- Email: [email protected]
Traumatic aortic rupture: diagnosis and management
Traumatic Aortic Rupture: Diagnosis and Management James S. Gammie, MD, Ashish S. Shah, MD, Brack G. Hattler, MD, PhD, Robert L. Kormos, MD, Andrew B. Peitzman, MD, Bartley P. Griffith, MD, and Si M. Pham, MD Division of Cardiothoracic Surgery, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, and Department of General and Thoracic Surgery, Duke University Medical Center, Durham, North Carolina
Background. Traumatic aortic rupture is a relatively uncommon lesion that presents the cardiothoracic surgeon with unique challenges in diagnosis and management. To address controversial aspects of this disease, we reviewed our experience. Methods. The study was performed by retrospective chart review. Results. Forty-two patients with traumatic thoracic aortic ruptures were managed between January 1988 and June 1997. Nine arrived without vital signs and died in the emergency department. Admission chest radiographs were normal in 3 patients (12 %) and caused significant delays in diagnosis. Four of 30 patients admitted with vital signs had rupture before thoracotomy and died. Twenty-six underwent aortic repair. In 1 patient repair was performed with simple aortic cross-clamping, whereas a second was managed with a Gott shunt. The
remaining 24 patients had repair with partial left heart bypass. In 1 patient hypothermic circulatory arrest was required. Two patients (7.7%) died. There were no cases of new postoperative paraplegia in the bypass group. There was no morbidity directly attributable to the administration of heparin for cardiopulmonary bypass. Conclusions. In a discrete group of patients with traumatic rupture of the aorta, the rupture will become complete during the first few hours of hospital admission; aggressive medical treatment with b-blockade and vasodilators in the interval before the operation is an essential aspect of management. Active distal circulatory support with partial left-heart bypass provides the optimal means of preventing spinal cord ischemia during repair of acute traumatic aortic rupture. (Ann Thorac Surg 1998;66:1295–300) © 1998 by The Society of Thoracic Surgeons
T
third of repairs in North America are done without bypass [9]. Because of the rarity of this injury, individual experience in its management is necessarily limited. We reviewed our experience at a level one trauma center over the past 9 years to evaluate outcome and to outline critical technical aspects of bypass management and aortic repair.
raumatic rupture of the thoracic aorta (TRA) remains a therapeutic challenge. Parmley and associates [1] defined the natural history of this disease in a classic autopsy series from the presurgical era, observing that 80% of patients die at the scene of injury of free rupture and exsanguination into the chest. When the mediastinal pleura, adventitia, and sometimes part of the aortic wall are spared, the victim will have a mediastinal hematoma of variable size and may survive to reach the hospital. Of these patients, more than 50% succumb to mediastinal hemorrhage over the ensuing week [1]. Aggressive diagnosis and timely operative intervention provide the opportunity to salvage most of these patients. The most feared complication of operative repair of TRA is spinal cord ischemia and paraplegia as a result of thoracic aortic cross-clamping. Despite a vast literature, controversy continues to surround the optimal method of spinal cord protection [2–10]. Although there is substantial evidence suggesting that the use of bypass to support the distal circulation during aortic cross-clamping is the safest approach [11], many surgeons continue to advocate the “clamp and sew” technique [12], and more than one Accepted for publication May 6, 1998. Address reprint requests to Dr Pham, Division of Cardiothoracic Surgery, University of Pittsburgh Medical Center, Suite C-700 PUH, 200 Lothrop St, Pittsburgh, PA 15213 (e-mail: [email protected]).
© 1998 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
Material and Methods Between January 1988 and June 1997 we treated 42 patients with blunt injuries to the thoracic aorta. There were 32 male and 10 female patients. Age ranged from 15 to 83 years, with a mean of 34 years. Three patients sustained falls, 1 patient was a pedestrian struck by a car, 1 patient was hit by a falling tree, and the remaining 37 patients were involved in motor vehicle crashes. Means are reported with standard deviations. Comparison of continuous variables was with Student’s t test.
Results Multiple injuries were the rule: injury severity scores ranged from 26 to 59, with a mean of 40 6 9 (an injury severity score of 40 predicts a mortality of 41%). Nine patients arrived without vital signs. Eight of them under0003-4975/98/$19.00 PII S0003-4975(98)00778-4
1296
GAMMIE ET AL TRAUMATIC AORTIC RUPTURE
Ann Thorac Surg 1998;66:1295–1300
Table 1. Clinical Characteristics of Patients With Free Rupture Early After Admission Patient No. 1 2
3 4 a
Clinical History 74-yo man, fall; 9 L fluid before arrival, L CT in ED: .2 L blood 77-yo man, fall; 2 L blood from L CT before arrival, 10 units PRBCs before arrival 75-yo man, MVA 47-yo man, MVA
Time to Arresta (min)
Medical Therapy
Operating room
75
No
Hemothorax
CAT scan
45
No
Wide Wide
Angiography suite Operating room
65 130
No No
Admission SBP (mm Hg)
Max SBP (mm Hg)
CXR
68
117
Hemothorax
82
114
74 134
144 230
Location of Arrest
Time from admission to arrest.
CAT 5 computed tomographic; MVA 5 motor vehicle accident;
CT 5 chest tube; CXR 5 chest radiograph; ED 5 emergency department; SBP 5 systolic blood pressure; yo 5 year-old.
went emergency thoracotomy in the emergency department and 1 had bilateral chest tubes placed; none survived. Families of 3 elderly patients with multiple injuries and substantial comorbidities declined operative intervention. Four of 30 patients (13%) admitted with vital signs progressed to complete aortic rupture within 2.5 hours of admission, before repair could be completed (Table 1). Three of 4 were hypotensive on arrival and probably had active leaks (2 had hemothoraces on radiographs). A fourth patient was normotensive (134/80 mm Hg) on admission and was resuscitated with 4 L of crystalloid and 2 units of packed red blood cells. Sustained hypertension persisted over the first 2 hours in the hospital (maximum systolic blood pressure 5 240 mm Hg), until he suddenly became hypotensive during angiography. Despite urgent transfer to the operating room, free rupture of the aorta was visualized at thoracotomy and he died. The remaining 26 patients underwent operative repair of their thoracic aortic tears and form the substance of this report. Time from hospital admission to repair ranged from 25 minutes to 127 hours, with a median of 5 hours. Four patients had delays of more than 24 hours to repair. In 3 patients, normal admission chest radiographs were responsible for the delays, whereas the fourth patient had repair postponed because of concern of anticoagulation in the setting of a traumatic brain injury. Diagnosis in all cases except 1 was confirmed with aortography. One patient was taken to the operating room within 25 minutes of arrival on the basis of a highly suggestive chest radiograph and transesophageal echocardiogram. Eight patients (8/26 5 31%) were evaluated with transesophageal echocardiography, which in all but one instance clearly demonstrated a tear. Three patients had normal chest radiographs on admission. Two of these films were available for retrospective review by a trauma radiologist and were confirmed to be normal. Two of these patients had development of Interval mediastinal widening developed 29 and 123 hours after admission in 2 of these patients, whereas the remaining patient was noted to have a new pulse deficit in his left arm 27 hours after arrival. In each case, aortography confirmed the presence of a TRA.
