Penetrating esophageal injuries: multicenter study of the American Association for the Surgery of Trauma.

OBJECTIVE: The purpose of this study was to define the period of time after which delays in management incurred by investigations cause increased morbidity and mortality. The outcome study is intended to correlate time with death from esophageal causes, overall complications, esophageal related complications, and surgical intensive care unit length of stay. METHODS: This was a retrospective multicenter study involving 34 trauma centers in the United States, under the auspices of the American Association for the Surgery of Trauma Multi-institutional Trials Committee over a span of 10.5 years. Patients surviving to reach the operating room (OR) were divided into two groups: those that underwent diagnostic studies to identify their injuries (preoperative evaluation group) and those that went immediately to the OR (no preoperative evaluation group). Statistical methods included Fisher's exact test, Student's T test, and logistic regression analysis. RESULTS: The study involved 405 patients: 355 male patients (86.5%) and 50 female patients (13.5%). The mean Revised Trauma Score was 6.3, the mean Injury Severity Score was 28, and the mean time interval to the OR was 6.5 hours. There were associated injuries in 356 patients (88%), and an overall complication rate of 53.5%. Overall mortality was 78 of 405 (19%). Three hundred forty-six patients survived to reach the OR: 171 in the preoperative evaluation group and 175 in the no preoperative evaluation group. No statistically significant differences were noted in the two groups in the following parameters: number of patients, age, Injury Severity Score, admission blood pressure, anatomic location of injury (cervical or thoracic), surgical management (primary repair, resection and anastomosis, resection and diversion, flaps), number of associated injuries, and mortality. Average length of time to the OR was 13 hours in the preoperative evaluation group versus 1 hour in the no preoperative evaluation group (p < 0.001). Overall complications occurred in 134 in the preoperative evaluation group versus 87 in the no preoperative evaluation group (p < 0.001), and 74 (41%) esophageal related complications occurred in the preoperative evaluation group versus 32 (19%) in the no preoperative evaluation group (p = 0.003). Mean surgical intensive care unit length of stay was 11 days in the preoperative evaluation group versus 7 days in the no preoperative evaluation group (p = 0.012). Logistic regression analysis identified as independent risk factors for the development of esophageal related complications included time delays in preoperative evaluation (odds ratio, 3.13), American Association for the Surgery of Trauma Organ Injury Scale grade >2 (odds ratio, 2.62), and resection and diversion (odds ratio, 4.47). CONCLUSION: Esophageal injuries carry a high morbidity and mortality. Increased esophageal related morbidity occurs with the diagnostic workup and its inherent delay in operative repair of these injuries. For centers practicing selective management of penetrating neck injuries and transmediastinal gunshot wounds, rapid diagnosis and definitive repair should be made a high priority.

Angiographic embolization of bilateral internal iliac arteries to control life-threatening hemorrhage after blunt trauma to the pelvis.

Angiographic embolization of bleeding pelvic vessels is increasingly used in patients with pelvic injuries. Temporary angiographic embolization of bilateral internal iliac arteries (TAEBIIA) is occasionally necessary. From November 1991 to March 1998, 30 consecutive patients (mean age of 43 years, mean Injury Severity Score of 25) with complex pelvic fractures underwent TAEBIIA to control severe hemorrhage not responding to subselective embolization. Angiography revealed multiple sources of pelvic bleeding in 28 (93%) patients. In the two remaining patients, no bleeding was identified but TAEBIIA was done empirically. Thirteen patients had laparotomies before TAEBIIA with unsuccessful bleeding control, and the remaining 17 had TAEBIIA as the primary treatment. After TAEBIIA 90 per cent of patients had successful clinical (27 of 30) and radiographic (25 of 28) control of bleeding. Of the three patients who continued to bleed after TAEBIIA two were successfully re-embolized and one died of acute cardiac failure before any further intervention was attempted. TAEBIIA had a success rate of 97 per cent (29 of 30) in controlling pelvic hemorrhage without significant complications related to it. TAEBIIA is a safe and effective alternative to subselective embolization in controlling retroperitoneal bleeding in selected patients with blunt pelvic trauma.

