- Email: [email protected]
Injury grade is a predictor of aortic-related death among patients with blunt thoracic aortic injury
Injury grade is a predictor of aortic-related death among patients with blunt thoracic aortic injury Gerald R. Fortuna Jr, MD, Alexa Perlick, BS, Joseph J. DuBose, MD, Samuel S. Leake, BS, Kristofer M. Charlton-Ouw, MD, Charles C. Miller III, PhD, Anthony L. Estrera, MD, and Ali Azizzadeh, MD, Houston, Tex Objective: The current Society for Vascular Surgery Clinical Practice Guidelines suggest urgent (<24 hours) thoracic endovascular aortic repair for grade (G) II to G IV blunt thoracic aortic injuries (BTAIs). The purpose of this study was to determine whether some patients may require more emergency treatment. Methods: We reviewed imaging variables of prospectively collected BTAI patients between 1999 and 2014. We used computed tomographic angiography to classify BTAIs into four categories: G I, intimal tear; G II, intramural hematoma; G III, aortic pseudoaneurysm; and G IV, free rupture. Specific examination of G III injuries was undertaken in an effort to predict aortic-related mortality (ARM) before repair. For this subset, we examined pseudoaneurysm size, lesion/ normal aortic diameter ratio, and mediastinal hematoma location and size. Results: Among 331 patients with BTAIs, 86 died before imaging. Admission computed tomographic angiography was available for 205 patients (71.2% male; mean age, 39.3 years) with BTAIs (24 G I, 49 G II, 124 G III, 8 G IV). The mean Injury Severity Score was 35.6, and 22.4% had hypotension (<90 mm Hg). Overall mortality was 11.2% (G I/G II, 4.1%; G III/G IV, 15.3%; P [ .02). ARM was 2.4% (G I/G II, 0%; G III/G IV, 3.8%; P [ .09). ARM was significantly greater in G IV (3 of 8 [37.5%]) than G III (2 of 124 [1.6%]) vs G I/II (0 of 73 [0%]) injuries (P < .0001). Medical management alone was used in 53 (20 G I, 18 G II, 13 G III, and 2 G IV). Open repair was performed in 51 (3 G I, 9 G II, 36 G III, and 3 G IV) at a mean time to repair (TTR) of 10.6 hours. Thoracic endovascular aortic repair was conducted for 101 patients (1 G I, 22 G II, 75 G III, and 3 G IV) at a mean TTR of 9.4 hours. Median TTR for the overall population of BTAI patients was 24.0 hours from admission. (G I, 64.5 hours; G II, 24.0 hours; G III, 19.7 hours; and G IV, 3.5 hours). ARM occurred in four of five patients before planned repair (2 G III and 2 G IV), 7.0 6 3.6 hours from admission. No G I/II ARM occurred. Among eight G IV injuries, there were three ARMs. Focus on G III injuries through regression analysis demonstrated that early clinical/imaging variables (eg, mediastinal hematoma dimensions and lesion/normal aortic diameter ratio) were not significant predictors of ARM. Conclusions: Injury grade is a predictor of ARM among patients with BTAIs. Aggressive use of the current Society for Vascular Surgery Clinical Practice Guidelines at a busy level I trauma center resulted in low rates of ARM. In this setting, identification of additional physiologic and radiographic criteria indicating the need for emergency (vs urgent) repair of aortic pseudoaneurysms remains elusive. (J Vasc Surg 2016;-:1-7.)
Mortality from thoracic aortic injuries remains a leading cause of death after blunt trauma.1,2 Despite the lethal potential progression of blunt thoracic aortic injuries (BTAIs), the optimal timing to repair aortic injuries remains controversial. Delayed repair is frequently necessary because of significant concomitant injuries that preclude safe operative intervention.3 The use of anti-impulse therapy protocols have facilitated delayed approaches, and reports from the literature have demonstrated improved From the Department of Cardiothoracic and Vascular Surgery, The University of Texas Medical School at Houston and the Memorial Hermann Heart &Vascular Institute. Author conflict of interest: A.L.E. discloses a financial relationship with Gore and Maquet. A.A. discloses a financial relationship with Gore and Medtronic. Correspondence: Ali Azizzadeh, MD, 6400 Fannin St, Ste 2850, Houston, TX 77030 (e-mail: [email protected]). 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 conflict of interest. 0741-5214 Copyright Ó 2016 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvs.2015.11.046
results with delayed repair.4-11 The emergence of thoracic endovascular aortic repair (TEVAR), however, has raised new questions about the optimal timing of repair to provide the best results while limiting the potential for rupture before repair.12 The Society for Vascular Surgery (SVS) Clinical Practice Guidelines (CPG) suggest urgent (<24 hours) TEVAR for grade (G) II to IV BTAIs (Fig 1).13,14 Recent multicenter study data, however, suggest that G II injuries are unlikely to rupture during the initial hospitalization.15 At the other end of the clinical spectrum, patients with G IV injuries (Fig 2) with active extravasation most commonly undergo emergency repair. Between these injury grades lies a wide spectrum of traumatic pseudoaneurysms (G III) with potentially variable rupture rates depending on pseudoaneurysm size, associated radiographic findings, and other potential risk factors.12 Several additional radiographic grading systems have been developed to augment the current grading scale to help answer questions regarding the timing of repair for patients at the highest risk for rupture.6,16-18 Only one study to date, however, has tried to combine current grading schemes with radiographic and physiologic parameters in an effort to 1
JOURNAL OF VASCULAR SURGERY --- 2016
2 Fortuna et al
Fig 1. Classification of traumatic aortic injury. Adapted with permission from the Journal of Vascular Surgery, Vol 49, Azizzadeh A, et al, Blunt traumatic aortic injury: initial experience with endovascular repair, pages 1403-1408. Copyright Society for Vascular Surgery, 2009.
repair timing suggested by the SVS CPG.12,14 The purpose of our present study was to evaluate the BTAI outcomes of a busy level I trauma center, specifically examining the effect of early physiologic and radiographic criteria on aorticrelated mortality (ARM) due to rupture in a setting that aggressively uses the SVS CPG for BTAI care. METHODS
Fig 2. Image of a grade (G) IV blunt thoracic aortic injury (BTAI).
predict rupture risk in patients with G III BTAIs.12 Recent studies have suggested that a subset of patients with G III injuries may harbor early clinical/radiographic findings predictive of the need for emergency repair rather than the urgent
This study was approved by The University of Texas Medical School at Houston Committee for the Protection of Human Subjects, which serves as our local Institutional Review Board. Informed consent was waived for analysis of the prospectively collected trauma registry. Our database was queried to identify all patients with the diagnosis of BTAI admitted from September 1999 to December 2014. Demographics, physiologic status, and admission injury were abstracted. A trauma radiologist, a trauma surgeon, and a vascular surgeon interpreted available computed tomographic angiography (CTA) imaging to ensure agreement on BTAI grading. Once characterized, specific measurements of individual injuries were made. Examined variables included (1) diameter of the normal aorta proximal and distal to injury; (2) width and length of injury; (3) presence of associated radiographic findings, including mediastinal hematoma; and (4) measured depth of the mediastinal hematoma at the level of the aortic arch and the descending thoracic aorta.
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Fortuna et al 3
Fig 3. Blunt thoracic aortic injury (BTAI) treatment type over time at the University of Texas Houston. TEVAR, Thoracic endovascular aortic repair.