L 5 left;
Max 5 maximum;
All injuries with the exception of two were located in the descending thoracic aorta at the level of the ligamentum arteriosum, just distal to the takeoff of the left subclavian artery. One tear extended proximally to the transverse aorta and necessitated placement of an interposition graft to reconstruct the origin of the left subclavian artery. Another patient had two tears: one at the ligamentum arteriosum and a second on the underside of the aortic arch. All repairs were performed via a left posterolateral thoracotomy. Interposition grafts were used in all patients but 1, in whom a primary repair was done. Ten patients underwent laparotomy and 4 patients underwent orthopedic procedures in addition to thoracotomy. In 1 patient early in the series repair was performed with the “clamp and sew” technique, whereas a passive (aorta to aorta) heparin-bonded (Gott) shunt was used in another. Since then all patients have had distal circulatory support with partial left-heart bypass as an adjunct to aortic repair. Fourteen patients were managed with left atrial to distal (distal aorta or femoral artery) bypass, and 10 were managed with venoarterial bypass (pulmonary artery to distal bypass in 9, femoral venous to femoral artery bypass in 1). Heparin requirements for each of these strategies are listed in Table 2. The heparin dose was higher for the venoarterial bypass group (mean, 316 units/kg) versus the left atrial to distal group (mean, 73 units/kg). In general, patients given heparin for cardiopulmonary bypass had some form of assessment of their intracranial status before the operation. Among 24 pa-
Table 2. Method of Distal Circulatory Support Method LA to distal Venoarterial bypass PA to distal Femoral to femoral Gott shunt (Ao to Ao) Clamp and sew
No. of Patients
Heparin Dose (units/kg, mean [range])
14
73 (0 –125)
9 1 1 1
316 (50 – 480) 300 0 0
Ao 5 aorta; distal 5 distal aorta or femoral artery; PA 5 pulmonary artery.
LA 5 left atrial;
Ann Thorac Surg 1998;66:1295–1300
tients placed on bypass, 14 had negative head computed tomographic scans before the operation. Of the remaining 10 patients, 1 had a small subdural hematoma and was given minimal-dose heparin for partial left-heart bypass, 5 had nonfocal neurologic examinations with a Glasgow coma scale score of 15, 2 had Glasgow coma scale scores less than 15 in association with a nonfocal examination, and 2 were not fully assessed. Comparison of the low-dose heparin group (left atrial to distal bypass) and the full-dose heparin group (venoarterial bypass) revealed no differences in outcome as assessed by duration of mechanical ventilation (median, 9 versus 4; p 5 0.9), hospital length of stay (19 versus 13 days; p 5 0.8), or mortality (there was one death in each group). Median cross-clamp time was 49 minutes. Median cardiopulmonary bypass time was 54 minutes. Femoral artery mean arterial pressure during bypass ranged from 40 to 88 mm Hg, with a mean of 65 6 12 mm Hg. Proximal mean arterial pressure during bypass ranged from 52 to 115 mm Hg, with a mean of 82 6 18 mm Hg. Gradients (proximal 2 femoral mean arterial pressures) ranged from 26 to 53 mm Hg, with a mean of 117 6 24 mm Hg. Bypass circuit flows ranged from 0.9 to 2.4 L/m2 per minute, with a mean of 1.4 6 0.4 L/m2 per minute. Overall hospital mortality for patients undergoing repair was 2 of 26 (7.7%). One death occurred in a 58-yearold woman involved in a motor vehicle accident. Associated injuries included multiple rib fractures and a malleolar fracture. Aortic repair was performed using venoarterial (pulmonary artery to aorta) bypass. Crossclamping of the aorta precipitated hemodynamic deterioration with bradycardic arrest. Postoperatively she suffered two additional cardiac arrests and died on postoperative day 7 of multiple organ failure. Autopsy revealed significant two-vessel coronary artery disease and a small myocardial contusion. The second death was in a 30-year-old man who underwent repair using left atrial to distal aortic bypass. At the time of crossclamping, massive bleeding occurred from a second, previously unrecognized, tear in the underside of the aortic arch. Before control could be secured, he exsanguinated. Two patients suffered paraplegia. One was a 28-yearold man with multiple injuries including a ruptured spleen and a liver laceration necessitating splenectomy and hepatorrhaphy. He also sustained a closed head injury with a subdural hematoma and multiple rib and extremity fractures. Initial diagnosis of his thoracic aortic injury was delayed as a result of a normal admission chest radiograph. A widened mediastinum developed 5 days after admission. He was repaired without heparinization with the use of a Gott (aorta to aorta) shunt. Lower extremity paralysis was noted postoperatively in the intensive care unit, and somatosensory evoked potentials confirmed a spinal cord lesion. The second patient was a 19-year-old man injured in a motor vehicle accident who also had a normal admission chest radiograph. Suspicion of an aortic tear was raised when a marked pulse deficit developed in the left arm 27 hours
GAMMIE ET AL TRAUMATIC AORTIC RUPTURE
1297
Table 3. False-Negative Admission Chest Radiograph Rates
Author
Year
Normal Radiographs/Total (%)
Hilgenberg et al [3] Higgins et al [14] Walls et al [5] Read et al [6] Hunt et al [2] Fabian et al [9] Gammie et al (present study)
1992 1992 1993 1993 1996 1997 1998
3/48 (6.2) 0/19 (0) 4/27 (14.8) 3/16 (8) 6/144 (4) 19/259 (7) 3/26 (12)
after admission. Aortography confirmed the tear at the aortic isthmus. He was noted on arrival to the operating room to be without femoral pulses and to have developed paraplegia. Aortic repair was performed with left atrial to femoral bypass. Postoperative somatosensory evoked potentials identified a lesion at the level of the spinal cord. Median intensive care unit length of stay was 7 days, with a range of 2 to 35 days. Total hospital stays ranged from 7 to 51 days, with a median of 16 days. There were 26 complications in 15 patients: prolonged respiratory failure (vent . 4 days), 10; pneumonia, 7; pancreatitis, 4; vocal cord paresis, 1; acute renal failure (temporary hemodialysis), 1; missed duodenal injury, 1; candidemia, 1; and cortical blindness, 1.
