Secondary Logo

Prospective Study of Blunt Aortic Injury: Multicenter Trial of the American Association for the Surgery of Trauma

Fabian, Timothy C. MD; Richardson, J. David MD; Croce, Martin A. MD; Smith, J. Stanley Jr., MD; Rodman, George Jr., MD; Kearney, Paul A. MD; Flynn, William MD; Ney, Arthur L. MD; Cone, John B. MD; Luchette, Fred A. MD; Wisner, David H. MD; Scholten, Donald J. MD; Beaver, Bonnie L. MD; Conn, Alasdair K. MD; Coscia, Robert MD; Hoyt, David B. MD; Morris, John A. Jr., MD; Harviel, J. Duncan MD; Peitzman, Andrew B. MD; Bynoe, Raymond P. MD; Diamond, Daniel L. MD; Wall, Matthew MD; Gates, Jonathan D. MD; Asensio, Juan A. MD; McCarthy, Mary C. MD; Girotti, Murray J. MD; VanWijngaarden, Mary MD; Cogbill, Thomas H. MD; Levison, Marc A. MD; Aprahamian, Charles MD; Sutton, John E. Jr., MD; Allen, C. F. MD; Hirsch, Erwin F. MD; Nagy, Kimberly MD; Bachulis, Ben L. MD; Bales, Charles R. MD; Shapiro, Marc J. MD; Metzler, Michael H. MD; Conti, Vincent R. MD; Baker, Christopher C. MD; Bannon, Michael P. MD; Ochsner, M. Gage MD; Thomason, Michael H. MD; Hiatt, Jonathan R. MD; O'Malley, Keith MD; Obeid, Farouck N. MD; Gray, Perry MD; Bankey, Paul E. MD; Knudson, M. Margaret MD; Dyess, Donna Lynn MD; Enderson, Blaine L. MD

The Journal of Trauma: Injury, Infection, and Critical Care: March 1997 - Volume 42 - Issue 3 - p 374-383
Article
Free

Background  Blunt aortic injury is a major cause of death from blunt trauma. Evolution of diagnostic techniques and methods of operative repair have altered the management and posed new questions in recent years.

Methods  This study was a prospectively conducted multicenter trial involving 50 trauma centers in North America under the direction of the Multi-institutional Trial Committee of the American Association for the Surgery of Trauma.

Results  There were 274 blunt aortic injury cases studied over 2.5 years, of which 81% were caused by automobile crashes. Chest computed tomography and transesophageal echocardiography were applied in 88 and 30 cases, respectively, and were 75 and 80% diagnostic, respectively. Two hundred seven stable patients underwent planned thoracotomy and repair. Clamp and sew technique was used in 73 (35%) and bypass techniques in 134 (65%). Overall mortality was 31%, with 63% of deaths being attributable to aortic rupture; mortality was not affected by method of repair. Paraplegia occurred postoperatively in 8.7%. Logistic regression analysis demonstrated clamp and sew (p = 0.002) and aortic cross clamp time of > or = to30 minutes (p=0.01) to be associated with development of postoperative paraplegia.

Conclusions  Rupture after hospital admission remains a major problem. Although newer diagnostic techniques are being applied, at this time aortography remains the diagnostic standard. Aortic cross clamp time beyond 30 minutes was associated with paraplegia; bypass techniques, which provide distal aortic perfusion, produced significantly lower paraplegia rates than the clamp and sew approach.

From the University of Tennessee (T.C.F., M.A.C.), Memphis Tennessee, University of Louisville (J.D.R.), Louisville, Kentucky, The Pennsylvania State University (J.S.S.Jr.), Hershey, Pennsylvania, Methodist Hospital of Indiana, Inc. (G.R.Jr.), Indianapolis, Indiana, University of Kentucky Chandler Medical Center (P.A.K.), Lexington, Kentucky, Erie County Medical Center (W.F.), Buffalo, New York, Hennopin County Medical Center (A.L.N.), Minneapolis, Minnesota, University of Arkansas (J.B.C.), Little Rock, Arkansas, University of Cincinnati Medical Center (F.A.L.), Cincinnati, Ohio, University of California at Davis (D.H.W.), Sacramento, California, Butterworth Regional Trauma Center (D.J.S.), Grand Rapids, Michigan, Ohio State University (B.L.B.), Columbus, Ohio, Massachusetts General Hospital (A.K.C.), Boston, Massachusetts, St. John's Regional Health Center (R.C.), Springfield, Missouri, University of California (D.B.H.), San Diego, California, Vanderbilt University (J.A.M.Jr.), Nashville, Tennessee, Washington Hospital Center (J.D.H.), Washington, DC, University Medical Center (A.B.P.), Pittsburgh, Pennsylvania, University of South Carolina (R.P.B.), Columbia, South Carolina, Allegheny General Hospital (D.L.D.), Pittsburgh, Pennsylvania, Baylor College of Medicine (M.W.), Houston, Texas, Brigham and Women's Hospital (J.D.G.), Boston, Massachusetts, University of Southern California (J.A.A.), Los Angeles, California, Wright State University (M.C.M.), Dayton, Ohio, Victoria Hospital (M.J.G.), London, Ontario, University of Alberta (M.V.), Edmonton, Alberta, Gundersen Clinic, Ltd. (T.H.C.), LaCross, Wisconsin, John C. Lincoln Hospital & Health Center (M.A.L.), Phoenix, Arizona, Medical College of Wisconsin (C.A.), Milwaukee, Wisconsin, Dartmouth-Hitchcock Medical Center (J.E.S.Jr.), Lebanon, New Hampshire, Good Samaritan Regional Medical Center (C.F.A.), Phoenix, Arizona, Boston University Medical Center (E.F.H.), Boston, Massachusetts, Cook County Hospital (K.N.), Chicago, Illinois, Sparrow Hospital (B.L.B.), Lansing, Michigan, State University of New York (C.R.B.), Syracuse, New York, St. Louis University (M.J.S.), St. Louis, Missouri, University of Missouri (M.H.M.), Columbia, Missouri, University of Texas Medical Branch (V.R.C.), Galveston, Texas, University of North Carolina (C.C.B.), Chapel Hill, North Carolina, Mayo Clinic (M.P.B.), Rochester, Minnesota, Memorial Medical Center (M.G.O.), Savannah, Georgia, Carolinas Medical Center (M.H.T.), Charlotte, North Carolina, Cedars-Sinai Medical Center (J.R.H.), Los Angeles, California, UMDNJ (K.O.), Camden, New Jersey, Henry Ford Hospital (F.N.O.), Detroit, Michigan, Health Sciences Centre (P.G.), Winnipeg, Manitoba, University of Texas Southwestern Medical Center (P.E.B.), Dallas, Texas, University of California (M.M.K.), San Francisco, California, University of South Alabama (D.L.D.), Mobile, Alabama, and University of Tennessee (B.L.E.), Knoxville, Tennessee.