Angiographic embolization for intraperitoneal and retroperitoneal injuries.

Angiographic embolization (AE) has been used extensively for bleeding control after injuries to the face and neck. Its role in abdominal trauma requires further exploration. We reviewed the medical records of 137 consecutive patients who underwent angiography with the intent to embolize bleeding sites within the abdomen. Of them, 97 (71%) had blunt and 40 (29%) had penetrating trauma. AE was performed for hemorrhage associated with pelvic fractures (97 patients), liver lacerations (n = 26), renal lacerations (n = 12), splenic lacerations (n = 5), other injuries (n = 9), and multiple injuries (n = 12). On angiography, 102 patients were found to have bleeding sites and underwent AE, with angiographic and clinical bleeding control in 93 (91%). The rate of successful hemostasis by AE was identical in blunt and penetrating trauma patients. There was no major morbidity after AE. No factors predicted patients with a high likelihood to have a positive angiogram. Patients who had AE before or after a period of attempted hemodynamic stabilization in the intensive care unit were no different with respect to hemodynamic parameters immediately before AE or effectiveness of AE for bleeding control. AE is a safe and effective method for controlling bleeding after blunt and penetrating intra- and retroperitoneal injuries. Early AE may be used in selected patients as a front-line therapeutic intervention that offers expeditious hemostasis and prevents delays in definitive bleeding control.

Nonskeletal cervical spine injuries: epidemiology and diagnostic pitfalls.

BACKGROUND: Cervical spine injuries are the most commonly missed severe injuries with serious implications for the patient and physician. The diagnosis of subluxations or spinal cord injuries in the absence of vertebral fractures, especially in unevaluable patients, poses a major challenge. The objective of this study was to study the incidence and type of cervical spine trauma according to mechanism of injury; identify problems and pitfalls in the diagnosis of nonskeletal cervical spine injuries. METHODS: Retrospective study of all C-spine injuries caused by traffic accidents or falls admitted over a 5-year period at a large Level I trauma center. Data were obtained from the trauma registry, review of patient charts, and radiology reports. RESULTS: During the study period, there were 14,755 admissions due to traffic injuries or falls who met trauma center criteria. There were 292 patients with C-spine injuries, for an overall incidence of 2.0% (3.4% in car occupants, 2.8% for pedestrians, 1.9% for motorcycle riders, and 0.9% for falls). The incidence of C-spine injuries in patients with a Glasgow Coma Scale score of 13 to 15 was 1.4%, 9 to 12 was 6.8%, and in < or =8 was 10.2% (p < 0.05). Of C-spine injuries, 85.6% (250 patients) were a vertebral fracture, 10.6% of the injuries (31 patients) were subluxation without fractures, and 3.8% (11 patients) were an isolated spinal cord injury without fracture or subluxation. Of the 31 patients with isolated subluxations, one-third required an early endotracheal intubation before clinical evaluation of the spine, because of associated severe head injury or hypotension. Adequate lateral C-spine films diagnosed or suspected 30 of the 31 subluxations (96.8%). The combination of plain films and computed tomographic (CT) scan diagnosed or suspected all injuries. Of the 11 patients with isolated cord injury, 27.3% required early intubation before clinical evaluation of the spine. The diagnosis of cord injury was made on admission in only five patients (45.5%). In three patients, the neurologic examination on admission was normal and neurologic deficits appeared a few hours later. In the remaining three patients (two intubated, one intoxicated), the diagnosis was missed clinically and radiologically. CONCLUSIONS: Isolated nonskeletal C-spine injuries are rare but potentially catastrophic because of the high incidence of neurologic deficits and missed diagnosis. In subluxations, the combination of an adequate lateral film and CT scan was reliable in diagnosing or highly suspecting the injury. A large prospective study is needed to confirm these findings, before a recommendation is made to remove the cervical collar if the findings of these investigations are normal. However, in isolated cord injuries, the diagnosis was often missed because of associated severe head trauma and the low sensitivity of the plain films and CT scans.