Comment This report details a recent clinical experience with traumatic rupture of the thoracic aorta. This is an uncommon injury: among an average of 1,600 blunt trauma admissions per year at our center during the study period, 4.4 TRAs were diagnosed and 3.2 were repaired. A recent survey of 50 trauma centers in North America documented an average of 2.2 cases of TRA per center per year [9]. The relatively infrequent occurrence of this entity means that few surgeons are able to accrue large personal experiences. We have reviewed our experience to emphasize critical aspects of diagnosis, medical management, techniques of distal circulatory support, and operative repair. Three of 26 patients (12%) in this series had normal admission chest radiographs, even when reviewed retrospectively. These-false negative radiographs caused significant delays in treatment of the underlying thoracic injury. These patients had 30-, 33-, and 127-hour intervals from injury to repair, in comparison with a median time to repair of 5 hours for the entire series. Delays in treatment in 1 case almost certainly and in another likely contributed to paraplegia. Woodring and King [13] reported two normal chest radiographs in a series of 32 patients with known thoracic aortic transections, for a false-negative rate of 6%. Other recent surgical series [3, 5, 6, 14] have reported false-negative rates between 0% and 15% (Table 3). Our experience is at odds with the radiology literature, which suggests that a normal chest radiograph is strong
1298
GAMMIE ET AL TRAUMATIC AORTIC RUPTURE
evidence for the absence of an underlying aortic injury. Mirvis and colleagues [15] reviewed the admission chest radiographs of 205 patients with blunt trauma undergoing aortography, 41 of whom were found to have aortic tears. They concluded that a normal interpretation of an erect chest radiograph has a negative predictive value of 98%. Marnocha and Maglinte [16] concluded that a normal aortic arch combined with the absence of tracheal and nasogastric tube deviation ruled out an aortic injury. Our series, in accordance with previous reports in the surgical literature, demonstrates a surprisingly high incidence (12%) of normal admission chest radiographs in patients with TRA. A high level of suspicion is necessary to avoid the catastrophic results of a missed aortic injury. Although some reports have suggested that medical therapy and deliberate delayed operative therapy is acceptable management of a TRA [17], 4 of 30 patients (13%) in this series admitted with vital signs had rupture within 2.5 hours of arrival and died. Of these 4 patients experiencing free rupture during initial diagnostic and resuscitative measures, it is clear that at least 2 had substantial leaks before arrival. Three of 4 were hypotensive at the time of admission. The 1 patient with sustained hypertension during the first 2 hours after admission likely would have benefited from pharmacologic blood pressure control. Few patients in our series received medical therapy (b-blockade, antihypertensives), in accord with practice across North America: Fabian and associates [9] documented that only 17% of patients with TRA received medical therapy. Our experience suggests that the threat of completion of rupture of a thoracic aortic injury in the early hours after admission is real, and institution of aggressive blood pressure control during expeditious evaluation should not be overlooked. b-Blockade, possibly in combination with vasodilators and analgesics, should be given to all hemodynamically stable patients undergoing evaluation for a possible TRA. The presence of hypotension not readily explained by identified injuries should increase concern for imminent rupture. All patients except 2 in this series were supported intraoperatively with partial left-heart bypass. The optimal method of preventing spinal cord injury in the patient with TRA has been a source of considerable debate in the literature [3–10]. The approach to intraoperative protection of the spinal cord during repair of TRA has evolved. Initial reports of repair on full cardiopulmonary bypass emphasized prohibitive rates of hemorrhagic complications. Mortality was attributed to uncontrolled bleeding as a sequela of obligatory systemic heparinization for cardiopulmonary bypass. However, in many of these reports investigation of intraabdominal injuries was deferred until aortic repair was completed, with predictable adverse consequences. Gott [18] introduced the 9-mm heparin-bonded shunt with inflow from either the proximal aorta or left ventricular apex and outflow to the descending thoracic aorta or the femoral artery, as a means of supporting the spinal cord and viscera [19]. Experimental work by Molina and associates [20] suggested that the Gott shunt provided inadequate
Ann Thorac Surg 1998;66:1295–1300
Table 4. Summary of Morbidity and Mortality With Partial Left Heart Bypass Author
Year Patients Paraplegia
Benckart et al [23] Hess et al [7] Higgins et al [14] Read et al [6] Walls et al [5] Forbes and Ashbaugh [8] Nicolosi et al [24] Fabian et al [9] Gammie et al (present study)
1989 1989 1992 1993 1993 1994
21 16 10 16 8 21
0 0 0 0 0 0
3 1 0 2 1 4
1996 1997 1998
18 43 24
0 2 0
... 7 2
2 (1.1%)
20 (11.3%)
Total
177
Mortality
distal flow, and clinical data confirmed high rates of paraplegia. Mattox and colleagues [21] and others advocated simple cross-clamping of the aorta and repair of the injury without distal circulatory support. The “clamp and sew” technique emphasizes speed of repair, with many authors demonstrating a benefit to cross-clamp times less than 30 minutes [19]. Olivier and associates [22] introduced the technique of left atrial to femoral bypass with the centrifugal pump in 1984 and began the current trend favoring this technique for mechanical distal circulatory support. A recent metaanalysis of articles reporting outcomes of treatment of traumatic thoracic aortic tears in more than 1,000 patients provides strong support for the use of mechanical partial left heart bypass during repair of TRA [11]. Overall paraplegia rates were 19.2% for simple cross-clamping, 11.1% for passive (Gott) shunting, and 2.3% for active augmentation of distal perfusion. These data also emphasized the rapid increase in the incidence of paraplegia that occurs with cross-clamp times in excess of 30 minutes when the “clamp and sew” technique is used. Our experience with paraplegia after repair performed with active distal circulatory support is similar to recent series reporting minimal neurologic morbidity with partial left-heart bypass. In 177 patients undergoing repair of TRA with distal circulatory support, the incidence of paraplegia was only 1.1% (Table 4). Paraplegia did not develop in any patient in our series arriving in the operating room in hemodynamically stable condition and without evidence of preexisting neurologic compromise. Despite the compelling weight of published evidence supporting the role of distal circulatory support in minimizing the devastating complication of paraplegia after TRA, support for the “clamp and sew” technique remains strong [12]. In general, these authors acknowledge the importance of the 30-minute ceiling and express confidence in their ability to complete the repair in less than this time. Fabian and associates’ [9] survey of North American trauma centers documented a cross-clamp time in excess of 30 minutes in 67% of cases. Since the mid 1980s, our approach to distal perfusion during repair of traumatic transections of the aorta has
Ann Thorac Surg 1998;66:1295–1300
almost exclusively been one of partial left-heart bypass (see Table 2). We carry out partial left-heart bypass in one of two ways. Left atrial to femoral artery (or descending aorta) bypass is usually established with minimal heparin (5,000 units) and consists of a simple circuit with only a centrifugal pump. Recently we have cannulated the posterior surface of the inferior left pulmonary vein with a 24F cannula in lieu of the atrial appendage. This has provided flow of 2 to 3 L/min and is an effective nonpharmacologic means of controlling blood pressure in proximal and distal circulations. In contrast, venoarterial bypass requires full systemic heparinization and a complete cardiopulmonary bypass circuit, including an oxygenator, heat exchanger, and cardiotomy. For venoarterial bypass, we generally cannulate the pulmonary artery with a 32F right-angle cannula (TF-032-L-90; Research Medical Inc, Midvale, UT) directed toward the right ventricle for inflow and the femoral artery or descending aorta for outflow. The pulmonary artery is a reasonable option for inflow when venoarterial bypass is used because it is easily accessible via a left thoracotomy and offers adequate inflow and ease of cannulation. An excellent (and perhaps superior) alternative for inflow for venoarterial bypass is venous cannulation of the femoral vein (femorofemoral bypass). Venoarterial bypass offers several advantages over left atrial to femoral bypass. These include the ability to cool to 34°C (which may enhance spinal cord protection) and later rewarm the patient (of particular importance in the coagulopathic trauma population). An oxygenator in the circuit avoids difficulties in maintaining adequate oxygenation of the patient during single-lung ventilation. In our experience, intubation with a double-lumen tube is not always possible (eg, massive orofacial injuries) and the presence of venoarterial bypass permits discontinuation of ventilation to achieve critical exposure. Venoarterial bypass also affords the surgeon the option to perform hypothermic circulatory arrest should a complex aortic tear be encountered, as was necessary in 1 patient in this series. By using venoarterial bypass, the perfusionist can rapidly add volume to the circuit, a capability not present with the left atrial to femoral circuit without adjunctive rapid infusion systems. Cross-clamping of the aorta predictably leads to proximal hypertension, distal hypotension, and a markedly increased left ventricular afterload that may necessitate the concomitant administration of vasodilators, inotropic agents, and the administration of volume in the patient on left atrial to femoral bypass. Administration of vasodilators during aortic cross-clamping can initiate a deleterious steal of perfusion from the spinal cord. Venoarterial bypass mitigates these concerns by permitting complete unloading of the left ventricle and affording a means of rapidly replenishing volume. Arterial pressures are maintained at desired levels by simply adding or removing blood from the circuit. Finally, shed blood can be returned to the bypass circuit, and venoarterial bypass offers superior support to a failing heart. The principal disadvantage of venoarterial (pulmonary arterial to femoral) bypass is the requirement for full
GAMMIE ET AL TRAUMATIC AORTIC RUPTURE
1299
heparinization. Although long-standing dogma held that heparinization in this population of patients was highly undesirable, Pate and associates [10] have recently pointed out that there are few data to support this belief. In our experience, most patients were assessed for intracranial and intraabdominal injury before administration of heparin for bypass. We did not observe bleeding complications resulting from heparinization in any patient in this series. In addition, comparison of the lowdose (left atrial to femoral) and full-dose (venoarterial) heparin treatment groups revealed no relationship between heparin dose and outcome. Technical aspects of safe repair include a generous fourth intercostal space thoracotomy for optimal exposure and repair of thoracic aortic injuries. Double-lumen intubation should be performed when feasible to permit single-lung ventilation and provide wide exposure. Proximal control of the aortic arch should be between the left carotid and subclavian arteries, with separate control of the subclavian artery. Bypass is established and dissection commenced. Once the aorta has been clamped, the extent of the tear is assessed, and the distal clamp is repositioned as close as possible to the injury to allow maximal perfusion of the intercostal vessels from the distal pump flow. The adventitia surrounding the injury is typically distorted by hematoma and invariably has retracted. It is critically important to incorporate generous amounts of adventitia within the suture line. In conclusion, traumatic rupture of the aorta is an uncommon injury that requires aggressive diagnosis and treatment. Our review has emphasized several important aspects of the management of this disease. A TRA can be present despite a normal chest radiograph, and the resulting delays in treatment can lead to significant morbidity. Progression of TRA to complete rupture can occur within hours of hospital admission. Judicious administration of a b-blocker and vasodilators to patients with a suspected TRA should begin in the emergency department and continue to operation. Distal circulatory support is critical for effective spinal cord protection. Administration of heparin for bypass, in our experience, was not associated with significant morbidity. Management options for bypass include left atrial to distal and venoarterial bypass. The decision as to which to use should be tailored to the specific clinical situation.
