Traumatic aortic rupture: recent outcome with regard to neurologic deficit

Traumatic Aortic Rupture: Recent Outcome With Regard to Neurologic Deficit Safuh Attar, MD, Marcelo G. Cardarelli, MD, Stephen W. Downing, MD, Aurelio Rodriguez, MD, Douglas C. Wallace, MD, Robert S. West, MS, and Joseph S. McLaughlin, MD Division of Thoracic & Cardiovascular Surgery, Department of Surgery, Maryland Institute for Emergency Medical Services System, Baltimore, Maryland

Background. Traumatic aortic rupture is highly lethal, and its surgical treatment is complicated by a high rate of paraplegia. Methods. The charts of 263 patients with traumatic aortic rupture from vehicular accidents treated between 1971 and 1998 were reviewed. Patients were grouped according to four periods: group 1, 1971 to 1975, (n 5 31); group 2, 1976 to 1985, (n 5 83); group 3, 1986 to 1994, (n 5 82); and group 4, 1994 to 1998 (n 5 67). Seventy-one patients died of exsanguination before definitive care. One hundred-ninety two patients had surgical repair with the following techniques: clamp and sew, 6 in group 1, 22 in group 2, 54 in group 3, none in group 4; shunt, 23 in group 1, 39 in group 2, 2 in group 3; cardiopulmonary bypass, 2 in group 1, 1 in group 3. Forty-three patients had partial bypass with the centrifugal pump and heparin-coated circuits in group 4. Results. Operative mortality was 6 of 31 (19%) in group 1, 22 of 61 (36%) in group 2, 15 of 57 (26%) in group 3, and

7 of 43 (16%) in group 4. There was one case of paraplegia in group 1 (4%), ten in group 2 (18%), 11 in group 3 (26%), and none in group 4. This difference of paraplegia between the groups was significant (p < 0.002). Significant factors for paraplegia were intraoperative hypotension (p < 0.000002), cross-clamp time longer than 30 minutes (p < 0.008), pump versus no pump (p < 0.008), and younger age group (28 6 11 versus 39 6 17 years) (p < 0.03). Conclusions. There were no statistically significant improvements in mortality rate over the four periods, although, the mortality rate was lowest in the last period when partial bypass with the centrifugal pump was used exclusively. Further, the use of the centrifugal pump with heparin-coated circuits, with femoral vein cannulation into the right atrium and distal aortic perfusion, reduced paraplegia significantly. (Ann Thorac Surg 1999;67:959 – 65) © 1999 by The Society of Thoracic Surgeons

T

result of vehicular accidents, who were admitted to the University of Maryland Hospital and the Maryland Institute for Emergency Medical Services System Shock Trauma Center (MIEMSS) between 1971 and 1998. The Maryland Institute for Emergency Medical Services System encompasses nine regional trauma centers distributed throughout Maryland and a 100-bed shock trauma center located in Baltimore. All of the patients in this study were admitted to the latter clinical facility either directly from the scene of the accident or in transfer from a regional trauma center. Transport was preferentially by helicopter. At the scene of injury the paramedic was empowered with the triage decision to transport the patient to the nearest hospital, to a regional trauma center, or directly to the Shock Trauma Center. Patients were divided into four groups according to time period, group 1 (n 5 31) 1971 to 1975, group 2 (n 5 83) 1976 to 1985, group 3 (n 5 82) 1986 to 1994, and group 4 (n 5 67) 1995 to 1998. Groups 1 and 2 were previously

raumatic aortic rupture remains a formidable sequela of blunt chest injury. It accounts for about 30% of fatalities from blunt chest injuries caused by vehicular accidents. Related morbidity and mortality rates remain high despite rapid transportation to trauma centers, aggressive resuscitation, and expeditious surgical intervention. Issues continue to be raised concerning the optimal time for repair of thoracic aortic rupture, the technique used for repair, the type of repair, and risk factors contributing to morbidity and mortality. Paraplegia is the most serious complication of the surgical repair of aortic rupture. We studied the evolution of the surgical treatment of aortic rupture in a university trauma center.

