudinal arteries, the anterior spinal artery located in the anterior-median position and supplying 75% of the spinal cord, and the paired posterior spinal arteries located near the nerve roots. Segmental intercostal and lumbar arteries originating from the posterior aspect of the aorta supply a series of unpaired radicular arteries that subsequently contribute flow to the anterior spinal artery. The vertebral arteries also provide radicular branches to supply the anterior spinal artery; therefore, in addition to the risk of cord ischemia during aortic cross-clamp at the isthmus and associated paraplegia rates, clamping proximal to the left subclavian compromises flow into the left vertebral and further threatens the integrity of the spinal cord. The posterior spinal arteries are supplied by smaller radicular arteries that originate from the aorta at nearly every spinal level. The largest and most significant radicular artery, typically originating at the level of T10 and entering the vertebral column at the first lumbar vertebrae, is the arteria radicularis magna (or artery of Adamkiewicz) and this vessel is essential for spinal cord perfusion in nearly 25% of patients.
Some groups report low paraplegia rates exclusively using a "clamp-and-sew" technique; however, these results are not widely reproducible and rely on cross-clamp times of less than 30 minutes.80 Because of its simplicity, this technique may be preferentially employed by the non-cardiothoracic surgeon confronted with repair of a blunt aortic injury. Fragility of the aortic wall, anatomical distortion by the periaortic hematoma, and extension of the defect into the left subclavian artery pose significant obstacles and the average cross-clamp time reported in the literature is 41 minutes.81 Paraplegia rates are greatly reduced and may even approach zero when extracorporeal lower body perfusion techniques are paired with short cross-clamp times (Tables 56-6 and 56-7; Fig. 56-9).4,126,127,169
Table 56-6 Incidence of Postoperative Paraplegia in Relation to Surgical Management: Meta-Analysis ||Download (.pdf)
Table 56-6 Incidence of Postoperative Paraplegia in Relation to Surgical Management: Meta-Analysis
|Operative Technique||Patients, n||Paraplegia, %||Clamp Time, Minutes|
|Partial bypass*|| 71|| 1.4||39.5|
Table 56-7 Incidence of Postoperative Paraplegia in Relation to Surgical Management: AAST Prospective Trial ||Download (.pdf)
Table 56-7 Incidence of Postoperative Paraplegia in Relation to Surgical Management: AAST Prospective Trial
|Operative technique||Patients, n||Paraplegia, %|
|Gott shunt|| 4|| 0|
|Full bypass|| 22|| 4.5|
|Partial bypass|| 39|| 7.7|
|Centrifugal pump|| 69|| 2.9†|
|Clamp and sew|| 73||16.4*,†|
Probability of paraplegia in relation to aortic cross-clamp time with and without lower body perfusion in patients with traumatic aortic disruption at the isthmus. (Reproduced with permission from Katz NM, Blackstone EH, Kirklin JW, Karp RB: Incremental risk factors for spinal cord injury following operation for acute traumatic aortic transection. J Thorac Cardiovasc Surg 1981; 81:669-674.)
