Management of Hepatobiliary Trauma

Historical Perspective

As there is significant historical literature describing hepatobiliary injuries, understanding the historical approaches to the care of these patients is informative in appreciating recent evolutions in care. Some of the first descriptions of liver injuries and their treatment can be traced back to Greek and Arabic medical literature near the turn of the first millennium. The first successful treatment of a liver injury is attributed to Hildanus in the early 17th century, who described the care of a young man stabbed in the abdomen with resultant severe hemorrhage. A large piece of liver that eviscerated through the wound was removed and cauterized, and the patient subsequently recovered. Otis painstakingly reviewed Civil War injuries in which he documented 37 individuals who recovered after gunshot wounds (GSWs) to the liver. Twenty-three of these cases were complicated by injury of other viscera in the abdomen, foreshadowing the challenge of caring for patients with multiple intraabdominal injuries. In 1905, Tilton reported on a series of 189 injuries to the liver, emphasizing that hepatic wounds are very frequently associated with concomitant injuries to other visceral organs. In one of his most important observations, he noted that “there are many mild cases of laceration of the liver to go onto recovery without complications and with very few symptoms. The number of these cases is, I think, larger than is generally supposed.” Tilton's review of the literature at that time showed that injury of the liver was associated with a 78.1% mortality if the wound was caused by blunt forces, 39% if caused by GSWs, and 37.5% if caused by stab wounds. In addition, he reviewed all of the New York hospitals with large accident services over a 10-year period and found that there were 25 liver injuries: 12 were caused by blunt injuries, 9 by GSWs, and 4 by stab wounds, with an overall mortality of 44%. Twenty of the 25 patients were operated on, with a mortality of 40%. In his paper, he discussed current therapy of the day and acknowledged that some surgeons recommended nonoperative management. Tilton stated, “This seems a wrong principle to work on. Many cases might recover without interference but others will prove fatal from oversight, an intestinal perforation or foreign body or from insufficient drainage of the wound in the liver.” He also made the point that the surgeon “has no choice” but to operate on those patients with aggressive symptoms of internal hemorrhage. His recommended method of hemorrhage control was to use sutures or gauze packing. He stated that “the thermal cautery is of very little value in arresting hemorrhage from the liver.” His opinion on the treatment of these injuries would shape the management of hepatobiliary injuries for the next 75 years.

Despite these early advances, hepatobiliary trauma during the first half of the 20th century was marked by high morbidity and mortality, with reported mortality for patients with liver injuries during World War I being as high as 66%. Packing of liver wounds was frequently practiced and often resulted in perihepatic infection and abscess. However, during World War II, major advances in resuscitation, anesthetic techniques, early operation, hemorrhage control, establishment of liver drainage, and use of antibiotics reduced the mortality rate. In their book Trauma to the Liver , which referenced their World War II experience, Madding and Kennedy stated that “before the war, house surgeons advocated expectant or conservative treatment, or no treatment at all for the majority of wounds of the liver.…Peritonitis, hepatitis, fistulas and numerous other complications often followed this form of treatment.” In an 18-month period during the latter part of World War II, they cared for 829 wounds of the liver in 3154 patients with abdominal and thoracoabdominal wounds. Overall mortality was reduced to 27%, a significant improvement that they felt was due to the use of drainage and “aggressive resectional débridement” as needed. They further stated that “use of packing either with gauze or absorbable hemostatic agents should be avoided, except for temporary purposes.” The lessons that were advocated for the next four decades could be summarized as follows: (1) most patients with liver wounds require operation, (2) all wounds should be drained, (3) hepatic tissue should be judiciously débrided, and (4) liver packing should not be performed.

After World War II, the mortality from hepatic injuries decreased greatly, likely due to a decline in death from infection. Contributing factors included earlier transport, better resuscitation, advances in antibiotic therapy, and improved supportive care. A 36-year view of liver injury mortality from the Ben Taub Hospital in Houston, Texas, noted that the mortality declined from 20.6% in 1939 to 9.2% by the early 1970s. However, as highway speeds increased and civilian GSWs became more prevalent, deaths from infection were supplanted by deaths from hemorrhage, leading to increased aggressiveness in operative treatment for these injuries during the 1960 to 1990 time period, including major formal hepatic resections including hepatic lobectomy, hepatic artery ligation for hemorrhage control, atriocaval shunting, and tractotomy to expose deep bleeding. Although each of these techniques continues to be used in the treatment of major hepatic injuries, these methods are reserved for difficult cases rather than routine practice.