Materials and Methods We reviewed the medical records of 263 patients with acute blunt traumatic rupture of the thoracic aorta as a Presented at the Forty-fifth Annual Meeting of the Southern Thoracic Surgical Association, Orlando, FL, Nov 12–14, 1998. Address reprint requests to Dr Attar, University of Maryland Hospital, 22 South Greene St, Baltimore, MD 21201; e-mail: sattar@surgeryl. umaryland.edu.

© 1999 by The Society of Thoracic Surgeons Published by Elsevier Science Inc

This article has been selected for the open discussion on the STS Web site: http://www.sts.org/section/atsdiscussion/ 0003-4975/99/$20.00 PII S0003-4975(99)00174-5

960

ATTAR ET AL AORTIC RUPTURE

Ann Thorac Surg 1999;67:959 – 65

Table 1. Operative Mortality After Traumatic Aortic Rupture Repair Technique

1971–75

1976 – 85

1986 –94

1995–98

Clamp and sew Shunt Roller pump Centrifugal pump Total Mortality ratea

0/6 5/23 1/2 ... 6/31 19%

6/22 16/39 ... ... 22/61 36%

12/54 2/2 1/1 ... 15/57 26%

... ... ... 7/43 7/43 16%

a

No significant differences.

reported in detail [1, 2]. Group 1 comprised 31 patients ranging in age from 15 to 67 years (mean, 26 years). Group 2 comprised 83 patients ranging in age from 15 to 80 years (mean, 31.3 years). Patients in group 3 (n 5 82) ranged in age from 15 to 68 years (mean, 35 years) and in group 4 (n 5 67) from 19 to 73 years (mean, 42 years). These were 202 male and 61 female patients. Diagnosis was made in groups 1 and 2 by anteriorposterior chest roentgenograms and confirmed by aortography in all cases. Chest roentgenograms showed widening of the mediastinum, although the degree of widening was equivocal in eight cases. Loss of definition of the aortic knob was seen in all cases. In groups 3 and 4 there were four negative results of chest x-ray, which were confirmed to be false negatives by aortography, an incidence of 3%. Between 1986 and 1994, diagnosis was made by angiography in most cases, although computed axial tomographic scan was done in 44 cases. Currently, diagnosis is based on chest roentgenograms and helical computed tomographic scan; rarely is an aortogram done. Many patients with traumatic aortic rupture were either dead upon arrival or were hypotensive and unstable. The latter were resuscitated aggressively, and at times, emergency thoracotomy was done to control the bleeding. In group 2, 22 (26%) of 83 patients and in group 3, 25 (30%) of 82 patients died of exsanguination during resuscitation. In group 4, 24 patients (36%) died on admission. Stable patients had a complete evaluation of their multiple injuries, in addition to the work-up for aortic disruption. In hypotensive patients the source of bleeding was investigated. Abdominal bleeding was controlled by laparatomy or embolization. Patients with fractured pelvis had external fixation of the pelvis and embolization of the bleeders if needed. If head injury was detected and the patient was neurologically intact, treatment of traumatic aortic rupture (TAR) took priority. Except for three cases, the site of aortic rupture was the descending thoracic aorta, just distal to the origin of the left subclavian artery. There were two aortic ruptures just above the diaphragm. One patient had rupture of the ascending aorta, with massive right hemothorax. He was not operated on and died of multiple organ system failure. The tear was single in nearly all cases. In 7 patients, there were two areas of rupture within 2 or 3 cm of each other, and in 1 patient, four tears were identified. Major nonthoracic injuries were present in all cases.

