From the Society for Vascular Surgery
A new classi?cation scheme for treating blunt aortic injury
Benjamin W. Starnes, MD, FACS,a Rachel S. Lundgren, MD,a Martin Gunn, MBChB,b
Samantha Quade, MD,a Thomas S. Hatsukami, MD,a Nam T. Tran, MD,a Nahush Mokadam, MD,c and Gabriel Aldea, MD,c Seattle, Wash
Background: There are numerous questions about the treatment of blunt aortic injury (BAI), including the management of small intimal tears, what injury characteristics are predictive of death from rupture, and which patients actually need intervention. We used our experience in treating BAI during the past decade to create a classi?cation scheme based on radiographic and clinical data and to provide clear treatment guidelines. Methods: The records of patients admitted with BAI from 1999 to 2008 were retrospectively reviewed. Patients with a radiographically or operatively con?rmed diagnosis (echocardiogram, computed tomography, or angiography) of BAI were included. We created a classi?cation system based on the presence or absence of an aortic external contour abnormality, de?ned as an alteration in the symmetric, round shape of the aorta: (1) intimal tear (IT)—absence of aortic external contour abnormality and intimal defect and/or thrombus of <10 mm in length or width; (2) large intimal ?ap (LIF)—absence of aortic external contour abnormality and intimal defect and/or thrombus of >10 mm in length or width; (3) pseudoaneurysm—presence of aortic external contour abnormality and contained rupture; (4) rupture— presence of aortic external contour abnormality and free contrast extravasation or hemothorax at thoracotomy. Results: We identi?ed 140 patients with BAI. Most injuries were pseudoaneurysm (71%) at the isthmus (70%), 16.4% had an IT, 5.7% had a LIF, and 6.4% had a rupture. Survival rates by classi?cation were IT, 87%; LIF, 100%; pseudoaneurysm, 76%; and rupture, 11% (one patient). Of the ITs, LIFs, and pseudoaneurysms treated nonoperatively, none worsened, and 65% completely healed. No patient with an IT or LIF died. Most patients with ruptures lost vital signs before presentation or in the emergency department and did not survive. Hypotension before or at hospital presentation and size of the periaortic hematoma at the level of the aortic arch predicted likelihood of death from BAI. Conclusions: As a result of this new classi?cation scheme, no patient without an external aortic contour abnormality died of their BAI. ITs can be managed nonoperatively. BAI patients with rupture will die, and resources could be prioritized elsewhere. Those with LIFs do well, and currently, most at our institution are treated with a stent graft. If a pseudoaneurysm is going to rupture, it does so early. Hematoma at the arch on computed tomography scan and hypotension before or at arrival help to predict which pseudoaneurysms need urgent repair. ( J Vasc Surg 2012;55: 47-54.)
Since the seminal report by Parmley in 1958, much has changed in the diagnosis and treatment of blunt aortic injury (BAI). This was perhaps best described in the American Association for the Surgery of Trauma (AAST) 1 and 2 trials showing the transition from angiography to computed tomography (CT) for diagnosis and from open to endovascular repair as de?nitive treatment.1,2 AAST 2 and several meta-analyses have shown reduced mortality and paraplegia rates with endovascular repair of BAI compared with open repair, along with an increase in devicerelated complications.2-5
From the Division of Vascular Surgery,a Department of Radiology,b and Division of Cardiothoracic Surgery,c University of Washington, Harborview Medical Center. Competition of interest: none. Presented at the 2010 Vascular Annual Meeting of the Society for Vascular Surgery, Boston, Mass, June 10, 2010. Reprint requests: Benjamin W. Starnes, MD, FACS, Harborview Medical Center, 410 Ninth Ave, Box 359908, Seattle, WA 98104 (e-mail: firstname.lastname@example.org). The editors and reviewers of this article have no relevant ?nancial 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. 0741-5214/$36.00 Copyright ? 2012 by the Society for Vascular Surgery. doi:10.1016/j.jvs.2011.07.073
Numerous questions remain despite a growing body of experience in treating BAI with endovascular methods. These include management of small intimal tears and whether there are injury characteristics predictive of death from rupture that might persuade surgeons to choose one mode of therapy over another. These would include antiimpulse therapy with observation, endovascular repair, or open surgical repair. Our institution receives most of the severe trauma from a ?ve-state region in the United States and thus sees a relatively large volume of BAI. Our goal was to review our experience in treating BAI during the past decade to create a classi?cation scheme based on radiographic and clinical data and to provide clear treatment guidelines. MATERIALS AND METHODS This study was approved by the Institutional Review Board at the University of Washington, Harborview Medical Center. We retrospectively searched our hospital discharge database for International Classi?cation of Diseases codes 901.0 injury aorta (thoracic), 902.0 injury aorta (abdominal), 441.01 thoracoabdominal dissection, and 441.01 thoracic aneurysm ruptured. We also reviewed the 47
48 Starnes et al
JOURNAL OF VASCULAR SURGERY January 2012
Fig 1. Blunt aortic injury classi?cation scheme based on the presence or absence of aortic external contour abnormality on axial computed tomography imaging. Representative images are shown.