Adjuvant Perfusion Techniques
Elective repair of thoracic or thoracoabdominal aneurysms allow employment of several techniques to minimize spinal cord ischemia, but the preoperative preparation required to monitor somatosensory evoked potentials, provide lumbar cerebrospinal fluid drainage, or achieve epidural cooling is typically not available in the trauma setting.126,128–130 Hypothermic circulatory arrest techniques have been successfully applied to injuries involving the arch, but are not practical when partial bypass systems are employed.131,132
The system used by any one group should be routine; however, it is important to be well versed in the various lower body perfusion systems because distinct circumstances may require alterations in practice. Intra-arterial blood pressure monitoring of both the upper and lower limbs should be performed with a goal perfusion pressure of 60 to 70 mm Hg.133 Systemic heparinization poses a significant risk of hemorrhage in the trauma patient, particularly in those with severe lung or intracranial injuries. Use of a centrifugal pump with heparin-bonded tubing and active partial left heart bypass or use of a heparin-bonded passive shunt is an option that does not require systemic heparinization.133,134 Alternatively, safe use of partial left heart bypass with full systemic heparinization has been reported by many centers.4,80,89,90,135,136
For partial left heart bypass, pump inflow is established by cannulating the left atrium through the left inferior pulmonary vein using a small single- or dual-stage cannula (Fig. 56-10). Pulmonary venous cannulation near its confluence with the left atrium has demonstrated a lower complication rate than cannulation of the left atrial appendage.137 Arterial cannulation may occur through the distal descending aorta or the femoral artery. Distal aortic cannulation has the advantage of convenience and speed. Partial left heart bypass serves several purposes: (1) to unload the left heart and control proximal hypertension at the time of cross-clamping; (2) to maintain lower body perfusion; (3) to allow rapid infusion of volume; and (4) to control (remove) intravascular volume. Lower body mean arterial pressure should be maintained at 60 to 70 mm Hg and this can typically be accomplished with a perfusion flow rate of 2 to 3 L/min. Mean arterial pressures from 70 to 80 mm Hg are generated in the upper body by the native heart and ventricular arrhythmias remain a significant risk. The pump reservoir and/or cell saver are employed to return blood from the field and a heat exchanger can be used to maintain core temperatures above 35°C. If the system is used without systemic heparinization, however, heat exchangers and oxygenators should be removed from the circuit to minimize surface area and thrombotic risks.
Diagram showing a typical setup for partial left heart bypass in a patient with aortic disruption at the isthmus.
In full or partial cardiopulmonary bypass, a long venous catheter with multiple side holes may be introduced through the left common femoral vein and placed in the right atrium with a guidewire. Alternatively, direct right atrial cannulation at the inferior vena cava–right atrial junction may be accomplished from a left thoracotomy by simple, transverse, inferior pericardiotomy below the left phrenic nerve. Right atrial–femoral arterial bypass has also been used with or without an oxygenator like partial left heart bypass. A partial arterial oxygen pressure of 40 mm Hg has been reported in non-oxygenated circuits, a level shown to be sufficient for lower body tissue oxygenation when the hemoglobin is maintained at 10 g/dL.138 In cases of aortic arch injury, full cardiopulmonary bypass is beneficial.139,140
Partial or complete bypass may be established before entering the chest by pursuing right femoral venous to arterial bypass, a technique particularly advantageous when a concomitant right lung contusion inhibits oxygenation. In cases of aortic arch transection in proximity to the innominate or left common carotid, anterior exposure via sternotomy or thoracosternotomy may offer better exposure for total arch replacement under deep hypothermic circulatory arrest (HCA).139,140 Use of HCA in trauma patients poses significant bleeding risks; therefore, aortic injuries requiring this technique may be appropriate for anti-impulse therapy and a delay in repair until other concomitant injuries have been addressed. Additionally, aortic valvular insufficiency must be ruled out. When HCA is used within the left chest, the left ventricle should be vented through the left atrium.
A passive (Gott) shunt has been described where the proximal and distal ends of a heparin coated polyvinyl tube are placed in the ascending aorta or arch and the descending aorta or femoral artery, respectively. Ventricular cannulation had been used in the past; however, it was abandoned because of a high rate of ventricular dysrhythmias, reduced shunt flows, and a higher rate of paraplegia.141,142 Flow through this fixed tube is dependent on a pressure gradient and this inability to control flow is the main disadvantage of the Gott shunt.142 Moreover, it offers no left ventricular unloading or loading advantage, as partial bypass systems do; therefore, blood pressure control is left to pharmacology alone.
