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.

Interleukin-10 is associated with the development of sepsis in trauma patients.

Interleukin-10 (IL-10) is a potent regulator of proinflammatory cytokines, including tumor necrosis factor-alpha, IL-1, IL-6, and interferon-gamma. We retrospectively evaluated 66 severely injured patients for detectable plasma IL-10. the presence or absence of IL-10 was correlated with clinical parameters. Forty of 66 patients had detectable levels of IL-10. Plasma IL-10 was associated with admission hypotension (p < 0.01) and the development of sepsis (p < 0.05). There was no difference between IL-10-positive and -negative patients with respect to age, mechanism or severity of injury, blood transfusion, operative interventions, or the subsequent development of ARDS, hepatic dysfunction, or renal insufficiency. We conclude that IL-10 can be detected in the plasma of some severely injured patients and that it is associated with the development of sepsis. Further investigation of the immunoregulatory effects of IL-10 after trauma is indicated.

Pharmacokinetics of reconstituted human high-density lipoprotein in pigs after hemorrhagic shock with resuscitation.

OBJECTIVES: Reconstituted human high-density lipoprotein (HDL) can inhibit lipopolysaccharide effects in vivo. The major objectives of this study were to characterize the pharmacokinetics of reconstituted HDL in a stressed large-animal model and to provide preclinical tolerance information in support of use of reconstituted HDL in humans. DESIGN: A randomized, blinded, placebo-controlled trial where each animal received either reconstituted human HDL at a dose of 100 mg/kg (apolipoprotein A-I) or placebo, immediately after hemorrhagic shock and resuscitation. SETTING: Animal laboratory. SUBJECTS: Twelve immature female swine (18 to 25 kg) were studied. INTERVENTIONS: Six to 8 days before shock and study drug administration, animals were anesthesized and catheters were placed in the external jugular vein and abdominal aorta. These catheters were secured to the dorsal surface. On the day of shock, the animals were sedated (alpha-chloralose) and 50 mL/kg of arterial blood was removed over 0.5 hr. One half hour after blood removal, shed blood was infused, which was immediately followed by study drug (reconstituted HDL or placebo), and then by 1 L of lactated Ringer's solution. MEASUREMENTS AND MAIN RESULTS: Physiologic (arterial blood pressure, heart rate, respiratory rate) and laboratory (serum chemistries, hematologic and coagulation studies, and blood gases) measurements were determined intermittently for 96 hrs after the induction of shock. Blood was collected intermittently for 48 hrs after shock for assay of apolipoprotein A-I and phosphatidylcholine in plasma. Reconstituted HDL was well tolerated and did not appear to alter the physiologic responses to shock and resuscitation. HDL transient increase in aspartate aminotransferase concentration was noted in the reconstituted group but this increase normalized by 24 hrs after drug administration. Mean apolipoprotein A-I pharmacokinetic parameters were as follows: half-life 24.5+/-5.3 (SD) hrs; clearance 41.9+/-10 mL/hr; and volume of distribution 1.39+/-0.08 L. The apparent mean half-life of phosphatidylcholine was 5.4+/-0.8 hrs. CONCLUSIONS: Reconstituted human HDL was well tolerated in animals that had undergone hemorrhagic shock with resuscitation. The apolipoprotein component of reconstituted HDL had a relatively long half-life, with distribution limited to the vascular space. These findings support the investigational use of this product in humans.

Emergency cricothyrotomy: a reassessment.

The Medical College of Georgia Level I Trauma Center admitted 5603 adult trauma patients from January 1, 1989 through June 30, 1993. Cricothyrotomy was required in 66 of 525 patients who required emergency airway control but could not be intubated nonsurgically in an expeditious manner. There were three major complications (thyroid cartilage laceration, significant hemorrhage, and failure to obtain a surgical airway) involving two patients, but each resolved without sequelae. Twenty-six patients with cricothyrotomy survived their hospital course, of which seven had decannulation of the cricothyrotomy without further airway procedures, and 19 had conversion to tracheostomy. No patient had clinically significant morbidity from the cricothyrotomy, whether with or without a subsequent tracheostomy. Surgical cricothyrotomy remains an important technique with low morbidity for selected trauma victims needing emergency airway control.