References 1. Parmley LF, Mattingly TW, Manion WC, Jahnke EJ. Nonpenetrating traumatic injury of the aorta. Circulation 1958; 17:1086-101. 2. Hunt JP, Baker CC, Lentz CW, et al. Thoracic aorta injuries: management and outcome of 144 patients. J Trauma 1996;40: 547–56. 3. Hilgenberg AD, Logan DL, Akins CW, et al. Blunt injuries of the thoracic aorta. Ann Thorac Surg 1992;53:233–9. 4. DelRossi AJ, Cernaianu AC, Madden LD, et al. Traumatic disruptions of the thoracic aorta: treatment and outcome. Surgery 1990;108:864 –70. 5. Walls JT, Boley TM, Curtis JJ, Schmaltz RA. Experience with four surgical techniques to repair traumatic aortic pseudoaneurysm. J Thorac Cardiovasc Surg 1993;106:283–7.
1300
GAMMIE ET AL TRAUMATIC AORTIC RUPTURE
6. Read RA, Moore EE, Moore FA, Haenel JB. Partial left heart bypass for thoracic aorta repair. Survival without paraplegia. Arch Surg 1993;128:746 –52. 7. Hess PJ, Howe HR Jr, Robicsek F, et al. Traumatic tears of the thoracic aorta: improved results using the Bio-Medicus pump. Ann Thorac Surg 1989;48:6 –9. 8. Forbes AD, Ashbaugh DG. Mechanical circulatory support during repair of thoracic aortic injuries improves morbidity and prevents spinal cord injury. Arch Surg 1994;129:494 – 8. 9. Fabian TC, Richardson JD, Croce MA, et al. Prospective study of blunt aortic injury: Multicenter Trial of the American Association for the Surgery of Trauma. J Trauma 1997; 42:374 – 83. 10. Pate JW, Fabian TC, Walker WA. Acute traumatic rupture of the aortic isthmus: repair with cardiopulmonary bypass. Ann Thorac Surg 1995;59:90 –9. 11. Von Oppell UO, Dunne TT, De Groot MK, Zilla P. Traumatic aortic rupture: twenty-year metaanalysis of mortality and risk of paraplegia. Ann Thorac Surg 1994;58:585–93. 12. Sweeney MS, Young DJ, Frazier OH, Adams PR, Kapusta MO, Macris MP. Traumatic aortic transections: eight-year experience with the “clamp-sew” technique. Ann Thorac Surg 1997;64:384 – 8. 13. Woodring JH, King JG. The potential effects of radiographic criteria to exclude aortography in patients with blunt chest trauma. Results of a study of 32 patients with proved aortic or brachiocephalic arterial injury. J Thorac Cardiovasc Surg 1989;97:456 – 60. 14. Higgins RS, Sanchez JA, DeGuidis L, et al. Mechanical circulatory support decreases neurologic complications in the treatment of traumatic injuries of the thoracic aorta. Arch Surg 1992;127:516 –9. 15. Mirvis SE, Bidwell JK, Buddemeyer EU, et al. Value of chest
Ann Thorac Surg 1998;66:1295–1300
16. 17. 18. 19.
20.
21.
22.
23. 24.
radiography in excluding traumatic aortic rupture. Radiology 1987;163:487–93. Marnocha KE, Maglinte DD. Plain-film criteria for excluding aortic rupture in blunt chest trauma. AJR Am J Roentgenol 1985;144:19 –21. Maggisano R, Nathens A, Alexandrova NA, et al. Traumatic rupture of the thoracic aorta: should one always operate immediately? Ann Vasc Surg 1995;9:44 –52. Gott VL. Heparinized shunts for thoracic vascular operations. Ann Thorac Surg 1972;14:219 –20. 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 –74. Molina JE, Cogordan J, Einzig S, et al. Adequacy of ascending aorta-descending aorta shunt during cross-clamping of the thoracic aorta for prevention of spinal cord injury. J Thorac Cardiovasc Surg 1985;90:126 –36. Mattox KL, Holzman M, Pickard LR, Beall AC Jr, DeBakey ME. Clamp/repair: a safe technique for treatment of blunt injury to the descending thoracic aorta. Ann Thorac Surg 1985;40:456 – 63. Olivier HF Jr, Maher TD, Liebler GA, Park SB, Burkholder JA, Magovern GJ. Use of the BioMedicus centrifugal pump in traumatic tears of the thoracic aorta. Ann Thorac Surg 1984;38:586 –91. Benckart DH, Magovern GJ, Liebler GA, et al. Traumatic aortic transection: repair using left atrial to femoral bypass. J Card Surg 1989;4:43–9. Nicolosi AC, Almassi GH, Bousamra M II, Haasler GB, Olinger GN. Mortality and neurologic morbidity after repair of traumatic aortic disruption. Ann Thorac Surg 1996;61:875– 8.