The patient is positioned in the right lateral decubitus position with the left groin prepped for arterial and venous access. A right radial arterial line is preferred to avoid losing arterial pressure tracing if occlusion of the left subclavian artery is required during the repair. Use of a pulmonary arterial catheter is optional. Selective ventilation of the right lung is required. A standard fourth interspace posterolateral thoracotomy with or without fifth rib notching usually provides excellent exposure to the aortic isthmus and proximal descending aorta. The incision should be long enough to facilitate dissection of the descending aorta below the level of the inferior pulmonary vein and dissection of the arch of the aorta between the left common carotid and left subclavian arteries. In a patient with a prior left thoracotomy, the associated scarring offers both an advantage and disadvantage to the patient. The adhesions between the lung and mediastinum help contain the rupture making it less likely to exsanguinate. Dissection near the isthmus or tear should be avoided until both proximal and distal aortic control is established. Depending on the stability of the patient, lower body perfusion can be established before aortic exposure by gaining access to the left groin. If cannulation is planned in the chest, the left inferior pulmonary vein–left atrial junction is dissected and cannulated, in addition to arterial cannulation of the distal descending thoracic aorta or left common femoral vein. Excessive compression or traction of the lung should be avoided, particularly when dissecting out the aortic arch, because the left pulmonary artery may be easily disrupted at this location.
Distal control is obtained first, usually by fairly simple passage of a blunt instrument or finger around the descending aorta at the distal margin of the hematoma. Care must be taken not to avulse intercostal arterial branches with this maneuver. The subclavian artery is isolated next. Great care is taken to avoid injury to either the phrenic or vagus nerves as they pass over the aortic arch, which can be difficult because they are often obscured by the hematoma. They should be reflected off the aorta with the overlying pleura and retracted medially by attaching stay sutures to the pleura just lateral to the vagus nerve. Loops around the nerves themselves should be avoided, as even stretch of these nerves can result in paresis. This reflection exposes the arch of the aorta between the left common carotid and left subclavian arteries, which is the point needed for proximal aortic control in the majority of cases. Inferiorly, the vagus nerve and its branching left recurrent laryngeal nerve are reflected medially as well. This exposes the ligamentum arteriosum, which can be sharply divided, but usually this step is not required. Careful dissection is then undertaken between the left common carotid artery and left subclavian artery using a combination of sharp and gentle finger dissection to completely encircle the aortic arch with an umbilical tape. As with distal aortic control, the peri-aortic hematoma facilitates this dissection considerably. There should be no dissection distal to either the left subclavian or the ligamentum in order to avoid free disruption of the hematoma.
Lower body perfusion is initiated, and once systemic blood pressure is stabilized the left subclavian artery is clamped (if necessary) followed by the proximal aorta. With modern imaging techniques, it is usually possible to predict, to the millimeter, exactly where the aortic tear is in order to avoid involving it in the proximal clamp. The distal aorta is clamped last. Traumatic aortic disruptions that occur in close proximity (<1 cm) from the left subclavian artery portend a higher mortality risk and greater operative difficulty than injuries further away from the left subclavian ostium.139 Upper and lower body pressures are stabilized with the bypass circuit to maintain upper body mean arterial pressures of 70 to 80 mm Hg and a lower body pressure of 60 to 70 mm Hg with flows of 2 to 3 L/min.
The periaortic hematoma is then entered, and the edges of the transected aorta identified. Usually the aorta is completely transected, and the edges are separated by 2 to 4 cm.5,62 Less frequently the transection is partial. Some authors advocate primary repair at this point;141,143 however, we advocate placing a short interposition graft after debridement of the torn edges in all cases.3,80,83,90,136,144,145 Collagen-coated woven polyester grafts or gelatin-impregnated grafts are used most commonly. Use of intraluminal prostheses has been abandoned by most groups.146 Grafts are sewn using a running polypropylene suture with the proximal anastomosis performed first, followed by the distal. Generous amounts of adventitial tissue are included in each bite. If the proximal anastomosis is done under HCA, cardiopulmonary bypass and reperfusion of the arch should be reinstituted immediately after completion of the proximal anastomosis for optimal neurocerebral protection. This requires cannulation of the graft just beyond the proximal anastomosis (branched grafts are useful here for re-cannulation), and then the distal anastomosis is completed using a dual arterial-inflow perfusion setup perfusing the arch and lower body simultaneously. The left subclavian can either be incorporated into the proximal anastomosis or grafted separately as appropriate.
