Abdominal Trauma

The abdomen is frequently injured after both blunt and penetrating trauma. Approximately 25% of all trauma victims will require abdominal exploration. Clinical evaluation of the abdomen by means of physical examination is inadequate to identify intra-abdominal injuries because of the high number of patients with altered mental status secondary to head trauma, alcohol, or drugs and because of the inaccessibility of the pelvic, upper abdominal, and retroperitoneal organs to palpation. For these reasons, several diagnostic modalities have evolved during the past 3 decades, including diagnostic peritoneal lavage (DPL), ultrasound, CT, and laparoscopy, all of which have advantages, disadvantages, and limitations. The development of more modern technology, experience, and invasiveness have been the most important determinants of the use of diagnostic methods for abdominal trauma. In modern trauma centers in the 21st century, better noninvasive technology favors the use of ultrasound and CT in the evaluation of trauma victims.
Mechanism of Injury

Blunt trauma secondary to motor vehicle accidents, motorcycle accidents, falls, assaults, and striking of pedestrians remains the most frequent mechanism of abdominal injury. Penetrating abdominal wounds are usually caused by either gunshot or stab wounds and by a significantly smaller number of shotgun wounds. Based on the high frequency of intra-abdominal organ injury after gunshot wounds, mandatory abdominal exploration, with the rare exception of tangential and superficial wound trajectories restricted to the right upper quadrant, remains the standard form of management. Stab wounds to the abdomen, however, carry a significantly lower risk of intra-abdominal organ injury than do gunshot wounds, and several studies have recently favored a more selective approach, as opposed to mandatory exploratory laparotomy. The impetus for nonoperative management of solid organ injury in stable blunt trauma patients has expanded to penetrating trauma as well. With improved imaging, more stable patients sustaining a single solid organ injury after stab and gunshot wounds to the abdomen will be treated conservatively. In children, besides the aforementioned mechanisms of injury, child abuse and trauma secondary to recreational activities such as bicycling, swimming, and roller skating should also be considered.
Diagnosis

The history of the traumatic event is particularly important in determining the likelihood of an intra-abdominal organ injury. All possible information should be obtained from the prehospital personnel, including the mechanism of injury, the height of a fall, damage to the interior and exterior of a vehicle in a motor vehicle accident, other deaths at the scene, ejection, vital signs, mental status, the presence of external bleeding, the type of weapon, and other pertinent data.

On arrival at the hospital, the history and physical examination are usually accurate in determining intra-abdominal injury in an awake and responsive patient, although the limitations of physical examination are significant. Many patients with moderate intra-abdominal bleeding will be in a compensated hemodynamic condition and will not have peritoneal signs. Furthermore, retroperitoneal and pelvic injuries cannot be ruled out on the basis of only physical findings. We believe that an objective abdominal evaluation is necessary and should be performed by any of the available diagnostic modalities, in addition to the physical examination. The test of choice depends on the hemodynamic stability of the patient and the severity of associated injuries. Hemodynamically stable patients sustaining blunt trauma are adequately evaluated by abdominal ultrasound or CT, unless other severe injuries take priority and the patient needs to go to the operating room before the objective abdominal evaluation. In such instances, DPL or focused abdominal sonography for trauma (FAST) is usually performed in the operating room to rule out intra-abdominal bleeding requiring immediate surgical exploration. Hemodynamically stable blunt trauma patients are evaluated by ultrasound in the resuscitation room, if available, or by DPL to rule out intra-abdominal injuries as the source of blood loss and hypotension. Hypotensive patients with isolated penetrating abdominal trauma who are hypotensive or in shock or have peritoneal signs should go to the operating room despite the mechanism of injury. Stab wound victims without peritoneal signs, evisceration, or hypotension benefit from wound exploration and DPL. Gunshot wound victims should generally undergo exploration.
Plain Radiographs

The chest radiograph is a useful test to reveal pneumoperitoneum, abdominal contents in the chest (ruptured hemidiaphragm), or lower rib fractures. This later sign increases the probability of splenic and hepatic injury. IV pyelography (IVP) and retrograde cystography, useful tests in the past in the evaluation of a trauma patient with hematuria, have largely been substituted by contrast-enhanced CT. With the current frequent use of CT to objectively evaluate the abdomen after blunt trauma in stable patients, the routine use of AP pelvic radiographs, as recommended by the ATLS course, has been questioned. Stable patients undergoing CT of the abdomen and pelvis do not need a pelvic radiograph. Unstable patients, however, may continue to benefit from a pelvic radiograph because other priorities may take place that will require prompt diagnosis of a pelvic fracture in the trauma resuscitation room. Other studies suggest that clinical factors could accurately identify patients at high risk for pelvic fractures, thus making routine films unnecessary.[32]
Diagnostic Peritoneal Lavage

DPL is a rapid and accurate test used to identify intra-abdominal injuries after blunt trauma in a hypotensive or unresponsive patient without obvious indication for abdominal exploration. Standard criteria for positive DPL findings in blunt trauma include aspiration of at least 10 mL of gross blood, a bloody lavage effluent, a red blood cell count greater than 100,000/mm3, a white blood cell count greater than 500/mm3, amylase level greater than 175 IU/dL, or detection of bile, bacteria, or food fibers. The indications and contraindications for DPL are listed in Box 202 . DPL is highly sensitive to the presence of intraperitoneal blood; however, its specificity is low, and because positive DPL findings prompt surgical exploration, a significant number of explorations will be nontherapeutic. Box 20-2 Indications and Contraindications for Diagnostic Peritoneal Lavage Indications Equivocal physical examination Unexplained shock or hypotension Altered sensorium (closed head injury, drugs, etc.) General anesthesia for extra-abdominal procedures Cord injury Contraindications Clear indication for exploratory laparotomy Relative contraindications: Previous exploratory laparotomy Pregnancy Obesity Significant injuries may also be missed by DPL. Diaphragmatic tears, retroperitoneal hematomas, and renal, pancreatic, duodenal, minor intestinal, and extraperitoneal bladder injuries are frequently underdiagnosed by DPL alone. Complications are infrequent and mostly related to iatrogenic injuries caused during insertion of the catheter into the abdominal cavity. A semiopen or open technique should be the preferred method to avoid or reduce the incidence of such complications. DPL results can be misleading in the presence of a pelvic fracture. False-positive findings are expected with bleeding from the retroperitoneum into the peritoneal cavity. Anterior abdominal and flank wounds can be accurately evaluated by DPL. False-positive results are frequent after DPL because of bleeding of the abdominal wall, thus increasing the number of negative explorations. Another potential disadvantage of DPL is its low accuracy in the diagnosis of hollow viscus injuries. Debate still exists regarding the most appropriate positive criteria to determine the threshold for surgical exploration after stab wounds to the abdomen. If a red blood cell count of 1000/mm3 is considered, the number of negative explorations may be higher than

20%. If 100,000/mm3 is considered, the missed injury rate will approach 5%. There is no consensus on this matter, although most trauma centers use a low threshold (cell count between 1000 and 5000/mm3) for exploration.
Ultrasound

Ultrasound has been used more frequently in recent years in the United States for evaluation of blunt abdominal trauma patients. The objective of ultrasound evaluation is to search for free intraperitoneal fluid. It can be done expeditiously and is as accurate as DPL in detecting hemoperitoneum. It can also evaluate the liver and the spleen once free fluid is identified; however, this is not its main purpose. Portable machines can be used in the resuscitation area or in the emergency department in a hemodynamically unstable patient without delaying the resuscitation. Another advantage of ultrasound over DPL is its noninvasiveness. No further workup is necessary after a negative ultrasound in a stable patient. CT of the abdomen usually follows positive ultrasound findings in a stable patient. The advantages and disadvantages of abdominal ultrasound are listed in Box 20-3 . Its sensitivity ranges from 85% to 99% and its specificity from 97% to 100%.[33] Box 20-3 Advantages and Disadvantages of Ultrasound Advantages Noninvasive Does not require radiation Useful in the resuscitation room or emergency department Can be repeated Used during initial evaluation Low cost Disadvantages Examiner dependent Obesity Gas interposition Lower sensitivity for free fluid <500 mL False-negatives: Retroperitoneal and hollow viscus injuries
Abdominal Computed Tomography

CT is the most frequently used method to evaluate a stable blunt abdominal trauma patient. The retroperitoneum is best evaluated by CT. The indications and contraindications for abdominal CT are listed in Box 20-4 . The drawback of CT is the need to transport the patient to the radiology department. Additionally, it is more expensive than other tests. CT also evaluates solid organ injury, and in a stable patient with positive ultrasound findings, it is indicated to grade organ injury and to evaluate contrast extravasation. If contrast extravasation is seen, even with

minor hepatic or splenic injuries, exploratory laparotomy or, more recently, angiography and embolization are indicated. Another indication for CT is in the evaluation of patients with solid organ injuries initially treated nonoperatively who have a falling hematocrit. The most important disadvantage of CT is its inability to reliably diagnose hollow viscus injury ( Box 20-5 ). Usually, the presence of free abdominal fluid on CT without solid organ injury should raise suspicion for mesenteric, intestinal, or bladder injury, and exploratory laparotomy is often warranted. Box 20-4 Indications and Contraindications for Abdominal Computed Tomography Indications Blunt trauma Hemodynamic stability Normal or unreliable physical examination Mechanism: Duodenal and pancreatic trauma Contraindications Clear indication for exploratory laparotomy Hemodynamic instability Agitation Allergy to contrast media Box 20-5 Advantages and Disadvantages of Abdominal Computed Tomography Advantages Adequate assessment of the retroperitoneum Nonoperative management of solid organ injuries Assessment of renal perfusion Noninvasive High specificity Disadvantages Specialized personnel Hardware Duration: Helical versus conventional Hollow viscus injuries Cost