J. Kent Trinkle, MD 1934 –1998 It is with great sadness that we report the death of Dr Kent Trinkle, member of the Editorial Board of The Annals of Thoracic Surgery for 15 years, and Associate Editor for 8 years, first in charge of Book Reviews, and later the Section on Collective and Current Reviews. In addition to his duties as Chief of the Division of Cardiothoracic Surgery at the University of Texas Health Science Center in San Antonio, plus his numerous other commitments at the local, state, and national level, Kent was always a faithful supporter of our journal. He de-
© 1998 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
voted a great deal of time and effort to improving The Annals, and for this we are grateful beyond measure. We shall miss him as an Editorial Board member, but much more as a superb human being. I invite you to read the homilies about Dr Trinkle written by his two long-time associates, Dr Fred Grover and Dr John Calhoon, in the Transitions Section of the CTSNet. Tom Ferguson Editor, The Annals of Thoracic Surgery
Ann Thorac Surg 1998;66:1300
•
0003-4975/98/$19.00
Background. Traumatic aortic rupture is a relatively uncommon lesion that presents the cardiothoracic surgeon with unique challenges in diagnosis and management. To address controversial aspects of this disease, we reviewed our experience. Methods. The study was performed by retrospective chart review. Results. Forty-two patients with traumatic thoracic aortic ruptures were managed between January 1988 and June 1997. Nine arrived without vital signs and died in the emergency department. Admission chest radiographs were normal in 3 patients (12 %) and caused significant delays in diagnosis. Four of 30 patients admitted with vital signs had rupture before thoracotomy and died. Twenty-six underwent aortic repair. In 1 patient repair was performed with simple aortic cross-clamping, whereas a second was managed with a Gott shunt. The
remaining 24 patients had repair with partial left heart bypass. In 1 patient hypothermic circulatory arrest was required. Two patients (7.7%) died. There were no cases of new postoperative paraplegia in the bypass group. There was no morbidity directly attributable to the administration of heparin for cardiopulmonary bypass. Conclusions. In a discrete group of patients with traumatic rupture of the aorta, the rupture will become complete during the first few hours of hospital admission; aggressive medical treatment with b-blockade and vasodilators in the interval before the operation is an essential aspect of management. Active distal circulatory support with partial left-heart bypass provides the optimal means of preventing spinal cord ischemia during repair of acute traumatic aortic rupture. (Ann Thorac Surg 1998;66:1295–300) © 1998 by The Society of Thoracic Surgeons
T
third of repairs in North America are done without bypass [9]. Because of the rarity of this injury, individual experience in its management is necessarily limited. We reviewed our experience at a level one trauma center over the past 9 years to evaluate outcome and to outline critical technical aspects of bypass management and aortic repair.
raumatic rupture of the thoracic aorta (TRA) remains a therapeutic challenge. Parmley and associates [1] defined the natural history of this disease in a classic autopsy series from the presurgical era, observing that 80% of patients die at the scene of injury of free rupture and exsanguination into the chest. When the mediastinal pleura, adventitia, and sometimes part of the aortic wall are spared, the victim will have a mediastinal hematoma of variable size and may survive to reach the hospital. Of these patients, more than 50% succumb to mediastinal hemorrhage over the ensuing week [1]. Aggressive diagnosis and timely operative intervention provide the opportunity to salvage most of these patients. The most feared complication of operative repair of TRA is spinal cord ischemia and paraplegia as a result of thoracic aortic cross-clamping. Despite a vast literature, controversy continues to surround the optimal method of spinal cord protection [2–10]. Although there is substantial evidence suggesting that the use of bypass to support the distal circulation during aortic cross-clamping is the safest approach [11], many surgeons continue to advocate the “clamp and sew” technique [12], and more than one Accepted for publication May 6, 1998. Address reprint requests to Dr Pham, Division of Cardiothoracic Surgery, University of Pittsburgh Medical Center, Suite C-700 PUH, 200 Lothrop St, Pittsburgh, PA 15213 (e-mail: [email protected]).
© 1998 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
Material and Methods Between January 1988 and June 1997 we treated 42 patients with blunt injuries to the thoracic aorta. There were 32 male and 10 female patients. Age ranged from 15 to 83 years, with a mean of 34 years. Three patients sustained falls, 1 patient was a pedestrian struck by a car, 1 patient was hit by a falling tree, and the remaining 37 patients were involved in motor vehicle crashes. Means are reported with standard deviations. Comparison of continuous variables was with Student’s t test.
Results Multiple injuries were the rule: injury severity scores ranged from 26 to 59, with a mean of 40 6 9 (an injury severity score of 40 predicts a mortality of 41%). Nine patients arrived without vital signs. Eight of them under0003-4975/98/$19.00 PII S0003-4975(98)00778-4
1296
GAMMIE ET AL TRAUMATIC AORTIC RUPTURE
Ann Thorac Surg 1998;66:1295–1300
Table 1. Clinical Characteristics of Patients With Free Rupture Early After Admission Patient No. 1 2
3 4 a
Clinical History 74-yo man, fall; 9 L fluid before arrival, L CT in ED: .2 L blood 77-yo man, fall; 2 L blood from L CT before arrival, 10 units PRBCs before arrival 75-yo man, MVA 47-yo man, MVA
Time to Arresta (min)
Medical Therapy
Operating room
75
No
Hemothorax
CAT scan
45
No
Wide Wide
Angiography suite Operating room
65 130
No No
Admission SBP (mm Hg)
Max SBP (mm Hg)
CXR
68
117
Hemothorax
82
114
74 134
144 230
Location of Arrest
Time from admission to arrest.
CAT 5 computed tomographic; MVA 5 motor vehicle accident;
CT 5 chest tube; CXR 5 chest radiograph; ED 5 emergency department; SBP 5 systolic blood pressure; yo 5 year-old.
went emergency thoracotomy in the emergency department and 1 had bilateral chest tubes placed; none survived. Families of 3 elderly patients with multiple injuries and substantial comorbidities declined operative intervention. Four of 30 patients (13%) admitted with vital signs progressed to complete aortic rupture within 2.5 hours of admission, before repair could be completed (Table 1). Three of 4 were hypotensive on arrival and probably had active leaks (2 had hemothoraces on radiographs). A fourth patient was normotensive (134/80 mm Hg) on admission and was resuscitated with 4 L of crystalloid and 2 units of packed red blood cells. Sustained hypertension persisted over the first 2 hours in the hospital (maximum systolic blood pressure 5 240 mm Hg), until he suddenly became hypotensive during angiography. Despite urgent transfer to the operating room, free rupture of the aorta was visualized at thoracotomy and he died. The remaining 26 patients underwent operative repair of their thoracic aortic tears and form the substance of this report. Time from hospital admission to repair ranged from 25 minutes to 127 hours, with a median of 5 hours. Four patients had delays of more than 24 hours to repair. In 3 patients, normal admission chest radiographs were responsible for the delays, whereas the fourth patient had repair postponed because of concern of anticoagulation in the setting of a traumatic brain injury. Diagnosis in all cases except 1 was confirmed with aortography. One patient was taken to the operating room within 25 minutes of arrival on the basis of a highly suggestive chest radiograph and transesophageal echocardiogram. Eight patients (8/26 5 31%) were evaluated with transesophageal echocardiography, which in all but one instance clearly demonstrated a tear. Three patients had normal chest radiographs on admission. Two of these films were available for retrospective review by a trauma radiologist and were confirmed to be normal. Two of these patients had development of Interval mediastinal widening developed 29 and 123 hours after admission in 2 of these patients, whereas the remaining patient was noted to have a new pulse deficit in his left arm 27 hours after arrival. In each case, aortography confirmed the presence of a TRA.