If it is discovered that the tear extends above the proximal aortic clamp, an attempt can be made to dissect more proximally and place a second clamp provided the left common carotid artery is not compromised. If this is not possible, the best recourse is to cannulate the aortic arch in addition to the distal arterial cannula, commence full cardiopulmonary bypass through both arterial routes, and perform the proximal anastomosis during a brief period of profound HCA. We cool to 20°C and use cerebral protection adjuncts, including packing the head in ice and 15 mg/kg sodium thiopental intravenously. One must be prepared to vent the left ventricular apex if distention occurs as a result of cooling induced ventricular fibrillation. A continuous short axis view by transesophageal echocardiography is very useful is this circumstance.
If the aorta is already ruptured with bleeding into the hemithorax, proximal aortic dissection between the left carotid and subclavian arteries is rapidly performed, and a cross-clamp quickly applied. The descending aorta is then clamped below the injury, and the hematoma opened. No attempt is made to establish lower body perfusion, but every attempt is made at maintaining adequate mean arterial pressure during clamping. The aortic repair is done as expeditiously as possible to minimize clamp time. Repair sutures are placed accordingly after clamps are removed. Hemostasis is achieved after continuity of the aorta is re-established.
Occasionally, the aortic tear will extend into the left subclavian orifice. In this case the proximal clamp may have to partially or totally occlude the left common carotid. The left subclavian can then be completely detached from the aorta, the proximal anastomosis completed, and the clamp then moved distally onto the graft. The left subclavian is then reattached to the aortic graft with an interposition graft after completing the distal aortic anastomosis. The left subclavian interposition graft is fashioned with an end-to-end anastomosis distally and an end-to-side anastomosis proximally.
Following open repair of blunt thoracic aortic injury, the in-hospital mortality rates range from 0 to 20% and the complication rates range from 40 to 50% with pneumonia being most common.4,83,86,92 Bacteremia, renal insufficiency, paraplegia, left vocal cord paralysis, and aortobronchial fistula have also been reported (Table 56-8).4,83,86,92,147,148 A meta-analysis of 1492 patients who underwent open aortic repair, 13.5% died in the postoperative period and 9.9% suffered paraparesis or paraplegia.81 Numerous studies have demonstrated the variation in paraplegia rates, however, and the determining factor appears to be the operative technique used (see Tables 56-6 and 56-7; Fig. 56-9).3,4,64,65,80,81,90–92,117,127 According to current data, perfusing the lower body via partial left heart bypass in association with a cross-clamp time less than 30 minutes will provide the lowest risk for postoperative paraplegia.4,91
Table 56-8 Major Postoperative Complications ||Download (.pdf)
Table 56-8 Major Postoperative Complications
|Schmidt et al141n = 73, %||Cowley et al83n = 51, %||Kodali et al168n = 50, %||Fabian et al4n = 207, %|
|Renal failure|| 9.6|| 9.8|| 4||8.7|
|Left vocal cord paralysis|| 4.1||13.7||14||4.3|
|Left phrenic nerve palsy|| 1.4|| 5.9||—||—|
|Stroke|| 2.7||—|| 4||—|
|Re-exploration for bleeding|| 1.4|| 9.8|| 4.0||—|
|Pulmonary embolism|| 1.4|| 3.9||—||—|
|Deep vein thrombosis|| 2.7||—||—||—|
|Wound infection||—|| 3.9||—||—|
|Deaths|| 8.8||43.1||28.0||14 (31.3)*|
Use of endovascular techniques to treat abdominal aortic aneurysms began in 1991 and has subsequently been expanded to degenerative thoracic aortic aneurysms. In trauma patients with blunt aortic injury, this technology was initially applied to patients considered extremely high risk for open repair, such as those with a head injury, abdominal visceral injury, or severe pulmonary contusions.65,66,121,124 The safety and efficacy of this procedure has since been demonstrated and endovascular stent-grafting (EVSG) is considered the preferred method of treatment by many groups.65,66,87,88,117,121–125,149–152 Theoretical advantages of stent-grafting include avoidance of thoracotomy, one-lung ventilation, the systemic effects of cardiopulmonary bypass, aortic cross-clamping, and spinal cord ischemia, which should decrease perioperative mortality and complications.121,122,124,125
Several anatomical considerations must be addressed when deciding whether a patient is a candidate for EVSG of a blunt aortic transection. A proximal landing zone of at least 1.5 cm is considered necessary to achieve a reliable seal, raising the concern of left arm ischemia if the left subclavian artery needs to be covered; however, this has not been observed.123,151 In the rare case of symptomatic left arm ischemia, elective carotid to subclavian bypass may be performed.123,151 The left subclavian artery origin is also a good marker for an area of acute angulation of the proximal descending aorta and the wall stents used in early EVSG procedures often became distorted in this region.123,153,154 Newer flexible stents have overcome this issue.