One of the most intriguing problems regarding the objective evaluation of blunt abdominal trauma by CT is what to do when free fluid without signs of solid organ or mesenteric injury is found. Coupled with the relatively marginal sensitivity of CT to diagnose hollow viscus injury, it creates a dilemma for most trauma surgeons. The options are either to surgically explore all patients and accept a significant rate of nontherapeutic laparotomy or to observe and act when peritoneal signs develop while keeping in mind that a delay in the diagnosis of bowel injury may be catastrophic. A recent survey in which trauma surgeons were asked what would be the appropriate management of patients in this circumstance showed a variety of responses: 42% would perform DPL, 28% would observe the patient, 16% would surgically explore, and 12% would repeat an abdominal CT scan. As technology evolves, diagnosis of mesenteric and hollow viscus injury by CT will be facilitated. Two- and three-dimensional reconstructions may help in the identification of bowel thickening, small bubbles of free air in the proximity of the area of injury, and small amounts of free fluid between loops of bowel or in the mesentery. The accuracy of CT ranges from 92% to 98%, with low false-positive and false-negative rates. Although the use of abdominal CT for the evaluation of penetrating abdominal trauma has been limited because of low sensitivity in diagnosing bowel and diaphragmatic injury, newer technology (multislice spiral CT) has been evaluated in selected cases when nonoperative management is being considered.[34] Nonoperative management of stab wounds to the anterior aspect of the abdomen has been emphasized because of the high morbidity rate after nontherapeutic laparotomy.
Other Diagnostic Modalities

Despite the initial enthusiasm, the use of diagnostic laparoscopy in blunt trauma patients is very limited. It is an invasive and expensive method and does not seem to be superior to other methods used for decision making. Missed small bowel, splenic, and retroperitoneal injuries have been reported. It seems that laparoscopy is the best method for evaluating diaphragmatic injuries after thoracoabdominal penetrating injuries. Angiography is used to evaluate renal artery thrombosis and to manage pelvic hemorrhage in patients with pelvic fractures and bleeding from minor hepatic and splenic injuries.
Gastric Injuries

Gastric injuries often result from penetrating trauma. Less than 1% of such wounds are due to blunt trauma secondary to motor vehicle accidents, falls, cardiopulmonary resuscitation, or interpersonal violence. The stomach is partially protected by the rib cage, thus making blunt injuries rare and relatively difficult to diagnose. Causes of blunt gastric rupture include vigorous ventilation with inadvertent placement of an endotracheal tube in the esophagus, crushing of the stomach against the spine, cardiopulmonary resuscitation, the Heimlich maneuver, and other causes leading to a sudden increase in intraluminal pressure.

Blunt gastric trauma includes a wide range of injuries, from mucosal lacerations to full-thickness disruption and gastric necrosis secondary to avulsion of the vascular pedicles. Other intraabdominal and extra-abdominal injuries are frequently present. DPL or CT of the abdomen may confirm the diagnosis; however, in most instances the diagnosis will be made during surgical exploration. Any penetrating abdominal injury, particularly in the upper part of the abdomen, should be suspected of causing injury to the stomach. During initial evaluation a nasogastric tube should be inserted, and if the aspirate is positive for blood, injury to the stomach should be suspected. The intraoperative evaluation includes good visualization of the esophagogastric junction, examination of the anterior gastric wall, opening of the gastrocolic ligament, and complete visualization of the posterior gastric wall. Minor injuries may not be identified and require distention of the organ with saline or methylene blue to evaluate for leak. Most penetrating wounds are treated by dbridement of the wound edges and primary closure in layers. Injuries with major tissue loss may best be treated by gastric resection. Postoperative complications include bleeding, usually from submucosal vessels; intra-abdominal abscesses; and, more rarely, gastric fistula with peritonitis. Because of its proximity to the diaphragm, the stomach is frequently injured after thoracoabdominal wounds. Depending on the severity of contamination from spillage of gastric contents, empyema is another frequent complication. The role of extravasation of gastric contents in the genesis of postoperative complications is closely related to the dynamics of the gastric flora. Usually, many microorganisms originating from the nasopharynx and oropharynx reach the stomach through saliva and nasal mucus. Changes in gastric pH are frequent after eating, drinking, and ingestion of saliva, which act in an attempt to neutralize gastric acidity. When gastric pH is below 4, gastric juice has bactericidal properties that act by inhibiting bacterial enzymatic activity. In this situation, microorganisms such as Streptococcus salivarius, Streptococcus viridans, Lactobacillus, Bacteroides, Veillonella, Micrococcus, Staphylococcus, and Neisseria are found in very low concentrations, usually below 1000/mL. Inversely, when gastric pH is neutralized, the bactericidal properties of gastric juice are extremely suppressed, which leads to prompt bacterial growth. Concentrations can reach as high as 106/mL and remain there for approximately 1 hour before returning to normal levels. If the neutralization occurs for prolonged periods, bacteria from the lower digestive tract, such as Bacteroides fragilis, Escherichia coli, Streptococcus faecalis, and enterobacteria, can be found inside the stomach. This fact is especially important in trauma patients, who frequently have great amounts of food and liquid inside the stomach. Morbidity and mortality rates after penetrating abdominal injuries associated with gastric wounds have been reported to be close to 27% and 14%, respectively, in most cases because of the presence of associated injuries, although the risk for morbidity from gastric injury itself is close to 6%.
Injuries to the Duodenum

The majority of duodenal injuries are caused by penetrating trauma; however, blunt injuries, though infrequent, are difficult to diagnose because patients may have subtle findings on admission. The incidence of duodenal injuries varies from 3% to 5%. Most duodenal injuries are accompanied by other intra-abdominal injuries because of the close anatomic relationship of the duodenum with other solid organs and major vessels. A motor vehicle accident causing impact of the steering wheel on the epigastrium is the most common mechanism of blunt duodenal injuries. Other mechanisms such as assault and falls also cause duodenal injuries. Closed-loop compression from a direct blow to an air-filled loop can account for duodenal rupture. The retroperitoneal location of the duodenum (second and third portions) exerts a protective effect against injuries, but it also prevents early diagnosis. Isolated injury to the duodenum is rare and does not usually cause significant clinical signs of peritonitis or hemodynamic instability. A thorough search based on the mechanism of injury is necessary to prevent delays in diagnosis. Failure to recognize this injury is associated with the development of intra-abdominal abscesses and sepsis and high mortality rates. Hyperamylasemia occurs in about 50% of patients with blunt injury to the duodenum, and although it is not diagnostic of an injury, its presence should raise suspicion and further diagnostic studies should be obtained. Plain films of the abdomen may suggest duodenal injury by showing mild scoliosis, obliteration of the right psoas shadow, absence of air in the duodenal bulb, or air in the retroperitoneum outlining the kidney. Definitive diagnosis requires a diatrizoate meglumine (Gastrografin) upper gastrointestinal series or a CT scan of the abdomen with oral and IV contrast in hemodynamically stable patients ( Fig. 20-17 ). Extravasation of contrast material is an absolute indication for laparotomy. The radiographic finding of a duodenal hematoma (coiled spring or stacked coin sign) is not an indication for surgical exploration. If a hematoma is causing obstruction that fails to resolve, operative management is indicated. If CT findings are equivocal, an upper gastrointestinal series with diluted barium is the test of choice.

Figure 20-17 Computed tomography scan showing a blunt duodenal injury with retroperitoneal air.

The utility of duodenography in the diagnosis of blunt duodenal injury was recently evaluated and compared with that of abdominal CT. Duodenography in patients with CT findings suggestive of duodenal injury was found to be of minimal utility. The most important sign of duodenal perforation was retroperitoneal extraluminal air seen on CT. Intraoperative evaluation of the duodenum requires adequate mobilization of the duodenum by means of a Kocher maneuver. The hepatic flexure of the colon is also mobilized to provide adequate exposure of the anterior wall of the second portion, and examination of the third and fourth portions of the duodenum should also be done. The presence of retroperitoneal hematomas around the duodenum should raise suspicion of an associated pancreatic injury. Appropriate repair of duodenal injuries depends on the severity of the injury ( Table 20-8 ) and elapsed time from injury to treatment. Approximately 80% to 85% of duodenal wounds can be repaired primarily. The remaining 15% to 20% are severe injuries that require more complex procedures.

Table 20-8 -- Duodenum Injury Scale TYPE OF GRADE[*] INJURY DESCRIPTION OF INJURY I II III IV V Hematoma Laceration Hematoma Laceration Laceration Laceration Laceration Involving a single portion of the duodenum Partial thickness, no perforation Involving more than one portion Disruption <50% of the circumference Disruption 50%-75% of the circumference of D2 Disruption 50%-100% of the circumference of D1, D3, D4 Disruption >75% of the circumference of D2 and involving the ampulla or distal common bile duct Massive disruption of the duodenopancreatic complex

Vascular Devascularization of the duodenum From Moore EE, Cogbill TH, Malangoni MA, et al: Organ injury scaling: II. Pancreas, duodenum, small bowel, colon, and rectum. J Trauma 30:1427-1429, 1990, with permission. D1, first portion of the duodenum; D2, second portion of the duodenum; D3, third portion of the duodenum; D4, fourth portion of the duodenum.
* Advance one grade for multiple injuries up to grade III.