There was a total of 59 associated injuries in the period of 1971 to 1975, 70 injuries between 1986 and 1994 and 72 between 1994 and 1998. Associated injuries for the period 1976 to 1985 are not available. All patients admitted from the scene of injury for whom data are available arrived within 2 hours of injury, whereas, those received in transfer arrived later. The mean time interval from admission to operation was 6 hours for the period 1986 to 1994 compared with 9 hours for the period 1995 to 1998. The longer interval in the latter group was because 3 patients spent 62, 140, and 312 hours in the hospital before aortic rupture was diagnosed. The techniques used as adjunct to surgery were clamp and sew, shunts, and pump, either with left heart bypass with the roller pump with total systemic heparinization in the earlier groups or with the centrifugal pump and heparin-coated oxygenator and cannulas without heparinization in the later groups. The technique of repair in each period is summarized in Table 1. The Morris silicone shunt or the Gott shunt were used from the ascending to the descending aorta. In left heart bypass, a left atrial femoral artery bypass was used initially; however, because of problems associated with ischemia of the cannulated extremity, including the necessity for amputation in one case, the descending thoracic aorta became the preferred site for arterial cannulation. A double lumen endotracheal tube is inserted. Percutaneous femoral venous cannulation of the right atrium is done with a 19F Carmeda heparin-coated cannula (Medtronic Inc, Minneapolis, MN). The patient is turned right-side down, and a left thoracotomy in the fourth intercostal space is done. The aorta is inspected, and the extent of the hematoma is assessed without disturbing it. Distal aortic control is obtained below the hematoma, and a pursestring suture is inserted for cannulation. The left subclavian artery is isolated well above the hematoma. The pleura is divided medial to the subclavian artery, and the aorta is identified between the subclavian and the carotid arteries. The vagus nerve is identified and isolated, as its path extends downward between the carotid and subclavian arteries. It is gently retracted laterally as it crosses the aorta. The aorta between the carotid and subclavian arteries is isolated sufficiently for clamping. The distal aorta is cannulated. Aortic bypass is done using the Carmeda Coated Maxima Plus Hollow Fiber Oxygenator with an integral heat exchanger and Carmeda coated tubing (Medtronic Inc, Minneapolis, MN). Flow is maintained above an index of 1.2 L and at a rate of more than 2 L per minute to prevent thrombosis. Body temperature is maintained at 37°C as determined by a bladder probe. Blood pressure is not measured routinely in the lower body, but if measured, it is usually kept above 60 mm Hg. Clamps are applied to the aortic arch, the subclavian artery, and the distal aorta. The hematoma is entered and the aortic damage evaluated. The recurrent laryngeal nerve can be traced as it passes anterior to the aorta to dive beneath the fascia inferior to the aorta and lateral to

Ann Thorac Surg 1999;67:959 – 65

the ductus arteriosus. Debridement of the aortic tear is done, and the proximal aorta is freed from its posterior attachments. The ductus might require division to develop a sufficient cuff of proximal aorta. Graft replacement is done when complete or nearly complete separation of the aorta is encountered. A Dacron graft was used in the early groups. Currently we use a Hemashield graft (Meadox; Boston Scientific Corp, Oakland, NJ). In cases in which the aorta was partially torn, approximation and primary suture is done using 3-0 polypropylene sutures. Statistical analysis was done on extracted data in files formatted for the Statistical Analysis System (SAS Institute Inc, Cary, NC). Continuous data were converted to nonparametric dichotomous form by selection of suitable ranges, and significance was tested where possible by x2 analysis, applying Fisher’s exact test (one-tailed) where appropriate. Numeric data were analyzed using the Wilcoxon rank sum test. The patients who died on admission because of exsanguination were excluded from the analysis.

Results The four groups of patients with traumatic aortic rupture between 1971 and 1998 were comparable with respect to age, sex, extent of major nonthoracic injuries, initial blood pressure, and time interval from admission to operation. In group 1 (1971 to 1975), 23 of 31 patients had passive shunt, with five deaths (21.7%) and one case of postoperative paraplegia (4%). Six had the clamp and sew technique without death or paraplegia, and 2 had left heart bypass after systemic heparinization, with one intraoperative death (50%) and no paraplegia. In group 2 (1976 to 1985), 39 of 61 patients had shunt procedures, with 16 (41%) deaths and six (17%) cases of postoperative paraplegia. Twenty-two patients had the clamp and sew technique, with six deaths (27%) and four (23%) cases of postoperative paraplegia. No patients in this series had a centrifugal pump used. In group 3 (1986 to 1994), 54 of 57 patients had the clamp and sew technique; there were 12 deaths (18%) and 11 (23.4%) cases of postoperative paraplegia. One patient who had left heart bypass died. Two patients with shunts did not survive. In group 4 (1994 to 1998), 43 of 67 patients had partial bypass through the femoral vein and right atrium to the descending thoracic aorta using heparin-coated circuits without systemic heparinization. There were seven deaths (16%) and no paraplegia. The surgical complications during the period of 1971 to 1975 were hypertension in 18 patients, paraplegia in 1, renal failure in 3, left recurrent nerve palsy in 8, exsanguination in 2, and shunt technique errors in 2. The major complications in the second period (1976 to 1985) were paraplegia in 10 patients, adult respiratory distress syndrome in 9, renal failure in 5, sepsis in 5, pseudoaneurysm in 3, hypertension in 9, left recurrent nerve palsy in 7, left phrenic nerve injury in 3, and severe jaundice in 4. The complications of the period 1986 to 1994 were paraplegia in 11 patients, renal failure in 7, adult respiratory distress syndrome in 10, sepsis in 7, prolonged ventila-