Approach to the management of complex hepatic injuries.

BACKGROUND: Complex hepatic injuries American Association for the Surgery of Trauma Organ Injury Scale grades IV and V incur high mortality rate ranging from 40 to 80%, respectively. The objective of this study is to assess the clinical experience with an aggressive approach to the management of these, the most complex of hepatic injuries. METHODS: This is a retrospective 6-year study (1992-1997) at an American College of Surgeons urban Level I trauma center of patients sustaining complex hepatic injuries whose interventions included surgery, angiographic embolization, endoscopic retrograde cholangiopancreatography plus biliary stenting and percutaneous computed tomographic-guided drainage. The main outcome measure was survival. RESULTS: A total of 22 patients sustaining complex hepatic injuries; mean age of 26 years (range, 10-52 years), mean Revised Trauma Scale score of 9.9, mean Injury Severity Score of 32 (range, 16-75), American Association for the Surgery of Trauma - Organ Injury Scale grade IV (13 cases); grade V (9 cases). Mean estimated blood loss was 4,600 mL; mean number of units of blood transfused was 15. The patients underwent the following interventions: surgery (n = 22), re-operated (n = 13), mean number of operations 1.6 (range, 1-4), extensive hepatotomy and hepatorrhaphy (n = 17), nonanatomic resection (n = 7), formal hepatectomy (n = 4), packing (n = 10), direct approach to hepatic veins (n = 3); angiographic embolization (n = 15); endoscopic retrograde cholangiopancreatography and stenting (n = 5); computed tomographic guided drainage (n = 6). Mean length of stay in the intensive care unit was 21 days (range, 2-134 days), mean hospital length of stay was 40 days (range, 2-147 days). Overall mortality rate was 14% (3 of 22 cases), hepatic mortality rate was 9% (2 of 22 cases), mortality rate by injury grade was 8% grade IV (1 of 13 cases) and 22% grade V (2 of 9 cases). CONCLUSION: In this select patient population, improvements in mortality rates can be achieved with an aggressive approach to the management of complex hepatic injuries, including surgery, early packing, angiographic embolization, endoscopic retrograde cholangiopancreatography and stenting of biliary leaks, and drainage of hepatic abscesses.

Operative management and outcome of 302 abdominal vascular injuries.

BACKGROUND: Abdominal vascular injuries incur high mortality rates. The purposes of this study are (1) review institutional experience, (2) determine additive effect on mortality of multiple vessel injuries, (3) determine mortality of combined arterial and venous injuries, and (4) correlate mortality with American Association for the Surgery of Trauma-Organ Injury Scale (AAST-OIS) for abdominal vascular injury. METHODS: A retrospective 6-year study was made at an urban level I trauma center of patients with abdominal vascular injuries. Main outcome measure was survival. RESULTS: (1) There was a total of 302 patients, mean age 28, mean Injury Severity Score (ISS) 25 (range 4 to 75). Mechanism of injury was penetrating in 266 (88%), blunt in 36 (12%). Emergency Department thoracotomy was done in 43 of 302 (14%), 504 vessels were injured: arteries 238(47%), veins 266(53%). Surgical management was ligation 245, primary repair 141, prosthetic interposition grafts 24, autogenous 2. Overall mortality was 162 of 302 (54%). (2) Mortality multiple vessels injured: 1 vessel 160 (45%), 2 vessels 102 (60%), 3 vessels 33 (73%), >4 vessels 5 (100%). Mortality arterial injuries: aorta isolated (I) 78% versus combined with other arterial injuries (C) 82.4%, superior mesenteric artery (SMA) I 47.6% versus C 71.4%, iliac I 53% versus C 72.7%, renal I 37.5% versus C 66.7%. Venous injuries: inferior vena cava (IVC) isolated (I) 70% versus combined with other venous injuries (C) 77.7%, superior mesenteric vein (SMV) I 52.7% versus C 65%, IMV I 16% versus C 50%. (3) Specific mortality combined arterial and venous injuries: aorta plus IVC 93%, SMA plus SMV 43%, iliac artery plus vein 45.5%. (4) Mortality versus AAST-OIS: grade II 25%, grade III 32%, grade IV 65%, grade V 88%. CONCLUSION: Abdominal vascular injuries are highly lethal. Multiple arterial and venous injuries increase mortality. Mortality correlates with AAST-OIS for abdominal vascular injury.