Temporary abdominal wall closure in trauma patients: indications, technique, and results.

From 1988 to 1992 more than 5300 patients were admitted to a level I trauma center, with 36 of these patients requiring Silastic abdominal closure. Patients ages ranged from 13 to 75 years with a mean Injury Severity Score (ISS) of 30 (range 13-50). Nineteen patients (53%) suffered penetrating injuries, and 17 (47%) were victims of blunt trauma. Silastic closure was performed at admission laparotomy in 15 patients (42%) due to inability to close the fascia primarily. Twenty-one patients (58%) underwent Silastic closure at reexploration for inability to close primarily (12 patients), elevated intraabdominal pressure with deteriorating renal function (6 patients), and wound sepsis with fascial necrosis (3 patients). Twenty-six patients (72%) survived and 10 patients (28%) expired. The causes of death were multiple organ failure (9 patients) and exsanguination (1 patient). Of the 26 survivors, 8 patients (31%) underwent fascial closure at initial hospitalization, and 18 patients (69%) required split-thickness skin grafting to visceral granulation tissue. Of these 18 patients, 13 (72%) have already undergone ventral herniorrhaphy at subsequent admission. No patient developed a complication attributable to the technique.

Single-agent versus combination antibiotic therapy in the management of intraabdominal infections.

For the treatment of intraabdominal infection, single-agent antimicrobial regimens such as beta-lactams with good antianaerobic activity are frequent alternatives to combination regimens such as aminoglycosides or aztreonam plus an antianaerobic agent such as clindamycin or metronidazole. The major issues in selecting a regimen are relative efficacy, potential for adverse drug effects, and cost. Single agents are clearly equivalent to combinations in preventing infectious complications after penetrating abdominal trauma and in treating established intraabdominal infections of mild to moderate severity or in relatively low-risk patients. A few trials demonstrated their equivalency in patients at high risk of mortality, although experience is limited. Single-agent regimens may reduce the risks of adverse drug effects compared with combination regimens, but they are not always less expensive.

Reconstituted high density lipoprotein inhibits physiologic and tumor necrosis factor alpha responses to lipopolysaccharide in rabbits.

OBJECTIVE: To determine the effect of reconstituted human high density lipoprotein (rHDL) on physiologic and cytokine responses to infusion of lipopolysaccharide. DESIGN: A blinded, randomized trial of three preparations of a purified human rHDL with apolipoprotein A-I-phosphatidyl choline-cholesterol molar ratios of 1:100:10, 1:150:10, and 1:200:0 and placebo in a rabbit lipopolysaccharide intravenous infusion model. INTERVENTIONS: Groups of six New Zealand white rabbits received either placebo or one of the three human rHDL preparations above as a single, 75-mg/kg (apolipoprotein A-I equivalent) dose intravenously over 10 minutes ending 5 minutes before the start of a 3-hour infusion of lipopolysaccharide. MAIN OUTCOME MEASURES: Mean arterial pressure, base excess, and plasma tumor necrosis factor alpha (TNF-alpha) production were determined. RESULTS: The human rHDL suppressed TNF-alpha production with the products having the highest fraction of phosphatidyl choline producing the greatest suppression of TNF-alpha production. The human rHDL 1:200:0 group maintained a low, near-baseline TNF-alpha concentration and minimal decline in mean arterial pressure and base excess throughout the lipopolysaccharide infusion in contrast to the placebo group. CONCLUSION: Reconstituted human high density lipoprotein appears to be useful in inhibiting the physiologic effects and cytokine release associated with endotoxemia and may provide adjunctive treatment for patients with gram-negative sepsis.

Does blood transfusion or hemorrhagic shock induce immunosuppression?