ATTAR ET AL AORTIC RUPTURE

961

tion in 7, pneumonia in 4, and hypertension in 4. The complications of the current period (1994 to 1998) were renal failure in 2 patients, sepsis in 3, adult respiratory distress syndrome in 2, multiple organ system failure in 2, prolonged ventilation in 8, injury to vagus nerve in 1, deep vein thrombosis in 1, and above the knee amputations in 3 patients, one from ischemia, one from thrombosis of the venous cannula, and the third from extensive trauma. There were six perioperative deaths in group 1; causes were renal failure in 2 patients, exsanguination due to intrapleural rupture of the aortic hematoma during dissection in 2, and error in shunt technique in 2. In group 2, 22 (26%) of the 83 patients died of exsanguination during resuscitation attempts in the admitting or operating room. Seven others died during surgical repair, and 12 died within 30 days of operation. In group 3, 25 (30%) of 82 patients died during initial resuscitation. There were 15 operative deaths, eight from exsanguination in the operating room, 2 from myocardial infarction during opening, 3 from cardiac arrest after release of the aortic clamps, and four from multiple system organ failure. In group 4, 24 of 67 patients died during resuscitation, there were seven surgical deaths, two deaths in the operating room, one from hypotension, and the other from arrhythmia. The other five deaths were from brain ischemia in 2 patients, multiple organ failure in 1, adult respiratory distress syndrome in 1, and pneumonia in 1. A total of 192 patients with TAR had surgical repair between 1971 and 1998. In 167 patients the aortic rupture was repaired with tubular graft interposition, and 25 patients had primary suture repair. Four patients died during the repair. There was no statistical difference in the mortality rate between the two methods of repair, graft interposition and primary suture repair ( p . 0.05). There was no significant difference in the mortality among the four groups (Table 1). There was no significant difference in the rate of paraplegia (Table 2) among the first two groups when shunts were used. There was no significant difference in the incidence of paraplegia with clamp and sew techniques in the first three groups; this technique was not used in the fourth group. There was a significant difference in the last two groups when a pump was used ( p 5 0.001), although the numbers are too small to draw definitive conclusions. The incidence of paraplegia between group 4 and the previous groups was significant ( p , 0.002). There were 22 cases of paraplegia in 167 operating room survivors (13%) (Tables 2 and 3). The risk factors thought to be contributory to this complication were analyzed. The one patient in group 1 who developed paraplegia was a 23-year-old man who had an aorta-toaorta shunt. The aortic cross-clamp time was 36 minutes. There was no perioperative hypotension and no other complications. In group 2, paraplegia was present preoperatively in 3 patients. It was associated with a thoracic spine fracture in 1 patient who died intraoperatively, associated with a preexisting hemiparesis from radiation arachnoiditis in 1 patient who survived, and associated with massive injuries in a patient who died 4 days

962

ATTAR ET AL AORTIC RUPTURE

Ann Thorac Surg 1999;67:959 – 65

Table 2. Incidence of Paraplegia After Traumatic Aortic Rupture Variable

1971–75

Patients To operating room Operating room survivors Paraplegia Incidence a

1976 – 85

1986 –94

83 61 54 10 18.5%

82 57 47 11 23.4%

31 31 25 1 4%

1995–98 67 43 41 0 0%a

Total 263 192 167 22 13%

p # 0.002.