Our treatment algorithm, including our techniques for open repair (OR) and TEVAR have been previously reported.8,13,19 We instituted TEVAR as an alternative to OR for anatomically suitable patients in 2005. The early procedures were conducted with the off-label use of the TAG (W. L. Gore & Assoc, Flagstaff, Ariz), the first thoracic aortic device approved by the U.S. Food and Drug Administration, which was available in diameters of 26 mm to 40 mm intended for patients with aortic diameters of 23 mm to 37 mm. During this phase, young trauma patients with aortic diameters <23 mm were treated with OR. The approval of additional smaller diameter devices in 2008 enabled us to apply TEVAR to a wider range of patients. As a result, TEVAR gradually replaced OR as the treatment of choice for all anatomically suitable patients with thoracic aortic injuries at our center. The progressive shift from OR to TEVAR for repair of BTAI at our institution is demonstrated in Fig 3. Information on management and outcomes was identified through trauma registry abstraction and individual record review. Contributors to cause of death were categorized as stroke, refractory hemorrhage with hemodynamic instability, traumatic brain injury/neurologic, cardiac arrest not related to ongoing hemorrhage, sepsis secondary to contamination or infection, respiratory failure, multiorgan failure, and ARM. For the purpose of our study, ARM was defined as death directly related to BTAI or complications of BTAI treatment. Autopsy results were not available in all patients; however, operative findings and independent record review by a trauma and vascular surgeon were sufficient to make an ARM determination. Retrospective review of medical records included available outpatient follow-up data. No ARM or adverse events were noted in the follow-up period for the reported series. Data were analyzed using IBM SPSS Statistics 22.0 (IBM Corp, Armonk, NY). Continuous variables with a
normal distribution assessed by skewness (between 1.0 and 1.0) are reported using means 6 standard deviation. Continuous variables that were not normally distributed are reported using median and interquartile range (IQR). For the univariate analysis, we used the c2 test with Yates correction for comparison of categoric risk factors and the Student t-test or the Mann-Whitney test for comparison of continuous risk factors. Logistic regression analysis was used to identify factors independently associated with the development of ARM. Adjusted odds ratios and 95% confidence intervals (CIs) were derived. RESULTS Trauma registry review revealed 332 patients treated at our level I trauma center for BTAIs between 1999 and 2014 (Fig 4). The subsequent analysis excluded 86 patients who died before the CTA was obtained. Of the remaining 246 patients, CTAs for 205 were available for review in the archived radiology system. BTAI injuries identified included 24 G I, 49 G II, 124 G III, and 8 G IV injuries (Fig 4). BTAI patients (71.2% male) were a mean age of 39.3 years (Table). Emergency transfers from other facilities constituted 18.5% of the population, with the remainder arriving to our facility directly from the point of injury. The predominant mechanism of injury was motor vehicle collision (72.2%). On presentation, 22.4% were hypotensive (systolic blood pressure < 90 mm Hg), and 30.2% had a Glasgow Coma Score of #8. Median Injury Severity Score (ISS) was 35.0 (IQR, 14). CTA was followed by traditional diagnostic angiogram in 81 patients, with an additional 46 undergoing joint traditional angiogram combined with intravascular ultrasound (IVUS) evaluation. Four false-positive CTAs (1 G I, 1 G II, and 2 G III) were identified among these patients. After diagnosis, all patients underwent medical management using anti-impulse control to maintain systolic
4 Fortuna et al
JOURNAL OF VASCULAR SURGERY --- 2016
Fig 4. Flow chart shows demographics, treatment, and outcomes of blunt thoracic aortic injury (BTAI). CT, Computed tomography; LSAC, left subclavian artery coverage; SVS, Society for Vascular Surgery; TBI, traumatic brain injury; TEVAR, thoracic endovascular aortic repair.
blood pressure <120 mm Hg. Medical management alone was used in 53 BTAI patients (83.3% [20 of 24] G I, 36.7% [18 of 49] G II, 10.4% [13 of 124] G III, and 25% [2 of 8] G IV). Four patients included in the medical management-only group died on medical therapy while en route to planned urgent or emergency repair (2 G III and 2 G IV). Among patients selected for repair (TEVAR or OR), median time from admission to repair was 24.0 hours (IQR, 86 hours). OR was performed for 50 patients (2 G I, 9 G II, 36 G III, and 3 G IV) with an overall mortality of 8% (4 of 50). The median time to OR for patients with G III BTAIs was 9.5 hours (IQR, 62 hours) and 4.0 hours (IQR, 50 hours) for G IV injuries. No patients sustained ARM after successful OR. Additional complications occurring after OR included stroke in 4% (2 of 50), paraplegia in 8% (4 of 50), and acute renal failure in 22% (11 of 50). Median lengths of stay (LOS) in the intensive care unit and hospital among survivors of OR were 16.0 days (IQR, 21.0; mean, 22.9 6 24.1 days) and 25.5 days (IQR, 28.8; mean, 36.0 6 31.6 days), respectively.
TEVAR was performed in 101 patients (1 G I, 22 G II, 75 G III, 3 G IV), with an overall mortality rate of 5.0%. During the study period, TEVAR increased over time to supplant OR as the most common intervention for BTAI (Fig 3). One ARM occurred after TEVAR due to persistent thoracic hemorrhage from the aorta. Left subclavian artery coverage was required during TEVAR in 37 patients (36.6%). Complications after TEVAR included stroke in 2.0% (2 of 101), paraplegia in 2% (2 of 101), and acute renal failure in 8.9% (9 of 101). Median ICU and hospital LOS among survivors of TEVAR were 8.0 days (IQR, 0; mean, 9.3 6 9.1 days) and 8.0 days (IQR, 0; mean, 9.3 6 9.1 days), respectively. Overall mortality among the BTAI population surviving to obtain CTA was 11.2% (23 of 205; 2 G I, 1 G II, 17 G III, and 3 G IV). Cause of death identified included stroke in 8.7% (2 of 23), refractory hemorrhage with hemodynamic instability in 30.4% (7 of 23), traumatic brain injury/neurologic in 47.8% (11 of 23), cardiac arrest not directly related to ongoing hemorrhage in 8.7% (2 of 23), sepsis secondary to contamination or infection
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Fortuna et al 5
Table. Continued.