For most minor injuries (grades I and II) diagnosed within 6 hours of injury, simple primary repair is suitable. After 6 hours the risk of leakage increases, and any form of duodenal decompression (transpyloric nasogastric tube, tube jejunostomy, or tube duodenostomy) is advisable. Grade III injuries involving major disruption of the duodenal circumference are best treated by primary repair, pyloric exclusion, and drainage or, alternatively, by Roux-en-Y duodenojejunostomy. Grade IV injuries (involving the ampulla or distal common bile duct) are difficult to repair. In this situation, primary repair of the duodenum, repair of the common bile duct, and placement of a T-tube with a long transpapillary limb or a choledochoenteric anastomosis may be attempted when possible. If repair of the common bile duct is impossible, ligation and a second intervention for a biliary enterostomy can be done. Pancreaticoduodenectomy, though rarely needed, is reserved for grade V injuries, including massive disruption of the duodenum and pancreatic head or massive devascularization of the duodenum.

Duodenal hematomas are expected to resolve in 10 to 15 days, and management consists of nasogastric suction until peristalsis resumes and the slow introduction of solid food. Exploration is indicated in the event of persistent duodenal obstruction. The incidence of complications after duodenal injuries is high and ranges from 30% to more than 100%.[35] The most significant complication after duodenal injury is the development of a duodenal fistula, which occurs in 5% to 15% of patients. Duodenal fistulas are generally managed nonoperatively with nasogastric suction, IV nutritional support, and aggressive stoma care. Usually, closure will occur within 6 to 8 weeks. Abscesses develop in 10% to 20% of patients and may or may not be associated with a duodenal fistula. Abscesses are initially managed by percutaneous drainage. Surgical drainage is indicated if multiple abscesses are present or when located between small bowel loops.
Pancreatic Injuries

Pancreatic injury is rare and accounts for approximately 10% to 12% of all abdominal injuries. The great majority of such injuries are caused by penetrating mechanisms and are often associated with significant injuries involv-ing other intra-abdominal organs. Blunt trauma to the abdomen caused by a direct blow or seat belt injury may compress the pancreas over the vertebral column and result in pancreatic disruption. Major abdominal vascular injuries are present in more than 75% of cases of penetrating pancreatic trauma, and injuries to solid organs and hollow viscera are common after blunt trauma. Mortality rates range from 10% to 25%, mostly secondary to associated intra-abdominal injuries. Approximately 50% of the overall mortality after a pancreatic injury is caused by associated major abdominal vascular injuries. Sepsis and multiple organ failure account for most of the late deaths. The incidence of pancreatic-related mortality ranges from 2% to 5% in large urban trauma series. Diagnosis of a pancreatic injury is made by having a high index of suspicion based on the history, mechanism of injury, and associated clinical findings. However, because of its retroperitoneal location, the pancreas is a well-protected organ, and signs and symptoms may appear late, thus delaying diagnosis. Increased levels of serum and urinary amylase after a blunt injury are not diagnostic, but a persistent elevation suggests pancreatic injury. Contrast-enhanced duodenography may reveal widening of the C-loop. DPL is not sensitive enough for the diagnosis of retroperitoneal injuries, but this test may be positive because of the high frequency of associated injures and should prompt abdominal exploration. Abdominal CT is of potential value, but its role is still unclear. The diagnosis of a pancreatic injury with the use of newergeneration CT scanners has improved significantly; however, some injuries may be identified only during follow-up scans obtained because of changes in clinical status ( Fig. 20-18 ).

Figure 20-18 Computed tomography scan showing pancreatic transection.

Isolated pancreatic injuries are rare. The diagnosis is difficult to make, and patients may complain of vague abdominal pain radiating to the back several hours after the incident. Frequently, mild abdominal tenderness develops, with peritoneal signs eventually developing in some patients. A delay in diagnosis correlates with an increased incidence of severe complications. Patients are generally operated on because of intraperitoneal blood loss or peritonitis, and the diagnosis of a pancreatic injury is generally an incidental finding. However, patients with questionable CT scan findings, persistent abdominal pain, or elevated serum amylase may benefit from a repeat CT scan. At laparotomy, careful examination of the pancreas should be carried out. All retroperitoneal hematomas surrounding the pancreas should be explored. The presence of a pancreatic duct injury appears to be a key factor in postoperative morbidity.[36] There is still controversy regarding the best method to evaluate the main pancreatic duct during laparotomy. Several authors are in favor of aggressively pursuing the identification of ductal injury by performing intraoperative pancreatography. Others favor a more conservative approach

and base their management on the location and surgical identification of ductal injuries. In fact, a more conservative approach to the diagnosis and management of pancreatic injuries has been advocated recently. Pancreatic injuries are divided into proximal or distal according to the location on the right or left of the superior mesenteric vessels. Classification of pancreatic injuries according to injury severity is presented in Table 20-9 .

Table 20-9 -- Pancreas Injury Scale TYPE OF GRADE[*] INJURY DESCRIPTION OF INJURY I II III IV Hematoma Laceration Hematoma Laceration Laceration Laceration Minor contusion without duct injury Superficial laceration without duct injury Major contusion without duct injury or tissue loss Major laceration without duct injury or tissue loss Distal transection or parenchymal injury with duct injury Proximal transection or parenchymal injury involving the ampulla[]

V Laceration Massive disruption of the pancreatic head From Moore EE, Cogbill TH, Malangoni MA, et al: Organ injury scaling: II. Pancreas, duodenum, small bowel, colon, and rectum. J Trauma 30:1427-1429, 1990, with permission. www.lww.com
* Advance one grade for multiple injuries up to grade III. The proximal portion of the pancreas is to the patient's right of the superior mesenteric vein.

Most distal pancreatic injuries with suspected ductal injury are treated by distal resection with or without splenectomy. If there is any evidence that the pancreas has been contused, it should be drained independent of the location of the contusion. Complete transection of the midportion of the pancreas can theoretically be treated by pancreaticojejunostomy, but there are no data to support this approach over distal pancreatectomy. Penetrating wounds to the right of the superior mesenteric vein should be treated by dbridement and direct suture ligation of areas of bleeding. Extensive injuries to the pancreatic head or to the right of the superior mesenteric vessels are usually associated with a greater than 40% probability of a temporary pancreatic fistula. In these circumstances, drainage seems to be the best treatment option, particularly when compared with proximal pancreatectomy or a Whipple

procedure, which is reserved for severe combined pancreaticoduodenal injuries. Severe trauma to the duodenum and head of the pancreas may be treated by dbridement of the pancreas, closure of the duodenal wound, and pyloric exclusion. Wide drainage is also mandatory in this situation. The most frequent complications after pancreatic trauma are pancreatic fistula and peripancreatic abscess. These complications occur in approximately 35% to 40% of patients sustaining pancreatic injuries. Pancreatic fistulas, if well drained, will close spontaneously in the majority of patients. Somatostatin has been used to expedite the healing of pancreatic fistulas, but the results are controversial. Peripancreatic abscesses are treated by surgical dbridement and drainage. The incidence of pancreatitis after a pancreatic injury is 8% to 18%. Pancreatic pseudocysts are infrequent.
Small Intestinal Injuries

The small bowel is the most frequently injured organ after penetrating injuries. After blunt trauma, the incidence of small bowel injuries ranges from 5% to 20% in patients who require surgical exploration. The postulated mechanisms involved in blunt intestinal injury include the following: 1. 2. 3. Crushing injury of the bowel between the vertebral bodies and the blunt object, such as a steering wheel or handlebars Deceleration shearing of the small bowel at fixed points, such as the ligament of Treitz and the ileocecal valve and around the mesenteric artery Closed-loop rupture caused by a sudden increase in intra-abdominal pressure

The presence of a seat belt sign should raise suspicion for enteric and mesenteric injuries. The majority of patients with blunt intestinal trauma will have signs of peritoneal irritation; however, small lacerations may be accompanied by mild abdominal pain without peritoneal signs. If peritoneal signs or hemodynamic instability are present, the patient should be taken to the operating room for surgical exploration, and the diagnosis of intestinal rupture will be made intraoperatively. Hollow viscus injuries are often characterized by a delay in diagnosis after blunt abdominal trauma. Delay in the diagnosis and management of blunt hollow viscus injuries is associated with increased morbidity and mortality, as shown by recent studies. Several tests may help in the diagnosis of blunt intestinal trauma in patients without a clear indication for surgical exploration. Plain films of the abdomen may reveal free air; however, this finding is uncommon. DPL is not a reliable test to identify small bowel injuries, particularly small injuries with minimal leakage. CT with IV and oral contrast also carries a significant falsenegative rate, and suggestive findings include free fluid in the abdomen without solid organ injury, free air, and thickening of the small bowel wall or mesentery. Negative abdominal CT results are inadequate to safely rule out a perforated small bowel injury.[37] Occasionally, a large tear in the mesentery occurs without bowel involvement. In these instances bowel necrosis and subsequent perforation occur hours or even days after the initial injury, and the patients may have frank peritoneal signs, acidosis, and sepsis.