postoperatively. Paraplegia (six cases) or paraparesis (four cases) developed postoperatively in 10 of 54 survivors (18.5%); six cases occurred in 39 patients (15%) in whom a shunt was used and four occurred in 17 patients (11%) in whom no shunt was used. The difference was not statistically significant. Hypotension in this series was not significant with regard to paraplegia because the number of cases in this category was too small. All ten cases of paraplegia or paresis occurred in the 42 patients with aortic cross-clamp times exceeding 30 minutes ( p 5 0.18). In group 3, of the 43 patients who survived the operation, 11 had paraplegia. One patient had paraplegia preoperatively associated with a compression fracture of the L5 vertebra. The remaining 10 patients had aortic repair using the clamp and sew technique. There were 6 patients with episodes of hypotension intraoperatively, with systolic pressure between 60 and 70 mm Hg. The aortic cross-clamp time exceeded 30 minutes in all ranging from 32 to 64 minutes. The effect of hypotension on paraplegia was highly significant ( p 5 0.000002). Age was also found to be a significant factor. The younger age groups had a slightly higher risk of paraplegia. The mean age of patients who had paraplegia was 28 6 11 years compared with 39 6 17 years in patients without paraplegia ( p , 0.03).

Comment This study presents our experience with blunt traumatic rupture of the aorta during the 27-year period, 1971 to 1998. Since 1994, we have used a heparin-coated circuit and the centrifugal pump as an adjunct to repair. The mortality rate for this period is the best we have achieved, and there was zero incidence of paraplegia. Thoracic aortic rupture is a highly lethal injury. Dischinger and associates [3] studied 739 traffic fatalities in Maryland in 1985. Of those who had complete autopsy, 105 (23.5%) had a ruptured aorta. Ninety-six (91.4%) of the ruptures were located at the isthmus of the aortic Table 3. Incidence of Paraplegia by Operative Technique Technique Clamp and sew Shunt Roller pump Centrifugal pump Total

1971–75

1976 – 85

1986 –94

1995–98

0/6 1/18 0/1 ... 1/25

4/16 6/23 ... ... 10/39

11/42 ... ... ... 11/42

... ... ... 0/36 0/36

arch, six were located on the ascending aorta, and three on the descending aorta. For 88 of the 105 cases, both time of accident and time of death were available: 47% of deaths occurred within 1 hour, and 95% within 2 hours of the accident. Approximately two thirds of the deaths occurred at the scene or before arrival at the hospital. Zieger and associates [4], in a review of 10 studies, noted that 112 (19%) of 591 patients with TAR died of exsanguination soon after arrival at medical facilities or while awaiting definitive tests and an operation. Aortic rupture is rarely a single lesion. Between 20% and 40% of patients with aortic rupture have associated intraabdominal injuries that necessitate laparotomy. The surgeon must decide which comes first, thoracotomy or laparotomy. Richardson and associates [5] monitored the clinical course of 408 patients with blunt trauma who had a widened mediastinum on chest x-ray. Of these patients, 35 (9%) had TAR by aortography, 7 (20%) required laparotomy in addition to thoracotomy. On the basis of these findings, the authors recommended that in the unstable patient, laparotomy should be done first if diagnostic peritoneal lavage is grossly positive. If the diagnostic peritoneal lavage was negative, other causes of hemodynamic instability, such as pelvic or femur fractures, should be assessed. In the stable patient, if diagnostic peritoneal lavage is grossly positive, laparotomy was done before aortography. If diagnostic peritoneal lavage was positive only by cell count, aortography and thoracotomy should be done first. At our Shock Trauma Center, Rodriguez and colleagues [6] retrospectively reviewed the treatment priorities in 50 patients with combined thoracic and abdominal injuries. Laparotomy was done first in 19 patients (38%), one of whom died perioperatively of free aortic rupture. The other 31 patients (61%) had aortography and thoracotomy before laparotomy. Five patients (16%) died perioperatively, four from aortic rupture and 1 from massive abdominal injury. This experience led us to assess the abdominal injury first, unless instability was associated with the aortic rupture. Considerable controversy surrounds the details of aortic repair. Many surgeons support the clamp and sew technique because it is expeditious, does not require bypass, and avoids complications associated with cannulation, bypass, and heparinization. Results reflecting survival and normal spinal cord function have been shown to equal those that used mechanical adjuncts. Mattox and colleagues [7] studied 32 patients with TAR repair, 23 by

