Timothy C. Fabian and Tiffany K. Bee
INTRODUCTION
Liver injury occurs in approximately 5% of all trauma admissions.1 The liver’s size and anatomic location, directly under the right costal margin, make it the most susceptible organ for injury in blunt trauma and a frequently involved organ in penetrating trauma. The management of liver injury has evolved greatly over the last decade. There have been many technical advances in medicine, which now allow us to better diagnose and treat liver injuries both operatively and nonoperatively. However, the most severe liver parenchymal and venous injuries as well as those involving the portal triad continue to challenge even the most adept trauma or hepatobiliary surgeon and often lead to death. Therefore, despite our progress in liver injury management, many avenues for improvement remain to be explored.
HISTORY
Liver injury management has been described in many of the early surgical textbooks. We consider nonoperative management of hepatic injury a modern approach; however, a 1905 surgical text states, “If the evidences of a rupture of the liver, such as the signs of shock and hemorrhage…. the continuous increase in pain, due to progressive abdominal distention, and muscular rigidity, are absent, no operative intervention can be considered.”2 Mortality from liver injury was as high as 62.5% in these early years.3 Pringle wrote a landmark paper examining the management of severe liver injury in 1908.4 Although many authors previous to this paper had described suturing methods of liver parenchyma as well as gauze packing into the liver laceration, Pringle described a maneuver of occluding the porta hepatis with the surgeon’s fingers and thus decreasing the amount of hemorrhage from a severely injured liver. This procedure continues to be a useful tool in the management of liver trauma.
During World War II, new ideas in the management of severe liver injury surfaced. Madding et al. used the principles of early laparotomy, drainage procedures, advances in anesthetic and aseptic care, as well as transfusion technology to improve mortality to 27.7%.5 The techniques of hemorrhage control adopted at that time incorporated parenchymal reapproximation with large blunt liver needles, resection, and direct vessel ligation. These methods prevailed until approximately 10 years ago. Trends in management have now led to an emphasis on nonoperative treatment for those patients who remain hemodynamically stable and liver packing with damage control for those who are unstable.
ANATOMY
Comprehensive knowledge of hepatic anatomy is essential to the proper management of traumatic liver injuries. The understanding of the ligamentous attachments, parenchyma, and intraparenchymal and extraparenchymal vascularity of the liver is key to the effective application of methods for control and repair in liver injuries (Fig. 29-1).
FIGURE 29-1 Surgical anatomy of the liver: (1) inferior vena cava; (2) right hepatic vein; (3) middle hepatic vein; (4) left hepatic vein; (5) portal vein; (6) right branch portal vein; (7) left branch portal vein; (8) right triangular ligament; (9) coronary ligament; (10) left triangular ligament; (11) falciform ligament; (12) ligamentum teres.
Lobes
Cantlie first described the lobar anatomy in 1898. The liver is divided into two lobes by a 75° angle traversing from the gallbladder fossa posteriorly to the left side of the inferior vena cava. This is the so-called line of Cantlie. Therefore, the left lobe includes the hepatic tissue to the left of the falciform ligament along with the quadrate and caudate lobes. The right lobe consists of the remaining parenchyma.
Functional Anatomy
The functional anatomy of the liver separates the liver into segments pertinent to resection. In 1953, Couinaud provided the basis of modern resection planes by dividing the liver based on the distribution of the hepatic veins and glissonian pedicles.6 The right hepatic vein traverses between the right posterolateral (VI and VII) and right anteromedial (V and VIII) segments. On the left, the left hepatic vein delineates the anterior (III and IV) and posterior (II) segments. The caudate lobe (I) drains directly into the inferior vena cava (Fig. 29-2).
FIGURE 29-2 Functional division of the liver, according to Couinaud’s nomenclature. (Reproduced with permission from Blumgart LH, ed. Surgery of the Liver and Biliary Tract. New York: Churchill Livingstone; 1988. © Elsevier.)