Presented at the 56th Annual Meeting of the American Association for the Surgery of Trauma, September 19-21, 1996, Houston, Texas.

Address for reprints: Timothy C. Fabian, MD, 956 Court Avenue, Room G210, Memphis, TN 38163.

Key Words: Blunt aortic injury, management.

Blunt aortic Injury (BAI) was recently found to be the second most common cause of death behind head injury in a study of 387 blunt trauma deaths. [1] It is somewhat of a disease of modern society in that the overwhelming majority of cases are caused by automobile crashes. The first successful repair by Klassen was performed only 37 years ago. [2] Before that advance, one report had admonished physicians to avoid attempts at repair of traumatic aortic aneurysms. [3]

The lethality of BAI was first emphasized by Parmley when he noted from a combined autopsy and clinical study that 85% of patients died at the accident scene. [4] That report also noted that of the 15% arriving at a medical facility alive, most died from aortic rupture within a few days. Since that time, considerable attention has been paid to BAI. Initial interest was largely devoted to rapid diagnosis to avoid rupture. In recent years, significant attention has been directed toward operative management in an effort to reduce the 10-20% incidence of paraplegia after otherwise successful repair. [5-7]

The present multicenter study was conducted to obtain a large sample size over a short time span to delineate presentday management of BAI in North American trauma centers. In addition to establishing the current presentation and associated findings of BAI, an effort was made to investigate the areas of diagnosis and operative management. Those latter two areas are where significant change has occurred over the past decade. Change naturally breeds controversy.

Back to Top | Article Outline

METHODS

The Multi-institutional Trial Committee of the American Association for the Surgery of Trauma (AAST) elected to conduct a North American study of the current presentation and management of blunt thoracic aortic injury (BAI) at the annual meeting of the organization in September of 1993. The study was prospectively conducted over a 30-month period ending June 30, 1996. Participating institutions were recruited by a letter sent to the membership of the AAST inviting participation. Fifty trauma centers from the United States (47) and Canada (3) participated in the study (Figure 1). An initial data collection sheet was sent out to the participating investigators for evaluation and a final data collection sheet was designed following that input. After completion, the data collection sheets were returned to the site of the principal investigator for entry into a computer data base. Criteria for entry required that all study patients be admitted to the hospital with vital signs. The diagnosis of BAI was made by autopsy, thoracotomy, or a definitive diagnostic technique. The study was designed to be descriptive. No diagnostic or therapeutic randomizations were performed. Data was prospectively collected on basic demographics, mechanism of injury, associated injuries, diagnostic procedures, aortic repair, and outcomes. Diagnostic procedures were evaluated to determine the incidence of aortography, transesophageal echocardiography (TEE), and chest computed tomography. Details of operative repair focused primarily on the use of bypass versus clamp and sew techniques as well as on aortic cross clamp time, and the impact of these in the development of postoperative paraplegia.

Figure 1

Figure 1

Univariate and multivariate analyses were performed. chi2 analyses were performed on dichotomous variables, whereas Student's t tests were performed for continuous variables. A logistic regression analysis was done to determine the independent significance of multiple variables on the development of postoperative paraplegia. Statistical significance was set at the p < 0.05 level.

Patients were divided into four groups for various parts of data analysis and reporting in the Results section: (1) in extremis, those patients presenting with some vital signs who undergo emergent thoracotomy; (2) stable, those patients who undergo planned thoracotomy after diagnostic evaluation; (3) rupture, those patients who are admitted with hemodynamic stability and develop aortic rupture before planned thoracotomy; (4) nonoperative, those patients who undergo diagnostic evaluation but do not have a thoracotomy because of either associated injuries/advanced age.

Back to Top | Article Outline

RESULTS

There were 274 patients enrolled, including 199 men and 75 women. The mean age was 38.7 years (range, 8-88). Motor vehicle crashes were responsible for 222 (72% head-on, 24% side impact, 4% rear impact), motorcycle crashes 19, autopedestrian 19, falls 7, and other mechanisms 7. The numbers of patients in the four groups were: 22 patients in extremis, 207 stable, 24 ruptured, and 21 nonoperative.

Total associated injuries are detailed in Table 1. Half of the population had a brain injury. One hundred seventy had some significant associated chest injury. Sixty had some intra-abdominal injury, 93 had pelvic or long-bone fractures. The mean Injury Severity Score (ISS) for the entire population was 42.1. Figure 2 illustrates the frequency distribution according to ranges of ISS and the corresponding mortalities. Excluding in extremis patients, the mean Glasgow Coma Scale (GCS) score was 12.1.