L 5 left;
Max 5 maximum;
All injuries with the exception of two were located in the descending thoracic aorta at the level of the ligamentum arteriosum, just distal to the takeoff of the left subclavian artery. One tear extended proximally to the transverse aorta and necessitated placement of an interposition graft to reconstruct the origin of the left subclavian artery. Another patient had two tears: one at the ligamentum arteriosum and a second on the underside of the aortic arch. All repairs were performed via a left posterolateral thoracotomy. Interposition grafts were used in all patients but 1, in whom a primary repair was done. Ten patients underwent laparotomy and 4 patients underwent orthopedic procedures in addition to thoracotomy. In 1 patient early in the series repair was performed with the “clamp and sew” technique, whereas a passive (aorta to aorta) heparin-bonded (Gott) shunt was used in another. Since then all patients have had distal circulatory support with partial left-heart bypass as an adjunct to aortic repair. Fourteen patients were managed with left atrial to distal (distal aorta or femoral artery) bypass, and 10 were managed with venoarterial bypass (pulmonary artery to distal bypass in 9, femoral venous to femoral artery bypass in 1). Heparin requirements for each of these strategies are listed in Table 2. The heparin dose was higher for the venoarterial bypass group (mean, 316 units/kg) versus the left atrial to distal group (mean, 73 units/kg). In general, patients given heparin for cardiopulmonary bypass had some form of assessment of their intracranial status before the operation. Among 24 pa-
Table 2. Method of Distal Circulatory Support Method LA to distal Venoarterial bypass PA to distal Femoral to femoral Gott shunt (Ao to Ao) Clamp and sew
No. of Patients
Heparin Dose (units/kg, mean [range])
14
73 (0 –125)
9 1 1 1
316 (50 – 480) 300 0 0
Ao 5 aorta; distal 5 distal aorta or femoral artery; PA 5 pulmonary artery.
LA 5 left atrial;
Ann Thorac Surg 1998;66:1295–1300
tients placed on bypass, 14 had negative head computed tomographic scans before the operation. Of the remaining 10 patients, 1 had a small subdural hematoma and was given minimal-dose heparin for partial left-heart bypass, 5 had nonfocal neurologic examinations with a Glasgow coma scale score of 15, 2 had Glasgow coma scale scores less than 15 in association with a nonfocal examination, and 2 were not fully assessed. Comparison of the low-dose heparin group (left atrial to distal bypass) and the full-dose heparin group (venoarterial bypass) revealed no differences in outcome as assessed by duration of mechanical ventilation (median, 9 versus 4; p 5 0.9), hospital length of stay (19 versus 13 days; p 5 0.8), or mortality (there was one death in each group). Median cross-clamp time was 49 minutes. Median cardiopulmonary bypass time was 54 minutes. Femoral artery mean arterial pressure during bypass ranged from 40 to 88 mm Hg, with a mean of 65 6 12 mm Hg. Proximal mean arterial pressure during bypass ranged from 52 to 115 mm Hg, with a mean of 82 6 18 mm Hg. Gradients (proximal 2 femoral mean arterial pressures) ranged from 26 to 53 mm Hg, with a mean of 117 6 24 mm Hg. Bypass circuit flows ranged from 0.9 to 2.4 L/m2 per minute, with a mean of 1.4 6 0.4 L/m2 per minute. Overall hospital mortality for patients undergoing repair was 2 of 26 (7.7%). One death occurred in a 58-yearold woman involved in a motor vehicle accident. Associated injuries included multiple rib fractures and a malleolar fracture. Aortic repair was performed using venoarterial (pulmonary artery to aorta) bypass. Crossclamping of the aorta precipitated hemodynamic deterioration with bradycardic arrest. Postoperatively she suffered two additional cardiac arrests and died on postoperative day 7 of multiple organ failure. Autopsy revealed significant two-vessel coronary artery disease and a small myocardial contusion. The second death was in a 30-year-old man who underwent repair using left atrial to distal aortic bypass. At the time of crossclamping, massive bleeding occurred from a second, previously unrecognized, tear in the underside of the aortic arch. Before control could be secured, he exsanguinated. Two patients suffered paraplegia. One was a 28-yearold man with multiple injuries including a ruptured spleen and a liver laceration necessitating splenectomy and hepatorrhaphy. He also sustained a closed head injury with a subdural hematoma and multiple rib and extremity fractures. Initial diagnosis of his thoracic aortic injury was delayed as a result of a normal admission chest radiograph. A widened mediastinum developed 5 days after admission. He was repaired without heparinization with the use of a Gott (aorta to aorta) shunt. Lower extremity paralysis was noted postoperatively in the intensive care unit, and somatosensory evoked potentials confirmed a spinal cord lesion. The second patient was a 19-year-old man injured in a motor vehicle accident who also had a normal admission chest radiograph. Suspicion of an aortic tear was raised when a marked pulse deficit developed in the left arm 27 hours
GAMMIE ET AL TRAUMATIC AORTIC RUPTURE
1297
Table 3. False-Negative Admission Chest Radiograph Rates
Author
Year
Normal Radiographs/Total (%)
Hilgenberg et al [3] Higgins et al [14] Walls et al [5] Read et al [6] Hunt et al [2] Fabian et al [9] Gammie et al (present study)
1992 1992 1993 1993 1996 1997 1998
3/48 (6.2) 0/19 (0) 4/27 (14.8) 3/16 (8) 6/144 (4) 19/259 (7) 3/26 (12)
after admission. Aortography confirmed the tear at the aortic isthmus. He was noted on arrival to the operating room to be without femoral pulses and to have developed paraplegia. Aortic repair was performed with left atrial to femoral bypass. Postoperative somatosensory evoked potentials identified a lesion at the level of the spinal cord. Median intensive care unit length of stay was 7 days, with a range of 2 to 35 days. Total hospital stays ranged from 7 to 51 days, with a median of 16 days. There were 26 complications in 15 patients: prolonged respiratory failure (vent . 4 days), 10; pneumonia, 7; pancreatitis, 4; vocal cord paresis, 1; acute renal failure (temporary hemodialysis), 1; missed duodenal injury, 1; candidemia, 1; and cortical blindness, 1.