Hepatic Artery
The common hepatic artery branches from the celiac artery. This provides about 25% of the hepatic blood flow and 50% of hepatic oxygenation. The artery then branches into the gastroduodenal, right gastric, and proper hepatic. The proper hepatic is found in the porta hepatis usually to the left of the common bile duct and anterior to the portal vein. At the hilum of the liver, the artery bifurcates into a right (the longer branch) and a left hepatic artery. There are a number of anatomic variances. The most frequent (11%) is the aberrant superior mesenteric origin of the right hepatic artery traversing behind the duodenum. Other variants include a left hepatic artery origin from the left gastric artery (8%) and the left and right hepatic arteries arising from a superior mesenteric artery origin (9%). With these multiple variants, great care must be taken when controlling the traumatic hemorrhage.
Hepatic Veins
The hepatic veins develop from within the hepatocytes’ central lobar veins. The superior, middle, and inferior vein branches originating from the right lobe form the right hepatic vein. The middle hepatic vein derives from the two veins arising from segments IV and V and frequently includes a branch from the posterior portion of segment VIII. In 90% of patients the middle hepatic vein joins the left hepatic vein just before draining into the inferior vena cava. The left hepatic vein is more variable in its segmental origin. Most important is the posterior positioning of the vein when dissecting the left coronary ligament; great caution must be used in this area to avoid inadvertent injury.
The retrohepatic vena cava is about 8–10 cm in length. It receives the blood of the hepatic veins and also multiple small direct hepatic vessels. Exposure to this area can be very difficult, especially when an injury and accompanying hemorrhage make visualization very difficult.
Portal Vein
The portal vein is formed from the confluence of the splenic and superior mesenteric veins directly behind the pancreatic head. It provides about 75% of hepatic blood flow and 50% of hepatic oxygen. The portal vein lies posteriorly to the hepatic artery and bile ducts as it ascends toward the liver. At the parenchyma, the portal vein divides into a short right and a longer left extrahepatic branch.
Ligaments
When operating on the liver, it is crucial to understand the ligamentous attachments. The coronary ligaments attach the diaphragm to the parietal surface of the liver. The triangular ligaments are at the lateral extensions of the right and left coronary ligaments. The falciform ligament with the underlying ligamentum teres attaches to the anterior peritoneal cavity. The medial portion of the coronary ligaments is where the hepatic veins traverse and therefore dissection in this area must be done with caution.
LIVER INJURY INCIDENCE AND CLASSIFICATION
Liver injury occurs in approximately 5% of all trauma admissions. Since the liver is the largest intra-abdominal organ, it is not surprising that the liver is the most commonly injured solid organ in blunt and penetrating injury. Data from the author’s institution over the past 5 years illustrate the frequency of liver injury compared to other abdominal solid organ injury (Table 29-1). Motor vehicle collision is by far the most common etiology for a blunt liver injury. This is followed by pedestrian/car collisions, falls, assaults, and motorcycle crashes. Liver injury in penetrating trauma is also frequent, ranging from 13% to 35% of penetrating admissions, and is dependent on the weapon utilized.
TABLE 29-1 Trauma Admissions, 2000 to May 2009 (N = 39,722)
Uniform classification of liver injury is essential to compare the efficacy of management techniques (Fig. 29-3). The American Association for the Surgery of Trauma established a detailed classification system that has been widely utilized7 (Table 29-2). This classification provides for uniform comparisons of both nonoperative and operatively managed hepatic injury.
TABLE 29-2 Liver Injury Scale (1994 Revision)
FIGURE 29-3 Hepatic injury grading is important to compare outcome.
INITIAL MANAGEMENT
Care for the patient with possible liver injury should proceed by the tenants of Advanced Trauma Life Support (ATLS). Of utmost importance is the initial evaluation, including attention to airway, breathing, and circulation (see Chapter 10). Other life-threatening injury may take precedence over possible internal injury in the primary survey. However, liver injury may indeed be a cause of hemorrhagic shock and cannot be overlooked. Resuscitation strategies are evolving in the care of trauma patients (see Chapter 12). The prospects of permissive hypotension, hypertonic saline resuscitation, and other strategies are the subject of many recent investigations.