Table 1

Table 1

Figure 2

Figure 2

Back to Top | Article Outline

Diagnostics

Chest radiographs at the time of admission were taken on 259 patients. The x-ray findings are shown in Table 2. Mediastinal widening was the predominant finding with 85% of x-ray films demonstrating such. Negative x-ray findings were reported in 7%; further diagnostic studies were pursued owing to either mechanism of injury or findings on subsequent chest x-ray films. Of the 8% of patients who had neither widened mediastinum nor negative examination, most had an indistinct aortic knob, followed by first and/or second rib fractures and deviation of the nasogastric tube.

Table 2

Table 2

Aortogram was performed on 220 patients. Transesophageal echocardiography (TEE) was performed on 30 patients: 24 (80%) positive, five (17%) equivocal (hematoma), one (3%) negative. Computed tomographic (CT) scans were performed on 88 patients: 65 (74%) positive, 20 (23%) equivocal (hematoma), three (3%) negative. The anatomic distribution of aortic injury was ascending seven (3%), arch 12 (4%), proximal descending 254 (93%), and diaphragm 1 (0.3%). Ten (3.6%) patients had associated great vessel injuries.

Back to Top | Article Outline

Operative Repair and Outcome

Time intervals for the 207 stable patients are as follows: injury to trauma center arrival 2.8 +/- 2.8 hours, trauma center arrival to diagnosis 4.7 +/- 6.9 hours, diagnosis to thoracotomy 10.1 +/- 73.3 hours, and injury to thoracotomy 16.5 +/- 70.8 hours. Of the 207 stable patients who underwent planned thoracotomy, 73 had clamp and sew technique whereas 134 had bypass. Table 3 lists the method of operative management with attendant paraplegia and mortality rates. Of the bypass patients, 79% received systemic heparin and 26% heparin bonded tubing. Eighty percent of the patients had primary aortic repair and 92% had a graft placed. Eighteen patients (8.7%) developed paraplegia after repair (does not include four stable patients with initial spinal cord injury); 12 clamp and sew versus six bypass (p < 0.004).

Table 3

Table 3

The paraplegia rate was also evaluated according to the duration (minutes) of aortic cross clamping (XCT). Evaluation of the paraplegic patients compared with those without paraplegia found XCT of 46.6 +/- 14.3 minutes versus 39.6 +/- 22.7 minutes (p = 0.22). Table 4 illustrates the salutary effect of short (<30 minutes) compared with long (> or = to30 minutes) XCT in those where the time was recorded. The effect of cross clamp times on paraplegia according to method of repair is shown in Table 5. Paraplegia is also correlated with short versus long cross clamp times in both methods of operative repair in Table 6. These data suggest bypass techniques allow for longer cross clamp times without increasing the paraplegia rate.

Table 4

Table 4

Table 5

Table 5

Table 6

Table 6

Data entered into the multivariate analysis for risks associated with paraplegia included age, admission systolic blood pressure, ISS, GCS, method of repair, and cross clamp times. Only the 183 of 207 stable with complete data bases for all of the parameters were included in that analysis. The results of the analysis are shown in Table 7. That analysis demonstrates the independent protective effect of both bypass techniques and short XCT.

Table 7

Table 7

Of the 207 planned operations, 91% were performed primarily by cardiothoracic surgeons and 9% by general surgeons. The paraplegia rates for cardiothoracic and general surgeons were 9 and 11%, respectively. The mortality rates (non-head injury) were 13 and 0%, respectively.

There were 93 laparotomies required (66% before aortic repair). There were 124 orthopedic operations, eight craniotomies, and 28 maxillofacial reconstructions performed.

Complications are listed in Table 8. There were a total of 86 (31%) deaths. All 22 in extremis and all 24 rupture patients died because of aortic rupture. The mortality rate for the 207 stable patients undergoing planned thoracotomy was 14%. The mortality for the 20 nonoperative patients was 55%. Table 9 lists the cause of mortality among the different patient categories. Mortality is stratified by age, ISS, GCS, and method of repair in Table 10. Only age was noted to significantly correlate with death in patients undergoing planned repair. Figure 3 demonstrates the age range and mortality according to age (excludes in extremis group).

Table 8

Table 8

Table 9

Table 9

Table 10

Table 10

Figure 3

Figure 3

Of the 24 patients with rupture, 22 (92%) died within 24 hours. One died 6 days after injury, and one died 30 hours from injury. Precise time of rupture data was available in 13 patients. Of these, 46% died within 4 hours, and an additional 38% died within 8 hours. There were 8 stable patients who died in the operating room. All died of exsanguination after loss of proximal aortic control. The mean intensive care unit (ICU) and hospital lengths of stay for survivors were 15.6 and 27.2 days, respectively. Those corresponding lengths of stay for nonsurvivors (excluding in extremis patients) were 12.0 and 12.0 days.

Back to Top | Article Outline

DISCUSSION

This study was designed to define the current presentation and management of BAI in the North American environment. It was hoped that in addition to describing the scenario, that novel findings may result, as well as an elucidation of advantages to diagnostic and management techniques. It does not appear there were any novel findings. But, it does appear that analysis of clinical management has provided some important new light.

The major strengths of the design are that it provided (1) the largest number of cases of BAI injury reported, (2) a broad based experience (trauma centers in 22 states, three Canadian provinces, and the District of Columbia), and (3) prospective data collection. Relative to other large series that have retrospectively reported over 6-15 years, [5-7] this study was conducted over 2.5 years, thus minimizing the effects of confounding variables over time impacting on interpretation of results. It has been estimated that 7,500 to 8,000 cases of BAI occur annually. [8,9] Furthermore, 80-85% of those patients die at the scene. [1,4] Thus, approximately 1,000-1,500 cases of BAI arrive at hospitals alive annually. The 274 reported cases reported here over 2.5 years likely represent approximately 10% of all cases of BAI surviving the accident scenes during the study period. This sample size should be sufficient to demonstrate a reasonable reflection of presentation and management today. The wide range of geographic representation of study sites should minimize regional bias and provide a broad perspective for interpreting results of how BAI is being managed. This study represents the first attempt at a large prospective study of BAI. Prospective data collection has the obvious advantages of providing more complete and accurate data than retrospective studies.