Duration of antibiotic prophylaxis in high-risk patients with penetrating abdominal trauma: a prospective randomized trial.

To evaluate the effect of varying durations of antibiotic prophylaxis in trauma patients with multiple risk factors for postoperative septic complications, a prospective randomized trial was undertaken at an urban level I trauma center. The inclusion criteria were full-thickness colon injury and one of the following: (1) Penetrating Abdominal Trauma Index > 25, (2) transfusion of 6 units or more of packed red blood cells, or (3) more than 4 hours from injury to operation. Patients were randomly assigned to a short course (24 hours) or a long course (5 days) of antibiotic therapy. All patients received 2 g cefoxitin en route to the operating room and 2 g intravenously piggyback every 6 hours for a total of 1 day vs. 5 days. Sixty-three patients were equally divided into short-course (n = 31) and long-course (n = 32) therapy. This was a high-risk patient population, as assessed by the mean Penetrating Abdominal Trauma Index (33), number of patients with multiple blood transfusions (51 of 63; 81%), number of patients with an Injury Severity Score greater than 15 (37 of 63; 59%), number of patients with destructive colon wounds requiring resection (27 of 63; 43%), and number of patients requiring postoperative critical care (37 of 63; 59%). Differences in intra-abdominal (1-day, 19%; 5-days, 38%) and extra-abdominal (1-day, 45%; 5-days, 25%) infection rates did not achieve statistical significance. There continues to be no evidence that extending antibiotic prophylaxis beyond 24 hours is of benefit, even among the highest risk patients with penetrating abdominal trauma. A large, multi-institutional trial will be necessary to condemn this common practice with statistical validity.

Angiographic embolization for arrest of bleeding after penetrating trauma to the abdomen.

BACKGROUND: Angiographic embolization is an effective technique to control bleeding after blunt trauma to the liver or pelvis. Its role in penetrating trauma to the abdomen has not been studied. METHODS: From January 1992 to May 1998, 40 patients underwent angiography for bleeding resulting from intra-abdominal penetrating injuries (33 gunshot wounds, 7 stab wounds). Angiographic embolization of intraperitoneal or retroperitoneal vessels was performed by standard angiographic techniques with gelatin sponge and/or coils. Data were extracted from medical records, radiology data bank, trauma registry, and morbidity/mortality records, and compared by Student's t test and chi-square test. The main outcome measures were failure of angiographic embolization to control bleeding and complications of angiographic embolization. RESULTS: Angiography was performed during a course of nonoperative management in 6 patients (group A), because of failure to control bleeding surgically in 23 (group B), and because of late vascular complications after an initially successful operation in 11 more (group C). In 32 patients, angiography revealed active bleeding; 29 (91 %) underwent successful angiographic embolization. Of the remaining 3 patients, 2 were successfully managed surgically (1 each from groups A and B) and 1 died despite multiple surgical maneuvers (group B). One patient who developed postoperatively a large, bleeding superior mesenteric artery pseudoaneurysm, suffered extensive bowel necrosis after angiographic embolization. No other significant complication was related to angiographic embolization. CONCLUSIONS: Angiographic embolization after penetrating injuries to the abdomen is safe and effective for a small number of selected patients. It is a valuable tool for bleeding control when surgery has failed. It may be ideal for control of late vascular complications when reoperation is not desirable. It may prove to be a useful adjunct in the nonoperative treatment of selected injuries.

Traumatic injury to the superior mesenteric artery.