Physical exam of the patient remains a critical component of the initial evaluation. However, physical exam of a trauma patient may indeed miss significant internal injury. A study by Olsen et al. found that trauma patients with a “benign” physical exam had a 43% incidence of significant intra-abdominal injury.8 Therefore, it is justified that patients with benign physical exams are further evaluated by either serial exams or radiologic methods. Plain radiographs and ultrasound obtained in the trauma bay may give clues to possible liver injury if lower right rib fractures, hemothorax, hemoperitoneum, or a ruptured diaphragm is diagnosed. Although nonoperative management has become routine, a patient exhibiting clear peritoneal signs and instability requires immediate celiotomy.
Important but often overlooked points include keeping the patient warm and collecting appropriate laboratory data. Hypothermia can have detrimental effects on coagulation and cardiac rhythm. Appropriate laboratory data should include type and cross, hematocrit, coagulation profile, amylase, and base deficit. Abnormalities can alert the clinician to possible internal injury and its severity.
DIAGNOSIS OF LIVER INJURY
Hemodynamically Unstable Patient
After primary survey and resuscitation have been initiated, the patient may still be hemodynamically unstable. In these cases it is necessary to immediately determine the possible causes of the continued shock state. This can be difficult in patients with multiple injuries involving many organ systems.
Intra-abdominal injury can be an obvious cause of instability if physical exam reveals peritoneal signs, penetrating injury, or increasing distention. More often, a rapid diagnostic modality must be employed. The two most pertinent modalities for these situations are diagnostic peritoneal lavage (DPL) and focused abdominal sonography for trauma (FAST).
Diagnostic Peritoneal Lavage
DPL is a very accurate method for determining the presence of intraperitoneal blood. Many reports have replicated the work of Root et al. that indicated up to a 98% accuracy of determining the presence of intra-abdominal blood.9DPL is rapid and safe if performed with the semi-open or open technique. It remains a very useful tool in those patients who have altered sensorium and remain hemodynamically unstable. A positive DPL is defined as a gross aspiration of 10 mL of blood or greater than 100,000 RBC/mm3 in at least 300 mL of irrigant. A finding of gross blood in an unstable patient leads to immediate operative intervention. DPL does have limitations. It is not useful in determining the origin of the bloody aspirate and can actually be too sensitive since it is positive with minimal hemoperitoneum. Therefore, though it has its place in rapid determination of hemoperitoneum and subsequent immediate operative intervention, DPL has been replaced in most trauma centers by ultrasound and in more stable patients by computed tomography (CT) scanning.
Focused Abdominal Sonography for Trauma
The FAST exam has superseded DPL in many institutions for the determination of hemoperitoneum in the unstable bluntly injured patient (see Chapter 16). Surgeons have become very adept and familiar with this diagnostic modality. Richards et al. reported a 98% sensitivity of ultrasound for hemoperitoneum in grade III and higher liver injury.10 However, they were not able to identify the anatomic location of the hepatic parenchymal injury in 67% of these severely damaged livers. A multi-institutional study by Rozycki et al. concluded that the RUQ area is the most common site of hemoperitoneum accumulation in blunt abdominal trauma.11 This information was reiterated in another study that discovered that the “two most common patterns of fluid accumulation after hepatic injuries were the RUQ only and the RUQ and lower recesses.”12 If the initial exam of the RUQ is negative, it is recommended that the pericardial, LUQ, and pelvic areas also be examined. The FAST exam is about 97% sensitive when 1 L of peritoneal fluid is present, but the examiner can rarely see volumes less than 400 mL with current technology.13 A repeat FAST exam can be beneficial after the initial resuscitation. Further resuscitation may promote further bleeding that then leads to more intraperitoneal blood on FAST exam. FAST exam is very beneficial in those unstable patients in whom the diagnosis of hemoperitoneum requires emergent surgery.
Hemodynamically Stable Patient
Ultrasound and CT scanning are the mainstays of diagnosing hepatic injury in the hemodynamically stable but bluntly injured patient. Once the primary and secondary surveys have been completed, the patient at risk for intra-abdominal injury should undergo further radiologic evaluation for definitive diagnosis.