Radiology Information System (IDX9.9.7, General Electric Healthcare, Waukesha, Wisc) from 2004 to 2008. The analysis included all patients with a BAI diagnosis con?rmed operatively or radiographically by echocardiogram, CT, or angiography between 1999 and 2008. Medical, operative, and radiology records were evaluated. Clinical notes and radiographic studies were reviewed for demographic information, aortic injury details, operative details, Injury Severity Score (ISS), survival, hospital outcomes, and late outcomes ( 6 months). When possible, medical records from other hospitals were obtained for outcome data. After a review of the types of injuries commonly seen, we created a classi?cation system (Fig 1) based on the presence or absence of aortic external contour abnormality. This is de?ned as a change in the symmetric, round shape of the aorta. Various imaging protocols were used throughout the study period, and aortic measurements were made orthogonal to the aortic ?ow channel:
Injury characteristic de?nitions Location of aortic injury. Ascending aorta: from the aortic valve to the proximal origin of the innominate artery. Aortic arch: from the proximal origin of the innominate artery to the distal aspect of the left subclavian artery. Aortic isthmus: from the distal aspect of the left subclavian artery to 2 cm caudal to the left pulmonary artery (on axial imaging). Descending aorta: from 2 cm caudal to the left pulmonary artery to the diaphragmatic hiatus. Abdominal aorta: from the diaphragmatic hiatus to the aortic bifurcation. Death from blunt aortic injury. Death directly resulting from injury from exsanguination or cardiac arrest in the emergency department (ED) or in the operating room. Death caused by exsanguination or prolonged hypotension for which there was no other obvious cause. Hypotension. Systolic blood pressure (SBP) 90 mm Hg. Measurements. Pseudoaneurysm size: Length—full length of injury on axial imaging. Width—width of closest normal aortic diameter distal to the injury subtracted from maximal width of pseudoaneurysm (Fig 2). Size of hematoma. Aortic arch: measured in arch at the proximal origin of the left subclavian artery (Fig 3). Left pulmonary artery. Measured from the edge of the aorta to the edge of the hematoma antero-posterior at the level of and toward the left pulmonary artery. Lung. Measured from the edge of the aorta to the edge of the lung laterally at the level of the left pulmonary artery. Length. Measured on axial imaging. Diameters. Aortic diameter: largest normal aortic diameter just distal to the area of injury. Aortic arch curvature
Intimal tear—absence of aortic external contour abnormality and intimal defect and/or thrombus of 10 mm in length or width. Large intimal ?ap (LIF)—absence of aortic external contour abnormality and intimal defect and/or thrombus of 10 mm in length or width. Pseudoaneurysm—presence of aortic external contour abnormality and contained rupture. Rupture—presence of aortic external contour abnormality and free contrast extravasation or hemothorax found upon thoracotomy.
JOURNAL OF VASCULAR SURGERY Volume 55, Number 1
Starnes et al 49
Fig 2. Pseudoaneurysm width measurements. A, A width measurement of the maximum aortic diameter at the site of injury is shown. B, A width measurement of the normal aorta just distal to the site of injury is shown. The pseudoaneurysm size is calculated by subtracting the measurement in B from that in A.