BACKGROUND: Superior mesenteric artery (SMA) injuries are rare and devastating injuries incurring very high mortality rates. It is the purpose of this study to review our experience with these injuries, to analyze Fullen's classification based on anatomical zone and injury grade for its predictive value, and to correlate the American Association for the Surgery of Trauma-Organ Injury Scale (AAST-OIS) for abdominal vascular injury with mortality. METHODS: Retrospective study was made over a 65-month period of all patients sustaining SMA injuries in an urban level I trauma center. RESULTS: Thirty-five patients, mean age 31, had a mean Revised Trauma Score of 5.86 and a mean Injurity Severity Score of 23. Mechanisms of injury were penetrating 27 (77%) and blunt 8 (23%). Mean admission systolic blood pressure was 85 mm Hg. Mean estimated blood loss was 8,500 mL and mean total fluid replacement 17,000 mL. Operating room findings were retroperitoneal hematoma in 34 (97%) and "black bowel" in 2 (6%). Number of associated injuries was nonvascular, mean 4.2, and vascular, mean 1.5. Surgical management consisted of ligation in 18 (51%), primary repair in 14 (40%), and interposition graft in 2 (6%). Overall mortality was 19 of 35 (54%). Mortality versus Fullen's zones was zone I, 100%, zone II, 43%, and zones III and IV, 25%. Mortality versus Fullen's ischemia grade was grade 1, 89%, grade 2, 58%, grade 3, 100%, and grade 4, 19%. Mortality versus AAST-OIS: was grade 1, 0%, grade II, 20%, grade III, 0%, grade IV, 59%, and grade V, 88%. CONCLUSIONS: SMA injuries are highly lethal. Most deaths are due to exsanguination. A higher number of associated vascular injuries increases mortality. "Black bowel" is an uncommon finding. Both Fullen's anatomical zones and the AAST-OIS for abdominal vascular injuries correlate with mortality. Fullen's ischemia grade does not.

Gunshot injuries to the liver: the role of selective nonoperative management.

BACKGROUND: Selective nonoperative management of blunt liver injuries has become standard practice in most trauma centers. We evaluated the role of selective nonoperative management of gunshot wounds to the liver. STUDY DESIGN: This was a retrospective review of gunshot wounds to the liver treated in a level I trauma center. Patients with peritoneal signs or hemodynamic instability were operated on without delay. Patients with a soft, nontender abdomen and no signs of heavy bleeding were selected for nonoperative management. Liver injury was diagnosed by CT scan. If peritonitis or signs of substantial internal bleeding developed, an operation was performed; otherwise the patient was discharged within a few days of admission. Analysis was restricted to the group of patients with isolated liver injuries. RESULTS: During a 42-month period, 928 patients were admitted with abdominal gunshot injuries, 152 of whom (16%) had a liver injury. In 52 patients (34% of all liver injuries), the liver was the only injured intraabdominal organ (4 patients had associated kidney or splenic injuries that did not require surgical repair). Thirty-six of the patients (69%) with isolated liver injuries had an emergent operation because of signs of peritonitis or hemodynamic instability. The remaining 16 patients (31%) were selected for nonoperative management (3 patients had associated right kidney injury). Five patients in the observed group required delayed operation because of development of signs of peritonitis (4 patients) or abdominal compartment syndrome (1 patient). The remaining 11 patients (7% of all liver injuries or 21% of isolated liver injuries) were managed successfully without operation. One patient with delayed operation developed multiple complications from abdominal compartment syndrome, and 1 patient in the nonoperative group had a biloma, which was treated with percutaneous drainage. CONCLUSIONS: Selected patients with isolated grades I and II gunshot wounds to the liver can be managed nonoperatively.

Penetrating injuries to the subclavian and axillary vessels.