Table. Demographics, presentation, management, and outcomes of patients with blunt thoracic aortic injury (BTAI) Variables Age, mean 6 SD, years Male, No. (%) Transferred from another facility, No. (%) Mechanism Motor vehicle crash, No. (%) Motorcycle crash, No. (%) Auto vs pedestrian, No. (%) Other blunt, No. (%) Presentation and overall injury burden Hypotension (SBP <90 mm Hg) on arrival, No. (%) Admission Glasgow Coma Score #8, No. (%) ISS, mean 6 SD Admission hemoglobin, mean 6 SD, g/dL Admission lactate, median (IQR), g/DL SVS injury grades and characteristics G I, No. (%) Length of G I injury, median (IQR) cm G II, No. (%) Length of G II injury, median (IQR) cm G III, No. (%) Length of G III injury, median (IQR) cm Diameter of G III injury, median (IQR) cm G IV, No. (%) Normal aortic diameter Proximal to injury, mean 6 SD, mm Distal to injury, mean 6 SD, mm Imaging used for diagnosis in addition to CT CTA þ angiography, No. (%) CTA þ angiography þ IVUS, No. (%) Associated radiographic findings Any mediastinal hematoma, No. (%) Mediastinal hematoma at the level of the aortic arch, No. (%) Max mediastinal hematoma diameter at aortic arch level, median (IQR), mm Mediastinal hematoma at the level of the descending thoracic aorta, No. (%) Max mediastinal hematoma diameter at descending thoracic aorta level, median (IQR), mm Management Time from admission to OR or TEVAR, median (IQR), hours OR, No. (%) TEVAR, No. (%) Diameter of device used, median (IQR), mm Length of aortic coverage, median (IQR), cm LSA coverage, No. (%) Stroke after LSA coverage, No. (%)
Variables
Total (N ¼ 205) 39.3 6 18.6 146 (71.2) 38 (18.5) 148 27 13 17
(72.2) (13.2) (6.3) (8.3)
46 (22.4) 62 (30.2) 35.6 6 10.5 12.9 61.9 3.8 (3) 24 (11.7) 1.1 (1.9) 49 (23.9) 1.5 (1.3) 124 (60.5) 3.3 (2.4) 3.0 (0.7) 8 (3.9) 23.1 6 4.3 21.1 6 3.7 81 (39.5) 46 (22.4) 127 (62.0) 90 (43.9) 18.1 (11) 121 (59.0) 11.9 (5)
24.0 (86) 50 (24.4) 101 (49.3) 26 (6) 10 (1) 37/101 (36.6) 2/37 (0.05) (Continued)
Complications and outcomes Paraplegia, No. (%) Stroke, No. (%) Acute renal failure, No. (%) LOS, median (IQR), days ICU Hospital In-hospital mortality, No. (%) Time from admission to death, median (IQR), hours ARM, No. (%)
Total (N ¼ 205) 8 (3.9) 10 (4.9) 26 (12.7) 7.0 17.0 23 28.0
(16.5) (23.0) (11.2) (216)
5 (2.4)
ARM, Aortic-related mortality; CTA, computed tomography angiography; G, grade; ICU, intensive care unit; IQR, interquartile range; ISS, Injury Severity Score; IVUS, intravascular ultrasound; LOS, length of stay; LSA, left subclavian artery; OR, open repair; SBP, systolic blood pressure; SD, standard deviation; SVS, Society for Vascular Surgery; TEVAR, thoracic endovascular aortic repair.
in 17.4% (4 of 23), respiratory failure in 17.4% (4 of 23), multiorgan failure in 26.1% (6 of 23), and ARM in 21.7% (5 of 23). Regardless of management, no patients with G I or II injuries died of ARM. Two ARMs occurred in patients with G III injuries; one died #1 hour of CTA before repair attempt and the other died 2.5 hours after admission while in the ICU awaiting planned repair #24 hours. Three ARMs occurred among patients with G IV injuries. One patient was transferred from another hospital and died in the emergency department shortly after arrival. A second was brought directly to our facility but died in the ICU w3.5 hours after admission and before planned OR. The third ARM occurred after TEVAR in an 84-year-old woman who sustained severe multiorgan trauma. A postoperative CTA revealed persistent extravasation from the left subclavian artery. The patient developed cardiac arrest in the operating theater before attempted TEVAR revision. There were no significant differences between patients with G I and II injuries treated with medical management alone vs OR or TEVAR. No ARMs occurred in the G I or II treatment group during the initial hospitalization. Only one CTA-documented progression of injury was notedda patient with G II injury progressed to G III at 3 days and underwent TEVAR. There was no difference between medical management only and repair with regard to observed renal failure (8.6% vs 10.5%; P ¼ .77), paraplegia (2.9% vs 2.6%; P ¼ .95), or stroke (5.7% vs 5.3%; P ¼ .93). Patients undergoing repair had trends toward longer LOS in the ICU (11.0 6 12.3 days vs 6.4 6 8.5 days; P ¼ .06) and the hospital (20.5 6 13.5 days vs 18.4 6 20.1 days; P ¼ .46), but these differences were not statistically significant. Among the two patients with G III injury who had ARM, there was no difference in the presence of mediastinal hematoma at the level of the aortic arch (50% of ARM, 46.7% of survivors; P ¼ .93) or the descending thoracic aorta (50% of ARM, 80% of survivors; P ¼ .35)
6 Fortuna et al
compared with survivors. Likewise, the maximal depth of mediastinal hematoma measured at these locations was not different between those sustaining ARM and survivors (12.8 vs 26.0 6 11.6 mm at the arch and 14.0 vs 12.9 6 4.5 mm at the descending thoracic aorta, respectively). The size and length of the pseudoaneurysm was also not different between patients who sustained ARM and survivors (length: 3.9 6 1.6 cm vs 4.0 6 2.2 cm, P ¼ .92; width: 3.7 6 0.6 vs 3.0 6 0.7 cm, P ¼ .22). Median time from admission to repair for G III injuries was 11.6 hours (IQR, 17; mean, 11.4 6 8.6 hours). All patients with G IV injuries had mediastinal hematoma with active contrast extravasation and CTA findings consistent with contained aortic rupture. Four patients underwent emergency repair in <3 hours. Two patients underwent planned delayed repair for BTAIs. The first of these two patients had a severe liver laceration and underwent angiographic hepatic embolization, followed by OR, at 51 hours after admission. The second patient had active pelvic bleeding and underwent TEVAR after angiographic embolization 55 hours after admission. Multivariate logistic regression analysis was used to identify independent predictors of all-cause mortality among G III BTAI patients. Nonoperative management (odds ratio, 29.65; 95% CI, 5.62-156.49; P < .0001) was an independent predictor of all-cause mortality. ISS (continuous), although not a significant factor on regression, was noted to have an odds ratio of 1.02 (95% CI, 0.96-1.09) and a value of 0.5302 within the model. When multivariate logistic regression analysis was used in an attempt to identify ARM for G III injuries, no independent predictors could be identified. CTA measurements did not prove predictive, including diameter of the normal aorta (odds ratio, 0.88; 95% CI, 0.66-1.17; P < .3817) maximal pseudoaneurysm diameter (odds ratio, 1.04; 95% CI, 0.85-1.28; P ¼ .6798), and the ratio of these two values (odds ratio, 0.15; 95% CI, 0.01-1.96; P ¼ .1470). In a similar fashion, the presence of mediastinal hematoma was not predictive of ARM among G III injuries (odds ratio, 1.31; 95% CI, 0.32-5.32; P ¼ .7059). DISCUSSION The optimal treatment of BTAI remains a challenge of modern trauma care. Expanding data suggest that G I and II injuries represent low risk for rupture.14,15 The dramatic appearance and frequent associated instability of G IV injuries most commonly elicits emergency repair. Recent investigations have continued to focus on the optimal treatment of those injuries falling between these categoriesdthe G III BTAIs. Our present series illustrates that G III injuries may be among the most common types of BTAIs encountered in busy trauma centers (Fig 4). This prevalence, along with an uncertain natural history, warrants additional investigation to determine optimal treatment timing. Harris et al,12 at the University of Maryland, used bootstrap risk analysis modeling to identify a cohort of 49 BTAI patients during a 13-year period
JOURNAL OF VASCULAR SURGERY --- 2016