Ann Thorac Surg 1999;67:959 – 65

the clamp and sew technique, 8 by standard cardiopulmonary bypass, and 1 with a Gott shunt. The mortality rate was higher with the clamp and sew technique than with bypass (23% versus 12.5%), but all deaths were due to associated injuries. The clamp and sew group had a 4.5% permanent paraplegia rate, compared with 25% for the bypass group. The authors concluded that the clamp and sew method was the preferred approach to repair of TAR. Young and associates [8] studied 20 patients who had TAR repair. Eight repairs used the clamp and sew technique, and in the other 12, heparinless femoral vein-femoral artery bypass was used. Two deaths occurred postoperatively secondary to concomitant head injury, but no patient had postoperative renal dysfunction or paraplegia. Sweeny and colleagues [9] studied 71 patients with TAR who were treated with the clamp and sew method. Aortic cross-clamp time averaged 24 minutes (range, 14 to 36 minutes), with four in excess of 30 minutes. One patient (clamp time, 28 minutes) became paraplegic. The authors concluded the clamp and sew technique was a safe and effective method for the treatment of TAR. Hilgenberg and associates [10] studied 42 patients with TAR. Nineteen ruptures were repaired using the clamp and sew technique, with one case of paraplegia (5.2%). In 23 ruptures a Gott shunt was used, with three cases of paraplegia (13%). Logistic regression analysis showed that the only factor significantly associated with paraplegia was cross-clamp time. A logistic model of probability relating the risks of paraplegia to cross-clamp time was developed; the calculated probability of paraplegia was less than 1% with a clamp time of 30 minutes and less than 5% with a clamp time of 45 minutes, but the curve rises steeply at 50 minutes. Analysis indicated that a shunt did not offer protection from paraplegia or extend the safe cross-clamp time. Aortic repair with a mechanical adjunct uses the clamp and sew technique but additionally provides perfusion to the lower part of the body. The presumed advantages of this type of repair include avoidance of a hurried aortic repair, a lower incidence of paraplegia and renal dysfunction, a decreased preload, and prevention of reperfusion injury to the viscera. Zeiger and associates [4] reviewed 409 cases of TAR. Shunts were used in 165 cases, with 13 cases of paraplegia (7.9%) and 23 deaths. The clamp and sew technique was used in 108 cases with 22 cases of paraplegia (20.4%) and no deaths. Pumps were used in 136 cases, with four cases of paraplegia (2.9%) and 36 deaths (26.5%). The authors concluded that shunt or bypass provided the safer methodology. McCrosky and colleagues [11] endorsed aortic repair with the centrifugal flow pump in a left atrial femoral artery circuit, citing the advantages of a high flow rate for distal perfusion without the need for systemic heparinization. In their series of 12 patients, six ruptures were repaired with the centrifugal pump and six were repaired by the clamp and sew technique. One patient in each group died intraoperatively of concomitant blunt cardiac injury. Paraplegia developed in none of the remaining 5 patients in the bypass group (mean aortic cross-clamp time, 39 minutes), whereas it developed in 2 patients (40%) in the