Fig 3. Size of hematoma measurement at the level of the aortic arch just distal to the left subclavian artery is shown on (left) axial and (right) oblique images.
diameter: measured from mid-ascending aorta to middescending aorta at the level of the left pulmonary artery on axial imaging. External iliac artery: smallest diameter of each external iliac artery. Statistics. We used the 2 test to compare proportions for dichotomous and categoric variables. When the expected cell sizes were 5, we used Fisher’s exact test. We used t-tests for continuous variables. Because this analysis was exploratory, we did not adjust for multiple comparisons. To further explore contributors to death from BAI, we performed logistic regression comparing individuals with death from BAI to all other individuals, combining both individuals who were still alive and those that died from other causes. All P values reported are two-sided. Analysis was performed using SPSS (IBM, Somers, NY) and Stata software (StateCorp, College Station, Tex).
RESULTS Demographics. During the 10-year period from 1999 to 2008, 140 patients (27% female) with radiographically diagnosed BAI were treated at our institution. The average age was 40 (range, 6-89 years), and 69% had been in motor vehicle collisions. Of the patients with multiple injuries, 34% were hypotensive before or in the ED, and 10 patients (7.1%) lost vital signs en route or upon arrival. Most patients had multiple associated injuries, with a mean ISS of 39.5, most commonly chest, abdomen, and extremity injuries. Hypotensive patients had signi?cantly higher ISS (P .005). ISS was signi?cantly inversely correlated with age, such that younger patients had a higher ISS (P .01). ISS was higher for patients with pseudoaneurysms and ruptures (40 and 56, respectively) than for intimal tears and large intimal ?aps
50 Starnes et al
JOURNAL OF VASCULAR SURGERY January 2012
Table I. Type of blunt aortic injury (BAI) and corresponding Injury Severity Score (ISS)
ISS BAI type Intimal tear Large intimal ?ap Pseudoaneurysm Rupture Mean (range) 36.5 (20-70) 36.6 (16-57) 40.4 (20-75) 55.6 (25-86)
(36 and 37) after adjusting for age and sex (P .001; Table I). Location of injury. Table II reports the type of aortic injury and relative location within the aorta. Seventy percent of all injuries were near or at the aortic isthmus. There were few injuries in the ascending aorta or aortic arch, and most were without aortic contour abnormality. Most of the intimal tears were at the isthmus or in the descending thoracic aorta, whereas most large intimal ?aps were in the abdominal aorta. Half of the ruptures were also in the abdominal aorta. Most of the BAIs were pseudoaneurysms (71% of injuries) with 16.4% intimal tears. Pseudoaneurysm size. Pseudoaneurysms measured 8.7 4.3 mm in width (range, 0-31 mm) and 28 14.1 mm in length (range, 7-84 mm). Size of hematoma. Hematomas were at the level of the aortic arch (maximal diameter range, 1-27 mm) in 46 patients (433%), at the level of the left pulmonary artery (range, 2-30 mm) in 60 (43%), and at the same level extending laterally toward the lung (range, 2-19 mm) in 53 (38%). The average length of hematoma was 138 60.8 mm, and the average distance of injury from the left subclavian artery (for patients in whom the data was available) was 16.4 14.1 mm. With the exception of one patient with an intimal tear, all patients with hematoma around the aorta had a pseudoaneurysm. Aortic diameter. The average aortic diameter was 21.6 4.6 mm (range, 13-45 mm). The average aortic arch curvature diameter ranged widely, from 43 to 110 mm. The external iliac artery diameters averaged 7.5 mm (range, 4-13 mm). Age correlated signi?cantly with aortic diameter and aortic arch curvature diameter, such that both increased as patients aged (P .0001 for each). Outcomes by type of injury Outcomes by type of injury are reported in Table III. Intimal tears. There were 23 patients with an intimal tear (16.4% of the total population) located throughout the aorta but with most in the isthmus and descending aorta. Of these, 20 patients were treated nonoperatively, one underwent open repair, and two underwent endovascular repairs. Operative intervention was chosen early in our endovascular experience when the benign course of patients with minimal aortic injury was not yet well elucidated. Those who underwent surgery had no operative complications and all survived. Survival was 85% in those