BACKGROUND: Subclavian and axillary vascular injuries are notorious for their mortality and their difficult surgical exposure. In the present study we analyze our experience with 79 patients and describe the techniques used for surgical access to these vessels. STUDY DESIGN: Retrospective review of the medical records of all patients with penetrating injuries to the subclavian or axillary vessels who had been admitted to 2 Los Angeles trauma centers during a 4-year, 3-month period. RESULTS: Seventy-nine patients were admitted during the study period January 1993 to March 1997 (58 gunshot injuries, 21 other penetrating injuries). The artery was injured in 59 patients and the vein in 40 (20 patients had both arterial and venous injuries). Eighteen patients (23%) were admitted with no signs of life or were in extremis and underwent an emergency room thoracotomy without any survivors. Fifty-eight patients underwent exploration in the operating room, 1 patient with an arteriovenous subclavian fistula was successfully managed with a radiologically placed endovascular stent, and 2 patients with minimal subclavian artery injuries were managed nonoperatively. Overall mortality was 34.2%. Excluding the ER thoracotomies the overall mortality was 14.8%. The mortality for isolated arterial injuries was 20.5%, for isolated venous injuries 50%, and for both vessels 45.0%. The mortality in venous injuries was significantly higher than in arterial injuries (p < 0.05). The standard clavicular incision provided adequate exposure in 32 (50.0%) of the operating room cases. In the other 50% of operating room cases a combination of a clavicular incision with a median sternotomy or thoracotomy was necessary. Proximal subclavian injuries may be accessed through a clavicular incision combined with a median sternotomy irrespective of left or right site location. CONCLUSIONS: Subclavian and axillary vascular injuries remain lethal. A clavicular incision provides satisfactory surgical exposure in about half the patients. In patients with proximal injuries addition of a median sternotomy provides adequate surgical access in both right and left subclavian vessels.

Double intussusception in an adult following a laparotomy for trauma: a case report and review of the literature.

Intussusception is an unusual occurrence in adults. Postoperative intussusception may represent a separate pathophysiologic entity from primary intussusception in adults, occurring more frequently in the small bowel and without a specific underlying cause. Surgical resection is not mandatory and in those patients with viable bowel in whom the intussusception can successfully be reduced, manual reduction alone is an appropriate surgical strategy. Intussusception after laparotomy for trauma is a recently recognized phenomenon. For as yet unexplained reasons there seems to be a high association with liver injuries. A case of double intussusception is presented after a penetrating injury to the liver with discussion of the possible etiology and pathophysiology of this entity.

The fate of colonic suture lines in high-risk trauma patients: a prospective analysis.

BACKGROUND: Some authors have stated that virtually all patients with penetrating colon injuries can be safely managed with primary repair. The purpose of this study is to test the applicability of this statement to all trauma patients by evaluating a protocol of liberal primary repair applied to a group of patients at high risk of septic complications. STUDY DESIGN: We performed a prospective analysis of a liberal policy of primary repair applied to patients at high risk of developing postoperative septic complications admitted to a Level I urban trauma center. Inclusion criteria were full-thickness colon injury and at least one of three additional risk factors: 1) Penetrating Abdominal Trauma Index (PATI) of 25 or more; 2) 6 U or more of blood transfused; and 3) 6 hours or longer elapsed between injury and surgery. RESULTS: Of 56 patients studied (55 male, 1 female, average age 28.8 years, mean PATI 35.3), the vast majority had gunshot wounds as the mechanism of injury (89%), PATI 25 or more (95%), multiple blood transfusions (77%), an Injury Severity Score greater than 15 (66%), and a need for postoperative ventilatory support in the surgical intensive care unit (61%). Of 56 patients, 49 (88%) had at least one colonic suture line, and 25 patients (45%) had destructive colon injuries requiring resection. Intraabdominal infections occurred in 15 (27%) of 56 patients and colon suture line disruption occurred in 3 (6%) of 49. Two of these patients developed multisystem organ failure, and death was directly related to breakdown of their colonic anastomosis. CONCLUSIONS: On the basis of these data and the relative infrequency of patients in prospective randomized trials with destructive colon injuries, we believe there is still room for consideration of fecal diversion in patients in high-risk categories with destructive colon injuries requiring resection.

Initial evaluation and management of gunshot wounds to the face.