(2000 to 2013), including 18 ruptures. Among these 18 patients, however, three were G IV with active extravasation. The remaining 15 progressed to aortic rupture with clinical decompensation after a median of only 64 minutes (IQR, 44-124 minutes). Three ruptured during exposure for planned OR, and 12 ruptured before aortic intervention. Through modeling, they identified that size of the aortic lesion and the presence of mediastinal hematoma were effective predictors of ARM due to rupture. In our present report, we were unable to identify any independent predictors of ARM among G III injuries. As with the Maryland report,12 we noted that ARMs were prone to occur very early in the hospital course (median 3.5 hours in our series), at intervals that challenge the marshaling of resources. Our results do demonstrate, however, that the effective application of the SVS CPG at a busy level I trauma center results in a very low ARM, with only five of 205 ARMs (2.4%) occurring after the opportunity for BTAI CTA diagnosis and characterization. Consistent with recent literature on the topic,14,15 we identified no ruptures in patients with G I or II BTAIs, regardless of management strategy. Consistent with SVS CPG for BTAI treatment, the median time in our experience to OR was 9.5 hours (IQR, 62 hours) for patients with G III and 4.0 hours (IQR, 50 hours) for G IV injuries, with no deaths in patients undergoing OR techniques. Our endovascular results included a median time to repair (TTR) for G III injuries of 3.0 hours (IQR, 53 hours), whereas G IV injuries had a median TTR of 28.9 hours (IQR, 115 hours). This longer TTR for G IV TEVAR was, perhaps, a result of the complicated management of associated injuries, because the three patients in this subset had ISS of 29, 38, and 75, respectively. There was, however, one patient death related to TEVAR, and that was in a patient with a G IV injury and a persistent type Ia endoleak who died upon returning to the OR. Only four patients died from ARM before receiving repair. We were unable to identify any independent risk factors that were able to accurately predict ARM among G III injuries. Using a rapid response TTR resulted in a low ARM. Although early deaths after G III and IV injuries continue to occur, aggressive use of the SVS CPG, in our experience, appears to mitigate the relative usefulness of aortic diameter measurements and the identification of mediastinal hematoma on initial CTA. Although SVS grade has been shown to be a predictor of overall death among BTAI patients,15 our present study is the only one to focus explicitly on the risk of ARM among G III injuries. With regard to all-cause mortality among G III BTAIs in our present series, nonoperative management alone was the only independent predictor identified. This finding suggests that some BTAI patients sustain patterns of injury that preclude their selection for open or endovascular repair. Although ISS did not factor out to be an independent risk factor for mortality, the trend was toward positive predictability. Increased ISS was frequently associated with poor candidates for any repair.
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Our present report has several limitations that must be acknowledged, beginning with the inherent limitations associated with retrospective design. Our study extended over a dynamic period during which TEVAR replaced OR as the treatment of choice for patients with G II to IV BTAIs, which should also be noted. This transition represents changes in care that have the potential to confound our findings. It is also important to note that our early treatment approach, although successful in sustaining only one intervention-related death and resulting in a low rate of ARM, is a function of a high-volume trauma center with a readily available vascular surgical team. Extrapolation to all care environments should, therefore, be undertaken with caution. Finally, because of our low ARM, we were unable to use modeling methodologies used in prior studies on this topic.12 Management of patients with G III BTAIs remains a clinical challenge that requires additional investigation. The uncertain natural history of these lesions demands early evaluation for treatment. Most aortic ruptures occur very early in the hospital course, challenging the ability to effectively marshal the resources required for expedient intervention. Further study is required, ideally in a prospective multicenter fashion, to better elucidate subsets of patients who may require treatment of G III BTAIs earliest in the SVS CPG suggested period of 24 hours. CONCLUSIONS Injury grade is a predictor of ARM among patients with BTAIs. Adherence to the current SVS CPG for management of BTAI resulted in very low rates of ARM. Without further discriminating parameters, extended periods of medical management in patients with G III BTAIs appears unnecessary. In our experience, however, identification of additional physiologic and radiographic criteria indicating the need for emergency (vs urgent) repair of patients with aortic pseudoaneurysms (G III BTAIs) remains elusive.
Fortuna et al 7
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
We thank Troy Brown for document editing, Chris Akers for artwork, and Dr John Holcomb and Edmundo Dipasupil for the trauma registry.
15.
AUTHOR CONTRIBUTIONS Conception and design: GF, AP, JD, SL, KC, CM, AE, AA Analysis and interpretation: GF, KC, CM, AE, AA Data collection: GF, AP, JD, SL Writing the article: GF, AP, JD Critical revision of the article: GF, AP, JD, SL, KC, CM, AE, AA Final approval of the article: GF, AP, JD, SL, KC, CM, AE, AA Statistical analysis: CM, JD Obtained funding: Not applicable Overall responsibility: AA REFERENCES 1. Arthurs ZM, Starnes BW, Sohn VY, Singh N, Martin MJ, Andersen CA. Functional and survival outcomes in traumatic blunt
16.
17.
18.
19.
thoracic aortic injuries: an analysis of the National Trauma Databank. J Vasc Surg 2009;49:988-94. Teixeira PG, Inaba K, Barmparas G, Georgiou C, Toms C, Noguchi TT, et al. Blunt thoracic aortic injuries: an autopsy study. J Trauma 2011;70:197-202. Rabin J, Harris DG, Crews GA, Ho M, Taylor BS, Sarkar R, et al. Early aortic repair worsens concurrent traumatic brain injury. Ann Thorac Surg 2014;98:46-52. Demetriades D, Vemahos GC, Scalea TM, Jurkovich GJ, KarmyJones R, Teixeira PG, et al. Association for the Surgery of Trauma Thoracic Aortic Injury Study Group. Operative repair or endovascular stent graft in blunt traumatic thoracic injuries: results of an American Association for the Surgery of Trauma Multicenter Study. J Trauma 2008;64:561-70; discussion: 570-1. Osgood MJ, Heck JM, Rellinger EJ, Doran SL, Garrard CL, Guzman RJ, et al. Natural history of grade I-II blunt traumatic aortic injury presenting over 10 years at a single institution. J Vasc Surg 2013;59:334-42. Rabin J, DuBose JJ, Sliker CW, O’Connor JV, Scalea TM, Griffith BP. Parameters for successful nonoperative management of traumatic aortic injury. J Thorac Cardiovasc Surg 2014;147:143-50. Marco Di Eusanio M, Folesani G, Berretta P, Petridis FD, Pantalco A, Russo V, et al. Delayed management of blunt traumatic aortic injury: open surgical versus endovascular repair. Ann Thorac Surg 2013;95: 1591-7. Estrera AL, Gochnour DC, Azizzadeh A, Miller CC, Coogan S, Charlton-Ouw K, et al. Progress in the treatment of blunt thoracic aortic injury: 12-year single-institution experience. Ann Thorac Surg 2010;90:64-71. Paul JS, Neideen T, Tutton S, Milia D, Tolat P, Foley D, et al. Minimal aortic injury after blunt trauma: selective nonoperative management is safe. J Trauma 2011;71:1519-23. Holmes JH, 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. 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. Harris DG, Rabin J, Kufera JA, Taylor BS, Sarkar R, O’Connor JV, et al. A new aortic injury score predicts early rupture more accurately than clinical assessment. J Vasc Surg 2015;61:332-8. Azizzadeh A, Keyhani K, Miller CC, Coogan SM, Safi HJ, Estrera AL. Blunt traumatic aortic injury: initial experience with endovascular repair. J Vasc Surg 2009;49:1403-8. Lee WA, Matsumura JS, Mitchell RS, Farber MA, Greenberg RK, Azizzadeh A, et al. Endovascular repair of traumatic thoracic aortic injury: clinical practice guidelines of the Society for Vascular Surgery. J Vasc Surg 2011;53:187-92. DuBose JJ, Leake SS, Brenner M, Pasley J, O’Callaghan T, LuoOwen X, et al. Aortic Trauma Foundation. Contemporary management and outcomes of blunt thoracic aortic injury: a multicenter retrospective study. J Trauma Acute Care Surg 2015;78:360-9. Gunn ML, Lehnert BE, Lungren RS, Narparla CB, Mistumori L, Gross JA, et al. Minimal aortic injury of the thoracic aorta: imaging appearance and outcome. Emerg Radiol 2014;21:227-33. Lamarch Y, Berger FH, Nicolaou S, Bilawich AM, Louis L, Inacio JR, et al. Vancouver simplified grading system with computed tomographic angiography for blunt aortic injury. J Thorac Cardiovasc Surg 2012;144:347-54. 354.e1. Starnes BW, Lundgren RS, Gunn M, Quade S, Hatsukami TS, Tran NT, et al. A new classification scheme for treating blunt aortic injury. J Vasc Surg 2012;55:47-54. Azizzadeh A, Ray HM, Dubose JJ, Charlton-Ouw KM, Miller CC, Coogan SM, et al. Outcomes of endovascular repair for patients with blunt traumatic aortic injury. J Trauma Acute Care Surg 2014;76: 510-6.