The major weaknesses of the study are that there were (1) a large number of institutions and individual surgeons involved and (2) there were no randomizations to directly compare issues surrounding diagnosis and treatment. Thus, the broad based nature is clearly a two-edged sword. It likely reflects the reality of BAI management on this continent, but also somewhat confounds because of the inherent heterogeneity of the study design. Although there were no direct comparisons of modes of management, the size of the trial did permit for some reasonably focused analyses, particularly with regard to issues of operative management.

The mechanisms of injury noted were similar to most recent series, with 81% caused by motor vehicle crashes. BAI has generally been thought to be associated with rapid decelerating injury, usually interpreted as being attributable to head-on impact. This study found 28% were attributable to other vectors (24% side and 4% rear). A recent autopsy study of 142 individuals with BAI found 72% resulted from motor vehicle crashes, and that 58% were associated with head-on collisions and 42% broadside collisions. [10] The associated injury pattern was similar to other studies with the exception that half of the patients were reported to have some degree of closed head injury, whereas others have reported closed head injuries in approximately 30%. [1,5,7]

The newer diagnostic modalities of TEE and chest computed tomography have received increasing attention over the past decade for definitive diagnosis of BAI. In this report, not surprisingly, chest x-ray technology has remained the main screening tool. The small number of cases in which it was not obtained were in patients presenting in extremis. Although multiple ancillary chest x-ray findings have been described, which can be helpful in calling attention to the diagnosis, mediastinal widening continues to be the dominant trigger for further work-up. Although there was a significant use of TEE and computed tomography, aortography remains the predominant approach for definitive diagnosis. A focused interest in TEE has demonstrated sensitivity and accuracy equal to aortography. [11] However, this generally requires an interested group of cardiologists and is operator dependent. In this study, only 11% had TEE, but with fairly good results. This technique may be more broadly used in the future, [12] especially considering the increasing maturation of trauma centers as well as the changes in health care financing. Chest CT scans were performed in a higher number of patients (32%). The results were similar to those of TEE, with both being less definitive than aortography. The higher utilization of CT scanning is probably explained by it being more accessible. Some reports of CT scans for BAI have found it to be inaccurate and perhaps dangerous. [13,14] However, with the advancing technology of helical scanners, others have recently found the sensitivity to equal aortography. [15]

Outcomes according to patient groups and methods of operative management will be discussed next. There were no survivors among patients either arriving in extremis or of those arriving stable but subsequently rupturing the pseudoaneurysm. This finding is in accord with other reports. [5,6] The degree of aortic injury is a continuum from subintimal hemorrhage to total aortic disruption. This pathology was nicely detailed in the classic study by Parmley et al. [4] That was a study of 296 BAI from the Armed Forces Institute of Pathology and the Walter Reed Army Hospital with only 38 (13%) patients arriving alive. Autopsies were done in all deaths with close attention paid to the aortic pathology. The lesions were classified as (1) intimal hemorrhage, (2) intimal hemorrhage with laceration, (3) medial laceration, (4) complete laceration of the aorta, (5) false aneurysm formation, and (6) periaortic hemorrhage. [4] The degree of aortic disruption is undoubtedly the key factor in the likelihood of rupture. Those with full-thickness injury and false aneurysm confined by a thin layer of adventitia and the mediastinal pleura likely represent the patients presenting in extremis or those who develop spontaneous rupture. Controlling blood pressure and aortic wall tension with vasodilators and beta-blockers has been used to diminish the risk of patients before planned thoracotomy both acutely as well as those who have planned nonoperative management because of associated injuries. [5,16] This control is probably an important concept that will decrease rupture and should be more widely applied. These measures were reported to be applied in 17% of patients in this series. We were unable to correlate admission blood pressures with subsequent rupture, but these were only isolated measurements. At the other end of the spectrum are those patients with isolated intimal injuries, and with intimal injuries into, but not through, the media. Some of these may have minimal to no hematoma and may be missed. It is probable that as computed tomography and TEE are more widely applied, more of those lesions will be discovered. That will require further study of the natural history of those minimal lesions. Undoubtedly some heal spontaneously as is the case with intimal injuries in large peripheral arteries.

Of the 38 patients from the Parmley series who arrived at the hospital alive, only two underwent operation. [4] Aortography was not used at that time. The remaining 36 patients are those in whom the natural history of spontaneous rupture after hospitalization was defined: of those surviving 1 hour or more, 30% died within 6 hours, 40% within 24 hours, 72% by 8 days, 83% by 3 weeks, and 90% by 10 weeks. Of the spontaneous ruptures in the current series, all but one occurred within 30 hours of injury. The difference in timing of rupture in this series versus the Parmley report of 1958 is that most of the patients who died beyond 12 hours in the early report would be diagnosed and operated upon today because of the advances in diagnostics. Nonetheless, the interval analysis of injury to repair time in stable patients in this study found a number of patients were operated upon a few days after injury. Thus, we must continue to improve our diagnostic capabilities. A significant proportion of those surviving for longer intervals without repair are likely the selected population with aortic tears less than full-thickness.

