Contemporary results of standard open repair of acute traumatic rupture of the thoracic aorta
Article Outline
Background
Open repair of acute traumatic rupture of the thoracic aorta has been the standard of care for the past half century. Traditional criteria of operative success have been patient survival and prevention of spinal cord ischemia. Historical series have reported a variability of surgical results with relation to the variety of operative approaches. This study aims to update the results obtained with a uniform surgical technique based on a systematic utilization of distal perfusion during aortic cross-clamping.
Methods
During a 35-year period (1974-2009), 138 consecutive patients with an acute traumatic rupture of the thoracic aorta were repaired with a Dacron graft interposition through a standard left thoracotomy. All patients received a method of circulatory support. A passive 9-mm Gott shunt inserted between the ascending and the descending aorta and delivering a median flow of 3 L/min was used in the first 40 cases. A partial left heart bypass realized from the left atrium to the descending aorta and driven with a centrifugal pump was used in the last 98 consecutive cases. A median flow of 4 L/min was recorded. Mean age of the patients was 27 years and 90.6% of them had associated injuries for a calculated mean ISS of 44.
Results
Two outcome variables were analyzed: hospital mortality and postoperative spinal cord ischemic injury. Overall hospital mortality is 5% (7/138 patients). This was improved from 7.5% (3/40) in patients perfused with the Gott shunt to 4% (4/98) in patients protected with the left heart bypass and lowered to 1.5% (1/68) in the last 68 patients. Among 134 cases with an intact preoperative spinal cord, one patient (0.7%) developed a new paraplegia due to a nonfunctional Gott shunt. Among 98 patients perfused with a centrifugal pump-driven left heart bypass, none of the 97 patients (0%) with a preoperative intact spinal cord developed a spinal cord ischemic deficit.
Conclusion
Conventional open surgical repair of acute traumatic rupture of the thoracic aorta performed with an orderly monitored circulatory support can be accomplished with a very low rate of mortality and spinal cord injury. Compared with the Gott shunt, a left heart bypass propelled with a centrifugal pump is technically a more versatile method of perfusion, and it provides higher hemodynamic performance.
Acute traumatic rupture of the thoracic aorta affects mainly a very young patient population with basically unlimited physiological reserves. Those who survive on arrival may generally be submitted to an immediate open surgical repair unless associated lethal injuries take priority and dictate an orderly delayed aortic reconstruction. Issues continue to be raised concerning the ideal technique of organ protection and repair, including endovascular grafting proposed more recently as a first-line approach and as a solution to the variability of surgical results. The present study revisits a uniform open-surgical approach based on the principle of organ protection refined over years during an accumulative personal experience in 1048 descending thoracic aortic aneurysm resections applied rigorously to the repair of acute thoracic aortic injury.
Methods
Institutional review board was obtained for the review of all patient records both relevant to this retrospective series. From 1974 to 2009, 138 consecutive survivors on arrival with an acute traumatic rupture of the thoracic aorta were surgically repaired by the author except for two cases. Ages of patients varied from 16 to 77 years (mean, 27 years). The incidence of aortic disruption as a single lesion was 9.4% (13/138 cases). Associated injuries were distributed in three major body regions (Table I). The mean injury severity score ISS was evaluated at 44 with a median of 41.
Table I. Distribution of associated injuries
| Body region | Number of cases | % |
|---|---|---|
| Head | 42 | 30.5 |
| Abdomen | 56 | 41 |
| Pelvis and limbs | 83 | 61 |
A mean 12-hour interval between the accident and the aortic repair reflects a large provincial territory covered by our trauma center and the priority also given to the treatment of associated life-threatening lesions over the aortic repair. The orderly evaluation and operative planning of all associated injuries were established by a group of surgical specialists involved in each patient's management along with a clear leadership exerted by a cardiovascular surgeon.