At laparotomy, careful examination of the entire small bowel should be performed. Bleeding should initially be controlled and clamps or sutures applied to prevent further leakage of intestinal contents into the peritoneal cavity. Penetrating injuries caused by firearms should be dbrided, and small tears are usually closed primarily. If two adjacent holes are found, they can be connected across the bridge of normal bowel and closed transversely to avoid narrowing of the intestinal lumen. Extensive lacerations, devascularized segments, or multiple lacerations in a short segment of bowel are better treated by resection and reanastomosis. All mesenteric hematomas should be explored because they can hide small bowel injuries. During the initial postoperative period, patients are usually maintained with a nasogastric tube for decompression until peristalsis resumes and feeding is started. Postoperative complications include intra-abdominal abscess and sepsis, anastomotic leakage, wound infection, enteric fistulas, and intestinal obstruction. Enteric leakage caused by suture breakdown is rare and manifested by fever, leukocytosis, leak through the surgical wounds, or peritonitis. Short-bowel syndrome is a devastating complication after extensive resection of the small bowel. It is characterized by persistent diarrhea, loss of protein and fat in stool, and weight loss. Ileal resections are less tolerated than jejunal resections because the ileum is the site for absorption of bile salts and vitamin B12 and the jejunum has greater adaptive capacity. The presence of the ileal cecal valve is also of paramount importance because it slows intestinal transit time, thereby providing prolonged exposure of the intestinal mucosa to nutrients. Treatment of short-bowel syndrome includes adequate fluid intake, parenteral hyperalimentation, vitamin B12 replacement, cholestyramine to reduce diarrhea, H2 blockers to reduce gastric secretion, and oral narcotics to reduce intestinal motility.
Injuries to the Colon

Colon injuries are generally the result of penetrating trauma. The colon is the second most frequently injured organ after gunshot wounds and the third after stab wounds to the abdomen. Colon injuries are relatively infrequent after blunt trauma, which accounts for only 5% of such injuries. Recent studies have shown that morbidity rates after colonic injury vary from 20% to 35% and mortality rates range from 3% to 15%. The incidence of infectious complications after a colonic injury is related to inadequate treatment or delay in diagnosis, and several reports have confirmed that repair of a colonic injury within 2 hours dramatically reduces the incidence of infectious complications. Physical examination is particularly useful to establish that laparotomy is necessary after a stab wound to the abdomen if peritoneal signs are present; however, a negative physical examination does not rule out the presence of a colonic injury, particularly in patients with stab wounds to the back and flanks. An objective evaluation of the abdomen is warranted after stab wounds and may include DPL or a triple-contrast (oral, IV, and rectal) CT scan. Gunshot wounds to the abdomen usually indicate the necessity for laparotomy, and with few exceptions, no further workup is necessary and the colonic injury will be diagnosed during abdominal exploration.

Laboratory studies are not generally helpful, and plain abdominal films may eventually show pneumoperitoneum. Rectal examination may show the presence of blood, which is strong evidence of colon or rectal injury. Patients undergoing abdominal exploration should receive preoperative antibiotics, an important adjunct for decreasing infectious complications. Operative management of colonic injury is still controversial. During World War I, primary repair of colonic injuries was the treatment of choice; however, mortality rates reached 60%. During World War II, surgeons concerned about high rates of postoperative infection considered diversion of fecal contents by means of a colostomy and delayed colostomy closure to be safer than primary repair. In fact, the mortality rates reported during World War II for colonic injuries were approximately 35%. These results influenced the way that colonic injuries were treated until recent years. Recently, this concept has been challenged because colonic injuries in civilian practice are caused by low-velocity missiles and stab wounds, mechanisms different from those in military practice. This led to a resurgence of primary repair as an adequate alternative to colostomy for the treatment of most (but not all) colonic injuries. Primary repair can be selected when known associated complicating factors have been excluded. General criteria for primary repair include early diagnosis (within 4-6 hours), absence of prolonged shock or hypotension, absence of gross contamination of the peritoneal cavity, absence of associated colonic vascular injury, less than 6 units of blood transfused, and no requirement for the use of mesh to permanently close the abdominal wall. Increased complication rates after primary repair are due to prolonged hypotension, massive intraperitoneal hemorrhage, more than two associated organs injured, significant fecal spillage, or delayed diagnosis. Most patients with low-risk penetrating colonic injuries can be treated by primary closure or resection and primary anastomosis by following these guidelines. High-risk colon injuries or those associated with severe injuries will benefit from resection and colostomy. Exteriorization of the colonic repair has been performed infrequently because of extremely high rates of failure, repair breakdown, and infectious complications. Some surgeons use different approaches to treat injuries on the right side than on the left side of the colon; however, no prospective randomized data are available to compare primary repair performed on right-sided colonic injuries with end-colostomy for left-sided injuries. A comparison of results between primary repair and colostomy for colonic injuries should include complications that occur during or after colostomy takedown. In analyzing complications and deaths after colostomy takedown, some studies have reported an overall 10% to 50% incidence of complications. Penetrating colon injuries requiring resection (colostomy versus primary anastomosis) were recently evaluated in a prospective multicenter study. The type of colon management was not found by multivariate analysis to be a risk factor for abdominal complications. The authors concluded that once resection is necessary, the surgical method of colon management does not affect the incidence of abdominal complications, irrespective of associated risk factors, and that primary anastomosis should be considered in all patients.[4] The type of anastomosis used in the colon (hand sewn versus stapled) has been also the subject of controversy. A multicenter prospective study comparing stapled with hand-sewn anastomoses

after penetrating colon trauma concluded that the method of colonic anastomosis does not affect the incidence of abdominal complications. In summary, stab and low-velocity wounds to the colon with minimal contamination and hemodynamic stability can be managed by primary repair. Postoperative complications include abscess formation, anastomotic leak, peristomal hernia, and the morbidity and mortality associated with closure of the colostomy.
Rectum

Rectal injuries are uncommon. Most rectal injuries result from gunshot wounds; however, other causes, such as a foreign body, impalement, pelvic fractures, and iatrogenic (after proctosigmoidoscopy), should be considered. Transpelvic gunshot wounds, as well as any penetrating injury to the lower part of the abdomen and buttocks, should raise suspicion for a rectal injury even if the physical examination is unremarkable. Rectal injuries can be intraperitoneal or extraperitoneal. The rectal examination may reveal blood, or an injury may be palpable. Workup of rectal injuries includes anoscopy and rigid proctosigmoidoscopy. Primary closure of extraperitoneal rectal injuries, particularly those located in the inferior third of the rectum, should be attempted, although such closure is not always possible. A diverting colostomy, washout of the distal rectal stump, and wide presacral drainage are mandatory. Rectal stump irrigation in this setting decreases the incidence of pelvic abscess, rectal fistulas, and sepsis. Intraperitoneal rectal injuries are usually managed by primary closure and a diverting colostomy. Primary abdominal perineal resection is indicated for extensive rectal injuries. Complications after rectal injuries include sepsis, pelvic abscesses, urinary or rectal fistulas, rectal incontinence and stricture, loss of sexual function, and urinary incontinence. A recent study evaluated whether the use of a clinical pathway based on precise anatomic characterization of rectal injuries improves outcomes. In this study, intraperitoneal rectal injuries were treated by primary repair. Injuries to the proximal two thirds and accessible distal third of the extraperitoneal rectum were treated by repair and selective fecal diversion. Inaccessible distal rectal injuries were treated by diversion and presacral drainage. Use of a clinical pathway decreased the overall infectious complication rate from 31% to 13%. No retrorectal abscesses were found in the groups treated by using the clinical pathway. This study reinforces the importance of location and accessibility in the decision to perform diversion and presacral drainage in extraperitoneal rectal injuries.[38]
Liver Injuries

Because of its size and location in the abdominal cavity, the liver is frequently injured in both blunt and penetrating trauma. Despite progress in the management of trauma patients in the last 2 decades, mortality after hepatic trauma has remained stable.[39] Spontaneous hemostasis is observed in more than 50% of small hepatic lacerations at the time of laparotomy. In fact, most liver injuries require only documentation and no drainage. Although most liver injuries can be properly managed with simple procedures, control of profuse bleeding from deep hepatic

lacerations remains a formidable challenge for trauma surgeons. The overall mortality rate ranges from 8% to 10%, and the overall morbidity rate varies from 18% to 30%, depending on the number of associated injuries and the severity of the injury. The classification of liver injuries is shown in Table 20-10 . In less severe hepatic injuries (grades I-III), mortality is related to associated injuries, which are more frequently seen after blunt trauma, although in high-grade liver injuries mortality is related to the injury itself, regardless of the mechanism. The mortality rate associated with isolated liver trauma is 3%; it increases to 24% in the presence of three associated injuries.

Table 20-10 -- Liver Injury Scale 9 (1994 Revision) TYPE OF GRADE[*] INJURY DESCRIPTION OF INJURY I II Hematoma Laceration Hematoma Laceration III Hematoma Subcapsular, <10% surface area Capsular tear, <1 cm in parenchymal depth Subcapsular, 10%-50% surface area; intraparenchymal, <10 cm in diameter Capsular tear, 1-3 cm in parenchymal depth; <10 cm in length Subcapsular, >50% surface area of ruptured subcapsular or parenchymal hematoma; intraparenchymal hematoma, >10 cm or expanding 3 cm in parenchymal depth Parenchymal disruption involving 25%-75% of the hepatic lobe or 1-3 Couinaud segments Parenchymal disruption involving >75% of the hepatic lobe or >3 Couinaud segments within a single lobe Juxtahepatic venous injuries, i.e., retrohepatic vena cava/central major hepatic veins

Laceration IV V Laceration Laceration Vascular

VI Vascular Hepatic avulsion From Moore EE, Cogbill TH, Jurkovich GJ, et al: Organ injury scaling: Spleen and liver (1994 revision). J Trauma 38:323-324, 1995, with permission. www.lww.com
* Advance one grade for multiple injuries up to grade III.