Over the past 20 years, it has become recognized that aggressive operative strategies often failed to prevent deaths from hemorrhage and that, even if operative treatment was prompt and efficient, the vascularity of the liver and overall physiologic state of the patient was such that ongoing bleeding not amenable to control by standard maneuvers often occurred, leading to mortality from ongoing bleeding and coagulopathy. Recent advances in management include updated imaging, improved resuscitation strategies, damage control operative strategies, better hemostatic agents, and the use of percutaneous techniques for angiography, embolization, and drainage. Together, these advances allow a customized approach to the care of these complex patients.

Classification of Liver Injuries

Although the Couinaud and Bismuth anatomic classifications of hepatic segmental anatomy have proven useful in planning elective operations and therapy, most liver injuries are nonanatomic in nature or cross multiple anatomic segments. To characterize these injuries and allow development of treatment strategies as well as ongoing research, the American Association for the Surgery of Trauma developed and validated a liver injury scoring system ( Table 123.1 ). This scoring system was originally based on operative findings so that the severity of injuries and mortality increases with higher grade. One known concern with this system is that, while it is valuable as a general guide, it is not precise enough to predict which patients will require intervention and which will not, as some patients will present with low-grade lesions that bleed significantly and many will present with high-grade lesions that do not require operative intervention.

Initial Evaluation

The initial assessment and resuscitation of trauma patients undergoing evaluation for hepatobiliary injury is no different than for any other injured patient. General principles of the Advanced Trauma Life Support Program as taught by the American College of Surgeons should be followed, including initial evaluation of airway, breathing, and circulation; evaluation for additional injuries; and monitoring the response to resuscitation. Special attention is paid to the patient's abdominal examination, vital signs, and physiologic and hemostatic response to treatment. The specific goals of the initial evaluation in all trauma patients is to efficiently determine the presence of potentially life-threatening injuries, assess the hemodynamic stability of the patient, and initiate a therapeutic plan that is based on the initial response to resuscitation and findings on the initial surgeon-performed ultrasound, diagnostic peritoneal lavage, or abdominal computed tomography (CT) scan. The response to resuscitation serves as an early decision point in the treatment of patients with hepatobiliary injuries.

Further evaluation of the patient will be dependent on the mechanism of injury, hemodynamic instability, and the ongoing response to resuscitation. In general, patients who are hemodynamically unstable due to hemorrhage from abdominal trauma should be treated with exploratory laparotomy in the operating room. In a similar fashion, urgent operation should be considered for most patients with penetrating injuries. In a patient with blunt trauma who is hemodynamically stable and responsive to resuscitation, the nature of the injury can be more thoroughly evaluated with a complete physical exam and imaging.

In most trauma centers, the initial diagnostic modality of choice is the focused assessment with sonography for trauma (FAST exam). FAST is an effective and accurate technique for rapid imaging in hemodynamically stable or unstable patients with possible hepatic injuries. FAST involves four views of the abdomen, including cardiac (to evaluate for pericardial effusion), right and left upper quadrants, and suprapubic (to evaluate for the presence of fluid). With training and practice, the sensitivity and specificity for detection of fluid in these views is 83% and 99.7%, respectively. Of note, when using FAST for patients with blunt abdominal trauma who are hypotensive, the sensitivity and specificity may increase to 100%. FAST has proven extremely useful as a decision making tool for hypotensive blunt trauma patients. In our practice, an unstable patient with a positive FAST is taken for laparotomy. If the patient is hemodynamically unstable and has a negative FAST, a search for extraabdominal injuries contributing to shock should be undertaken. If extraabdominal sources of shock are not present or if hemoperitoneum remains a concern in an unstable patient with a negative FAST, the FAST may be repeated or diagnostic peritoneal lavage may be performed. If concern for hemoperitoneum persists, the patient may require an exploratory laparotomy as a diagnostic and potentially therapeutic modality.