BACKGROUND: The literature on early management of gunshot wounds (GSWs) to the face is scant, with only six series reported in the English-language literature in the last 12 years. In the current study, we present a large series from a busy trauma center in an effort to identify early diagnostic and therapeutic problems and recommend management guidelines. METHODS: Retrospective analysis was done for all GSWs of the face during a 4-year period. Data were obtained from the Trauma Registry and Trauma Patient Summary hard copies. RESULTS: During the study period, there were 4,139 admissions for GSWs, with 247 (6%) involving the face. An associated brain trauma was found in 42 patients (17.0%), and cervical spine fracture was found in 20 patients (8.1%) with GSWs to the face. In 43 patients (17.4%), there was a need for emergency airway control because of local hematoma or edema. Angiography was performed in 70 patients (28.3%) for evaluation of a large hematoma or continuous bleeding, and in 10 of these patients successful embolization of bleeders was achieved. No patient required operative control of bleeding from facial structures. Overall, only 96 patients (38.9%) underwent operation for soft-tissue repair or reduction of facial bone fractures. There were 36 deaths (14.5%) from severe brain injury or severe bleeding from associated chest or abdominal injuries. No death occurred in isolated GSWs to the face. CONCLUSION: Most civilian GSWs can safely be managed nonoperatively. Airway control is required in a significant number of patients and should be established very early. Bleeding from the face is best controlled angiographically. The brain and cervical spine should be aggressively assessed radiologically because of the high incidence of associated trauma.

One hundred five penetrating cardiac injuries: a 2-year prospective evaluation.

OBJECTIVES: To analyze the parameters measured in the field, during transport, and upon arrival of the physiologic condition of patients sustaining penetrating cardiac injuries, along with the Cardiovascular Respiratory Score (CVRS) component of the Trauma Score, the mechanism and anatomical site of injury, operative characteristics, and cardiac rhythm as predictors of outcome. We also set out to identify a set of patient characteristics that best predict mortality outcome and to correlate cardiac injury grade as determined by the American Association for the Surgery of Trauma-Organ Injury Scale (AAST-OIS) with mortality. METHODS: This report was a prospective study at American College of Surgeons Level I urban trauma center. Interventions included thoracotomy, sternotomy, or both, for resuscitation and definitive repair of cardiac injury. The main outcome measures used were those parameters measuring physiologic condition of patients, CVRS, mechanism and anatomical site of injury, mortality, and grade of injury. RESULTS: A total of 105 patients sustained penetrating cardiac injuries: 68 injuries (65%) were gunshot wounds and 37 injuries (35%) were stab wounds. The mean Injury Severity Score was 36. Of the 105 wounds, 23 wounds (22%) involved multiple-chamber injuries. The overall survival was 35 of 105 patients (33%): survival of gunshot wound victims was 11 of 68 patients (16%); survival of stab wound victims was 24 of 37 patients (65%). Emergency department thoracotomy was performed in 71 of the 105 patients (68%) with 10 survivors (14%). CVRS: 94% mortality (50 of 53) when CVRS = 0, 89% mortality (57 of 64) when CVRS = 0 to 3, and 31% mortality (12 of 39) when CVRS 4 to 11 (p < 0.001). The presence of sinus rhythm when pericardium was opened predicted survival (p < 0.001). Anatomical site of injury (injured chamber) and the presence of tamponade did not predict survival. Stepwise logistic regression analysis identified gunshot wound, exsanguination, and restoration of blood pressure as most predictive variables of mortality. AAST-OIS injury grade and mortality: grade I, 0 of 1 (0%); grade II, 1 of 2 (50%); grade III, 2 of 3 (66%); grade IV, 28 of 50 (56%); grade V, 29 of 38 (76%); grade VI, 10 of 11 (91%). Overall incidence: grades IV-VI, 99 of 105 (94%). CONCLUSIONS: Parameters measuring physiologic condition, CVRS, and mechanism of injury are significant predictors of outcome in penetrating cardiac injuries. AAST-OIS injury grades I-III are rare in penetrating cardiac trauma. AAST-OIS Injury grades IV-VI are common in penetrating cardiac trauma and accurately predict outcome.