Some small nondeep bleeding lacerations are easily controlled with simple suture or the use of hemostatic agents. More severe liver injuries require more complex procedures, including deep mattress sutures, packing, dbridement, resection, mesh hepatorrhaphy, and other measures. The resurgence of packing and the emergence of damage control, as an alternative for the treatment of severe hepatic injuries in patients with shock, metabolic derangement, and coagulopathy, have

been incorporated in the armamentarium of trauma surgeons in recent years. A 34% survival rate was reported when packing was used as an adjunct to other measures to control bleeding. Injuries vary from capsular tears and nonbleeding lacerations to large fractures and lobar destruction with extensive parenchymal disruption and hepatic artery and venous injuries. The type of injury dictates the surgical management. The principles of surgical management of liver injury are the same regardless of the severity of injury. They involve control of bleeding, removal of devitalized tissue, and establishment of adequate drainage. Simple lacerations that are not bleeding at the time of surgery do not require drainage unless they are deep into the parenchyma with the possibility of a postoperative biliary fistula. Subcapsular hematomas can simply be observed or surgically evacuated if there is no associated parenchymal injury. Lacerations that continue to bleed despite attempts at local control require a more extensive approach, usually opening the liver wound and directly approaching the bleeding vessels, a procedure known as tractotomy. Bleeding vessels and biliary radicles should be individually ligated. In the event that bleeding continues despite directly ligating small vessels, a vascular clamp or vessel loops can be placed around the porta hepatis (Pringle's maneuver) ( Fig. 20-19 ). If the bleeding stops after clamping the portal triad, it can be assumed to be from the portal veins or hepatic artery branches. If the bleeding continues despite clamping the portal triad, an injury to the hepatic veins or the retrohepatic vena cava is suspected. The portal triad can also be intermittently clamped to allow visualization during the placement of sutures as the parenchymal vessels are ligated. If a Pringle maneuver is applied, caution regarding the duration of inflow occlusion is necessary. Hypothermic patients do not tolerate liver ischemia for prolonged periods, and significant damage to the liver parenchyma may occur as a result of ischemia. The exact length of warm ischemia time tolerated by the human liver is not known; however, some authors have reported occlusion of inflow for up to 1 hour with the use of adjuvant steroid therapy without major consequences.

Figure 20-19 Diagram showing the Pringle maneuver. DU, duodenum; GA, gastric antrum; GB, gallbladder; HP, hepatic pedicle; LI, liver; PA, pancreas. (From Aunf B: Critical maneuvers in trauma surgery. In Editora Pedagogica e Universitaria [E. P. U.] LTDA. New York, Springer-Verlag, 1982.)

Packing the liver wound is performed when the techniques just described fail to control hemorrhage. The results of temporary packing must be analyzed in light of its relation to timing. Perihepatic packing was once condemned because of the high incidence of intra-abdominal abscesses. Temporary packing has recently been used, particularly in patients with hypothermia, coagulopathy, and severe acidosis with severe injuries in other intra-abdominal organs.

Usually, these patients are taken to the intensive care unit for rewarming and resuscitation. Reexploration for removal of the packing is performed within 48 to 72 hours after the initial operation. After hemostasis is achieved and the packs are removed, copious irrigation of the abdominal cavity is performed and closed suction drains are placed. Arteriography is a useful adjunct to locate the arterial bleeding, and embolization may be of benefit before re-exploration for removal of the packing. The incidence of intra-abdominal abscess in survivors of liver packing is generally less than 15%. Despite the use of any method to obtain hemostasis, all necrotic tissue should be dbrided before closure. If bleeding in the raw surface of the liver after resectional dbridement is not significant, an omental flap can be used to cover or fill the defect in the liver parenchyma. Deep liver lacerations should not be simply closed because of the risk for abscess formation and hemobilia. As an alternative approach for deep liver lacerations, some investigators propose extending the liver laceration to expose and directly ligate the bleeding vessel. This is achieved by performing a finger fracture hepatotomy along nonanatomic planes. This technique was used in patients with grade III to grade V liver injuries, with the remarkably low mortality rate of 10.7%. The advantage of this technique is that direct ligation of the bleeding vessels and biliary radicles is achieved; the disadvantage is that the defect in the liver parenchyma is usually bigger than the initial injury, and the technique should be performed only by experienced surgeons. Formal hepatic resection is unusual after liver injuries and has been largely abandoned in the past decade because of high mortality and morbidity rates after this procedure and because other more conservative approaches have proved to be as effective in controlling hemorrhage, with significantly lower complication rates and mortality. In a 5-year multi-institutional review of 1335 liver injuries, resectional dbridement was performed in 36 patients (2.7%), hepatotomy and vessel ligation in 50 patients (3.7%), and segmentectomy in 18 patients (1.3%). Formal hepatic lobectomy was performed in only 12 patients (0.9%). Another technique described recently encompasses the use of absorbable mesh, with each lobe of the liver wrapped individually and the mesh attached to the falciform ligament. This technique is useful for multiple superficial lacerations of the liver with active bleeding; however, it is not effective when major vascular injuries are present. The reported mortality rate in hepatic trauma patients managed by mesh wrapping is 25% to 37.5%. Major hepatic injuries, including retrohepatic and juxtahepatic venous injuries, are discussed in the later section on major abdominal vascular injuries. Ligation of the hepatic artery is also an alternative for continued bleeding; however, with the use of modern cauterization devices (electric or argon bean coagulators), topical hemostatic agents, and fibrin glue, this is seldom required. It should be reserved for the occasional stab wound or gunshot wound involving one lobe in which exposure of the wound will require extensive incision of the liver. The proper hepatic artery must never be ligated. Injudicious hepatic artery

ligation may result in liver infarction, particularly if associated with portal vein injury.[40] Packing the liver is a reasonable alternative to hepatic artery ligation.
Nonoperative Treatment

Blunt hepatic injuries in hemodynamically stable patients without other indications for exploration are best served by a conservative, nonoperative approach. [49] [50] [51] These stable patients without peritoneal signs are better evaluated by ultrasound, and if abnormalities are found, a CT scan with contrast should be obtained ( Fig. 20-20 ). In the absence of contrast extravasation during the arterial phase of the CT scan, most injuries can potentially be treated nonoperatively. The classic criteria for nonoperative treatment of liver injuries include hemodynamic stability, normal mental status, absence of a clear indication for laparotomy such as peritoneal signs, low-grade liver injuries (grade I-III), and transfusion requirements of less than 2 units of blood. Recently, these criteria have been challenged and a broader indication for nonoperative management has been used. It has been demonstrated that most of these patients are monitored by serial hematocrit and vital signs rather than by serial abdominal examinations, which is the reason why intact mental status is not the sine qua non for nonoperative management. Furthermore, if the hematocrit drops, most patients will undergo a repeat CT scan to evaluate and quantify the hemoperitoneum. The overall reported success of nonoperative management of blunt hepatic injuries is greater than 90% in most series. Breaking it down by injury grade, the success rate of nonoperative treatment of injury grades I to III approaches 95%, whereas for injury grades IV and V the success rate decreases to 75% to 80%. With the use of angiography and superselective embolization in patients with persistent bleeding, the success rate may in fact be higher.

Figure 20-20 Computed tomography scan showing a grade IV liver laceration.

Angiographic embolization has been added to the protocol for nonoperative management of liver injuries in some institutions in an attempt to decrease the necessity for blood transfusions and the number of operations. [52] [53] [54] Patients are admitted to the intensive care unit for monitoring of vital signs and hematocrit. Usually, after 48 hours patients are transferred to an intermediate care unit, where they are started on an oral diet; however, they remain on bed rest until postinjury day 5. A repeat CT scan before discharge does not seem to be neces-sary. Normal physical activity resumes 3 months after injury. A recent multicenter study attempted to determine early risk factors for hepatic-related morbidity after nonoperative management of severe (grades III-V) blunt hepatic injuries. The authors reported complication rates of 5%, 22%, and 52% for grade III, IV, and V liver injuries, respectively. Using multivariate analysis, they demonstrated that liver injury grade and 24-hour transfusion requirements predicted complications. Currently, no single selection criterion can predict which patients will fail nonoperative management. Croce and colleagues prospectively analyzed 112 patients treated nonoperatively over a 22month period. They reported a failure rate of 11% (12 patients), with five failures being liver related. No relationship between injury grade and increased failure rate was observed. The authors concluded that nonoperative management is safe regardless of injury severity in hemodynamically stable patients; it carries a lower incidence of abdominal septic complications and leads to decreased transfusion requirements. They also compared 70 patients with grade III to V liver injuries treated nonoperatively with 50 patients who underwent surgical intervention. Blood transfusion at 48 hours consisted of 2.2 and 5.8 units, and mortality was 7% and 4% for nonoperative and operative controls, respectively. Although the transfusion requirement was slightly lower in the nonoperative group, no difference in mortality was demonstrated.[41] Management of patients with contrast extravasation during the arterial phase of CT is still debatable. Fang and associates proposed a classification system based on the location and character of extravasation and pooling of contrast material from a liver laceration on CT. In type 1, there is contrast extravasation to the peritoneal cavity. All patients in this category required operative intervention.[47] Type 2 consisted of hemoperitoneum and extravasation of contrast material within the hepatic parenchyma. The authors recommend that patients in this category undergo angiography with embolization, although some will require operative intervention. Type 3 was characterized by no hemoperitoneum and extravasation of contrast material within the hepatic parenchyma. Angiography is required in this subgroup of patients, and the results are usually good.