Additionally, the graft diameter should be oversized 10 to 20% for appropriate seating.122,123 The size of the chosen graft will determine the route of placement. Early procedures used grafts designed as extension cuffs for abdominal aortic grafts, and these devices had delivery systems 65 cm in length, necessitating iliac or distal aortic access.122,125 Those frequently reported in the literature include the Gore Excluder (W.L. Gore and Associates, Flagstaff, AZ), AneuRx (Medtronic, Santa Rosa, CA), and Zenith (Cook, Inc., Indianapolis, IN). These cuff diameters range from 18 to 28 mm, and the lengths range from 3.3 to 3.75 cm, necessitating placement of several cuffs to achieve adequate coverage.36 Commercially available thoracic aortic stent-grafts have been designed for treatment of aneurysm disease; therefore, the diameters are often too large for the otherwise healthy aorta encountered in a patient with traumatic aortic transection. The Gore TAG (W.L. Gore and Associates, Flagstaff, AZ) fits vessels 23 to 37 mm, and the Talent Valiant (Medtronic, Santa Rosa, CA) may be applied to vessels 20 to 42 mm in diameter, excluding a large percentage of blunt aortic transection victims.154 The ideal device for repair of traumatic aortic injury would be 16 to 40 mm in diameter, be available in 5-, 10-, and 15-cm lengths, and conform to a 90-degree curvature without deformation. Additionally, the delivery system should be distally flexible and approximately 80 to 90 cm in length.36
Patients who will undergo EVSG for traumatic aortic disruption should be positioned supine in a hybrid operating room/angiosuite or conventional operating room table with fluoroscopic capabilities. General endotracheal anesthesia is typically employed. Conversion to open repair is rare, but has occurred because of postdeployment stent migration; therefore, the operative team must be prepared for rapid conversion.87 Retrograde aortic access is accomplished percutaneously or by cutdown on the femoral or iliac artery depending on the chosen stent-graft. A floppy tipped J-wire should be advanced into the aorta under fluoroscopic guidance with subsequent placement of marked catheter. An aortogram should be obtained in steep left anterior oblique projection to clearly visualize the arch. The details of the aortic injury and device measurements are obtained from a CTA during operative planning, but the intraoperative aortogram is vital to confirming appropriate anatomy for EVSG and selection of the correct device. Additionally, the length of graft coverage is determined by this intraoperative image. Deployment of stent-grafts have been successful with and without administration of systemic heparin (Fig. 56-11).155 Intravascular ultrasound may be a useful adjunct when determining coverage length and graft diameter.156,157 Based on the proximity to the aortic injury, the left subclavian artery may need to be covered, and if covered, it may need to be embolized and bypassed or transposed to the left common carotid artery to ensure a proximal graft seal and avoid problems of ischemia to the left arm or vertebrobasilar system.125,158–160
Endovascular stent-graft repair of the blunt traumatic aortic injury displayed in Figure 56-8 demonstrating coverage of the left subclavian artery.