FAST
FAST examination, as mentioned above, has proven to be a very good diagnostic tool in the evaluation of the blunt trauma patient. Some centers are using ultrasound for definitive diagnosis of intra-abdominal injury. Most examiners, though, are unable to distinguish between different grades of hepatic injury by ultrasound.10 Also, the source of free fluid in the peritoneal cavity is difficult to discern by ultrasound alone, especially if multiple injuries are present. A FAST exam has been reported to have a sensitivity as high as 83.3% and specificity of 99.7%.14 With these relatively low false-negative rates, some institutions are observing patients with negative FAST and not proceeding with CT scanning. However, Chiu et al. in 1997 reported a 29% incidence of abdominal injury following negative initial FAST.15They reported confounding clinical factors including contusion, pain, pelvic fracture, and lower rib fractures that were present in many of the false-negative patients. Also, 27% of these negative FAST patients underwent laparotomy for undetected splenic injury. A follow-up FAST exam was not performed on these patients prior to surgery and given the time for further hemoperitoneal fluid development these scans may now have been positive. Serial ultrasound exams are now used in many trauma centers if the initial scan is negative. Patients with pelvic ring-type fractures should undergo CT scan even if a negative FAST has been performed due to the more frequent occult injuries in these patients.16
Contrast-enhanced sonography shows some promise in the detection of liver injury. Contrast-enhanced ultrasound uses intravenously injected microbubbles containing gases other than air to produce the “contrasted” images. Valentino et al. reported a 100% sensitivity and specificity in seven liver injury patients with grade II–IV injuries.17 Similarly McGahan et al. reported 90% detection in liver injuries of the same grades.18 Another study described the ability of this modality to detect active extravasation from solid organs.19 With these advancements, patients may be subject to less risk from radiation or CT contrast. Also, this can be done at bedside instead of transporting a critical patient to a radiology suite. Overall, ultrasound is an excellent tool for the diagnosis of significant hepatic injury in the blunt trauma patient.
FAST also has an expanding role in penetrating abdominal trauma with an institutional series sensitivity of 46% and specificity of 94% in penetrating injury,20 therefore concluding that FAST can be used to triage patients more directly to surgery. However, the limitations of FAST in penetrating trauma are significant. In a patient with a possible tangential wound, the question is often if the peritoneum has been penetrated. A finding of fluid in Morison’s pouch confirms penetration and will result in immediate surgical intervention. A negative fluid accumulation, however, does not definitively rule out penetration. These results have been validated.21 Two interesting studies have demonstrated that fascial penetration can be verified by ultrasound examination.22,23 Again, the sensitivity of this modality is low but the specificity is high. Ultrasound may be a good screening tool for finding fascial penetration and a positive result could alleviate the patient of a painful bedside wound exploration and also contribute to operative decision making. Future study in this area may develop greater uses for ultrasound in select penetrating injuries.
CT Scanning
The advent of CT scanning and advances in that technology have resulted in tremendous changes in the management of liver injury. Since the first use of CT to diagnose intra-abdominal injury in the early 1980s, CT has become a routine part of the management of trauma patients.24 One recent study revealed that the specificity of the clinical examination with bedside radiologic investigations of plain x-ray and sonography in addition to laboratory values is not sufficient to preclude the blunt trauma patient from obtaining a CT scan for definitive diagnosis of injury.25 The advent of the helical CT scan has improved resolution as well as increased the speed of a head to pelvis scan to less than 10 minutes. Trauma surgeons now use CT scans for diagnosis and for management decisions in liver injuries. Being able to grade the extent of injury and to follow an existing injury can determine if nonoperative management is possible and successful (Fig. 29-4).
FIGURE 29-4 Algorithm for nonoperative management of blunt liver injury.
CT scanning is also being used in penetrating injury. Triple-contrast CT in back and flank wounds has been shown to have good sensitivity; however, the sensitivity for diaphragmatic and small bowel injury is poor.26 Therefore, a minor hepatic laceration can be evaluated and nonoperatively managed with CT guidance but continued frequent abdominal exam must also accompany this algorithm.
Laparoscopy
Laparoscopy has been successfully used to diagnose peritoneal penetration of penetrating trauma, thus saving the patient from a nontherapeutic exploratory laparotomy.27 Repair of hepatic injury found at laparoscopy has also been reported.28,29 In carefully selected patients, laparoscopy can be advantageous in the diagnosis and repair of hepatic injury.