Blood transfusions have been implicated in predisposing patients to infection by inducing immunosuppression. This study evaluated the effects of syngeneic (ST) and allogeneic (AT) blood transfusion with and without hemorrhagic shock (HS) to determine whether transfusion or the accompanying hemorrhage affected certain components of the immune response. Lewis rats received ST or AT at 10%, 20%, or 30% of blood volume. Hemorrhagic shock was induced in other animals, which were resuscitated with either shed blood or substituted 10%, 20%, or 30% ST or AT. Intradermal staphylococcal abscess size, peritoneal leukocyte elicitation, and peritoneal macrophage Ia receptor expression were selected to measure the immune system response. Hemorrhagic shock increased abscess size significantly (p less than 0.05), but ST or AT alone or in combination with HS had no effect. Both shock and transfusion per se increased macrophage Ia receptor expression (p less than 0.05), but no additive or synergistic effect was observed. Peritoneal leukocyte elicitation was not affected by HS, ST, or AT. These results suggest that HS and not blood transfusion is a major determinant of the risk of infection.

The critical relationship of antibiotic dose and bacterial contamination in experimental infection.

Even though the usefulness of prophylactic antimicrobial administration for potentially contaminated operations is widely accepted, infection continues to occur in a finite number of cases. This study examined whether potential infection due to an increasing bacterial inoculum can be prevented or controlled by increasing antimicrobial doses. In an initial set of experiments, Sprague-Dawley rats were given various doses of cefazolin (15, 30, 60, 120 mg/kg) intraperitoneally, then serum and tissue levels were quantified. Serum and tissue drug concentrations correlated with the dose given. In another set of experiments, rats were given doses of either 0, 30 (standard dose), 60, or 120 mg/kg of cefazolin 30 min prior to subcutaneous inoculation of various doses of Staphylococcus aureus. After 6 days, inoculum sites were examined for abscess formation and size. At low levels of contamination, increasing in antibiotic dose to 30, 60, and 120 mg/kg, abscess formation was eliminated at 50, 80, and 92% of inoculum sites, respectively. At moderate levels of contamination, abscesses formed at all inoculum sites, but abscess size was significantly smaller as the dose increased. When a high inoculum of S. aureus was given, large doses of antibiotics provided no additional benefit. These data suggest that the risk of infection in this model of experimental infection is significantly related to the size of the bacterial inoculum. Increasing the dose of an effective antimicrobial increases drug concentration at the site of contamination and reduces the risk of infection. Administration of higher doses of prophylactic antimicrobials may be more effective when larger amounts of bacterial contamination are anticipated.

Helicopter transport of trauma victims: does a physician make a difference?

We studied the impact of physician presence on helicopter transportation of trauma victims during two periods; when physicians were part of the flight team and when they were not. Our data failed to demonstrate that physician participation in flights had an impact on patient outcome. The groups were comparable in average distance traveled, initial Trauma Scores, number of organ systems injured, and the final Injury Severity Scores. Each group showed an improved survival over that predicted by comparison with the Multiple Trauma Outcome Study cohort. No difference was found in the number of procedures performed at the scene, en route, or on arrival at the hospital. Untreated injuries were slightly higher in the physician-present group. It appears that experienced nurses and paramedics, operating with well-established protocols, can provide aggressive care that yields equal outcome results compared with those of a flight team that includes a physician.

Shock, transfusion, and pneumonectomy. Death is due to right heart failure and increased pulmonary vascular resistance.

To determine the physiologic cardiopulmonary abnormalities leading to death when pneumonectomy is required to stop bleeding in patients in hemorrhagic shock, we compared cardiopulmonary responses to resuscitation in pigs undergoing hemorrhagic shock alone, pneumonectomy alone, and hemorrhagic shock plus pneumonectomy. Four shock-plus-pneumonectomy pigs died acutely from right heart failure. When the five remaining shock-plus-pneumonectomy pigs were compared to the two control groups, pulmonary vascular resistance (PVR) increased to significantly higher levels than would be expected from the increase in PVR noted with resuscitation from shock alone and pneumonectomy alone. Right ventricular compensation maintained cardiac index in the hemorrhage-alone group and the pneumonectomy-alone group but could not maintain cardiac index in the shock-plus-pneumonectomy group, despite maximal increases in right ventricular systolic pressure, heart rate, and right ventricular end diastolic volume. These data indicated that resuscitation from shock plus pneumonectomy cannot be effectively accomplished because increased PVR leads to right ventricular failure, which limits left ventricular preload to levels that are insufficient to maintain cardiac index.