In 54 patients (39.1%), major associated life-threatening injuries were given priority over the aortic repair (Table II) and in 10 patients, the operation was delayed from 6 days to 3 months (mean delay, 30 days) for three main reasons: a major brain trauma with an unpredictable outcome (five cases), a severe cardiopulmonary contusion with bronchial bleeding (two cases), a septicaemia respectively due to pneumonia (one case), and a subphrenic abscess secondary to a splenectomy (one case), and finally a severe hypovolemic shock following a pancreatectomy undertaken for a massive pancreatic trauma (one case). In five patients with brain trauma, a beta-blocker regimen was necessitated to treat hypertension during the necessary recovery time to obtain suitable conditions for a safe and definitive open surgical repair.1 In 14 patients, an acute aortic syndrome required an emergency thoracotomy under very poor physiological conditions.
Table II. Etiology of associated life-threatening injuries in 54 patients
| Injuries | Number of cases | % |
|---|---|---|
| Severe brain trauma | 21 | 15.2 |
| Primary laparotomy for hemorrhage | 25 | 18.1 |
| Old patients with angina | 2 | 1.4 |
| Severe bilateral lung contusion and hypoxia | 2 | 1.4 |
| Severe heart contusion | 2 | 1.4 |
| Preoperative embolization for bleeding pelvis | 1 | 0.7 |
| Preoperative embolization for a bleeding transected kidney | 1 | 0.7 |
In 11 patients, an acute coarctation syndrome was characterized clinically by an absence of femoral pulses with a complete paraplegia in four cases, an anuria in five cases, and a concomitant lower limb and intestinal ischemia in one case. At the time of the operation, the pressure gradient recorded between the right radial and right femoral arteries ranged from 52 mm Hg to 80 mm Hg (mean, 53 mm Hg).
Three other patients stable on admission required an in extremis thoracotomy for sudden hypovolemic shock, and one of them needed a cardiac massage on his way to the operating room. Before the aorta could be controlled, the amount of left chest drainage for each patient was respectively 5, 12, and 16 liters of blood.
Surgical technique
All but two patients were repaired by the author (AV), a cardiovascular surgeon who made a point to use all optimal operative resources available, including anesthesiologists, perfusionists, and cardiovascular nurses, all members of a team working together on a daily basis. Standard left posterolateral thoracotomy through the bed of the fourth or fifth resected rib was performed, and Dacron graft interposition was used for the aortic repair in all. A mandatory distal perfusion was established as a rule. In the first 40 cases, a heparin-coated 9-mm Gott shunt (Sherwood Medical Instruments, St. Louis, Mo) inserted between the ascending and the descending aorta without systemic heparinization was favored.
A flow meter was adapted on the shunt, and distal pressure monitoring was also added because, in one instance, malfunction of the Gott shunt was unable to be determined by visual inspection alone. Maximum shunt flow recorded was 4 L/min with a median of 3 L/min, mostly dependent on the proximal aortic pressure. Mean distal aortic pressure recorded was 61 mm Hg with a median of 50 mm Hg (Table III).
Table III. Comparative hemodynamic data
| Accepted significant P value < .05 | Number of cases: 40 9-mm Gott shunt | Number of cases: 98 left heart bypass |
|---|---|---|
| Flow (mL/min) | Maximal 4000 | Maximal6000 |
| P = .001 | Median3000 | Median4000 |
| Distal pressure mm Hg | Mean61 | Mean73 |
| P = .01 | Median50 | Median72 |
In 1994, passive shunting was abandoned and replaced by a standard left heart bypass driven with a Bio-Medicus pump (Medtronic, Inc, Minneapolis, Minn) routinely equipped with a flow meter and usable without systemic heparinization.2 In absence of associated bleeding injuries, a minimal bolus of 3000 units of heparin was given to reach an active clotting time around 150 seconds. Besides intraoperative monitoring of pump flow, simultaneous right radial and femoral artery pressures, cardiac output, venous oxygen saturation (SVO2), pulmonary and left atrial pressure were also routinely recorded.
The last 98 cases were protected with this type of circulatory support inserted between the left atrium and the descending thoracic aorta.