MANAGEMENT OF LIVER TRAUMA
Anatomic relationships are key to understanding the management of liver trauma. Blunt hepatic injury traverses almost exclusively along the segments of the liver. This most likely occurs due to the strength of the fibrous covering around the portal triad preventing injury from transecting these structures. However, the hepatic veins do not have a similar fibrous structure and therefore, having less resistance, are the primary structures injured in blunt trauma. Penetrating trauma, on the other hand, involves both venous and arterial injury with direct transection of any structure in the trajectory. These anatomic principles are key to understanding the rationale for making decisions in the management of liver trauma.
Hemodynamically Stable Patient with Blunt Injury
Nonoperative treatment of the hemodynamically stable patient with blunt injury has become the standard of care in most trauma centers (see Fig. 29-4). In 1995, Croce et al. published a prospective trial of nonoperative management of liver injury.1 In this study patients with all grades and volumes of hemoperitoneum were evaluated against operative controls. They found that they were able to successfully manage 89% of hemodynamically stable patients without celiotomy. Most blunt liver injuries produce hepatic venous injuries that are low pressure (3–5 cm H2O). Hence, hemorrhage usually stops once a clot forms on the area of disruption. Successful nonoperative therapy resulted in lower transfusion requirements, abdominal infections, and hospital lengths of stay. Hurtuk et al. have reported that indeed trauma surgeons “practice what they preach” in a recently published evaluation of the National Trauma Data Bank. They found that in the past 10 years there has been no effect on mortality in solid organ injury with prevalence of nonoperative management.30 Coimbra et al. reiterated these data by examining their experience in nonoperative treatment of grade III and IV hepatic injury.31 They reported no mortality in their nonoperatively managed patients and “discouraged” operative management of these injuries.
Approximately 85% of patients with blunt liver trauma are stable. Once stability has been established, the patient must be carefully analyzed for the appropriateness of nonoperative care. The patient cannot exhibit signs of peritonitis and must continue to be hemodynamically stable without a significant transfusion requirement. The authors are generally comfortable in nonoperatively managing a stable patient with 3–5 U of blood in his or her abdomen. A contrast-enhanced helical CT scan should be obtained to evaluate injury grade, amount of hemoperitoneum, evidence for enteric injury, active extravasation of contrast, and presence of pseudoaneurysm (Fig. 29-5).
FIGURE 29-5 CT scan demonstrating a “contrast blush,” indicative of active arterial bleeding in a patient with a grade IV blunt hepatic injury.
High-grade injury, large hemoperitoneum, contrast extravasation, and pseudoaneurysm are not contraindications for nonoperative management; however, these patients are at higher risk for nonoperative failure and may need a multimodality approach to stabilize their nonoperative injury. Stable patients with high-grade injury may be observed. However, Malhotra et al. noted that 14% of grade IV and 22.6% of grade V injuries fail nonoperative management, which was substantially higher than the 3–7.5% failure rate of more minor injuries.32 That same article reports large hemoperitoneum (blood around liver, pericolic gutter, and in the pelvis by CT) as a significant factor in failure of nonoperative management but that it could not predict which patients would ultimately fail nonoperative management. Richardson et al. speculated that many experienced trauma surgeons have taken stable but high-grade injury patients to the operating room only to find that “manipulation of venous injuries resulted in massive hemorrhage that resulted in the patient’s death.”33 They concluded that nonsurgical treatment has a “positive impact on survival.”
A CT finding of contrast blush or extravasation has previously meant that patients were not candidates for nonoperative therapy. However, with the assistance of interventional radiology, some patients may be candidates for embolization and nonoperative treatment. Successful embolization of hepatic arterial injury in patients who are hemodynamically stable but with CT scans demonstrating intrahepatic contrast pooling was reported in 1996.34Choosing the appropriate patient for embolization can be a challenge. One interesting study looked at 11 patients with hepatic injury and CT evidence of contrast extravasation who were stable “only with continuous resuscitation.”35These patients were evaluated by hepatic angiography and seven patients were successfully treated with hepatic embolization. The other four patients had no active extravasation seen by angiography and became hemodynamically stable not requiring surgery. Misselbeck et al. reviewed their 8-year experience with hepatic angioembolization and found that hemodynamically stable patients with contrast extravasation on CT scan were 20 times more likely to require embolization than those without extravasation.36 Arterial extravasation with blunt liver injury is much less common than venous injury. However, many centers are anecdotally noting excellent results with a multimodality approach.