If a patient is hemodynamically stable, the patient then undergoes CT scanning of the abdomen and pelvis to document the presence and magnitude of injuries to the liver and other intraabdominal organs. CT of the abdomen and pelvis in the hemodynamically stable patient performed after the intravenous injection of contrast remains the most sensitive and specific imaging modality for the evaluation of hepatobiliary trauma. In our institution, we do not use oral contrast for CT scans in trauma patients. CT findings of interest to the surgeon include the presence and magnitude of the hepatic parenchymal injury, the presence and magnitude of intraperitoneal blood, and the presence and magnitude of associated intraperitoneal and retroperitoneal visceral, mesenteric, and vascular injuries. Hepatic injuries typically noted on a CT study include parenchymal lacerations, intrahepatic hematoma, or subcapsular hematoma with or without active extravasation of intravenous contrast ( Fig. 123.1 ). In addition, CT studies worrisome for active IV contrast extravasation can be quickly repeated to survey for ongoing active bleeding on venous phase imaging.

Management

Similar to the evaluation phase, management of patients with hepatic trauma is driven by hemodynamic status, mechanism of injury, and associated injuries. Guidelines generated by the Western Trauma Association, the Eastern Association for the Surgery of Trauma, and, more recently, the World Society of Emergency Surgery each provide important direction in the care of these patients.

Nonoperative management of stab wounds or GSWs to the liver has been shown to be safe, but there is a high incidence of concomitant intraabdominal injuries, and these patients must be carefully and thoughtfully selected. Potential candidates for nonoperative management of penetrating injuries to the liver should undergo CT scans. In one series, 15% of patients (and 80% of those presenting with isolated liver injury) were managed nonoperatively. In our practice, patients with penetrating abdominal trauma undergo laparotomy under most circumstances, with the exception of those in whom a reliable abdominal exam is present, additional injuries can be reliably excluded, and serial abdominal exams and physiologic status can be monitored.

Nonoperative management is more common in patients with injuries from blunt trauma. A study of patients from the National Trauma Data Bank examined the role of nonoperative management in patients with high grade (Abbreviated Injury Scale of 4 or higher) liver injuries. Of these, 73% were managed nonoperatively with a 7% failure rate. Of note, failed nonoperative management was independently associated with higher mortality. Predictors of need for conversion to operative management included older age, female gender, higher injury severity score, and hypotension. Previous series have examined the relationship between solid organ injuries and hollow viscus injuries and, perhaps unsurprisingly, noted a positive relationship between the two injury patterns. Thus, in the presence of liver injuries, hollow viscus injuries must be strongly considered, even in the absence of supporting CT scan findings.

In our practice, patients who are hemodynamically unstable, fail to respond to resuscitation, have hollow viscus injuries, or have evidence of peritonitis on exam are taken for laparotomy. All other patients are considered for nonoperative management. Patients with American Association for the Surgery of Trauma (AAST) grade I, II, and III injuries as determined by CT scan are admitted to the surgical floor for observation with hemoglobin assessment every 8 hours for 24 hours. These patients are considered for discharge after 24 hours of stable hemoglobin measurements if they are tolerating a diet. If they become unstable or require a blood transfusion, they are transferred to the intensive care unit with consideration for additional imaging and operation. Patients with AAST grade IV and V injuries are admitted to the surgical intensive care unit and undergo hemoglobin measurement every 6 hours for 24 hours. Need for blood transfusion or the occurrence of hemodynamic instability prompts thorough reassessment and consideration of angiography or operation. If initial nonoperative management is successful, these patients undergo interval CT scan prior to discharge to evaluate for interval development of hepatic artery pseudoaneurysm or bile collections.