Packing and planned reexploration for hepatic and retroperitoneal hemorrhage: critical refinements of a useful technique.

We evaluated 35 consecutive patients treated with temporary intraabdominal packing for control of bleeding to determine factors that could improve hemorrhage control, morbidity from infection, and mortality. Twelve patients could not be resuscitated from hemorrhagic shock and died in the operating or recovery room. Bleeding was controlled in the remaining 23 patients; however, five (22%) died of complications other than hemorrhage. Intra-abdominal abscesses occurred in seven of the 21 patients who survived longer than 5 days and were more frequent in patients who had gastrointestinal perforation (50% versus 27%) and selective hepatic artery ligation (80% versus 19%). Four patients with either retrohepatic vena cava injury, hepatic vein injury, or both, were packed without attempted repair; three underwent delayed repair and survived. Coagulopathy occurred in 55% of patients who received greater than 15 units of blood before packing but in only 17% who received less than 15 units. The abdomens of ten patients were closed with a prosthetic mesh which did not prevent hemorrhage control, and only one patient developed a wound infection compared to 42% of patients with primary suture closure. We therefore conclude: 1) packing is more effective if instituted early (when less than 15 units of blood have been transfused) and is not contraindicated before either repair of retrohepatic vena cava injury, hepatic vein injury, or both; 2) selective hepatic artery ligation should be avoided if packing alone stops bleeding; 3) abdominal closure with a synthetic mesh decreases the incidence of wound infection; and 4) patients should be returned to the operating room for repacking if 24-hour postoperative blood requirements exceed 10 units.

A prospective, randomized comparison between open and closed peritoneal lavage techniques.

We randomized 327 blunt trauma patients to compare the open peritoneal lavage technique with the percutaneous (Seldinger wire) technique. The open and closed lavage groups were similar with respect to accuracy and safety. There were one complication in the percutaneous group and two in patients treated by the open method. The incidence of positive lavage was similar in each group. There was one false positive in the percutaneous group and none in the open method group. False negative results did not occur by either method. The percutaneous lavage method required less time for performance, had better patient tolerance, and only required one surgeon to perform the procedure. Percutaneous diagnostic peritoneal lavage (DPL), in the hands of trauma surgeons, is a safe and acceptable alternative to the open DPL method and actually had several advantages as mentioned above.

Evaluation of splenic injury by computed tomography and its impact on treatment.

We reviewed 37 consecutive, hemodynamically stable patients (16 adults, 21 children) who had splenic injuries diagnosed by computed tomography (CT) scan to compare the CT evaluation with operative assessment of injury and eventual treatment. Computed tomographic scans and operative findings were graded by a splenic injury scoring system. Two patients were classified as having grade 1, 21 as grade 2, 11 as grade 3, and 3 as grade 4 splenic injuries. Computed tomography underestimated the degree of injury in 9 of 17 (53%) operated patients (mean CT score, 2.6; mean operative score, 3.3; p less than 0.01). Six of sixteen adults and 19 of 21 children were intentionally treated by observation. There were 5 treatments failures (20%), 3 due to bleeding and 1 each due to pancreatic injury and splenic abscess. The failure rate of observation was lower in children (16%) than in adults (33%), even though children had a higher Splenic Injury Score (2.4 versus 1.8). Patients who underwent an operation received twice as much blood as the observed group. There was no significant difference in Injury Severity Score or total fluid requirements between operated and observed patients. Operations increased in frequency in both adults and children as the injury score increased. This experience suggests that CT scan accurately determines the presence of splenic injury but commonly underestimates its severity. While children with grades 1 through 3 injuries are likely to be treated successfully with observation, adults who have more minor splenic injuries often fail observation and may be treated better by prompt operation.

Solitary splenic abscess: a new complication of splenic salvage treated by percutaneous drainage.

A patient sustained a gunshot injury to the spleen. The spleen was left intact in an attempt to maintain normal immune function in the patient. The patient developed a splenic abscess as a result of the injury, a complication of splenic salvage that we have not found reported before. The abscess was treated successfully via CT-guided percutaneous drainage.