treated medically: three died of multiorgan system failure, but none died from the BAI. Of the three patients who received intervention, two who had endovascular repair had follow-up CT imaging that showed resolution of the injury. Of the 20 patients treated nonoperatively, 17 survived. Additional imaging in 16 showed 14 injuries had healed and two were stable. Follow-up averaged 71 days (range, 2-196 days). All of the healed injuries had done so 3 months and more than half were healed in 1 month (Fig 4). Large intimal ?aps. There were eight patients with LIF (5.7% of the total population), found primarily in the abdominal aorta (62.5%). Only one patient (12.5%) was hypotensive, and the average ISS was 37. Radiographic progression resulted in two open repairs and four endovascular repairs, and two patients were treated nonoperatively. No surgical patient had procedural-related complications and none of the patients with LIF died. Of those treated with an intervention, all endovascular repair patients had at least one follow-up CT scan. CT follow-up averaged 745 days (range, 6-1840 days), with all injuries resolved and no new complications found. Two patients with LIF were treated nonoperatively, and both survived. Follow-up CT angiography (CTA) was available for both, with limited follow-up of 5 to 9 days; neither injury changed within that time frame. Pseudoaneurysms. Pseudoaneurysms were present in 100 patients (71% of the entire cohort), and 87% of these were at the isthmus. The average ISS in these patients was higher than in those without an aortic external contour abnormality at 40. Hypotension was noted in 37 (37%), and two patients (2%) lost vital signs before or in the ED. One of those two patients underwent open repair and died of exsanguination from the injury, and the other underwent endovascular repair and survived. Mortality was 24%, of which nine deaths (37.5%) were due to BAI (con?rmed intraoperatively at time of death, or after de?nitive repair with clinical ?ndings suggestive of aortic rupture). CTA follow-up averaged 581 days (range, 3-2960). Eighty-six patients underwent surgery (43 open repairs and 43 thoracic endovascular repair). Of the 14 patients with pseudoaneurysms treated nonoperatively, seven survived, and ?ve received additional CT imaging. With an average follow-up after discharge of 75 days (range, 12-168 days), none of the pseudoaneurysms had worsened in size or in shape. All of the seven patients who died with nonoperative treatment, died of multiple organ or respiratory failure, and none died of the BAI. Ruptures. Nine patients had rupture as diagnosed on thoracotomy: four in the thoracic aorta, four in the abdominal aorta, and one in the aortic arch. These patients had the highest ISS at 56, 100% were hypotensive, and 90% lost vital signs en route or in the ED. All nine patients received a thoracotomy and attempt at open repair. Endovascular repair was not considered for these patients at the time of presentation in extremis. Eight patients (89%) died; six of the eight deaths (75%) were directly from the BAI, the other two were from brain death and multiorgan system
JOURNAL OF VASCULAR SURGERY Volume 55, Number 1
Starnes et al 51
Table II. Type of aortic injury, relative location within the aorta, and corresponding mortality
AECA (–), No. (%) Location Ascending No. Dead Arch No. Dead Isthmus No. Dead Descending No. Dead Abdominal No. Dead Total Dead Intimal tear 2 0 2 0 7 2 7 1 5 0 23 (16.4) 3 (13) Large intimal ?ap ... 2 0 ... 1 0 5 0 8 (5.7) 0 (0) AECA ( ), No. (%) Pseudoaneurysm ... 5 0 87 21 5 3 3 0 100 (71.4) 24 (24) Rupture ... 1 0 4 4 ... 4 4 9 (6.4) 8 (89) Total No. (%) 2 0 10 0 98 27 13 4 17 4 140 35 (25)
AECA, Aortic external contour abnormality.