Ciraulo and coworkers analyzed a group of 11 patients requiring continuous fluid resuscitation, with 7 requiring embolization. All embolization attempts were successful. The authors concluded that hepatic artery embolization is a viable alternative in the management of patients with severe liver injuries that require continuous fluid resuscitation, thereby bridging the therapeutic options of operative and nonoperative intervention.[44] The most important concern of nonoperative management is the potential for missed injuries, particularly hollow viscus perforations. Delay in diagnosing a hollow viscus injury is associated with significant morbidity and increased mortality.[48]
Porta Hepatis

A recent multicenter retrospective study that included data from eight trauma centers reported an incidence of portal triad injuries of only 0.07%.[40] Penetrating trauma is the most frequent mechanism associated with porta hepatis injuries, although 30% of porta hepatis injuries in the aforementioned study followed blunt trauma. Isolated injuries to the porta hepatis are uncommon. Because of the proximity of other organs, porta hepatis injuries are usually associated with hepatic, duodenal, gastric, colonic, and other major vascular injuries. The overall mortality rate is 50%, but it increases to 80% in patients with associated injuries. Management is difficult because of life-threatening hemorrhage and associated organ injury. If the patient survives the operation, complications such as biliary fistula, portal vein thrombosis, and hepatic ischemia may contribute to morbidity. Management of portal vein and hepatic artery injuries is discussed later under major abdominal vascular injuries. Management of common bile duct injury is challenging. Primary repair and placement of a Ttube should be attempted for partial or minor injuries involving less than 50% of the duct's circumference. Major injuries or complete transection of the common bile duct are best managed by choledochoenteric anastomosis. This procedure significantly reduces the incidence of late postoperative complications, in particular, the development of strictures. A closed suction drain should always be place in the vicinity of the repair to allow adequate drainage of an eventual biliary fistula. Missed extrahepatic bile duct injuries occurred in nine patients in a multicenter review, with a 75% complication rate in those who survived.[40] Gallbladder injury is also an uncommon injury after both blunt and penetrating trauma. Cholecystectomy is the procedure of choice.
Postoperative Complications

Significant complications after liver injury include pulmonary complications, postoperative bleeding, coagulopathy, biliary fistulas, hemobilia, and subdiaphragmatic and intraparenchymal abscess formation. Postoperative bleeding occurs in less than 10% of patients sustaining liver injuries. It may occur as a result of inappropriate hemostasis, postoperative coagulopathy, or both. If the patient is not hypothermic, coagulopathic, or acidotic, re-exploration should be undertaken. Bleeding vessels

should be directly visualized and ligated, even if more extensive disruption of the hepatic parenchyma is necessary for adequate exposure. If diffuse oozing is found, packing and a planned re-exploration are performed. Intra-abdominal abscesses have accounted for late deaths after hepatic trauma. A 7.2% incidence of perihepatic abscesses was reported in a prospective analysis of 482 injuries. The population at increased risk included patients with prolonged shock, extensive parenchymal disruption, associated hollow viscus injuries, hepatic ischemia from ligation of major vessels, and open drainage. Several studies have criticized the use of drains after liver injury because of the risk for intraperitoneal infection. It seems that injury grades I and II do not require drain placement. However, in severe injuries, drainage, though controversial, is frequently used.[39] The presence of nonviable hepatic tissue is also an important cause of postoperative abscess formation and points to the fact that adequate dbridement of all devitalized tissue is an important step before closing the abdomen. CT is the method of choice to diagnose intraabdominal abscesses. Percutaneous drainage is the treatment of choice for nonloculated abscesses. However, patients with peritoneal signs, persistence of fever despite percutaneous drainage, or multiple abscesses should be surgically re-explored through the same incision used for the initial operation. The incidence of biliary fistulas after hepatic trauma varies from 7% to 10%.[9] In a recent multicenter review of hepatic injuries, an 8% incidence of biliary fistula was noted in 210 patients with grade III, IV, and V hepatic injuries. The group at risk includes patients with severe hepatic injuries (grade III and higher) and those requiring hepatic resection or extensive dbridement. Usually, biliary fistulas close spontaneously after a period of 2 to 4 weeks of closed drainage. Hemobilia is a rare complication that usually occurs after blunt intrahepatic hematomas have formed and is manifested as bleeding into the bile ducts and subsequently into the small bowel. Patients generally complain of jaundice, right upper quadrant pain, malaise, and melena. Hemobilia can be diagnosed by upper gastrointestinal endoscopy and treated by angiographic embolization. Surgery is rarely required.
Splenic Injuries

The spleen is the intra-abdominal organ most frequently injured in blunt trauma. Suspicion of a splenic injury should be raised in any patient with blunt abdominal trauma. History of a blow, fall, or sports-related injury to the left side of the chest, flank, or left upper part of the abdomen is usually associated with splenic injury. The diagnosis is confirmed by abdominal CT in a hemodynamically stable patient or during exploratory laparotomy in an unstable patient with positive DPL findings. For several decades, splenectomy was considered the only acceptable surgical option for splenic injuries. With the significant experience in nonoperative management of splenic injuries in the pediatric population and with the recognition of overwhelming postsplenectomy syndrome as a serious postoperative threat, other options have emerged in the past decades. Initially, splenic

repair and, more recently, nonoperative management in the adult population have been considered adequate options in selected patients. Pediatric surgeons initiated this more conservative approach, and their experience was later applied to the adult trauma population. In 1952, a postsplenectomy syndrome of severe, sometimes fatal meningitis and sepsis in four of five children splenectomized before the age of 6 months for congenital hemolytic anemia was reported. The term overwhelming postsplenectomy infection (OPSI) was introduced in 1969. The true incidence of overwhelming postsplenectomy sepsis is not well defined, although a commonly used estimation of the incidence of OPSI is 0.6% in children and 0.3% in adults, which may be a low estimate. The courses of 688 patients (388 children, 300 adults) who underwent splenectomy for injury to the spleen were reviewed. Among these were 10 patients with sepsis (incidence of 1.45%), 4 of whom died, for a mortality rate of 0.58%. When combined with four deaths from sepsis after splenectomy for trauma in another series of 342 children, the incidence of mortality from sepsis is 0.78%, or 78 times the expected rate in the general population. The risk for postsplenectomy septicemia, pneumonia, and meningitis was estimated to be 8.3% in trauma patients, or 166 times the 0.05% rate expected in the general population. The longest follow-up of splenectomy patients included 740 World War II veterans who underwent splenectomy between 1939 and 1945. Six patients in this group (0.8%) died of pneumonia, whereas none of the 740 matched control patients died. This syndrome is unlike fulminating bacteremia and septicemia in patients with normal splenic function. OPSI syndrome is distinct from septicemia in patients with normal immune function and is characterized by a sudden onset of symptoms and a rapid and fulminating course that often lasts only 12 to 18 hours. Patients usually complain of fever, nausea, vomiting, headache, and altered mental status. It is mainly caused by pneumococci, but other bacteria such as E. coli, Haemophilus influenzae, meningococci, Staphylococcus, and Streptococcus may also be found in decreasing frequency. The disease is complicated by shock, electrolyte imbalance, hypoglycemia, and disseminated intravascular coagulation. The overall mortality rate is as high as 50% to 80%. Because of the severity of the disease process and high mortality rates, the universal use of polyvalent pneumococcal vaccine (Pneumovax 23, Merck; Pnu-Immune, Lederle) and close follow-up after splenectomy for trauma is routine. Patients should receive the vaccine before discharge. The effectiveness of the vaccine in splenectomized patients is unclear. The use of prophylactic antibiotics in asplenic patients is also controversial; however, minor infections in this group should be treated with antibiotics.
Management

Hemodynamically stable patients now undergo ultrasound examination. If the ultrasound findings are positive for free fluid and the patient remains stable, an abdominal CT scan is obtained to identify the source of bleeding, evaluate for contrast extravasation and other intraabdominal injuries that would require an operation, and grade the severity of the splenic injury ( Fig. 20-21 ). The finding of contrast extravasation or contrast blush during the arterial phase of the IV contrast on abdominal CT scanning is indicative of persistent bleeding. Some authors would argue that when present in the spleen, contrast blush should prompt operative intervention, whereas others argue that there is an opportunity for angiographic embolization and continuation of nonoperative management provided that the patient remains hemodynamically

stable. In a recent study the authors found contrast blush in 11% of patients with splenic injuries. No correlation between contrast blush and operative intervention was found. The authors concluded that the presence of contrast blush is not an absolute indication for operative or angiographic intervention.

Figure 20-21 Computed tomography scan showing a splenic laceration.

Some institutions advocate more routine use of angiography, but overall splenic salvage rates are similar to those in institutions following more selective use of angiography. Prospective studies should clarify this issue in the future.[49] More than 70% of all stable patients are currently being treated by means of a nonoperative approach. The classic criteria for nonoperative treatment include hemodynamic stability, negative abdominal examination, absence of contrast extravasation on CT, absence of other clear indications for exploratory laparotomy or associated injuries requiring surgical intervention, absence of associated health conditions that carry an increased risk for bleeding (coagulopathy, hepatic failure, use of anticoagulants, specific coagulation factor deficiency), and injury grade I to III.