Pump flow up to 6 L/min was reached with a median of 4 L/min (Table III) directly related to the preoperative preload volume and cardiac output. Adequate cardiac loading for optimal left heart bypass performance relies on a perfect synchronization between the anesthesiologist and the perfusionist. As the chest is open, proximal and distal aortic control is usually achieved first.
However, in two specific situations, the immediate installation of the left heart bypass has priority over a primary aortic approach: first, when one must anticipate the possibility of sudden rupture of the false aneurysm upon opening the chest. The pump is started with a minimal flow (less than 1 L/min) during the mediastinal dissection in order to prevent stasis in the pump circuit. Also, this facilitates rapidly going on bypass with optimal flow should the need of emergency proximal and distal cross-clamping arise.
Second, when the presence of a severe malperfusion syndrome in which immediate distal perfusion institution becomes an essential part of the resuscitation process allowing reversal of a potential vicious circle made of intractable metabolic acidosis, cardiovascular depression, and heart fibrillation.
When graft interposition is completed, meticulous de-airing of the aortic graft and of the left subclavian artery was performed before releasing respectively the aortic and the left subclavian artery clamps to prevent air embolization to the brain, heart, and spinal cord. Also, declamping shock is prevented by discontinuing the left heart bypass before proceeding to the aortic clamp release.
Results
Comparison of hemodynamic data obtained with each method of perfusion has been evaluated according to a P value less than .05 considered as significant with the Students' t test (Table III).
The left heart bypass driven with a centrifugal pump disclosed a clear hemodynamic superiority compared with the passive 9-mm Gott shunt, delivering larger flow rate with a superior distal perfusion pressure. With this type of circulatory support, three patients who underwent an emergency thoracotomy for active aortic bleeding survived without paraplegia. Their body temperature dropped down to 31°C without even minor arrhythmia.
Among 11 patients presenting a severe coarctation syndrome, one of them died from a massive intestinal ischemia nonresponsive to the re-establishment of distal aortic flow.
None of the four patients who had paraplegia prior to the aortic repair recovered. Five patients who had anuria before the operation resumed adequate urinary output as soon as the left heart bypass was initiated and all recovered normal kidney function postoperatively.
The overall hospital survival rate is 95% (131/138 patients). In-hospital mortality of 7.5% (3/40 patients perfused with the Gott shunt) was reduced to 4% (4/98 patients perfused with the left heart bypass) and further improved to 1.4% among the last 68 patients (1/68). Causes of death in seven cases were exclusively due to associated injuries (Table IV).
Table IV. Seven deaths
| Causes of death | Number |
|---|---|
| Irreversible brain injury | 3 |
| Stroke due to traumatic dissections of intracranial left carotid artery | 1 |
| Ischemic cardiac failure (77-year-old patient) | 1 |
| Massive intestinal gangrene (preoperative aortic embolization) | 1 |
| Septicemia from massive crushed leg injury | 1 |
Among 10 patients submitted to a delayed surgical strategy, no aortic exsanguination occurred. Paraplegia was already present on arrival in four patients with associated acute coarctation syndrome. Among 134 other patients with a preoperative intact spinal cord, paraplegia occurred in one case (0.7%) in which a nonfunctional Gott shunt was not recognized. With the exclusion of one patient who was paraplegic before the operation, no new paraplegia occurred in 97 patients perfused with the left heart bypass. In the overall series, the mean aortic cross-clamp time was 29.3 ± 13 minutes with a median of 25 minutes.
Discussion
The keystone of this series is a uniform approach regarding all operations performed in the descending thoracic aorta with a compulsory utilization of distal perfusion mandatory applied in traumatic aortic repair. Direct cross-clamping suggested many years ago by Mattox et al3 was not performed in our series. It is our conviction that distal perfusion is not only organ protective but also may be life-saving in many circumstances. It may prevent metabolic deterioration during routine aortic repair. In unstable patients with pre-existing severe metabolic acidosis due to either a hemorrhagic shock or a coarctation syndrome, it represents an essential adjunct to the resuscitation maneuvers during the operation. Ideally when 60% to 70% (mean, 61%) of the cardiac output was permissively derided with the pump, normal proximal pressure between 100 mm Hg and 125 mm Hg was preserved for heart and brain perfusion, and a mean distal pressure around 75 mm Hg was also maintained for distal organ and spinal cord requirements.