Complications of Nonoperative Blunt Hepatic Injury Management
Most patients with blunt nonoperative liver injuries heal without complication. Follow-up CT scans generally show resolution of severe injuries within 4 months and about 15% show complete resolution at hospital discharge.1However, complications can arise and management requires the surgeon to be prepared to deal with the possible adverse outcomes.37 A retrospective multi-institutional study included 553 patients with grade III–V injury.38 Of these patients, 12.6% developed hepatic complications that included bleeding, biliary problem, abdominal compartment syndrome, and infection. Significant coagulopathy and grade V injury were found to be predictors of complication. Therefore, with current nonoperative management strategies, complications must be dealt with appropriately.
Bile Leaks
One of the more frequent complications is bile leakage. Bilomas or bile leak can occur in 3–20% of nonoperatively managed patients.1 Hepatobiliary hydroxy iminodiacetic acid (HIDA) scan and MRCP have been used to localized bile leaks.39 Evidence of bile leak by HIDA scan does not mandate intervention. In fact, of the 14 patients found to have HIDA evidence of bile leak in a 1995 study, only 1 patient became symptomatic and required percutaneous drainage.1 Abnormal liver function tests, abdominal distention, and intolerance to feeding may all indicate a bile leak. CT scan evaluation with percutaneous drainage usually remedies the problem completely. However, large bile leaks can develop. Many authors have described management of bile peritonitis or large leaks not responsive to percutaneous drainage using percutaneous drainage techniques along with endoscopic retrograde cholangiography (ERC) and biliary stent placement.40 It has also been demonstrated that sphincterotomy can decrease the biliary pressure and allow healing of the bile leak.41 In some instances, actual stenting of a large ductal injury can be accomplished.42 Griffen et al. have reported success with a combined laparoscopic and ERC approach. They described patients with biliary ascitis taken to operating room for laparoscopic bile drainage and placement of drain near injury site with postoperative ERC and bile duct stenting. They report no septic complications and healing of the substantial biliary leaks.43 The authors have rarely experienced a persistent bile leak in the nonoperatively managed patient. Bile leaks or bilomas are drained percutaneously, sometimes for up to 4–6 weeks, and they nearly always resolved without ERC or other decompressive maneuvers.
Abscess
Perihepatic abscesses have also been uncommonly encountered with nonoperative management. The patient may exhibit signs of sepsis, abnormal liver function tests, abdominal pain, or food intolerance. Abscesses, like biliary collections, can often be managed by CT-guided drainage catheters. However, if the patient fails to improve with drainage and antibiotics, wide surgical drainage should be performed. This may involve merely incision and adequate drainage of the cavity or it may involve extensive debridement of the hepatic parenchyma.
Hemorrhage
Delayed hemorrhage after nonoperative management is a feared complication. Gates presented a review of the subject in 1994 and suggested an overall incidence of delayed hematoma rupture of 0–14%.44The 14% figure is well above current reports. He discussed 13 publications and determined that 69% of these delayed hemorrhage cases could have been successfully treated nonsurgically. Using the same criteria that were originally utilized to manage these patients nonoperatively, namely, hemodynamic stability without ongoing blood loss, patients with delayed hemorrhage can undergo hepatic angiographic embolization and observation with success. Therefore, it seems that delayed hemorrhage is actually a rare and manageable complication.