There has been interest expressed in recent literature concerning who should perform aortic repair. A study from Tampa reported a 50% mortality rate among 12 patients operated upon by ten different board certified general/thoracic surgeons compared with 7% mortality among 15 patients operated upon by two designated thoracic trauma surgeons. [17] They concluded that designated thoracic trauma surgeons should be operating upon these patients. The Denver group reported on 18 patients with an 11% mortality, all operated upon by experienced trauma surgeons without formal thoracic surgical residency. [18] They concluded that trauma surgeons can produce outcomes equivalent to cardiothoracic surgeons. They also believe more general surgeons should be adequately trained to handle such cases. Both reports prove important points that are valid. The first demonstrates that the occasional surgeon is not appropriately prepared to manage BAI, whereas the second demonstrate that general surgeons who are dedicated and take a very active interest and receive nonformalized training by cardiothoracic surgeons can produce excellent results. Although cardiothoracic surgeons performed 91% of the planned repairs in the current study, the outcomes obtained in the 9% managed by general surgeons demonstrate that interested general surgeons can indeed achieve excellent results.

The most important points addressed by this study are those concerning details of operative repair. A significant controversy has developed over whether there are differences in results comparing simple clamp exclusion and direct repair (clamp and sew) to techniques that provide distal aortic perfusion during repair. Specifically, is paraplegia, the most significant complication of repair, lowered by distal aortic perfusion techniques? Advantages of clamp and sew are simplicity and the lack of need for heparin. That approach has received significant support in recent literature. [5,6,9] Potential advantages of bypass techniques are that the distal aorta is perfused, which may decrease paraplegia rates, as well as reduce ischemia and reperfusion injuries associated with sacrifice of hepatic, renal, and mesenteric flow during clamp and sew repair. [19] Disadvantages of partial and full bypass are the need for systemic heparinization, which may predispose toward bleeding complications in patients with multiple injuries. The centrifugal pumps provide the advantages of distal perfusion while eliminating the potential hazards of heparin. [20] Gott shunts have been largely abandoned because of their somewhat awkward nature for insertion into the left ventricle or proximal aorta as well as a tendency to occlude.

Long cross clamp times have been demonstrated to be associated with paraplegia in several studies. The tenuous and variable distribution of blood supply to the spinal cord accounts for the susceptibility to injury as well as the unpredictability, i.e., some patients with very short cross clamp times develop cord injury. [21] Several investigators have reported statistically significant increases in paraplegia rates when cross clamp times exceeded 30, [6,22,23] or 35 minutes. [7,24] Although those reports are retrospective, the results do not appear to be biased by more severe injury patterns or more perturbed physiology. The logistic regression analysis from this study strongly confirms the impact of cross clamp times on paraplegia, while accounting for the confounding variables.

If cross clamp times below 30 minutes occur, perhaps the issue of clamp and sew versus bypass is of little consequence. However, it is often difficult to predict preoperatively which cases will have more complex injuries and will require more extended cross clamp time. Several groups have reported decreased paraplegia rates with bypass compared with clamp and sew. [7,22,24-26] Hunt, et al. have further demonstrated that bypass provided significant (p = 0.005) protection against spinal cord injury compared with clamp and sew when cross clamp times exceeded 35 minutes. [7] The results of the logistic regression analysis from this study also substantiate that observation; not only was cross clamp time predictive of spinal cord injury, but bypass was also independently predictive of protection against spinal cord injury.

The final issue dealing with operation is the optimal choice of bypass technique. A review of five published series using centrifugal pumps in 58 patients, found there was no instance of paraplegia. [27] In this report, although not statistically significant, the paraplegia rate was lowest (2.9%) with the centrifugal pump in 69 cases compared with partial/full cardiopulmonary bypass (6.6%) in 61 cases. There were a high number of associated laparotomies and orthopedic procedures. Although neither significant complications were noted with heparin in the partial/full bypass groups, nor were there differences in mortality compared with the centrifugal pump group, the centrifugal pump does eliminate heparin requirements. It is possible that concerns with short-term (generally 1 to 2 hours) heparinization are exaggerated. Prolonged heparinization for blunt carotid artery injury management has been undertaken in patients with multiple injuries with minimal complications. [28] Heparin does not cause clot lysis and would probably promote only ongoing bleeding. But, it is obvious that avoidance of heparin would be preferred.

The overall mortality of 31% for BAI is commensurate with recent series. Aortic rupture was the direct cause for 63% of those deaths. Forty percent of the ruptures were in patients presenting in extremis. Beyond prevention measures, and maximizing the expediency of transport to the definitive care facility, nothing will reduce that mortality. Twenty-four patients (8.8% of the study population and 44% of ruptures) arrived with stable vital signs and had rupture of the aorta within 30 hours of admission. That group probably represents those with full-thickness aortic disruption with the false aneurysm maintained by a thin layer of adventitia and surrounding tissue. They clearly illustrate the continued importance of early, aggressive diagnostic measures followed by prompt thoracotomy.

Back to Top | Article Outline

Acknowledgments

The authors would like to acknowledge the excellent study coordination by the following: Colleen Swartz, MSN (Lexington, Ky), Maureen Misinski, RN, MSN (Indianapolis, Ind), Karen S. Romano, RN (Buffalo, NY), Kathy Layton, RN (Sacramento, Calif), Carol S. Hill-Gulick, RN, MSN (Cincinnati, Ohio), Adale Derrick, RN (Springfield, Mo), Kathryn Jenkins, RN (Springfield, Mo), Grace McDonald-Smith, MEd (Boston, Mass), Joan L. Garcia, RN, CCRC (San Diego, Calif), Connie Mattice, RN, MSN (Grand Rapids, Mich), Christine Stinson, ART (Pittsburgh, Penn), Jay Hamm, RN (Columbis, SC), Susan Kennedy, RN (Washington, DC), Judy Schmidt, RN (Washington, DC), Glenn Raup, RN (Houston, Tex), Patricia J. Manni, RN (Pittsburgh, Penn), Eileen Corcoran, RN (Lebanon, NH), Ronald Diamond (Lebanon, NH), Theresa Christiansen, MS, RN, Pamela Lambert, BSN, RN, Melinda L. Kilbride, RN, Deborah Allen, RN (Syracuse, NY), Mary J. Keegan, RN, MSN (St. Louis, Mo), Mary Forcia, RN (Lansing, Mich), Tammy M. Bratton, RN, BSN (Savannah, Ga), Susan Schrage, RN (Rochester, Minn), Dennis Rowe, EMT-P (Knoxville, Tenn), Robert Marburger, RN (Camden, NJ), Bruce Bennett, MD (Minneapolis, Minn), Carol Hill-Gulick (Cincinnati, Ohio).