ATTAR ET AL AORTIC RUPTURE

963

group that did not have bypass (mean cross-clamp time, 30 minutes). Fabian and colleagues [12], recently reported a prospective multicenter trial of blunt aortic injury involving 50 trauma centers in North America. There were 274 aortic injuries over 2.5 years. Two hundred seven stable patients had planned thoracotomy and repair. Clamp and sew technique was used in 73 (35%) patients and bypass techniques in 134 (65%). Overall mortality rate was 31%, with 63% of deaths being attributable to aortic rupture. The mortality rate was not affected by method of repair. Paraplegia occurred postoperatively in 18 (8.7%) of 207 patients, in 12 of 73 (16.4%) in the clamp and sew group and in 6 of 134 (4.5%) in the bypass group. Logistic regression analysis showed that the clamp and sew technique ( p 5 0.002) and aortic cross-clamp time greater than 30 minutes ( p 5 0.01) were associated with postoperative paraplegia. They concluded that bypass techniques that provide distal aortic perfusion produce significantly lowered paraplegia rates than the clamp and sew technique. The controversy concerning the use of primary repair as opposed to graft repair continues. Proponents of primary repair point to a more rapid repair, less potential for infection and suture line dehiscence, and similar if not better results with respect to mortality and postoperative paraplegia rates. Most TAR patients are young and have healthy aortas that hold sutures well, once the damaged areas are debrided. McBride and associates [13] used direct repair in 15 of 22 patients with TAR. They noted a shorter aortic cross-clamp time, an incidence of postoperative paraplegia of 14%, and a mortality rate of 18%. There were no complications related to the direct repair. Kieny and Charpentier [14] reported the successful use of direct repair in 21 (62%) of 34 patients with TAR. Only one instance (3%) of paraplegia occurred in 34 patients. Schmidt and Smith [15] reported excellent results with direct repair. In an earlier series, 21 (57%) of 37 patients and in the later series, 54 (74%) of 73 patients had direct repair of their TAR without graft interposition. Aortic cross-clamp time for 43 patients who had direct repair without a mechanical adjunct averaged 19 minutes. Paraplegia occurred in 2 (3.7%) of the 54 patients who had direct repair and in 2 (10.5%) of the 19 patients receiving graft interposition. The mortality rate was 11% in the direct repair group and 42% in the group receiving a graft. Death was mostly a result of associated injury. Although aortic transection often was complete, and the margins often retracted, the author did not have difficulty approximating the edges by gentle traction along the continuous posterior row of polypropylene sutures. Schmidt and associates [16] also reported a successful direct repair in a 7-year-old child with a blunt injury to the aortic arch with avulsion of the left carotid and subclavian arteries. Direct repair seems to be the preferred method in children, because it obviates the need for reoperation and graft replacement as the child grows. The timing of repair of TAR has been challenged. Duhaylongsod and associates [17] stated that the natural history of patients with TAR suggests a self-selection among victims. Patients with significant residual adven-

964

ATTAR ET AL AORTIC RUPTURE

titia, termed late-type transection, followed a protracted course. They might remain hemodynamically stable provided aortic wall tension and elevated systolic ejection pressure are avoided. Consequently, in patients who are exceptional surgical risks, delayed aortic repair, as suggested by Stiles and associates [18] and Akins and associates [19], might be appropriate. Maggisano and colleagues [20] examined the hypothesis that TAR can be treated by selective delayed operative repair. Forty-four patients required initial admission and treatment in the intensive care unit because of concomitant injuries or sepsis. The delay ranged from 1 day to 7 months until the clinical status had improved sufficiently to allow operative repair. Eight patients were well at follow-up from 1 to 4 years later. Overall survival rate was 82%. There were three deaths (10%). Only 2 patients (4.5%) died as a result of ruptured aorta within 72 hours of admission. Similarly Galli and associates [21] studied 42 patients with TAR. Twenty-one patients had immediate repair, with a 19% mortality rate, compared with no deaths in 21 patients who had an operation after intensive medical therapy and treatment of the associated lesions. Pate and colleagues [22] supported this concept and indicated that pharmacologic reduction of wall stress appeared to decrease probability of rupture. Our experience in deferred operative repair of TAR is anecdotal, but delaying the operation in patients with severe intracranial injuries appears justified.

References 1. Turney SZ, Attar S, Ayella R, et al. Traumatic rupture of the aorta, five year experience. J Thorac Cardiovasc Surg 1976; 72:727–32. 2. Cowley RA, Turney SZ, Hankins JR, et al. Rupture of thoracic aorta caused by blunt trauma, a fifteen year experience. J Thorac Cardiovasc Surg 1990;108:652– 61. 3. Dischinger PC, Cowley RA, Shanker BS, et al. The incidence of ruptured aorta among vehicular fatalities. Proc Am Assoc Auto Med 1988;22:15–20. 4. Zieger MA, Clark DE, Morton JR. Reappraisal of surgical treatment of traumatic transection of the thoracic aorta. J Cardiovasc Surg 1990;31:607–10. 5. Richardson JD, Wilson ME, Miller FB. The widened mediastinum: diagnostic and therapeutic priorities. Ann Thorac Surg 1990;211:731–7.