Comment This report details a recent clinical experience with traumatic rupture of the thoracic aorta. This is an uncommon injury: among an average of 1,600 blunt trauma admissions per year at our center during the study period, 4.4 TRAs were diagnosed and 3.2 were repaired. A recent survey of 50 trauma centers in North America documented an average of 2.2 cases of TRA per center per year [9]. The relatively infrequent occurrence of this entity means that few surgeons are able to accrue large personal experiences. We have reviewed our experience to emphasize critical aspects of diagnosis, medical management, techniques of distal circulatory support, and operative repair. Three of 26 patients (12%) in this series had normal admission chest radiographs, even when reviewed retrospectively. These-false negative radiographs caused significant delays in treatment of the underlying thoracic injury. These patients had 30-, 33-, and 127-hour intervals from injury to repair, in comparison with a median time to repair of 5 hours for the entire series. Delays in treatment in 1 case almost certainly and in another likely contributed to paraplegia. Woodring and King [13] reported two normal chest radiographs in a series of 32 patients with known thoracic aortic transections, for a false-negative rate of 6%. Other recent surgical series [3, 5, 6, 14] have reported false-negative rates between 0% and 15% (Table 3). Our experience is at odds with the radiology literature, which suggests that a normal chest radiograph is strong
1298
GAMMIE ET AL TRAUMATIC AORTIC RUPTURE
evidence for the absence of an underlying aortic injury. Mirvis and colleagues [15] reviewed the admission chest radiographs of 205 patients with blunt trauma undergoing aortography, 41 of whom were found to have aortic tears. They concluded that a normal interpretation of an erect chest radiograph has a negative predictive value of 98%. Marnocha and Maglinte [16] concluded that a normal aortic arch combined with the absence of tracheal and nasogastric tube deviation ruled out an aortic injury. Our series, in accordance with previous reports in the surgical literature, demonstrates a surprisingly high incidence (12%) of normal admission chest radiographs in patients with TRA. A high level of suspicion is necessary to avoid the catastrophic results of a missed aortic injury. Although some reports have suggested that medical therapy and deliberate delayed operative therapy is acceptable management of a TRA [17], 4 of 30 patients (13%) in this series admitted with vital signs had rupture within 2.5 hours of arrival and died. Of these 4 patients experiencing free rupture during initial diagnostic and resuscitative measures, it is clear that at least 2 had substantial leaks before arrival. Three of 4 were hypotensive at the time of admission. The 1 patient with sustained hypertension during the first 2 hours after admission likely would have benefited from pharmacologic blood pressure control. Few patients in our series received medical therapy (b-blockade, antihypertensives), in accord with practice across North America: Fabian and associates [9] documented that only 17% of patients with TRA received medical therapy. Our experience suggests that the threat of completion of rupture of a thoracic aortic injury in the early hours after admission is real, and institution of aggressive blood pressure control during expeditious evaluation should not be overlooked. b-Blockade, possibly in combination with vasodilators and analgesics, should be given to all hemodynamically stable patients undergoing evaluation for a possible TRA. The presence of hypotension not readily explained by identified injuries should increase concern for imminent rupture. All patients except 2 in this series were supported intraoperatively with partial left-heart bypass. The optimal method of preventing spinal cord injury in the patient with TRA has been a source of considerable debate in the literature [3–10]. The approach to intraoperative protection of the spinal cord during repair of TRA has evolved. Initial reports of repair on full cardiopulmonary bypass emphasized prohibitive rates of hemorrhagic complications. Mortality was attributed to uncontrolled bleeding as a sequela of obligatory systemic heparinization for cardiopulmonary bypass. However, in many of these reports investigation of intraabdominal injuries was deferred until aortic repair was completed, with predictable adverse consequences. Gott [18] introduced the 9-mm heparin-bonded shunt with inflow from either the proximal aorta or left ventricular apex and outflow to the descending thoracic aorta or the femoral artery, as a means of supporting the spinal cord and viscera [19]. Experimental work by Molina and associates [20] suggested that the Gott shunt provided inadequate
Ann Thorac Surg 1998;66:1295–1300
Table 4. Summary of Morbidity and Mortality With Partial Left Heart Bypass Author
Year Patients Paraplegia
Benckart et al [23] Hess et al [7] Higgins et al [14] Read et al [6] Walls et al [5] Forbes and Ashbaugh [8] Nicolosi et al [24] Fabian et al [9] Gammie et al (present study)
1989 1989 1992 1993 1993 1994
21 16 10 16 8 21
0 0 0 0 0 0
3 1 0 2 1 4
1996 1997 1998
18 43 24
0 2 0
... 7 2
2 (1.1%)
20 (11.3%)
Total
177
Mortality
distal flow, and clinical data confirmed high rates of paraplegia. Mattox and colleagues [21] and others advocated simple cross-clamping of the aorta and repair of the injury without distal circulatory support. The “clamp and sew” technique emphasizes speed of repair, with many authors demonstrating a benefit to cross-clamp times less than 30 minutes [19]. Olivier and associates [22] introduced the technique of left atrial to femoral bypass with the centrifugal pump in 1984 and began the current trend favoring this technique for mechanical distal circulatory support. A recent metaanalysis of articles reporting outcomes of treatment of traumatic thoracic aortic tears in more than 1,000 patients provides strong support for the use of mechanical partial left heart bypass during repair of TRA [11]. Overall paraplegia rates were 19.2% for simple cross-clamping, 11.1% for passive (Gott) shunting, and 2.3% for active augmentation of distal perfusion. These data also emphasized the rapid increase in the incidence of paraplegia that occurs with cross-clamp times in excess of 30 minutes when the “clamp and sew” technique is used. Our experience with paraplegia after repair performed with active distal circulatory support is similar to recent series reporting minimal neurologic morbidity with partial left-heart bypass. In 177 patients undergoing repair of TRA with distal circulatory support, the incidence of paraplegia was only 1.1% (Table 4). Paraplegia did not develop in any patient in our series arriving in the operating room in hemodynamically stable condition and without evidence of preexisting neurologic compromise. Despite the compelling weight of published evidence supporting the role of distal circulatory support in minimizing the devastating complication of paraplegia after TRA, support for the “clamp and sew” technique remains strong [12]. In general, these authors acknowledge the importance of the 30-minute ceiling and express confidence in their ability to complete the repair in less than this time. Fabian and associates’ [9] survey of North American trauma centers documented a cross-clamp time in excess of 30 minutes in 67% of cases. Since the mid 1980s, our approach to distal perfusion during repair of traumatic transections of the aorta has
Ann Thorac Surg 1998;66:1295–1300
almost exclusively been one of partial left-heart bypass (see Table 2). We carry out partial left-heart bypass in one of two ways. Left atrial to femoral artery (or descending aorta) bypass is usually established with minimal heparin (5,000 units) and consists of a simple circuit with only a centrifugal pump. Recently we have cannulated the posterior surface of the inferior left pulmonary vein with a 24F cannula in lieu of the atrial appendage. This has provided flow of 2 to 3 L/min and is an effective nonpharmacologic means of controlling blood pressure in proximal and distal circulations. In contrast, venoarterial bypass requires full systemic heparinization and a complete cardiopulmonary bypass circuit, including an oxygenator, heat exchanger, and cardiotomy. For venoarterial bypass, we generally cannulate the pulmonary artery with a 32F right-angle cannula (TF-032-L-90; Research Medical Inc, Midvale, UT) directed toward the right ventricle for inflow and the femoral artery or descending aorta for outflow. The pulmonary artery is a reasonable option for inflow when venoarterial bypass is used because it is easily accessible via a left thoracotomy and offers adequate inflow and ease of cannulation. An excellent (and perhaps superior) alternative for inflow for venoarterial bypass is venous cannulation of the femoral vein (femorofemoral bypass). Venoarterial bypass offers several advantages over left atrial to femoral bypass. These include the ability to cool to 34°C (which may enhance spinal cord protection) and later rewarm the patient (of particular importance in the coagulopathic trauma population). An oxygenator in the circuit avoids difficulties in maintaining adequate oxygenation of the patient during single-lung ventilation. In our experience, intubation with a double-lumen tube is not always possible (eg, massive orofacial injuries) and the presence of venoarterial bypass permits discontinuation of ventilation to achieve critical exposure. Venoarterial bypass also affords the surgeon the option to perform hypothermic circulatory arrest should a complex aortic tear be encountered, as was necessary in 1 patient in this series. By using venoarterial bypass, the perfusionist can rapidly add volume to the circuit, a capability not present with the left atrial to femoral circuit without adjunctive rapid infusion systems. Cross-clamping of the aorta predictably leads to proximal hypertension, distal hypotension, and a markedly increased left ventricular afterload that may necessitate the concomitant administration of vasodilators, inotropic agents, and the administration of volume in the patient on left atrial to femoral bypass. Administration of vasodilators during aortic cross-clamping can initiate a deleterious steal of perfusion from the spinal cord. Venoarterial bypass mitigates these concerns by permitting complete unloading of the left ventricle and affording a means of rapidly replenishing volume. Arterial pressures are maintained at desired levels by simply adding or removing blood from the circuit. Finally, shed blood can be returned to the bypass circuit, and venoarterial bypass offers superior support to a failing heart. The principal disadvantage of venoarterial (pulmonary arterial to femoral) bypass is the requirement for full
GAMMIE ET AL TRAUMATIC AORTIC RUPTURE
1299
heparinization. Although long-standing dogma held that heparinization in this population of patients was highly undesirable, Pate and associates [10] have recently pointed out that there are few data to support this belief. In our experience, most patients were assessed for intracranial and intraabdominal injury before administration of heparin for bypass. We did not observe bleeding complications resulting from heparinization in any patient in this series. In addition, comparison of the lowdose (left atrial to femoral) and full-dose (venoarterial) heparin treatment groups revealed no relationship between heparin dose and outcome. Technical aspects of safe repair include a generous fourth intercostal space thoracotomy for optimal exposure and repair of thoracic aortic injuries. Double-lumen intubation should be performed when feasible to permit single-lung ventilation and provide wide exposure. Proximal control of the aortic arch should be between the left carotid and subclavian arteries, with separate control of the subclavian artery. Bypass is established and dissection commenced. Once the aorta has been clamped, the extent of the tear is assessed, and the distal clamp is repositioned as close as possible to the injury to allow maximal perfusion of the intercostal vessels from the distal pump flow. The adventitia surrounding the injury is typically distorted by hematoma and invariably has retracted. It is critically important to incorporate generous amounts of adventitia within the suture line. In conclusion, traumatic rupture of the aorta is an uncommon injury that requires aggressive diagnosis and treatment. Our review has emphasized several important aspects of the management of this disease. A TRA can be present despite a normal chest radiograph, and the resulting delays in treatment can lead to significant morbidity. Progression of TRA to complete rupture can occur within hours of hospital admission. Judicious administration of a b-blocker and vasodilators to patients with a suspected TRA should begin in the emergency department and continue to operation. Distal circulatory support is critical for effective spinal cord protection. Administration of heparin for bypass, in our experience, was not associated with significant morbidity. Management options for bypass include left atrial to distal and venoarterial bypass. The decision as to which to use should be tailored to the specific clinical situation.