Recent series have also indicated that nonoperative management should be performed in patients older than 55 years, those with a large hemoperitoneum, and patients with injury grades IV and V, which in the past have been relative contraindications. It has been shown that inclusion of high-risk patients increases the nonoperative management rate without changing the failure rate significantly. Age older than 55 years and grade IV and V splenic injuries are predictors but do not constitute contraindications to nonoperative management of splenic injuries. Patients are usually admitted to the intensive care unit and kept on bed rest with a nasogastric tube in place. Serial abdominal examinations and hematocrit determinations are performed during the initial 48 to 72 hours. The necessity for blood transfusion is recorded. At 48 to 72 hours, stable patients are transferred to an intermediate care unit, start walking and eating, and are monitored clinically. A repeat CT scan is obtained in the event of a falling hematocrit, hypotension, or persistent ileus. If contrast extravasation is observed or a pseudoaneurysm found, a select group of patients benefit from angiography and selective embolization. A repeat CT scan before discharge does not seem to be necessary. Patients are instructed to avoid intense physical activity and contact sports for 3 months. The success rate of nonoperative treatment is greater than 90%. Several reports have concluded that nonoperative treatment of splenic injuries is safe and effective.
Areas of Controversy

Currently, with the increased use of faster CT scanning, the critical question is whether a select group of patients with negative abdominal ultrasound findings should undergo CT scanning to diagnose intraparenchymal solid organ injury. There is no doubt that a small number of patients will sustain splenic (as well as liver) injuries without hemoperitoneum or with small amounts of blood in the peritoneal cavity undetectable by ultrasonography. These injuries are usually minor and their clinical relevance may be insignificant, so the cost of CT scanning is not justifiable. Another area of controversy in nonoperative management of splenic injuries is related to the use of angiographic embolization. The definition of extravasation and blush has been inconsistent in several studies, thus adding confusion to the topic. However, it is important to state that a technically adequate abdominal CT scan with IV and oral contrast (water could substitute for oral lipid-soluble contrast) is mandatory. Images should be obtained during the arterial phase and also during the excretory phase. In this way, possible areas of contrast extravasation and blush could be confirmed by the presence of radiolucency within the parenchyma when the contrast in the vascular system has been already washed out. Surgical treatment of a splenic injury depends on its severity ( Table 20-11 ), the presence of shock, and associated injuries.

Table 20-11 -- Spleen Injury Scale (1994 Revision)

TYPE OF GRADE[*] INJURY I II Hematoma Laceration Hematoma Laceration III Hematoma

DESCRIPTION OF INJURY Subcapsular, <10% surface area Capsular tear, <1 cm in parenchymal depth Subcapsular, 10%-50% surface area; intraparenchymal, <5 cm in diameter Capsular tear, 1-3 cm in parenchymal depth and not involving a trabecular vessel Subcapsular, >50% surface area or expanding, ruptured subcapsular or parenchymal hematoma; intraparenchymal hematoma, 5 cm or expanding >3 cm in parenchymal depth or involving the trabecular vessels Laceration involving the segmental or hilar vessels and producing major devascularization (>25% of spleen) Completely shattered spleen

Laceration IV V Laceration Laceration

Vascular Hilar vascular injury that devascularizes the spleen From Moore EE, Cogbill TH, Jurkovich GJ, et al: Organ injury scaling: Spleen and liver (1994 revision). J Trauma 38:323-324, 1995, with permission.
* Advance one grade for multiple injuries up to grade III.

The spectrum of injury may vary from a simple laceration or contusion without capsular disruption to total fragmentation of the spleen. During laparotomy, the spleen is evaluated for active bleeding. If active hemorrhage is present, the surgeon must decide to perform either total splenectomy or a splenic salvage procedure. Careful adequate mobilization of the spleen is essential to prevent further injury. Ongoing bleeding from the spleen during mobilization can be controlled by digital compression. Capsular tears of the spleen can be controlled by compression only or by using topical hemostatic agents. Deeper lacerations can be controlled with horizontal absorbable mattress sutures. Major lacerations involving less than 50% of the splenic parenchyma and not extending into the hilum can be treated by segmental or partial splenic resection. Resection is indicated only if the patient is stable and no other major injuries are present. More extensive injuries involving the hilum or the central portion of the spleen may be managed by splenectomy; however, alternative procedures have been described. The technique of implanting thin splenic fragments in an omental pouch (autotransplantation) remains experimental and controversial but may provide significant long-term splenic function. The use of a Dexon or Vicryl mesh to wrap the spleen or use of the argon beam coagulator and fibrin glue has been described in selected patients. The success rate of splenic salvage procedures varies from 40% to 60%. It increases to 90% if nonoperative treatment is included.
Complications

Inadequate hemostasis, massive transfusion, or coagulopathy may cause bleeding after splenectomy or splenic salvage procedures. Other complications include transient thrombocytosis, pancreatitis, and intra-abdominal abscess.
Urinary Tract Injuries

Injuries to the genitourinary tract are often clinically unsuspected and frequently overlooked. Gross hematuria is the most frequent sign associated with urinary tract injuries. An understanding of the mechanism of injury and the forces involved is essential to identify urologic trauma and avoid missed injuries. In blunt trauma, fractures of the lower ribs or spinous processes, abdominal or pelvic crush injuries, direct blows to the back and flanks, or decelerating injuries such as with falls or motor vehicles accidents have been associated with urologic injuries. Upper urologic tract injuries are frequently accompanied by gross or microscopic hematuria. Lower urinary tract injuries are usually manifested as blood in the urethral meatus, a floating or displaced prostate on rectal examination, bladder distention, inability to void, and large perineal hematomas or other perineal injuries. Penetrating injuries to the back or the flank have the potential to cause significant renal injury without obvious clinical manifestations. The workup of patients with suspected urinary tract injuries depends on hemodynamic status. Patients sustaining penetrating abdominal injuries requiring immediate exploratory laparotomy may undergo one-shot IVP. Victims of blunt trauma with blood at the urethral meatus should undergo urethrocystography to rule out the presence of a urethral injury before bladder catheterization ( Fig. 20-22 ). Once urethral injury has been ruled out, cystography is performed by injecting 250 to 300 mL of contrast medium through the Foley catheter to maximally distend the bladder. Films should be obtained after full distention and after emptying the bladder. This postvoid film is important to identify posterior extravasation of contrast that is not seen on AP films obtained when the bladder is maximally distended ( Fig. 20-23 ). Patients with pelvic fractures involving the anterior arch are particularly likely to have an associated bladder injury.

Figure 20-22 Urethrogram showing a complete urethral injury.

Figure 20-23 Cystogram showing a bladder injury

CT is as effective as IVP for evaluating the urinary tract; however, its major advantage over IVP is its ability to evaluate potential intra-abdominal injuries and the retroperitoneum. It is also useful for staging renal injuries and evaluating renal perfusion and function. Absence of kidney perfusion is an indication for renal artery angiography.
Specific Injuries Renal Injuries

The kidney is the most commonly injured part of the urinary tract. Penetrating wounds causing small parenchymal injuries are generally treated by dbridement, primary repair, and drainage. More extensive wounds may require partial or total nephrectomy. An important technical aspect to keep in mind is that in major perinephric hematomas, proximal control of the renal pedicle before opening Gerota's fascia is advisable. Injuries involving the hilum are seldom repaired

primarily, and in most circumstances total nephrectomy is necessary. More than 80% of patients sustaining penetrating renal injuries have other intra-abdominal injuries. Blunt renal injuries are generally divided into minor and major injuries ( Fig. 20-24 ). Minor injuries account for approximately 85% of cases. A classification system for renal injuries proposed by the American Association for the Surgery of Trauma was recently validated. It found that the severity of organ injury correlates with the need for operative intervention.

Figure 20-24 Different types of renal injury. A, Small renal laceration with contained subcapsular hematoma. B, Minor subcapsular and parenchymal hematoma. C, Parenchymal laceration extending through the renal cortex without involvement of the collecting system. D, Multiple parenchymal lacerations; the inferior one extends through the cortex and collecting system. E, Parenchymal laceration extending through the cortex, medulla, and collecting system, along with major subcapsular hematoma and urine extravasation. F, Injury to the renal vessels at the hilum. (From Peterson NE: Genitourinary trauma. In Feliciano DV, Moore EE, Mattox KL [eds]: Trauma, 3rd ed. Norwalk, CT, Appleton & Lange, 1996, p 667, with permission of the McGraw-Hill Companies.)