Devascularization
Disruption of vascular inflow to a hepatic segment following trauma or post-angioembolization can lead to necrosis of that segment of liver. The consequences of necrosis may include elevation of liver transaminases, coagulopathy, bile leaks, abdominal pain, feeding intolerance, respiratory compromise, renal failure, and sepsis.45 Many studies suggest that patients with significant necrosis should undergo hepatic resection before complications arise.46,47Devascularization can be identified by CT scan. It can be differentiated from intraparenchymal hemorrhage when follow-up CT scans reveal segments of liver that remain devascularized or have air within the devascularized area.45
Hemobilia
Hemobilia can occur after blunt hepatic injury. In 1871, Quincke described the triad of right upper quadrant pain, jaundice, and upper GI bleeding that indicated hemobilia. This triad may not be evident in the trauma patients with hemobilia.48 In a 1994 study, three patients developed hemobilia with massive upper gastrointestinal hemorrhage following blunt hepatic injury.49 The authors concluded that hepatic artery pseudoaneurysm with hemobilia is predisposed by bile leak and that angiographic embolization was appropriate for patients without sepsis and with small cavities. However, formal hepatic resection or drainage, after angiographic vascular control, may be necessary for septic patients or those with large cavities. Hemobilia is much less common with the prevalence of nonoperative management. With operative interventions of the past including large parenchymal suturing and vessel ligation, communications between vessels and bile ducts often occurred iatrogenically. Now that nonoperative care is practiced, we rarely see hemobilia.
Systemic Inflammatory Response
Nonoperatively treated patients with inadequately drained bile or blood collections may be susceptible to the development of a systemic inflammatory responses syndrome that may include respiratory distress. Recent articles from Franklin et al. and from Letoublon et al. advocate laparoscopic evacuation of undrained bile or hemoperitoneum at postinjury days 3–5.50,51 They report a marked decrease in the inflammatory response in many of these patients.
Unusual Complications
Large subcapsular hematomas have been described to elevate intraparenchymal pressures high enough to cause segmental portal hypertension and hepatofugal flow.52 This “compartment syndrome of the liver” was described in a patient managed non-operatively whose decreasing hematocrit and increasing liver function tests promoted angiographic examination revealing the hepatofugal flow in the right portal vein. After operative drainage of the tense hematoma, the patient did well with reversal of flow and viability of the right lobe liver tissue. This type of compressive complication has also been described causing a Budd–Chiari syndrome when hematoma results in intrahepatic vena cava compression or hepatic venous obstruction.53
Follow-Up CT Scanning of Blunt Hepatic Injury
Definitive data on the value of follow-up CT scanning of blunt hepatic injury are not available. Recent published reports suggest postobservation CT scans on those with more severe (grade III–V) injuries. Cuff et al. reported that of the 31 patients who received follow-up CT scans 3–8 days postinjury, only 3 patients’ scans affected future management.54 Additionally, the three scans that affected management were obtained due to a change in clinical picture and not merely routine. A 1996 report similarly concluded that follow-up CT did not change decision making in those with grade I–III injury.55 The authors’ institution concluded from their follow-up of 530 patients, including 89 grade IV or V, that follow-up CT scans are not indicated as part of the nonoperative management of blunt liver injuries.56 Follow-up CT scans are indicated only for those patients who develop signs or symptoms suggestive of hepatic abnormality. By scanning only those with clinical suspicion, there is a small inherent risk of missing unsuspected, possibly deleterious pseudoaneurysms that may result in delayed hemorrhage and require embolization. If a patient has had a follow-up CT that reveals significant healing, a postdischarge scan is not necessary. However, if significant healing has not occurred or if the patient had a grade IV or V injury, our practice is to obtain a postdischarge scan at 4–6 weeks after the injury.
Resumption of Activity
No steadfast rules apply to activity resumption in patients with uncomplicated hospital courses following blunt hepatic injury. The practice of keeping a patient from activity for 4 months has been commonly employed. This practice most likely resulted from the observation that most hepatic injury seems to have resolved by CT in 4 months. A contrary approach to this practice can be based on some interesting animal studies. Dulchavsky et al. found in animal studies that hepatic wound burst strength at 3 weeks was as great or greater than uninjured hepatic parenchyma.57 This is most likely a result of fibrosis throughout the injured parenchyma and Glisson’s capsule. Therefore, activity can be resumed about 1 month after injury if a follow-up CT (in grade III–V) has shown significant healing.