Back to Top | Article Outline

PAPER DISCUSSION

Dr. Wayne Meredith (Winston-Salem, North Carolina): I enjoyed this paper very much, Dr. Fabian, and I want to commend the Multi-institutional Trials Committee, and I want to commend you. This is one of those manuscripts, the more you read it, the more information you find and the better it gets. I highly recommend it to everyone. It reads easily, it's well-written, and it is full of pearls.

I want to emphasize a few of those that I don't think you covered in your discussion well, and end up with just a few questions.

This study is a snapshot of what's happening in the country. The authors admit and start out by saying it is not possible to recommend treatment methods with this study. The time to diagnosis in the trauma centers in this country, 50 different trauma centers is 4.7 hours from arrival to the trauma center to the diagnosis. Times from making the diagnosis to operating the patient, 10 hours mean. Those are very interesting statistics, and I think they set the standard, and it's a pretty low goalpost to overcome.

I would appreciate those times being reported as a median so we could get some picture of the distribution, because I think a few long times distort the mean.

It's very interesting that 25% of the patients are side impact collisions. That's very valuable information. So this injury is not confined to head-on collision patients.

The next subtle point that's hidden in these data is that these patients actually come in two groups, stable or dying. Patients who are hypotensive, patients who require recurrent transfusions, patients who can't keep their blood pressure up, it is not because their mediastinum is wide. If their aorta is ruptured, all of those patients die, and so it should not be assumed that the patient who is in shock is in shock from their widened mediastinum. Something else is causing shock, and it needs to be fixed, and these data strongly support that concept, although it's hidden in the data to be teased out.

The next part of the picture, TEE and CT scans are useful, but they're not diagnostic and they're not adequate.

Now, as for the paraplegia question, my personal series at North Carolina Baptist Hospital, as the surgeon, there's 35 patients, the mean cross-clamp time was 21 minutes. The range is eight minutes to 27 minutes. There's no paraplegia in those patients. It is possible to do this operation and repair this injury with the clamp and run technique and have very good results.

There is no question from this data that you can extend the length of time available to you to teach, to look around, to figure out what to do by putting the patient on bypass. I do not think we should take from this study in this highly litigious subject, the concept that one must repair this injury on bypass. There are some patients where that is not going to be possible. I think that we need to move to a selective use of bypass and would appreciate your recommendation as to which patients those would be.

A few questions, and then I'll close. Are there any patients too old to repair? In other words, how are the nonoperative patients chosen?

What are the trauma centers in the country doing to prevent hypertension between arrival of the patient and getting the patient to the operating room? It's very important to prevent these patients from becoming hypertensive, I believe. Can you tease from this mass of data any way to substantiate that concept or show us what should be done?

Did you look at the relationship between paraplegia and shock, or paraplegia and the number of intercostal vessels excluded? I think both of those factors are major contributors that are not considered in just looking at how the bypass is managed and perfusion is managed.

And last, can you see any relation between mortality or paraplegia to the volume of patients treated, i.e., to the experience of the institution? This is a very good model for regionalizing trauma care. The average number of patients seen with this injury in the 50 trauma centers all well respected is four patients over this 2.5-year period. It is clear that if you gain a lot of experience with this injury, you can do it better, and it seems to me that we ought to be able to look for the data in this study to show that regionalization would be helpful.

This is a very good paper. The Multi-Institutional Trials Committee needs to be very proud, and Dr. Fabian can be proud for how well it's written.

I thank the organization for the privilege of the floor.

Dr. Donald D. Trunkey (Portland, Oregon): Dr. Fabian, I, too, enjoyed this paper. I look forward to reading the manuscript.

There was one area, though, that you went over very briefly, and I was wondering if you would expand on it. You said that a fair number of patients had associated celiotomy: two-thirds before repair of the aorta; one-third after repair of the aorta.

I've always taught our residents to do the celiotomy first, and I'm curious about the one-third that you did after the thoracotomy. In your analysis, could you find anything that would help us in the decision-making process? Did these patients bleed during the time that the thoracotomy was being done? Why was the celiotomy done afterwards?

Dr. Aurelio Rodriguez (Baltimore, Maryland): Dr. Fabian, you have collected important data in this project.

On the basis of my personal surgical involvement with more than 100 patients with blunt aortic injury during the past 15 years, allow me to ask two questions: At the R. Adams Cowley Shock Trauma Center, we have not observed 100% correlation between clamping time and paraplegia. However, hypotension before, during, or after clamping seems to be a very important risk factor for paraplegia. Does anything in your data implicate perioperative hypotension as such a risk factor?

We cannot anticipate preoperatively, based on radiographic or clinical evidence, that repair of a blunt aortic injury will demand more or less than 30 minutes of clamping time. Therefore, for the past 2 years, we have been using the Bio-Medicus pump during these aortic repairs. Twelve patients have undergone surgery in this manner, and none of them developed paraplegia. Unfortunately, brain-injured patients with aortic injury repaired using this device tend to get worse neurologically. Have you found any correlation between use of a pump and neurologic outcome in patients with head injury?

Thank you very much. I am sorry that my institution could not participate in this valuable study.

Dr. John R. Hall (Kingsport, Tennessee): Dr. Fabian, I enjoyed your paper, too. I have several questions.