Despite whether single or multiple graft devices have been deployed, successful exclusion of the pseudoaneurysm has been accomplished in 90 to 100% of patients reported in the literature.51,65,88,123–125,161 Decreased operating room time, diminished physiologic derangement and hypothermia, decreased transfusion rates, and shorter intensive care unit and overall hospital stays have been documented in series directly comparing open and endovascular repair of blunt traumatic aortic injury.87,124,125 Procedure related paraplegia and mortality rates are also markedly reduced to nearly zero.87,124,125 Long-term surveillance is recommended to evaluate for endoleak, stent migration, or delayed pseudoaneurysm formation with annual CT angiography.125
Complications of EVSG repair have been rare (Table 56-9). Graft or procedure related mortality has only occurred in 2% of patients.36 The risk of paraplegia is almost entirely eliminated because the length of aorta covered with stent-grafting includes few intercostal vessels, thereby minimally affecting spinal cord perfusion.120,124 Early endoleak has occurred at a rate of 5.1% in stenting of a traumatic blunt aortic injury, with Type I endoleaks being most commonly reported, likely because of difficulty attaining adequate proximal seal in a short neck, heavily calcified arch, or small radius of curvature of the arch.36 Identification during the initial procedure allows application of adjunctive techniques to seal the stent, but many series also report spontaneous resolution.87,122,161,162 Early endoleak has been the cause of stent-related mortality in 0.9% of patients in published series.122,124 Delayed endoleak has been identified in less than 1% of patients on routine surveillance with CTA and successfully managed nonoperatively with spontaneous resolution within 3 to 6 months.36 Rousseau and coworkers reported no incidence of graft kinking, twisting, stenosis, thrombosis, migration, pseudoaneurysm expansion, or rupture; however, one patient in that series suffered acute compression of the left main bronchus with resultant atelectasis requiring placement of silicone endobronchial stent.65 Access related complications, including iliac artery rupture, may occur in about 2.5% of patients because of the large caliber of the stent-graft delivery systems, and this is consistent with the rate documented for elective repair of degenerative thoracic aortic disease.66,163 Although not without risks, current data have demonstrated the safety and efficacy of EVSG in blunt aortic trauma, suggesting it should be preferentially applied to anatomically appropriate injuries.
Table 56-9 Results of Endovascular Stent Grafting (EVSG) for Aortic Transection (with Comparison to Open Repair) ||Download (.pdf)
Table 56-9 Results of Endovascular Stent Grafting (EVSG) for Aortic Transection (with Comparison to Open Repair)
|# of Cases||Mortality||Paraplegia|
|Author||Study Period (years)||Open Repair||EVSG||Open Repair (%)||EVSG (%)||Open Repair (%)||EVSG (%)|
|Cook et al99||20||79||19||24.1||21.1||4||0|
|Pacini et al117||23||51||15|| 7.8||0||5.9||0|
|Rousseau et al65||18||35||29||17 ||0||5.7||0|
|Andrassy et al87||14||16||15||18.8||13.3||12.5||0|
|Ott et al124||11||12|| 6||17||0||17||0|
|Morishita et al88||3||11||18||9||17||0||5.6|
|Reed et al66||5||9||13||11||23||0||0|
|Lachat et al150||N/A||N/A||12||N/A||0||N/A||0|
|Peterson et al125||4||N/A||11||N/A||0||N/A||0|
Rarely, blunt aortic transections may go undiagnosed and form chronic pseudoaneurysms through fibrous reorganization and calcification of the periadventitial tissues.164,165 Ninety percent involve the aortic isthmus, reflective of the apparent protection afforded to this area by mediastinal periadventitial tissues.165–167 Interestingly, these chronic pseudoaneurysms have been diagnosed in patients who had fewer concomitant injuries at the time of their initial trauma and 35% had no associated injuries.164 The most common presenting symptom was chest pain, but other common complaints include dyspnea or cough secondary to compression of the left main stem bronchus, hoarseness owing to stretching of the recurrent nerve, hemoptysis, or dysphagia.164 In a review by Finkelmeier et al., overall mortality at 5 years was 70%. Patients were compared according to operative versus nonoperative management, and those who underwent aortic repair had an operative mortality of nearly 5%, with exsanguination being the leading cause of death.164 Alternatively, in the 60 patients followed nonoperatively, 20 deaths were attributable to the aortic lesion.164 Therefore, although there exists significant surgical risk, operative repair of chronic traumatic aortic pseudoaneurysms does provide a survival benefit.164 Occasionally, completely calcified pseudoaneurysms are identified. These must be completely resected to allow placement of a tube graft without kinking it. The dissection required for this puts the recurrent laryngeal nerve, thoracic duct, esophagus, and pulmonary artery at risk of being violated. These lesions are unlikely to grow significantly and can be followed with serial imaging.