If the initial CT scan reveals evidence of contrast blush with suspected hemorrhage or the formation of hepatic artery pseudoaneurysm, consideration is made for immediate selective angiography with embolization. In a hemodynamically unstable patient with no additional injuries, we pursue selective angioembolization primarily rather than proceeding to laparotomy. These situations are challenging in that they may require transporting a critically ill and unstable patient to the angiography suite. Our practice is to “bring the surgical intensive care unit to the patient” and continue active resuscitation with the intensive care unit and trauma team regardless of physical location. The advent of advanced hybrid operating rooms, with the capability for open operative intervention and high quality fluoroscopy, may alleviate this challenge, but these are not yet in widespread use.

Principles of Operative Treatment

With rare exception, operations for trauma should use a midline incision to permit rapid exploration of the entire abdomen as well as extension as needed to gain needed exposure. Transverse incisions, which are commonly used in elective hepatic procedures, may not permit a thorough abdominal evaluation and are not the standard incision for trauma patients. The use of an upper abdominal retraction system greatly facilitates exposure of the liver. If further exposure is required, the incision may be extended laterally in a subcostal direction or cephalad either partially or completely through the sternum. Although a median sternotomy may be required to control life-threatening bleeding, opening the sternum for a bicavitary procedure increases evaporative heat loss and may worsen coagulopathy. Therefore this should be done only if it confers a major advantage in exposure. The conduct of the operation will be determined to a considerable degree by the extent of hemoperitoneum encountered and the site of the presumed source of bleeding. If a large volume of free blood is encountered, packs should be placed in the four quadrants of the abdomen in an attempt to determine the primary site of bleeding. If liver bleeding can be controlled with packs initially, these can be left in place while other potential injuries are evaluated and addressed. A seriously injured spleen should be removed promptly and any concomitant gastrointestinal injuries rapidly evaluated and managed.

When addressing hemorrhage from the liver, a series of operative maneuvers may be employed to obtain hemostasis. A very useful algorithm to provide guidance on the approach to liver hemorrhage has been developed by the Western Trauma Association ( Fig. 123.2 ). Minor bleeding from hepatic trauma may be encountered in association with other intraabdominal injuries and is often due to low-grade (AAST I and II) liver injuries. These injuries can usually be managed by compression and temporary packing alone or with electrocautery, argon beam coagulation, with adjunct use of topical hemostatic agents. Packing with temporary abdominal closure will rarely be required for isolated low-grade injuries.

Severe hepatic injuries (AAST grade III to V) require a rapid but logical approach to hemorrhage control. In our practice, we typically initially perform manual compression and perihepatic packing then allow resuscitation of the patient and transfusion of blood products as needed. Once the patient's resuscitation is underway, we proceed with obtaining further hemostasis. Compression and packing often lead to temporary hemorrhage control and allow definitive hemorrhage control. If hemorrhage is not amenable to temporary packing, bimanual compression is held, resuscitation is performed by the anesthesia team, and the operation progresses immediately to definitive control. If significant bleeding is encountered, the institution's massive transfusion protocol should be activated. We typically transfuse in a balanced ratio of packed red blood cells, plasma, and platelets (1 : 1 : 1) with minimal infusion of crystalloid, as this strategy has been demonstrated to reduce mortality and morbidity from hemorrhage in trauma patients.

A number of operative tactics are available to control hepatic bleeding (summarized in Box 123.1 ). The technique employed will depend on the geometry and depth of the wound as well as the nature of the injured vessels. Bleeding hepatic parenchyma responds well to cautery, argon beam coagulation, and hemostatic agents. Bleeding parenchymal injuries are best managed by vessel ligation if feasible. In general, direct suture ligation of vessels is preferable to control with surgical clips due to the susceptibility of the latter to dislodge with manipulation of the liver. As needed, use of the finger fracture technique can be used to expose and allow ligation of deeper vessels. We also find the use of stapling devices useful in the event that removal of devitalized tissue is needed or to control hemorrhage with a limited hepatic resection. These are used in a similar fashion as in elective hepatic surgery. We also find that reapproximation of hepatic parenchyma with adjunct use of hemostatic agents will lead to cessation of bleeding in many injuries.