Your 7% normal chest x-rays, were those upright or flat, and who read them as being normal? We recently had a case which the radiologist read as normal with the later caveat as normal for technique, which we felt was abnormal and had a positive aortogram.

Second, with your normal CAT scans, I have a problem with papers now days saying the CAT scans were normal, because there's apples and oranges. Did you have a spiral CAT scan or the old-fashioned regular CAT scan, and was your contrast given by constant infusion or by bolus?

Dr. Matthew Wall (Houston, Texas): Dr. Fabian, I enjoyed your paper.

Three questions and a comment. First, did you control for patient temperature? Hypothermia has a protective effect on the spinal cord, and many of the bypass circuits have a heat exchanger where you can manipulate the patient's temperature.

Second, what number of patients were operated on without arteriography?

Third, were any of the patients treated with perioperative spinal fluid drainage?

And finally, I think one of the important messages of this paper is that a significant number of patients can be managed nonoperatively, particularly if they have concomitant injuries.

Thank you.

Dr. Ernest E. Moore, Jr. (Denver, Colorado): The reduction of postoperative paraplegia attributed to partial left heart bypass in patients requiring >30 min of aortic crossclamping may underestimate the full benefits of active bypass in the typical multisystem injured patient with a torn aorta. Specifically, bypass may provide critical ventricular unloading and attenuate hepatoenteric ischemia-reperfusion during aortic clamping. Considering this hypothesis, have you done a multiple logistic regression analysis to elucidate the potential benefits of bypass, as an independent factor, on outcome measures reflecting dysfunctional systemic hyperinflammation, i.e., ARDS or MOF?

Dr. Christopher C. Baker (Chapel Hill, North Carolina): Again, I would compliment Dr. Fabian, particularly on his presentation.

I think some of the discussants are pushing for a little bit too much from data bases like this given our experience from our study in North Carolina presented last year.

The question I have for Dr. Fabian is: Can you share with us the range of numbers of cases per institution in terms of the low number and the high number with the median, since I think that would help us look at the data base a little bit better?

Dr. A. Brent Eastman (La Jolla, California): Dr. Fabian, I also would add my congratulations to you on the multi-institutional trial study.

I have one question, and that's your patient that had the below-the-diaphragm rupture. How was that case managed? There are only approximately 32 of these cases in the literature. Was there a dissection associated with this injury? Was it managed nonoperatively?

Dr. Timothy C. Fabian (closing): I would start out by acknowledging the importance of what Chip Baker said. In a multicenter study such as this, you can't do it among 50 sites and get every kernel of data that you would like. The number of the questions that have been asked, I will not be able to answer which I am sure most of you already recognize. But, there are a number of questions that can be addressed and I will go through them individually. Wayne Meredith asked about the prolonged times from diagnosis to management. They do indeed seem long but that is what is happening in the U.S. and Canadian centers. Should we work on it? Yes, I think we should, and we would probably have fewer in-hospital ruptures in stable patients than occurred in this study. Of the ruptures which occurred in-hospital following stabilization, all of those patients died and all but one within 30 hours.

Dr. Meredith is to be complimented on his personal series with excellent results in 35 patients with short cross-clamp times. He then asked me who should be bypassed. I think that this is a crux of the question in current management of blunt aortic injury, and I do not want anyone to take this data to a court of law and use it one way or the other. It is also an issue of "don't shoot the messenger." I am simply reporting on the data that has been generated among the institutions from the multicenter trial and what is happening in operative management in those centers. But, I think if you get to centers where some individuals have huge experiences, it may not make any difference which surgical approach is taken. It may be that like most operations: the more you do the better you get, but the volume of aortic injuries is such that everyone cannot have a large experience. There are a lot of places that are going to have some cases to do but that simply will not have a large volume of experience. You cannot predict who is going to require longer cross-clamp times: the more complex injuries. Someone may be seeing three or four a year, and may not know preoperatively the extent of the disease because angiography generally does not illustrate which are the more complex injuries requiring longer cross clamp times. So I think with smaller volume institutions that they may probably should look more into bypass techniques. Also, interestingly, over the course of this study, I could see more and more institutions were going to the centrifugal pump. As far as prevention of hypertension, I think this is an excellent point and is not being used nearly enough. Once the diagnosis is considered in the patient who is not hypotensive, I believe all of these patients should be given agents to keep their mean pressure in the 70 to 80 range. Combinations of short acting beta-blockers with or without nitroprusside have been demonstrated to decrease blood pressure and wall tension. A more wide-spread use of this technique would likely decrease the 8% of patients who rupture in the hospital.

A number of questions were asked about details of paraplegia relative to shock, temperature, and the like. These data were not attempted to be addressed in this type of multicenter study. I would recommend for other people who are doing nonfunded, multicenter trials, with this organization or with other organizations in the future that it gets rapidly to the point of diminishing returns. If you ask for too many things, it gets to be a self-fulfilling prophecy that you are going to start failing and you will get less and less. So I think in retrospect, there are a lot of things to be looked at in the future but many of those should be addressed by focussed, prospective, randomized comparative trials among a few institutions.

Don Trunkey asked about celiotomies and why a third were done following aortic repair. This is simply a description of what occurred and not why. I don't know how those patients were at the end of the aortic repair, if they went into shock or had other indications. I am also quite certain, without having the data, that when we have 50 institutions everyone has got a different trigger for when they are going to do a celiotomy today with all the nonoperative management business. So it would be really difficult to address such a question without a more focussed, possibly comparative trial.

Aurelio Rodriguez asked about cross-clamp times and the contribution of perioperative hypotension to paraplegia. Those data simply are not available. There was also some suggestion that perhaps closed head injury would be made worse by bypass. That was not apparent.