Table III. Outcome by type of blunt aortic injury (BAI)
Injury Intimal tear Large intimal ?ap Pseudoaneurysm Rupture Total Total (No.) 23 8 100 9 140 Open repair (No.) 1 2 43 9 55 TEVAR (No.) 2 4 43 0 49 Medical treatment (No.) 20 2 14 0 36 Dead (No.) 3 0 24 8 35 Non-BAI (No.) 3 0 15 2 20 BAI-related (No.) 0 0 9 6 15
TEVAR, Thoracic endovascular repair.
Fig 4. A, Intimal tear is shown at the initial presentation on May 7, 2006. B, The tear was stable after 4 days (May 11). C, The tear was completely healed on follow-up imaging at 38 days (June 14).
failure (resulting in part from massive blood loss from the BAI). The patient with aortic arch rupture was the only rupture patient not to lose vital signs and the only survivor to hospital discharge. Predictors of death from BAI. The BAI was the direct cause of death in 15 patients (11% of the entire patient population and 43% of the deceased). Before or at
ED arrival, 36 patients (26%) were hypotensive, with an average ISS of 36. Hypotension recorded at the time of ED admission was signi?cantly associated with death from rupture. In a multivariate logistic regression, only SBP 90 mm Hg was an independent predictor adjusting for age, sex, and ISS (Table IV). Other variables appeared to be associated, but the number of potential variables was lim-
52 Starnes et al
JOURNAL OF VASCULAR SURGERY January 2012
Table IV. Multivariate logistic regression of four variables associated with death from blunt aortic injurya
Variable Sex Age Systolic blood pressure 90 mm Hg Injury severity score OR (95% CI) 0.43 (0.07-2.38) 1.01 (0.98-1.04) 11.37 (2.30-56.30) 1.00 (0.95-1.05) P .33 .38 .003 .91
DISCUSSION There are an estimated 8000 BAIs per year in the United States, and in 2008, moderate-to-severe thoracic injuries were the second leading cause of trauma-related death. BAI is found in 0.1% to 0.6% of patients with serious trauma admitted to major trauma centers.3,6-8 The ages, sex distribution, and ISSs of our patients were quite similar to previously published reports. The ISS was signi?cantly higher in patients with more severe BAI, supporting the concept that the presence and the severity of BAI both serve as markers of trauma to the patient as a whole. Parmley’s classi?cation of BAI in 1958 was seminal in its ?rst description of aortic injury and awareness of the lethality associated with this devastating injury.9 This classi?cation, however, is unfortunately not relevant for today’s clinicians because it is pathologically based. Most BAIs are now diagnosed using CTA; hence, a classi?cation scheme based on CT images is more applicable to currently triage patients and choose a treatment. We have created a classi?cation scheme based on aortic external contour abnormality of the aorta with four types of injury: intimal tears, LIF ?aps (both enterobacterial common antigen negative), and pseudoaneurysms and ruptures (both enterobacterial common antigen positive). The four types of injury are similar to those described in the AAST I and II trials.1,2,10,11 Azizzadeh et al12 also described a similar system with intramural hematoma instead of large intimal ?aps. Their ?ndings were similar, especially in that intimal tears did well and ruptures fared poorly. We believe that intramural hematoma is radiographically poorly de?ned and may often be associated with injury to adjacent structures without a true aortic injury. Malhotra et al13 are credited with the ?rst published use of the term “minimal aortic injury” (MAI) in 2001. Since that time, with increasing use of CTA for diagnosis of BAI, increasing numbers of MAI have been diagnosed. Several small case series suggest most injuries heal, but others have identi?ed patients who progressed.13-16 Our report includes 23 patients (16.4%) with intimal tears 10 mm ?tting the diagnosis of MAI. We arbitrarily chose 10 mm based on our institutional experience. Most patients had at least one follow-up CT scan, and the injury in most healed, none of the injuries worsened, and no patient died from the BAI. Azizzadeh et al12 monitored 10 patients with MAIs nonoperatively with no deaths and no ruptures. Our current standard of care is to treat these injuries nonoperatively with anti-impulse therapy ( -blockers) and follow them with repeat CTA until they have demonstrated resolution of the injury. Our follow-up imaging regimen occurs at 30 days, 6 months, 1 year, and every other year thereafter. Patients with LIF all did well with short-term followup. The two treated nonoperatively remained stable. The current standard of care for nontraumatic type B aortic dissections is nonoperative management with anti-impulse therapy. Patients with traumatic LIF may likewise not require intervention. That being said, we have limited
CI, Con?dence interval; OR, odds ratio. a Reference was nondeath from blunt aortic injury; therefore, an OR 1 indicates reduced risk of death and OR 1 indicates increased risk of death.