Submitted Sep 14, 2015; accepted Nov 12, 2015.
Mortality from thoracic aortic injuries remains a leading cause of death after blunt trauma.1,2 Despite the lethal potential progression of blunt thoracic aortic injuries (BTAIs), the optimal timing to repair aortic injuries remains controversial. Delayed repair is frequently necessary because of significant concomitant injuries that preclude safe operative intervention.3 The use of anti-impulse therapy protocols have facilitated delayed approaches, and reports from the literature have demonstrated improved From the Department of Cardiothoracic and Vascular Surgery, The University of Texas Medical School at Houston and the Memorial Hermann Heart &Vascular Institute. Author conflict of interest: A.L.E. discloses a financial relationship with Gore and Maquet. A.A. discloses a financial relationship with Gore and Medtronic. Correspondence: Ali Azizzadeh, MD, 6400 Fannin St, Ste 2850, Houston, TX 77030 (e-mail: [email protected]). 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 conflict of interest. 0741-5214 Copyright Ó 2016 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvs.2015.11.046
results with delayed repair.4-11 The emergence of thoracic endovascular aortic repair (TEVAR), however, has raised new questions about the optimal timing of repair to provide the best results while limiting the potential for rupture before repair.12 The Society for Vascular Surgery (SVS) Clinical Practice Guidelines (CPG) suggest urgent (<24 hours) TEVAR for grade (G) II to IV BTAIs (Fig 1).13,14 Recent multicenter study data, however, suggest that G II injuries are unlikely to rupture during the initial hospitalization.15 At the other end of the clinical spectrum, patients with G IV injuries (Fig 2) with active extravasation most commonly undergo emergency repair. Between these injury grades lies a wide spectrum of traumatic pseudoaneurysms (G III) with potentially variable rupture rates depending on pseudoaneurysm size, associated radiographic findings, and other potential risk factors.12 Several additional radiographic grading systems have been developed to augment the current grading scale to help answer questions regarding the timing of repair for patients at the highest risk for rupture.6,16-18 Only one study to date, however, has tried to combine current grading schemes with radiographic and physiologic parameters in an effort to 1
JOURNAL OF VASCULAR SURGERY --- 2016
2 Fortuna et al
Fig 1. Classification of traumatic aortic injury. Adapted with permission from the Journal of Vascular Surgery, Vol 49, Azizzadeh A, et al, Blunt traumatic aortic injury: initial experience with endovascular repair, pages 1403-1408. Copyright Society for Vascular Surgery, 2009.
repair timing suggested by the SVS CPG.12,14 The purpose of our present study was to evaluate the BTAI outcomes of a busy level I trauma center, specifically examining the effect of early physiologic and radiographic criteria on aorticrelated mortality (ARM) due to rupture in a setting that aggressively uses the SVS CPG for BTAI care. METHODS
Fig 2. Image of a grade (G) IV blunt thoracic aortic injury (BTAI).
predict rupture risk in patients with G III BTAIs.12 Recent studies have suggested that a subset of patients with G III injuries may harbor early clinical/radiographic findings predictive of the need for emergency repair rather than the urgent
This study was approved by The University of Texas Medical School at Houston Committee for the Protection of Human Subjects, which serves as our local Institutional Review Board. Informed consent was waived for analysis of the prospectively collected trauma registry. Our database was queried to identify all patients with the diagnosis of BTAI admitted from September 1999 to December 2014. Demographics, physiologic status, and admission injury were abstracted. A trauma radiologist, a trauma surgeon, and a vascular surgeon interpreted available computed tomographic angiography (CTA) imaging to ensure agreement on BTAI grading. Once characterized, specific measurements of individual injuries were made. Examined variables included (1) diameter of the normal aorta proximal and distal to injury; (2) width and length of injury; (3) presence of associated radiographic findings, including mediastinal hematoma; and (4) measured depth of the mediastinal hematoma at the level of the aortic arch and the descending thoracic aorta.
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Fortuna et al 3
Fig 3. Blunt thoracic aortic injury (BTAI) treatment type over time at the University of Texas Houston. TEVAR, Thoracic endovascular aortic repair.