These hemodynamic conditions are usually easily reproducible in this population of very young patients. The same observation was reported by McCroskey4 in 1990. However, older traumatized patients occasionally encountered may develop unsuitable high distal pressure (over 90 mm Hg) as soon as the left heart bypass is initiated. This requires an immediate pharmacologic intervention by the anesthesiologist in order to correct excessive arterial bed resistance. Otherwise, in order to reverse the situation, pump flow reduction is needed at the risk of jeopardizing spinal cord perfusion.
Basic requirements in oxygen saturation may at times be difficult to meet during left lung exclusion. In our own experience with most patients repaired in the first 12 hours after the accident, pulmonary contusion was a rare cause of intraoperative hypoxia. It was mostly attributable to circulatory overload secondary to overaggressive fluid resuscitation. This condition is easily managed by using diuretics and continuous positive airway pressure (CPAP). The surgeon may have to synchronize the aortic repair between periods of temporary left lung reinflation at the request of the anesthesiologist.
Two series representative of practices around the world offer a major point of interest for comparison. In each one, only stable patients on arrival were retained to compare adequately our own patients treated with a mean delay of 12 hours for various reasons. The first one, published by Van Oppell,5 comprises 1492 patients reaching the operating room in stable condition. The overall mortality was 21%, including an intraoperative mortality rate of 8%, and the paraplegia rate was 10%. Up to 30% of the patients did not receive any method of perfusion.
The second is a prospective multicenter series of 207 patients6 also stable on arrival at the operating room with an ISS of 42, very similar to an ISS of 44 observed in our own series. The hospital mortality was 15%, the intra-operative mortality was 4%, and the paraplegia rate was 9%. Up to 35% of patients were repaired with the “clamp-and-sew” technique. The intraoperative deaths reported in both series for stable patients on arrival at the operating theater raise unresolved questions.
In our own series, a single case of paraplegia occurred with a nonfunctional Gott shunt that escaped our visual attention. We then realized that any chosen mode of circulatory support should be accompanied by precise distal flow monitoring. This was a clear motivation to adjust a flow meter on the Gott shunt and subsequently to adopt a fully monitorized left heart bypass also usable without systemic heparinization.
Pate et al7 in 1994 opened the way to the safe use of partial cardiopulmonary bypass for spinal cord protection in trauma patients. In a series of 86 stable patients, a pump flow varying from 1.8 L/min to 2.8 L/min was obtained, as well as distal perfusion pressure of 60 mm Hg. Despite a longer cross-clamp time than ours (45 minutes vs 25 minutes), the paraplegia rate was low (2.3%).
In 2001, Bouchard et al8 reported a series of 97 patients protected with a partial right heart bypass equipped with an oxygenator. Pump flow varied from 1.1 L/min to 2.5 L/min, and averaged distal pressure recorded was 60 mm Hg. The average aortic cross-clamp time was 45 minutes. Despite two transient and late ischemic spinal cord deficit occurring on days 5 and 8 after the operation with complete recovery, no definitive paraplegia was encountered. Bouchard et al8 also had a very low overall hospital mortality rate of 4%, improved from 6.6% in the first 45 patients to 1.9% in the last 52 patients. These results, very similar to ours, represent another demonstration of model results obtained with an alternative standardized technique.
Questions still persist regarding optimal flow and pressure needed to ensure optimal spinal cord perfusion. In 1998, Crawford et al,9 when using a left heart bypass driven with a centrifugal pump for repair of degenerative aneurysms, found that larger flows (2 L/min to 4.6 L/min) associated with lower pressures (20 mm Hg to 79 mm Hg) resulted in less neurological injury than high pressures (80 mm Hg to 116 mm Hg) associated with lower flows (0.8 L/min to 1.9 L/min). This was confirmed by Downing et al10 in 2000 while using a centrifugal pump adapted on a partial cardiopulmonary circuit equipped with an oxygenator but without reservoir. With this simplified right heart bypass, 50 cases of acute traumatic aortic disruptions were repaired. Pump flow varied from 2 L/min to 3.5 L/min (mean flow, 3 ± 0.8 L/min), and the distal pressure generally recorded was over 60 mm Hg. No spinal cord ischemic deficit occurred.