References 1. Parmley LF, Mattingly TW, Manion WC, Jahnke EJ. Nonpenetrating traumatic injury of the aorta. Circulation 1958; 17:1086-101. 2. Hunt JP, Baker CC, Lentz CW, et al. Thoracic aorta injuries: management and outcome of 144 patients. J Trauma 1996;40: 547–56. 3. Hilgenberg AD, Logan DL, Akins CW, et al. Blunt injuries of the thoracic aorta. Ann Thorac Surg 1992;53:233–9. 4. DelRossi AJ, Cernaianu AC, Madden LD, et al. Traumatic disruptions of the thoracic aorta: treatment and outcome. Surgery 1990;108:864 –70. 5. Walls JT, Boley TM, Curtis JJ, Schmaltz RA. Experience with four surgical techniques to repair traumatic aortic pseudoaneurysm. J Thorac Cardiovasc Surg 1993;106:283–7.
1300
GAMMIE ET AL TRAUMATIC AORTIC RUPTURE
6. Read RA, Moore EE, Moore FA, Haenel JB. Partial left heart bypass for thoracic aorta repair. Survival without paraplegia. Arch Surg 1993;128:746 –52. 7. Hess PJ, Howe HR Jr, Robicsek F, et al. Traumatic tears of the thoracic aorta: improved results using the Bio-Medicus pump. Ann Thorac Surg 1989;48:6 –9. 8. Forbes AD, Ashbaugh DG. Mechanical circulatory support during repair of thoracic aortic injuries improves morbidity and prevents spinal cord injury. Arch Surg 1994;129:494 – 8. 9. Fabian TC, Richardson JD, Croce MA, et al. Prospective study of blunt aortic injury: Multicenter Trial of the American Association for the Surgery of Trauma. J Trauma 1997; 42:374 – 83. 10. Pate JW, Fabian TC, Walker WA. Acute traumatic rupture of the aortic isthmus: repair with cardiopulmonary bypass. Ann Thorac Surg 1995;59:90 –9. 11. Von Oppell UO, Dunne TT, De Groot MK, Zilla P. Traumatic aortic rupture: twenty-year metaanalysis of mortality and risk of paraplegia. Ann Thorac Surg 1994;58:585–93. 12. Sweeney MS, Young DJ, Frazier OH, Adams PR, Kapusta MO, Macris MP. Traumatic aortic transections: eight-year experience with the “clamp-sew” technique. Ann Thorac Surg 1997;64:384 – 8. 13. Woodring JH, King JG. The potential effects of radiographic criteria to exclude aortography in patients with blunt chest trauma. Results of a study of 32 patients with proved aortic or brachiocephalic arterial injury. J Thorac Cardiovasc Surg 1989;97:456 – 60. 14. Higgins RS, Sanchez JA, DeGuidis L, et al. Mechanical circulatory support decreases neurologic complications in the treatment of traumatic injuries of the thoracic aorta. Arch Surg 1992;127:516 –9. 15. Mirvis SE, Bidwell JK, Buddemeyer EU, et al. Value of chest
Ann Thorac Surg 1998;66:1295–1300
16. 17. 18. 19.
20.
21.
22.
23. 24.
radiography in excluding traumatic aortic rupture. Radiology 1987;163:487–93. Marnocha KE, Maglinte DD. Plain-film criteria for excluding aortic rupture in blunt chest trauma. AJR Am J Roentgenol 1985;144:19 –21. Maggisano R, Nathens A, Alexandrova NA, et al. Traumatic rupture of the thoracic aorta: should one always operate immediately? Ann Vasc Surg 1995;9:44 –52. Gott VL. Heparinized shunts for thoracic vascular operations. Ann Thorac Surg 1972;14:219 –20. 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 –74. Molina JE, Cogordan J, Einzig S, et al. Adequacy of ascending aorta-descending aorta shunt during cross-clamping of the thoracic aorta for prevention of spinal cord injury. J Thorac Cardiovasc Surg 1985;90:126 –36. Mattox KL, Holzman M, Pickard LR, Beall AC Jr, DeBakey ME. Clamp/repair: a safe technique for treatment of blunt injury to the descending thoracic aorta. Ann Thorac Surg 1985;40:456 – 63. Olivier HF Jr, Maher TD, Liebler GA, Park SB, Burkholder JA, Magovern GJ. Use of the BioMedicus centrifugal pump in traumatic tears of the thoracic aorta. Ann Thorac Surg 1984;38:586 –91. Benckart DH, Magovern GJ, Liebler GA, et al. Traumatic aortic transection: repair using left atrial to femoral bypass. J Card Surg 1989;4:43–9. Nicolosi AC, Almassi GH, Bousamra M II, Haasler GB, Olinger GN. Mortality and neurologic morbidity after repair of traumatic aortic disruption. Ann Thorac Surg 1996;61:875– 8.
J. Kent Trinkle, MD 1934 –1998 It is with great sadness that we report the death of Dr Kent Trinkle, member of the Editorial Board of The Annals of Thoracic Surgery for 15 years, and Associate Editor for 8 years, first in charge of Book Reviews, and later the Section on Collective and Current Reviews. In addition to his duties as Chief of the Division of Cardiothoracic Surgery at the University of Texas Health Science Center in San Antonio, plus his numerous other commitments at the local, state, and national level, Kent was always a faithful supporter of our journal. He de-
© 1998 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
voted a great deal of time and effort to improving The Annals, and for this we are grateful beyond measure. We shall miss him as an Editorial Board member, but much more as a superb human being. I invite you to read the homilies about Dr Trinkle written by his two long-time associates, Dr Fred Grover and Dr John Calhoon, in the Transitions Section of the CTSNet. Tom Ferguson Editor, The Annals of Thoracic Surgery
Ann Thorac Surg 1998;66:1300
•
0003-4975/98/$19.00