Table V. Size of periaortic hematoma correlated with death from blunt aortic injury (BAI)
Deathb Variablea Size of hematoma, mm At arch At left pulmonary artery At lung
Not BAI-related 7.7 (1-23) 6.9 (2-30) 6.9 (2-19)
BAI-related 20.3 (9-27) 6.7 (5-8) 9.7 (7-11)
P .001 .93 .15
Data are presented as mean (range). Alive or non-BAI death.
ited by small patient numbers. Rupture as the type of injury and loss of vital signs before or in the ED were also signi?cant predictors of death secondary to BAI. None of the pseudoaneurysm measurements signi?cantly correlated with death secondary to BAI. However, the size of the periaortic hematoma at the level of the aortic arch was signi?cantly correlated with death from BAI, with an average of 20.3 mm in those that died from their aortic injury vs 7.7 mm in those that did not (P .001). Patients with a 15-mm arch hematoma were signi?cantly more likely to die from their BAI than those with a smaller hematoma (P .011; Table V). Device-related complications. Device-related complications occurred in three patients. Stent graft collapse occurred in two patients, one in the previously mentioned patient at 334 days after the procedure and one on day 3. The latter was immediately recognized and corrected with implantation of a giant Palmaz stent. Intraoperative proximal migration of the stent graft in one patient required open explantation and subsequent open repair. This patient also survived this event. Endovascular operative details. Of the 43 endovascular repairs, eight (19%) required intentional coverage of the left subclavian artery. This was well tolerated by all eight patients, and subsequent subclavian revascularization was not required. One stroke (2.3%) occurred in a 61-year-old woman who also required an adjunctive iliac conduit procedure for delivery of the device. One access-related complication (2.3%) resulted in common femoral artery occlusion.
JOURNAL OF VASCULAR SURGERY Volume 55, Number 1
Starnes et al 53
follow-up available, and many of these injuries are concerning enough with evidence of progression on imaging that we have chosen to ?x some of them promptly. No patient with a normal aortic external contour died of their BAI. Those who received surgical repair did well, and those treated nonoperatively also did well, with stable or healed injuries. None of these injuries progressed to pseudoaneurysms or rupture. The overall mortality rate was also better for these injuries, as were the ISSs, suggesting a less injured patient. Patients with pseudoaneurysms were more likely to die of their aortic injury than those without a contour abnormality, with more than one-third of those who died succumbing from their BAI. Of interest, no patient initially treated nonoperatively had a worsening injury on follow-up imaging and none ruptured. This is the group in whom predictors of rupture would be most useful for triage and treatment decisions. Because pseudoaneurysms made up most of the BAI, and nearly all injuries with documented periaortic hematoma had associated pseudoaneurysm, the ?nding that hematoma at the level of the aortic arch correlated with BAI-related death may serve as a useful marker for choosing urgent repair. In addition, hypotension was a strong predictor of death from BAI. Although 37% of all patients with pseudoaneurysms were hypotensive before or at arrival, 75% of patients who died of their BAI were hypotensive. Of the two who were not, one died in the operating room and the other ruptured 334 days later after a collapse of his stent graft with enlarging pseudoaneurysm. Lancey et al17 also found hypotension was signi?cantly correlated with BAIassociated death. Hypotensive patients with larger hematomas may bene?t from early repair and intensive antiimpulse therapy. Ten patients (7.1%) lost vital signs either en route to or in the ED. Of these, eight were ruptures and two were pseudoaneurysms. Nine of these 10 patients ultimately died during the ?rst hospital day (eight ruptures, one pseudoaneurysm). It is a moral challenge to let a young patient die, but patients who present in extremis after blunt thoracic trauma and do not respond to resuscitative measures typically fare poorly, with the eventual result being death. We may choose in the future to prioritize scarce resources for other patients with a better chance of survival. Others have shown that patients who die from BAI do so within several hours of hospital admission. A recent autopsy study found that of all patients who died from BAI, 57% were dead at the scene or at arrival to the hospital, 37% died within the ?rst 4 hours, and only 6% died after 6 hours.18,19 Recent reports from Feezor et al6 and Azizzadeh et al12 show 15% to 27% of patients with BAI died in the ED even before de?nitive repair could begin. The corollary to this concept is that patients who survive the ?rst few hours in the hospital are a self-selected subset who have a more stable injury and are unlikely to rupture later. One study20 reported three groups of patients: those who die at the scene (70%-80%); those who are
Table VI. University of Washington clinical treatment guidelines for blunt aortic injury
1. All patients with radiographic evidence of blunt aortic injury (BAI) should undergo anti-impulse therapy with -blockade, if tolerated, coupled with antiplatelet therapy (81 mg aspirin). 2. Observation alone with interval follow-up computed tomography angiography (CTA) within 30 days is appropriate for all intimal tears 10 mm. 3. Selective management of large intimal ?aps ( 10 mm) is appropriate with repeat imaging within 7 days to assess for progression. Evidence of progression should be managed, when possible, with endovascular repair. 4. All patients with an aortic external contour abnormality should be considered for semielective ( 1 week) endovascular repair if there is a high likelihood of survival from other associated injuries. These patients should be monitored with CT imaging as follows: 1 month, 6 months, 1 year, and every other year thereafter. Patients with hypotension on presentation and aortic arch hematoma 15 mm should be repaired with endovascular methods on a more urgent basis. 5. Intentional left subclavian artery coverage without revascularization is well tolerated in a majority of patients with BAI. 6. Patients with traumatic brain injury and an aortic external contour abnormality should be considered for earlier repair if a deliberate increase in mean arterial pressure is deemed bene?cial for the patient.
unstable and die quickly upon admission, regardless of whether they receive an operative repair (2%-5%); and those who survive admission and rarely die of their BAI. Our ?ndings support this concept: all patients who died of their BAI did so en route, in the ED, or in the operating room (immediately after ED admission). The two exceptions were patients who received repair (one open and one endovascular) and ruptured afterward. CONCLUSION Our classi?cation scheme showed that no patient with a normal external contour of the aorta died of the BAI. Intimal tears can be managed nonoperatively. Patients with ruptures will die, and resources may be prioritized elsewhere. Patients with LIFs do well, and currently, most are being treated with a stent graft. If a pseudoaneurysm is going to rupture, it does so early (before or in the operating room). Our current practice is to treat all pseudoaneurysms with endovascular repair if the patient has a reasonable likelihood of survival. Hematoma at the arch on CT scan and hypotension before or at ED arrival help to predict which pseudoaneurysms need urgent instead of semielective repair. Longer-term follow-up of these patients is needed to determine the durability of an “endovascular ?rst” strategy. Our recommended clinical treatment guidelines for the management of BAIs are listed in Table VI. AUTHOR CONTRIBUTIONS Conception and design: BS, RL, NT, MG Analysis and interpretation: BS, RL, SQ, NM, GA, NT, MG Data collection: BS, RL, SQ, MG