In our series of young patients who needed a very short proximal aortic resection, a left heart bypass driven with a Bio-Medicus pump resulted in no paraplegia in 97 consecutive cases. Larger flows (median, 4 L/min) and higher distal perfusion pressures (median, 72 mm Hg) were obtained.
The prerequisite for optimal pump performance is a close collaboration between the anesthesiologist and the perfusionist, mostly required in hypovolemic situations.
Despite a comparative ISS of 44, very comparative to other reports, our series may be viewed by certain observers as somewhat selective with the exclusion of moribund patients. The reality is otherwise, with the inclusion of three patients who developed an acute aortic hemorrhage soon after arrival and 11 other patients who also needed an emergency repair for a severe malperfusion syndrome.
Conclusion
The ultimate goal of traumatic aortic disruption repair is patient survival without ischemic organ injury with great emphasis on spinal cord protection. The prerequisite for the success of this surgical procedure is to rely on an optimal operating room team with special emphasis on an expertise in anesthesia that should contribute to few or non-operative nor in-hospital deaths due to acute traumatic aortic rupture.
Our method of circulatory support has undergone significant hemodynamic refinements when a passive shunt was replaced by a fully monitorized left heart bypass driven with a centrifugal pump. Both distal flow and perfusion pressure have been enhanced. The safety and the protective effect of an active left heart bypass have been demonstrated. The aortic lesion is usually located in the proximal descending aorta with very few intercostal artery involvements. This anatomical factor combined with a carefully monitorized active method of perfusion may contribute to a very low incidence of paraplegia. Comparative low mortality and paraplegia rates have also been reported by other authors that have used a right heart bypass combined with an active pump component as another alternative. We anticipate these data, along with other contemporary series, will serve as an incentive to retain the conventional open technique as the standard of care in this young trauma population.
Author contributions
References
- . Is traumatic rupture of the aorta misunderstood?. Ann Thorac Surg. 1994;57:530–531
- . Évolution du support circulatoire au cours de 420 résections anévrismales de l'aorte thoracique descendante ([Circulatory support refinements during resection of 420 aneurysms of the descending thoracic aorta]). Ann Chir. 1966;50:619–625
- . Clamp/repair: a safe technique for treatment of blunt injury to the descending thoracic aorta. Ann Thorac Surg. 1985;40:456–463
- . A unified approach to the torn thoracic aorta. Am J Surg. 1991;162:473–476
- . Traumatic aortic rupture: twenty-year metaanalysis of mortality and risk of paraplegia. Ann Thorac Surg. 1994;58:585–593
- Prospective study of blunt aortic injury: multicenter trial of the American Association for the Surgery of Trauma. J Trauma. 1997;42:374–383
- . Acute traumatic rupture of the aortic isthmus: repair with cardiopulmonary bypass. Ann Thorac Surg. 1995;59:90–99
- . Ruptures traumatiques aiguës de l'aorte thoracique et de ses branches (Résultats du traitement chirurgical). [Acute traumatic rupture of the thoracic aorta and its branches. Results of surgical management] Ann Chir. 2001;126:201–211
- . The impact of distal aortic perfusion and somatosensory evoked potential monitoring on prevention of paraplegia after aortic aneurysm operation. J Thorac Cardiovasc Surg. 1988;95:357–367
- Heparinless partial cardiopulmonary bypass for the repair of aortic trauma. J Thorac Cardiovasc Surg. 2000;120:1104–1111
Competition of interest: none.
The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a competition of interest.
PII: S0741-5214(09)01691-7
doi:10.1016/j.jvs.2009.05.066
© 2010 Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.