In the event of major bleeding that is not controlled by parenchymal compression, our next maneuver is to obtain control of the porta hepatis and perform a Pringle maneuver followed by additional hemorrhage control. Gentle dissection to encircle the hepatoduodenal ligament is performed and the resident structures are compressed between a thumb and forefinger. This can then be exchanged for a gentle, noncrushing fine vascular clamp or umbilical tape with a Rumel tourniquet to obtain more secure control. The duration of safe occlusion of the porta hepatis in the setting of traumatic injury with concomitant hypotension and shock is unknown.

If arterial bleeding from deep within the parenchyma is present and this is not amenable to direct control, additional strategies must be employed. If the bleeding slows or ceases with portal triad compression, selective hepatic artery ligation (SHAL) may be performed. Although used frequently in the past, SHAL is rarely performed in our current practice. We frequently perform a damage control operation with the control of venous hemorrhage and placement of perihepatic packing and then transport the patient to the angiography suite for selective angiography and embolization. If the Pringle maneuver fails to even temporarily control arterial bleeding, one may consider an aberrant left hepatic artery in the lesser omentum as the bleeding source. In the event of penetrating trauma to the liver with uncontrolled hemorrhage, balloon occlusion of the wound may be lifesaving. To perform this, a red rubber catheter is placed inside a tied-off Penrose drain. This is passed through the path of the projectile and then inflated until hemorrhage control is achieved. In most cases, the patient undergoes resuscitation and hepatic packing, followed by selective angioembolization. The balloon is removed at subsequent laparotomy. There are also reports of the use of extensive tractotomy for exposure of deep arterial bleeding. In our present practice, we prefer to pursue angioembolization in these patients.

Hepatorrhaphy or reapproximation of the liver parenchyma may be useful for control of venous bleeding or bile leakage. The technique usually involves the placement of large absorbable sutures, often in horizontal mattress fashion, for the reapproximation of the liver ( Fig. 123.3 ). Critics of the procedure argue that the use of large sutures may cause necrosis of hepatic tissue, which can be minimized with control of tension on the sutures. Attempting to control oozing from nonviable liver is an inappropriate use of hepatorrhaphy, as this tissue should be débrided. However, small bleeding cracks can often be closed with the arrest of hemorrhage by this technique without causing tissue necrosis or unnecessarily extending the wound by tractotomy or finger fracture. We have also encountered large injuries with significant lacerations in the liver with little bleeding from the exposed surfaces. After the major vessels and ducts have been controlled by suture techniques, gentle reapproximation of the divided portions may decrease oozing from the raw surfaces. This technique may be augmented by an omental flap, packing with topical hemostatic agents, or both. If used appropriately, hepatorrhaphy can be a useful tool, but it will not control major arterial bleeding.

Omental pedicle flaps can be very useful for prevention of diffuse oozing from raw surfaces of the liver. This was commonly used, but in the era of primarily operative management, the need for this maneuver has diminished considerably. The omental flap is created in a standard fashion and is used in our practice to cover large exposed liver surfaces. Indications for its use vary but might include the aforementioned crevasses in hepatic tissues, coverage of the remaining surface following resection or débridement of hepatic tissue, and for any larger area where there is no liver capsule. In the unusual instance where a subcapsular hematoma has ruptured or been surgically entered, diffuse bleeding from the exposed liver surface usually ensues. The use of an omental flap in this circumstance may be extremely useful.

Formal anatomic resection of the liver is not commonly indicated for trauma. However, there are several instances where a major hepatic resection is indicated: devitalizing injuries along anatomic planes, completion lobectomy when the injury has transected the liver along lobar anatomy, and for exposure of major venous bleeding associated with a major hepatic parenchymal injury. The use of hepatic artery embolization has occasionally resulted in lobar hepatic necrosis that requires a formal lobectomy for resolution. Unlike a planned lobar resection for hepatic tumors, major resections for injury must often be performed under an emergent, less-than-optimal situation in a patient who is at grave risk for exsanguination. In this circumstance, operative speed is imperative. The use of techniques employed by elective hepatic surgeons such as vascular and hepatic staplers is a useful adjunct in such cases. Formal resection may also be performed in delayed fashion at a second look operation to delineate and débride necrotic hepatic tissue following the initial packing and correction of coagulopathy.