54 Starnes et al
JOURNAL OF VASCULAR SURGERY January 2012
Writing the article: BS, RL Critical revision of the article: BS, RL, NT, TH, GA, NM, MG Final approval of the article: BS Statistical analysis: RL, SQ Obtained funding: Not applicable Overall responsibility: BS
1. Fabian TC, Richardson JD, Croce MA, Smith JS Jr, Rodman G Jr, Kearney PA, 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. 2. Demetriades D, Velmahos GC, Scalea TM, Jurkovich GJ, Karmy-Jones R, Teixeira PG, et al. Operative repair or endovascular stent graft in blunt traumatic thoracic aortic injuries: results of an American Association for the Surgery of Trauma Multicenter Study. J Trauma 2008;64: 561-71. 3. Tang GL, Tehrani HY, Usman A, Katariya K, Otero C, Perez E, et al. Reduced mortality, paraplegia, and stroke with stent graft repair of blunt aortic transections: a modern meta-analysis. J Vasc Surg 2008;47: 671-5. 4. Hoffer EK, Forauer AR, Silas AM, Gemery JM. Endovascular stentgraft or open surgical repair for blunt thoracic aortic trauma: systemic review. J Vasc Interv Radiol 2008;19:1153-64. 5. Takagi H, Kawai N, Umemoto T. A meta-analysis of comparative studies of endovascular versus open repair for blunt thoracic aortic injury. J Thorac Cardiovasc Surg 2008;135:1392-4. 6. Feezor RJ, Hess PJ, Martin TD, Klodell CT, Beaver TM, Lottenberg L, et al. Endovascular treatment of traumatic thoracic aortic injuries. J Am Coll Surg 2009;208:510-6. 7. Yamane BH, Tefera G, Hoch JR, Turnipseed WD, Acher CW. Blunt thoracic aortic injury: open or stent graft repair? Surgery 2008;144: 575-80. 8. Schulman CI, Carvajal D, Lopez PP, Soffer D, Habib F, Augenstein J. Incidence and crash mechanisms of aortic injury during the past decade. J Trauma 2007;62:664-7.
9. Parmley LF, Mattingly TW, Manion WC, Jahnke EJ. Nonpenetrating traumatic injury of the aorta. Circulation 1958;XVII:1958. 10. Demetriades D, Velmahos GC, Scalea TM, Jurkovich GJ, Karmy-Jones R, Teixeira PG, et al. Diagnosis and treatment of blunt thoracic aortic injuries: changing perspectives. J Trauma 2008;64:1415-9. 11. Demetriades D, Velmahos GC, Scalea TM, Jurkovich GJ, Karmy-Jones R, Teixeira PG, et al. Blunt traumatic thoracic aortic injuries: early or delayed repair--results of an American Association for the Surgery of Trauma prospective study. J Trauma 2009;66:967-73. 12. Azizzadeh A, Keyhani K, Miller CC 3rd, Coogan SM, Sa? HJ, Estrera AL. Blunt traumatic aortic injury: initial experience with endovascular repair. J Vasc Surg 2009;49:1403-8. 13. Malhotra AK, Fabian TC, Croce MA, Weiman DS, Gavant ML, Pate JW. Minimal aortic injury: a lesion associated with advancing diagnostic techniques. J Trauma 2001;51:1042-8. 14. Hirose H, Gill IS, Malangoni MA. Nonoperative management of traumatic aortic injury. J Trauma 2006;60:597-601. 15. Holmes JH 4th, Bloch RD, Hall RA, Carter YM, Karmy-Jones RC. Natural history of traumatic rupture of the thoracic aorta managed nonoperatively: a longitudinal analysis. Ann Thorac Surg 2002;73: 1149-54. 16. Kepros J, Angood P, Jaffe CC, Rabinovici R. Aortic Intimal injuries from blunt trauma: resolution pro?le in nonoperative management. J Trauma 2002;52:475-8. 17. Lancey RA, Pezzella AT, Phillips DA. Factors affecting early exsanguination and death in blunt thoracic aortic trauma. Asian Cardiovasc Thorac Ann 2004;12:202-7. 18. Burkhart HM, Gomez GA, Jacobson LE, Pless JE, Broadie TA. Fatal blunt aortic injuries: a review of 242 autopsy cases. J Trauma 2001;50: 113-5. 19. Hemmila MR, Arbabi S, Rowe SA, Brandt MM, Wang SC, Taheri PA, et al. Delayed repair for blunt thoracic aortic injury: is it really equivalent to early repair? J Trauma 2004;56:13-23. 20. Cook J, Salerno C, Krishnadasan B, Nicholls S, Meissner M, KarmyJones R. The effect of changing presentation and management on the outcome of blunt rupture of the thoracic aorta. J Thorac Cardiovasc Surg 2006;131:594-600. Submitted Jan 12, 2011; accepted Jul 16, 2011.