Our treatment algorithm, including our techniques for open repair (OR) and TEVAR have been previously reported.8,13,19 We instituted TEVAR as an alternative to OR for anatomically suitable patients in 2005. The early procedures were conducted with the off-label use of the TAG (W. L. Gore & Assoc, Flagstaff, Ariz), the first thoracic aortic device approved by the U.S. Food and Drug Administration, which was available in diameters of 26 mm to 40 mm intended for patients with aortic diameters of 23 mm to 37 mm. During this phase, young trauma patients with aortic diameters <23 mm were treated with OR. The approval of additional smaller diameter devices in 2008 enabled us to apply TEVAR to a wider range of patients. As a result, TEVAR gradually replaced OR as the treatment of choice for all anatomically suitable patients with thoracic aortic injuries at our center. The progressive shift from OR to TEVAR for repair of BTAI at our institution is demonstrated in Fig 3. Information on management and outcomes was identified through trauma registry abstraction and individual record review. Contributors to cause of death were categorized as stroke, refractory hemorrhage with hemodynamic instability, traumatic brain injury/neurologic, cardiac arrest not related to ongoing hemorrhage, sepsis secondary to contamination or infection, respiratory failure, multiorgan failure, and ARM. For the purpose of our study, ARM was defined as death directly related to BTAI or complications of BTAI treatment. Autopsy results were not available in all patients; however, operative findings and independent record review by a trauma and vascular surgeon were sufficient to make an ARM determination. Retrospective review of medical records included available outpatient follow-up data. No ARM or adverse events were noted in the follow-up period for the reported series. Data were analyzed using IBM SPSS Statistics 22.0 (IBM Corp, Armonk, NY). Continuous variables with a
normal distribution assessed by skewness (between 1.0 and 1.0) are reported using means 6 standard deviation. Continuous variables that were not normally distributed are reported using median and interquartile range (IQR). For the univariate analysis, we used the c2 test with Yates correction for comparison of categoric risk factors and the Student t-test or the Mann-Whitney test for comparison of continuous risk factors. Logistic regression analysis was used to identify factors independently associated with the development of ARM. Adjusted odds ratios and 95% confidence intervals (CIs) were derived. RESULTS Trauma registry review revealed 332 patients treated at our level I trauma center for BTAIs between 1999 and 2014 (Fig 4). The subsequent analysis excluded 86 patients who died before the CTA was obtained. Of the remaining 246 patients, CTAs for 205 were available for review in the archived radiology system. BTAI injuries identified included 24 G I, 49 G II, 124 G III, and 8 G IV injuries (Fig 4). BTAI patients (71.2% male) were a mean age of 39.3 years (Table). Emergency transfers from other facilities constituted 18.5% of the population, with the remainder arriving to our facility directly from the point of injury. The predominant mechanism of injury was motor vehicle collision (72.2%). On presentation, 22.4% were hypotensive (systolic blood pressure < 90 mm Hg), and 30.2% had a Glasgow Coma Score of #8. Median Injury Severity Score (ISS) was 35.0 (IQR, 14). CTA was followed by traditional diagnostic angiogram in 81 patients, with an additional 46 undergoing joint traditional angiogram combined with intravascular ultrasound (IVUS) evaluation. Four false-positive CTAs (1 G I, 1 G II, and 2 G III) were identified among these patients. After diagnosis, all patients underwent medical management using anti-impulse control to maintain systolic
4 Fortuna et al
JOURNAL OF VASCULAR SURGERY --- 2016
Fig 4. Flow chart shows demographics, treatment, and outcomes of blunt thoracic aortic injury (BTAI). CT, Computed tomography; LSAC, left subclavian artery coverage; SVS, Society for Vascular Surgery; TBI, traumatic brain injury; TEVAR, thoracic endovascular aortic repair.
blood pressure <120 mm Hg. Medical management alone was used in 53 BTAI patients (83.3% [20 of 24] G I, 36.7% [18 of 49] G II, 10.4% [13 of 124] G III, and 25% [2 of 8] G IV). Four patients included in the medical management-only group died on medical therapy while en route to planned urgent or emergency repair (2 G III and 2 G IV). Among patients selected for repair (TEVAR or OR), median time from admission to repair was 24.0 hours (IQR, 86 hours). OR was performed for 50 patients (2 G I, 9 G II, 36 G III, and 3 G IV) with an overall mortality of 8% (4 of 50). The median time to OR for patients with G III BTAIs was 9.5 hours (IQR, 62 hours) and 4.0 hours (IQR, 50 hours) for G IV injuries. No patients sustained ARM after successful OR. Additional complications occurring after OR included stroke in 4% (2 of 50), paraplegia in 8% (4 of 50), and acute renal failure in 22% (11 of 50). Median lengths of stay (LOS) in the intensive care unit and hospital among survivors of OR were 16.0 days (IQR, 21.0; mean, 22.9 6 24.1 days) and 25.5 days (IQR, 28.8; mean, 36.0 6 31.6 days), respectively.
TEVAR was performed in 101 patients (1 G I, 22 G II, 75 G III, 3 G IV), with an overall mortality rate of 5.0%. During the study period, TEVAR increased over time to supplant OR as the most common intervention for BTAI (Fig 3). One ARM occurred after TEVAR due to persistent thoracic hemorrhage from the aorta. Left subclavian artery coverage was required during TEVAR in 37 patients (36.6%). Complications after TEVAR included stroke in 2.0% (2 of 101), paraplegia in 2% (2 of 101), and acute renal failure in 8.9% (9 of 101). Median ICU and hospital LOS among survivors of TEVAR were 8.0 days (IQR, 0; mean, 9.3 6 9.1 days) and 8.0 days (IQR, 0; mean, 9.3 6 9.1 days), respectively. Overall mortality among the BTAI population surviving to obtain CTA was 11.2% (23 of 205; 2 G I, 1 G II, 17 G III, and 3 G IV). Cause of death identified included stroke in 8.7% (2 of 23), refractory hemorrhage with hemodynamic instability in 30.4% (7 of 23), traumatic brain injury/neurologic in 47.8% (11 of 23), cardiac arrest not directly related to ongoing hemorrhage in 8.7% (2 of 23), sepsis secondary to contamination or infection
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Fortuna et al 5
Table. Continued.
Table. Demographics, presentation, management, and outcomes of patients with blunt thoracic aortic injury (BTAI) Variables Age, mean 6 SD, years Male, No. (%) Transferred from another facility, No. (%) Mechanism Motor vehicle crash, No. (%) Motorcycle crash, No. (%) Auto vs pedestrian, No. (%) Other blunt, No. (%) Presentation and overall injury burden Hypotension (SBP <90 mm Hg) on arrival, No. (%) Admission Glasgow Coma Score #8, No. (%) ISS, mean 6 SD Admission hemoglobin, mean 6 SD, g/dL Admission lactate, median (IQR), g/DL SVS injury grades and characteristics G I, No. (%) Length of G I injury, median (IQR) cm G II, No. (%) Length of G II injury, median (IQR) cm G III, No. (%) Length of G III injury, median (IQR) cm Diameter of G III injury, median (IQR) cm G IV, No. (%) Normal aortic diameter Proximal to injury, mean 6 SD, mm Distal to injury, mean 6 SD, mm Imaging used for diagnosis in addition to CT CTA þ angiography, No. (%) CTA þ angiography þ IVUS, No. (%) Associated radiographic findings Any mediastinal hematoma, No. (%) Mediastinal hematoma at the level of the aortic arch, No. (%) Max mediastinal hematoma diameter at aortic arch level, median (IQR), mm Mediastinal hematoma at the level of the descending thoracic aorta, No. (%) Max mediastinal hematoma diameter at descending thoracic aorta level, median (IQR), mm Management Time from admission to OR or TEVAR, median (IQR), hours OR, No. (%) TEVAR, No. (%) Diameter of device used, median (IQR), mm Length of aortic coverage, median (IQR), cm LSA coverage, No. (%) Stroke after LSA coverage, No. (%)
Variables
Total (N ¼ 205) 39.3 6 18.6 146 (71.2) 38 (18.5) 148 27 13 17
(72.2) (13.2) (6.3) (8.3)
46 (22.4) 62 (30.2) 35.6 6 10.5 12.9 61.9 3.8 (3) 24 (11.7) 1.1 (1.9) 49 (23.9) 1.5 (1.3) 124 (60.5) 3.3 (2.4) 3.0 (0.7) 8 (3.9) 23.1 6 4.3 21.1 6 3.7 81 (39.5) 46 (22.4) 127 (62.0) 90 (43.9) 18.1 (11) 121 (59.0) 11.9 (5)
24.0 (86) 50 (24.4) 101 (49.3) 26 (6) 10 (1) 37/101 (36.6) 2/37 (0.05) (Continued)
Complications and outcomes Paraplegia, No. (%) Stroke, No. (%) Acute renal failure, No. (%) LOS, median (IQR), days ICU Hospital In-hospital mortality, No. (%) Time from admission to death, median (IQR), hours ARM, No. (%)
Total (N ¼ 205) 8 (3.9) 10 (4.9) 26 (12.7) 7.0 17.0 23 28.0
(16.5) (23.0) (11.2) (216)
5 (2.4)
ARM, Aortic-related mortality; CTA, computed tomography angiography; G, grade; ICU, intensive care unit; IQR, interquartile range; ISS, Injury Severity Score; IVUS, intravascular ultrasound; LOS, length of stay; LSA, left subclavian artery; OR, open repair; SBP, systolic blood pressure; SD, standard deviation; SVS, Society for Vascular Surgery; TEVAR, thoracic endovascular aortic repair.