Renal contusions encompass the vast majority of minor renal trauma and can almost invariably be treated nonoperatively. Major renal trauma includes deep cortical medullary lacerations with extravasation, large perinephric hematomas, and vascular injuries of the renal pedicle. These injuries should be explored because of a high incidence of complications such as bleeding, abscess formation, and hypertension, among others. At laparotomy the major problem is the decision to explore a perinephric hematoma. It is our opinion that all perinephric hematomas caused by penetrating mechanisms that have not previously been evaluated by IVP should be explored. If preoperative IVP shows renal pedicle injury, extensive parenchymal laceration, or urinary extravasation, surgical exploration remains the best option.
Ureteral Injuries

Injury to the ureter is uncommon and occurs mostly after penetrating trauma. The presence of hematuria in ureteral injury is the exception rather than the rule. Ureteral injury is suspected preoperatively by the location of the entrance site of penetrating injuries or, in the case of blunt injury, by the presence of concomitant intra-abdominal or other genitourinary tract injuries. Nondiagnosed ureteral injuries may lead to complications such as fistulas, urinomas, and abscess formation. In the majority of cases, IVP will confirm the diagnosis. In approximately 15% to 20% of ureteral injuries, retrograde ureterography will be required to confirm the diagnosis. In hemodynamically unstable patients the diagnosis of ureteral injury may be made at the time of laparotomy by injecting 5 mL of methylene blue or indigo carmine dye IV. Extravasation of blue-stained urine confirms the presence of a ureteral injury. The principles of ureteral repair are adequate dbridement, tension-free repair, spatulated anastomosis, watertight closure, ureteral stenting, and drainage. Surgical options include ureteroureterostomy for injuries located in the upper and middle thirds of the ureter. The use of a double-J stent is indicated because it seems to decrease the incidence of postoperative fistulas. More distal injuries may require ureteral reimplantation in the bladder. Percutaneous nephrostomy is indicated to divert urinary flow in cases in which primary repair is not feasible, either because of the overall clinical condition of the patient or when a long segment of the ureter has been lost. Other options in the presence of extensive ureteral injuries include transureteroureterostomy or kidney autotransplantation into the iliac fossa.
Bladder Injuries

The majority of bladder injuries occur as a result of blunt trauma, and the association of bladder rupture and pelvic fractures is extremely high. In fact, approximately 70% of patients with bladder rupture have associated pelvic fractures. Hematuria is the most frequent sign and, in the presence of a pelvic fracture, should increase suspicion for bladder injury. Bladder rupture may be extraperitoneal or intraperitoneal. Extraperitoneal rupture usually results from perforation by adjacent bony fragments. Intraperitoneal rupture of the bladder results from injuries located in the dome, which occur when a full bladder sustains a direct blow. The diagnosis is made by

cystography. As stated previously, a postvoid film is necessary to identify lateral or posterior injuries. Intraperitoneal injuries are repaired primarily via a transabdominal approach, including a threelayer closure. Suprapubic cystostomy may be necessary with large wounds. Management of extraperitoneal rupture of the bladder is primarily nonoperative and consists of leaving the Foley catheter in place for 10 to 14 days, provided that the patient has no intraabdominal injuries requiring surgical exploration. Patients with severe pelvic fractures and massive retroperitoneal bleeding are always initially managed nonoperatively. Once the retroperitoneal bleeding is controlled and the patient is stable, delayed repair of the extraperitoneal rupture can be performed, if necessary. Complications of bladder rupture include hemorrhage, urinoma, abscess formation, and sepsis.
Injuries to the Urethra

Disruption of the urethra is a rare injury in women. It is found mostly in men, frequently after either pelvic fractures or straddle injuries. Posterior urethral injuries are present in approximately 10% of pelvic fractures. Anterior urethral injuries are generally associated with straddle injuries and are often isolated lesions. Urethral injuries should be suspected on the basis of the mechanism of injury, associated pelvic fracture, perineal hematoma or perineal injury, blood at the urethral meatus, and displacement of the prostate gland. A retrograde urethrogram is essential for diagnosis. Currently, patients sustaining urethral injuries should be managed initially by bladder decompression via suprapubic cystostomy and delayed urethroplasty. Complications of urethral injuries include stricture, incontinence, and impotence with disruptions of the urethra.
Pelvic Fractures

Pelvic fractures are the prototype of severe trauma and account for less than 5% of all fractures after trauma. The most frequent mechanisms causing pelvic fractures are motor vehicle accidents, motorcycle accidents, falls, and accidents involving pedestrians. Unstable pelvic fractures are accompanied, most of the time, by major retroperitoneal hemorrhage. The incidence of associated injuries is high, particularly intra-abdominal, thoracic, and head injuries. Mortality rates vary depending on the amount of bleeding and the number of associated injuries. The mortality rate directly attributed to pelvic fracture is less than 15% in most series. Pain is frequently present in an awake and alert patient. Urethral injury in males is also frequent and may be manifested as urethral bleeding or inability to void with a distended bladder. Careful examination of the perineum is imperative because the mortality associated with open pelvic fractures is severalfold higher than that for closed fractures. A rectal examination should be performed carefully to identify rectal bleeding, to evaluate the position of the prostate gland, and to assess for mucosal lacerations. If the prostate is misplaced or urethral bleeding is present, a retrograde urethrogram is mandatory before Foley catheter placement. An AP film of the pelvis usually shows the fracture and asymmetry of the pelvis if present. The radiographic evaluation in these circumstances should be complemented by inlet and outlet

views of the pelvis. CT scan of the pelvis provides information on displacement of the sacroiliac joint, acetabular fractures, and sacral fractures; however it should not be performed in a hemodynamically unstable patient. Pelvic fractures can be classified according to the resultant vector force (AP compression, lateral compression, and vertical shear), anatomy of the fracture lines, and pelvic stability. The problem that trauma surgeons face with pelvic fractures is related to retroperitoneal bleeding. Hemorrhage can be arterial, venous, or osseous in origin. Fractures involving the posterior ring are generally believed to have more associated injuries and complications, require more resuscitation fluid, and have a higher mortality rate than is the case with pure anterior fractures. Unstable pelvic fractures are generally associated with increased blood loss. The objectives of the initial management of pelvic fractures are directed to control of hemorrhage. For unstable fractures and particularly those known as the open-book type, this can be accomplished by external fixation in the acute setting. Posterior fractures with involvement of the sacroiliac joint are frequently associated with arterial bleeding, which can be controlled by embolization of the bleeding vessel, usually branches of the internal iliac artery. Indications for angiography are recurrent hypotension after initial resuscitation attributable to the pelvic fracture or transfusion requirements exceeding 4 to 6 units within the first 2 hours after injury. Blood infusion should be started early during resuscitation in hemodynamically unstable patients. As a temporizing measure, the abdominal component of the MAST suit can be inflated during transport or in the resuscitation room to stabilize the pelvis and control bleeding. Because of the high incidence of associated intra-abdominal injuries, supraumbilical DPL should be performed; however there is an approximately 35% rate of false-positive results with DPL in the presence of retroperitoneal hematoma. Stable patients are best evaluated with an abdominal CT scan. If there is a clear indication for abdominal exploration and the retroperitoneum is intact, the hematoma should not be entered because of an increased risk for uncontrolled bleeding. In this circumstance, stabilization of the pelvis and angiographic embolization should be performed. If the retroperitoneum is ruptured and active bleeding is found during exploratory laparotomy, packing the pelvis with temporary closure of the abdomen, followed by external fixation and angiographic evaluation, is appropriate.
Damage Control

The traditional approach to abdominal trauma is not applicable in devastating injuries. Repeated episodes of hypotension and organ hypoperfusion will lead to severe metabolic acidosis, coagulopathy, and hypothermia that persist during the postoperative period despite adequate surgical treatment of multiple injuries. Recently, a new approach has been proposed in these circumstances. Damage control includes an abbreviated laparotomy, temporary packing, and closure of the abdomen in an effort to blunt the physiologic response to prolonged shock and

massive hemorrhage. During the initial operation, bleeding and contamination are controlled with temporary measures. The abdomen is packed and temporarily closed, and reconstruction and repair are delayed. The patient is then transferred to an intensive care unit, where further resuscitation and rewarming are performed, acidosis and coagulopathy are corrected, and full physiologic support is instituted. When the patient is stable and organ function is maintained, usually 48 to 72 hours after the initial operation, the patient is taken back to the operating room for removal of the packing, dbridement of nonviable tissue, and definitive repair.[8]
Abdominal Compartment Syndrome

Abdominal compartment syndrome occurs predominantly in patients in profound shock, in patients requiring large amounts of resuscitation fluids and blood, and in those with major visceral or vascular abdominal injuries. Abdominal compartment syndrome is characterized by a sudden increase in intra-abdominal pressure, increased peak inspiratory pressure, decreased urinary output, hypoxia, hypercapnia, and hypotension secondary to decreased venous return to the heart. The diagnosis is confirmed by measuring bladder pressure, which ultimately represents intra-abdominal pressure. Treatment includes rapid decompression of the elevated intra-abdominal pressure by opening the abdominal wound and performing a temporary closure of the abdominal wall with mesh or a plastic bag (Bogota bag). The physiologic consequences of persistent elevated intra-abdominal pressure are listed in Box 20-6 . Box 20-6 Physiologic Consequences of Increased Intra-abdominal Pressure Decreased Cardiac output Central venous return Visceral blood flow Renal blood flow Glomerular filtration Increased Cardiac rate Pulmonary capillary wedge pressure Peak inspiratory pressure Central venous pressure Intrapleural pressure Systemic vascular resistance
Wound-Related Complications of Damage Control

Temporary closure of the abdomen in damage control procedures is accompanied by high rates of wound complications. Wound complications such as infection, abscess, or fistula formation occurred in 25%. Death after wound closure occurred in 12%, and deaths were directly related to the wound complication in 23%. Morbidity is associated with the timing and method of wound closure (primary closure, skin grafting or absorbable mesh [or both], and prosthetic material [absorbable mesh]) and transfusion volume, but unrelated to severity of the injury. Delayed primary closure during the first week after injury seems to be associated with lower complication rates.[50] Email to Colleague Print Version