Ann Thorac Surg 1999;67:959 – 65

6. Rodriguez A, Chiascione RM, Clemens JA. Blunt traumatic rupture of the aorta. In: Maull KI, Rodriguez A, Wiles CE III, eds. Complications in trauma and critical care. Philadelphia, Pennsylvania: WB Saunders 1996;289 –305. 7. Mattox KL, Holzman M, Pickard LR, Beall AC, DeBakey ME. Clamp/repair: a safe technique for treatment of blunt injury to the descending thoracic aorta. Ann Thorac Surg 1985;40: 456– 63. 8. Young JN, Stallone RJ, Iverson LIG, et al. Surgical management of traumatic disruption of the descending aorta. West J Med 1989;150:662– 4. 9. Sweeney MS, Young DJ, Frazier OH, et al. Traumatic aortic transections. Eight year experience with the “clamp-sew” technique. Ann Thorac Surg 1997;64:384–9. 10. Hilgenberg AD, Logan DC, Akins CW, et al. Blunt injuries of the thoracic aorta. Ann Thorac Surg 1992;58:233– 8. 11. McCroskey BL, Moore EE, Moore FA, et al. A unified approach to the torn thoracic aorta. Ann Surg 1991;162: 473– 6. 12. Fabian TC, Richardson JP, Grace 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– 80. 13. McBride LR, Tridik S, Stothert JC, et al. Primary repair of traumatic aortic disruption. Ann Thorac Surg 1987;43:65–7. 14. Kieny R, Charpentier A. Traumatic lessons of the thoracic aorta. A report of 73 cases. Cardiovasc Surg 1992;32:613–9. 15. Schmidt CA, Smith DC. Traumatic avulsion of arch vessels in a child. Primary repair using hypothermic circulatory arrest. J Trauma 1989;29:248–50. 16. Schmidt CA, Wood MA, Rayzock AJ, et al. Primary repair of traumatic aortic rupture. A preferred approach. J Trauma 1992;32:588–92. 17. Duhaylongsod FG, Glower DD, Wolfe WG. Acute traumatic aortic aneurysm. The Duke Experience from 1970 to 1990. J Vasc Surg 1992;15:331– 43. 18. Stiles OR, Cohlmia GS, Smith JH, et al. Management of injuries of the thoracic and abdominal aorta. Ann Surg 1985; 150:132– 40. 19. Akins CW, Buckley MJ, Daggett W, et al. Acute traumatic disruptions of the thoracic aorta: a ten year experience. Ann Thorac Surg 1981;31:305–9. 20. Maggisano R, Nathems A, Alexandrova NA, et al. Traumatic rupture of the thoracic aorta: should one always operate immediately? Ann Vasc Surg 1995;9:44–52. 21. Galli R, Pacini D, Di Bartalomeo R, et al. Surgical indications and timing of repair of traumatic rupture of the thoracic aorta. Ann Thorac Surg 1998;65:461– 4. 22. Pate JW, Fabian TC, Walker W. Traumatic rupture of the aortic isthmus: an emergency? World Surg 1995;19:119–26.

DISCUSSION DR JOSEPH W. COOK (Charlotte, NC): In Charlotte we have been interested in this problem; however, we have taken a different approach. We have preferred to use left atrial to distal thoracic aorta or femoral bypass and have had good success with this technique. We now have 45 patients over the past 10 years in whom we have used this type of bypass for traumatic transections and have had no incidence of paraplegia in our 40 survivors. We did have 7 patients in whom we used femoralfemoral bypass with an oxygenator. We reserved this circuit for patients who had severe pulmonary insufficiency from pulmonary contusions or adult respiratory distress syndrome. None of these patients had paraplegia. Two of those 7 patients died. My questions are, one, what proportion of your patients were

helped by having the oxygenator in the circuit because of pulmonary insufficiency in those patients? Number two, what advantages do you think your circuit has, including an oxygenator, over the simpler, less costly bypass circuit involving just the heparin-coated tubing and centrifugal pump? DR CARDARELLI (Baltimore, MD): Thank you for your comments. I believe that the use of an oxygenator does not make much difference except in patients who already have a severe contusion of the lung, and because the number is so small, I cannot comment on that. The incidence of adult respiratory distress syndrome has been the same, whether we used an oxygenator or not. And I think that what we have to emphasize