in 17.4% (4 of 23), respiratory failure in 17.4% (4 of 23), multiorgan failure in 26.1% (6 of 23), and ARM in 21.7% (5 of 23). Regardless of management, no patients with G I or II injuries died of ARM. Two ARMs occurred in patients with G III injuries; one died #1 hour of CTA before repair attempt and the other died 2.5 hours after admission while in the ICU awaiting planned repair #24 hours. Three ARMs occurred among patients with G IV injuries. One patient was transferred from another hospital and died in the emergency department shortly after arrival. A second was brought directly to our facility but died in the ICU w3.5 hours after admission and before planned OR. The third ARM occurred after TEVAR in an 84-year-old woman who sustained severe multiorgan trauma. A postoperative CTA revealed persistent extravasation from the left subclavian artery. The patient developed cardiac arrest in the operating theater before attempted TEVAR revision. There were no significant differences between patients with G I and II injuries treated with medical management alone vs OR or TEVAR. No ARMs occurred in the G I or II treatment group during the initial hospitalization. Only one CTA-documented progression of injury was notedda patient with G II injury progressed to G III at 3 days and underwent TEVAR. There was no difference between medical management only and repair with regard to observed renal failure (8.6% vs 10.5%; P ¼ .77), paraplegia (2.9% vs 2.6%; P ¼ .95), or stroke (5.7% vs 5.3%; P ¼ .93). Patients undergoing repair had trends toward longer LOS in the ICU (11.0 6 12.3 days vs 6.4 6 8.5 days; P ¼ .06) and the hospital (20.5 6 13.5 days vs 18.4 6 20.1 days; P ¼ .46), but these differences were not statistically significant. Among the two patients with G III injury who had ARM, there was no difference in the presence of mediastinal hematoma at the level of the aortic arch (50% of ARM, 46.7% of survivors; P ¼ .93) or the descending thoracic aorta (50% of ARM, 80% of survivors; P ¼ .35)
6 Fortuna et al
compared with survivors. Likewise, the maximal depth of mediastinal hematoma measured at these locations was not different between those sustaining ARM and survivors (12.8 vs 26.0 6 11.6 mm at the arch and 14.0 vs 12.9 6 4.5 mm at the descending thoracic aorta, respectively). The size and length of the pseudoaneurysm was also not different between patients who sustained ARM and survivors (length: 3.9 6 1.6 cm vs 4.0 6 2.2 cm, P ¼ .92; width: 3.7 6 0.6 vs 3.0 6 0.7 cm, P ¼ .22). Median time from admission to repair for G III injuries was 11.6 hours (IQR, 17; mean, 11.4 6 8.6 hours). All patients with G IV injuries had mediastinal hematoma with active contrast extravasation and CTA findings consistent with contained aortic rupture. Four patients underwent emergency repair in <3 hours. Two patients underwent planned delayed repair for BTAIs. The first of these two patients had a severe liver laceration and underwent angiographic hepatic embolization, followed by OR, at 51 hours after admission. The second patient had active pelvic bleeding and underwent TEVAR after angiographic embolization 55 hours after admission. Multivariate logistic regression analysis was used to identify independent predictors of all-cause mortality among G III BTAI patients. Nonoperative management (odds ratio, 29.65; 95% CI, 5.62-156.49; P < .0001) was an independent predictor of all-cause mortality. ISS (continuous), although not a significant factor on regression, was noted to have an odds ratio of 1.02 (95% CI, 0.96-1.09) and a value of 0.5302 within the model. When multivariate logistic regression analysis was used in an attempt to identify ARM for G III injuries, no independent predictors could be identified. CTA measurements did not prove predictive, including diameter of the normal aorta (odds ratio, 0.88; 95% CI, 0.66-1.17; P < .3817) maximal pseudoaneurysm diameter (odds ratio, 1.04; 95% CI, 0.85-1.28; P ¼ .6798), and the ratio of these two values (odds ratio, 0.15; 95% CI, 0.01-1.96; P ¼ .1470). In a similar fashion, the presence of mediastinal hematoma was not predictive of ARM among G III injuries (odds ratio, 1.31; 95% CI, 0.32-5.32; P ¼ .7059). DISCUSSION The optimal treatment of BTAI remains a challenge of modern trauma care. Expanding data suggest that G I and II injuries represent low risk for rupture.14,15 The dramatic appearance and frequent associated instability of G IV injuries most commonly elicits emergency repair. Recent investigations have continued to focus on the optimal treatment of those injuries falling between these categoriesdthe G III BTAIs. Our present series illustrates that G III injuries may be among the most common types of BTAIs encountered in busy trauma centers (Fig 4). This prevalence, along with an uncertain natural history, warrants additional investigation to determine optimal treatment timing. Harris et al,12 at the University of Maryland, used bootstrap risk analysis modeling to identify a cohort of 49 BTAI patients during a 13-year period
JOURNAL OF VASCULAR SURGERY --- 2016
(2000 to 2013), including 18 ruptures. Among these 18 patients, however, three were G IV with active extravasation. The remaining 15 progressed to aortic rupture with clinical decompensation after a median of only 64 minutes (IQR, 44-124 minutes). Three ruptured during exposure for planned OR, and 12 ruptured before aortic intervention. Through modeling, they identified that size of the aortic lesion and the presence of mediastinal hematoma were effective predictors of ARM due to rupture. In our present report, we were unable to identify any independent predictors of ARM among G III injuries. As with the Maryland report,12 we noted that ARMs were prone to occur very early in the hospital course (median 3.5 hours in our series), at intervals that challenge the marshaling of resources. Our results do demonstrate, however, that the effective application of the SVS CPG at a busy level I trauma center results in a very low ARM, with only five of 205 ARMs (2.4%) occurring after the opportunity for BTAI CTA diagnosis and characterization. Consistent with recent literature on the topic,14,15 we identified no ruptures in patients with G I or II BTAIs, regardless of management strategy. Consistent with SVS CPG for BTAI treatment, the median time in our experience to OR was 9.5 hours (IQR, 62 hours) for patients with G III and 4.0 hours (IQR, 50 hours) for G IV injuries, with no deaths in patients undergoing OR techniques. Our endovascular results included a median time to repair (TTR) for G III injuries of 3.0 hours (IQR, 53 hours), whereas G IV injuries had a median TTR of 28.9 hours (IQR, 115 hours). This longer TTR for G IV TEVAR was, perhaps, a result of the complicated management of associated injuries, because the three patients in this subset had ISS of 29, 38, and 75, respectively. There was, however, one patient death related to TEVAR, and that was in a patient with a G IV injury and a persistent type Ia endoleak who died upon returning to the OR. Only four patients died from ARM before receiving repair. We were unable to identify any independent risk factors that were able to accurately predict ARM among G III injuries. Using a rapid response TTR resulted in a low ARM. Although early deaths after G III and IV injuries continue to occur, aggressive use of the SVS CPG, in our experience, appears to mitigate the relative usefulness of aortic diameter measurements and the identification of mediastinal hematoma on initial CTA. Although SVS grade has been shown to be a predictor of overall death among BTAI patients,15 our present study is the only one to focus explicitly on the risk of ARM among G III injuries. With regard to all-cause mortality among G III BTAIs in our present series, nonoperative management alone was the only independent predictor identified. This finding suggests that some BTAI patients sustain patterns of injury that preclude their selection for open or endovascular repair. Although ISS did not factor out to be an independent risk factor for mortality, the trend was toward positive predictability. Increased ISS was frequently associated with poor candidates for any repair.