Gene Moore wondered about multiple organ insults and the advantages of unloading the heart with left heart bypass and the possibility of decreased visceral organ ischemic/reperfusion injury. We did not look specifically at organ failure scores. Others have suggested that indeed organ failure may be reduced by implementing distal aortic perfusion. That is clearly an area that should be investigated more fully in the future.

There were a couple of questions about range of cases. There was a range of one to 30. It was further asked about differences in results with high volume versus low volume institutions. The one way we looked at that issue was grouping institutions that had 10 or more cases compared with those that had less than 10 for both mortality and paraplegia. My suspicion was that we would see differences. But, in fact, we saw no differences in either mortality or paraplegia according to volume analyzed in this fashion in this particular study.

Brent Eastman asked about the aortic disruption at the diaphragm. I am not sure exactly what triggered the diagnosis. It was repaired rather than nonoperatively managed.

Again, the authors would like to thank the Association for the privilege of presenting the data, which were really produced by the Association, and the honor to conduct the trial. Thank you.

Back to Top | Article Outline

REFERENCES

1. Smith RS, Chang FC: Traumatic rupture of the aorta: Still a lethal injury. Am J Surg 152:660, 1986
2. Passaro E, Pace WB: Traumatic rupture of the aorta. Surgery 46:787, 1959
3. Steinberg I: Chronic traumatic aneurysm of the thoracic aorta: Report of five cases, with a plea for conservative treatment. N Engl J Med 257:913, 1957
4. Parmley LF, Mattingly TW, Manion TW, et al: Nonpenetrating traumatic injury of the aorta. Circulation XVII:1086, 1958
5. Hilgenberg AD, Logan DL, Akins CW, et al: Blunt injuries of the thoracic aorta. Ann Thorac Surg 53:233, 1992
6. Cowley RA, Turney SZ, Hankins JR, et al: Rupture of the thoracic aorta caused by blunt trauma, a fifteen-year experience. J Thorac Cardiovasc Surg 100:652, 1990
7. Hunt JP, Baker CC, Lentz CW, et al: Thoracic aorta injuries: Management and outcome of 144 patients. J Trauma 40:547, 1996
8. Jackson DH: Of TRAs and ROCs. Chest 85:585, 1984
9. Mattox KL: Fact and fiction about management of aortic transection. Ann Thorac Surg 48:1, 1989
10. Feczko JD, Lynch L, Pless JE, et al: An autopsy case review of 142 nonpenetrating (blunt) injuries of the aorta. J Trauma 33:856, 1992
11. Smith MD, Cassidy JM, Souther S, et al: Transesophageal echocardiography in the diagnosis of traumatic rupture of the aorta. N Engl J Med 332:356, 1995
12. Cohn SM, Burns GA, Jaffe C, Milner KA: Exclusion of aortic tear in the unstable trauma patient: The utility of transesophageal echocardiography. J Trauma 39:1087, 1995
13. Miller FB, Richardson JD, Thomas HA, et al: Role of CT in diagnosis of major arterial injury after blunt thoracic trauma. Surgery 106:596, 1989
14. Durham RM, Zuckerman D, Wolverson M, et al: Computed tomography as a screening exam in patients with suspected blunt aortic injury. Ann Surg 220:699, 1994
15. Gavant ML, Menke PG, Fabian T, et al: Blunt traumatic aortic rupture: Detection with helical CT of the chest. Radiology 197:125, 1995
16. Pate JW, Fabian TC, Walker W: Traumatic rupture of the aortic isthmus: An emergency? World J Surg 19:119, 1995
17. Albrink MH, Rodriguez E, England GJ, et al: Importance of designated thoracic trauma surgeons in the management of traumatic aortic transection. South Med J 87:497, 1994
18. Kim FJ, Moore EE, Moore FA, et al: Trauma surgeons can render definitive surgical care for major thoracic injuries. J Trauma 36:871, 1994
19. Van Norman GA, Pavlin EG, Eddy AC, et al: Hemodynamic and metabolic effects of aortic unclamping following emergency surgery for traumatic thoracic aortic tear in shunted and unshunted patients. J Trauma 31:1007, 1991
20. Walls JT, Boley T, Curtis J, et al: Centrifugal pump support for repair of thoracic aortic injury. Mo Med 88:811, 1991
21. Gharagozloo F, Larson J, Dausmann MJ, et al: Spinal cord protection during surgical procedures on the descending thoracic and thoracoabdominal aorta. Chest 109:799, 1996
22. Higgins RSD, Sanchez JA, DeGuidis L, et al: Mechanical circulatory support decreases neurologic complications in the treatment of traumatic injuries of the thoracic aorta. Arch Surg 127:516, 1992
23. Katz M, Blackstone E, Kirklin J, et al: Incremental risk factors for spinal cord injury following operation for acute traumatic aortic transection. J Thorac Cardiovasc Surg 81:669, 1981
24. Zeiger A, Clark DE, Morton JR: Appraisal of surgical treatment of traumatic transection of the thoracic aorta. J Cardiovasc Surg 31:607, 1990
25. Forbes AD, Ashbaugh DG: Mechanical circulatory support during repair of thoracic aortic injuries improves morbidity and prevents spinal cord injury. Arch Surg 129:494, 1994
26. Nicolosi AC, Almassi GH, Bousamra M, et al: Mortality and neurologic morbidity after repair of traumatic aortic disruption. Ann Thorac Surg 61:875, 1996
27. Read RA, Moore EE, Moore FA, et al: Partial left heart bypass for thoracic aorta repair, survival without paraplegia. Arch Surg 128:746, 1993
28. Fabian TC, Patton JH Jr, Croce MA, et al: Blunt carotid injury: Importance of early diagnosis and anticoagulant therapy. Ann Surg 233:513, 1996
© Williams & Wilkins 1997. All Rights Reserved.