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Our present report has several limitations that must be acknowledged, beginning with the inherent limitations associated with retrospective design. Our study extended over a dynamic period during which TEVAR replaced OR as the treatment of choice for patients with G II to IV BTAIs, which should also be noted. This transition represents changes in care that have the potential to confound our findings. It is also important to note that our early treatment approach, although successful in sustaining only one intervention-related death and resulting in a low rate of ARM, is a function of a high-volume trauma center with a readily available vascular surgical team. Extrapolation to all care environments should, therefore, be undertaken with caution. Finally, because of our low ARM, we were unable to use modeling methodologies used in prior studies on this topic.12 Management of patients with G III BTAIs remains a clinical challenge that requires additional investigation. The uncertain natural history of these lesions demands early evaluation for treatment. Most aortic ruptures occur very early in the hospital course, challenging the ability to effectively marshal the resources required for expedient intervention. Further study is required, ideally in a prospective multicenter fashion, to better elucidate subsets of patients who may require treatment of G III BTAIs earliest in the SVS CPG suggested period of 24 hours. CONCLUSIONS Injury grade is a predictor of ARM among patients with BTAIs. Adherence to the current SVS CPG for management of BTAI resulted in very low rates of ARM. Without further discriminating parameters, extended periods of medical management in patients with G III BTAIs appears unnecessary. In our experience, however, identification of additional physiologic and radiographic criteria indicating the need for emergency (vs urgent) repair of patients with aortic pseudoaneurysms (G III BTAIs) remains elusive.
Fortuna et al 7
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
We thank Troy Brown for document editing, Chris Akers for artwork, and Dr John Holcomb and Edmundo Dipasupil for the trauma registry.
15.
AUTHOR CONTRIBUTIONS Conception and design: GF, AP, JD, SL, KC, CM, AE, AA Analysis and interpretation: GF, KC, CM, AE, AA Data collection: GF, AP, JD, SL Writing the article: GF, AP, JD Critical revision of the article: GF, AP, JD, SL, KC, CM, AE, AA Final approval of the article: GF, AP, JD, SL, KC, CM, AE, AA Statistical analysis: CM, JD Obtained funding: Not applicable Overall responsibility: AA REFERENCES 1. Arthurs ZM, Starnes BW, Sohn VY, Singh N, Martin MJ, Andersen CA. Functional and survival outcomes in traumatic blunt
16.
17.
18.
19.
thoracic aortic injuries: an analysis of the National Trauma Databank. J Vasc Surg 2009;49:988-94. Teixeira PG, Inaba K, Barmparas G, Georgiou C, Toms C, Noguchi TT, et al. Blunt thoracic aortic injuries: an autopsy study. J Trauma 2011;70:197-202. Rabin J, Harris DG, Crews GA, Ho M, Taylor BS, Sarkar R, et al. Early aortic repair worsens concurrent traumatic brain injury. Ann Thorac Surg 2014;98:46-52. Demetriades D, Vemahos GC, Scalea TM, Jurkovich GJ, KarmyJones R, Teixeira PG, et al. Association for the Surgery of Trauma Thoracic Aortic Injury Study Group. Operative repair or endovascular stent graft in blunt traumatic thoracic injuries: results of an American Association for the Surgery of Trauma Multicenter Study. J Trauma 2008;64:561-70; discussion: 570-1. Osgood MJ, Heck JM, Rellinger EJ, Doran SL, Garrard CL, Guzman RJ, et al. Natural history of grade I-II blunt traumatic aortic injury presenting over 10 years at a single institution. J Vasc Surg 2013;59:334-42. Rabin J, DuBose JJ, Sliker CW, O’Connor JV, Scalea TM, Griffith BP. Parameters for successful nonoperative management of traumatic aortic injury. J Thorac Cardiovasc Surg 2014;147:143-50. Marco Di Eusanio M, Folesani G, Berretta P, Petridis FD, Pantalco A, Russo V, et al. Delayed management of blunt traumatic aortic injury: open surgical versus endovascular repair. Ann Thorac Surg 2013;95: 1591-7. Estrera AL, Gochnour DC, Azizzadeh A, Miller CC, Coogan S, Charlton-Ouw K, et al. Progress in the treatment of blunt thoracic aortic injury: 12-year single-institution experience. Ann Thorac Surg 2010;90:64-71. Paul JS, Neideen T, Tutton S, Milia D, Tolat P, Foley D, et al. Minimal aortic injury after blunt trauma: selective nonoperative management is safe. J Trauma 2011;71:1519-23. Holmes JH, 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. 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. Harris DG, Rabin J, Kufera JA, Taylor BS, Sarkar R, O’Connor JV, et al. A new aortic injury score predicts early rupture more accurately than clinical assessment. J Vasc Surg 2015;61:332-8. Azizzadeh A, Keyhani K, Miller CC, Coogan SM, Safi HJ, Estrera AL. Blunt traumatic aortic injury: initial experience with endovascular repair. J Vasc Surg 2009;49:1403-8. Lee WA, Matsumura JS, Mitchell RS, Farber MA, Greenberg RK, Azizzadeh A, et al. Endovascular repair of traumatic thoracic aortic injury: clinical practice guidelines of the Society for Vascular Surgery. J Vasc Surg 2011;53:187-92. DuBose JJ, Leake SS, Brenner M, Pasley J, O’Callaghan T, LuoOwen X, et al. Aortic Trauma Foundation. Contemporary management and outcomes of blunt thoracic aortic injury: a multicenter retrospective study. J Trauma Acute Care Surg 2015;78:360-9. Gunn ML, Lehnert BE, Lungren RS, Narparla CB, Mistumori L, Gross JA, et al. Minimal aortic injury of the thoracic aorta: imaging appearance and outcome. Emerg Radiol 2014;21:227-33. Lamarch Y, Berger FH, Nicolaou S, Bilawich AM, Louis L, Inacio JR, et al. Vancouver simplified grading system with computed tomographic angiography for blunt aortic injury. J Thorac Cardiovasc Surg 2012;144:347-54. 354.e1. Starnes BW, Lundgren RS, Gunn M, Quade S, Hatsukami TS, Tran NT, et al. A new classification scheme for treating blunt aortic injury. J Vasc Surg 2012;55:47-54. Azizzadeh A, Ray HM, Dubose JJ, Charlton-Ouw KM, Miller CC, Coogan SM, et al. Outcomes of endovascular repair for patients with blunt traumatic aortic injury. J Trauma Acute Care Surg 2014;76: 510-6.
Submitted Sep 14, 2015; accepted Nov 12, 2015.