Hemodynamically Stable Patient with Penetrating Injury
Nonoperative Management of Penetrating Injury
Peritoneal penetration has mandated operative exploration for many years. However, many trauma centers have adopted selective nonoperative management of knife stab wounds to the right upper quadrant. The work of Nance and Cohn in 1969 supported this nonoperative care in patients with stab wounds who were hemodynamically stable and had no evidence of peritoneal irritation.58 Since then, reports of successful nonoperative management of gunshot wounds (GSW) have been published. Renz and Feliciano prospectively treated 13 patients with right thoracoabdominal GSW nonoperatively.59 The rationale behind this management is that these wounds of small caliber weapons may have injury to diaphragm and liver only, sparing any intestinal injury. The authors stressed the importance of serial abdominal exams and contrast CT scanning in their successful nonoperative management of penetrating injury. Other center experience has concurred with this selective nonoperative management.60,61 Demetriades et al. even reported successful nonoperative management of penetrating grade III and IV liver injuries that required angioembolization.62 The criteria for non-operative management include those patients who are hemodynamically stable, have no peritoneal signs, and are not mentally impaired. These patients then undergo contrast-enhanced CT scan to rule out other abdominal visceral injury. Serial abdominal exams as well as close hemodynamic monitoring are also implemented. Triple-contrast CT of 86 abdominal GSW, as reported by Shanmuganathan et al., had a sensitivity and specificity of 97% and 98%, respectively.63Velmahos et al. do not use triple-contrast CT at their center. They report a sensitivity and specificity of 90.5% and 96%, respectively, in diagnosing intra-abdominal organ injuries requiring surgical intervention.64
All trauma surgeons do not accept nonoperative management of GSW. Missed or deliberate nonrepair of small diaphragmatic lesions may lead to long-term adverse sequelae, not only of diaphragmatic herniation but also of possible biliopleural fistula.65 Late intervention for other missed injury (e.g., duodenal injury) may also lead to substantial morbidity. Nonoperative management of RUQ penetrating trauma must be performed under the care of a center that has not only the capability of close continuous monitoring but also CT radiology accessibility and immediate operating room availability.
Operative Management of Patients with Minor Liver Injury
The decision for operative intervention of incidental liver injury may develop due to laparotomy for penetrating injury, patient instability, or concomitant internal injury. The incision of choice is the midline incision in a trauma patient. Not only will the operating surgeon be able to gain access to the hepatic region but the entire peritoneal cavity will also be able to be inspected and manipulated. On opening the peritoneal cavity, attention should first be focused on stopping uncontrolled hemorrhage. Laparotomy pads should be used to clear the peritoneal cavity of clot. In minor liver injury, the bleeding from the liver can initially be managed with packing of the hemorrhagic area. Before dealing with a minor liver injury, the remainder of the peritoneal cavity should be inspected for injury, including bowel injury and other solid organ injury. Many minor liver injuries do not require operative fixation and nonbleeding wounds should not be probed or otherwise manipulated. Small wounds of the liver parenchyma with minimal bleeding may be able to be controlled with electrocautery or argon beam coagulation. Small to moderate bleeding cavities may first be inspected for any obvious bleeding vessels that can be ligated. Next, packing a tongue of omentum, with its vascular supply intact, into the wound and securing it into place halts most moderate bleeding. Stone and Lamb first described this technique in 1975.66 Wrapping a column of absorbable gelatin sponge with oxidized regenerated cellulose makes another beneficial device (Fig. 29-6). This is then inserted like a plug into deeper bleeding cavities. Omentum is often then brought up into the wound and secured to increase hemostasis. These maneuvers are very successful in the management of minor liver injury.
FIGURE 29-6 Hepatic injury plugs may be useful for tamponading deep parenchymal wounds.
Operative Management of Patients with Major Liver Injury
Initial Management
Patients with major hepatic trauma may present with hemodynamic instability and are therefore taken urgently to the operating room. As in minor injury the most optimal incision for expected major liver injury is the midline incision. Once the peritoneum is entered in these patients, a large amount of blood may be evacuated, which decreases the natural tamponade of a large hemoperitoneum. Adequate resuscitation is the key at this time. Manual compression of obvious injury will decrease bleeding (Fig. 29-7). It is imperative that the anesthesia team is allowed to catch up with fluid loss prior to proceeding. Fluids should be warmed and coagulopathy corrected, keeping in mind current recommendations for coagulation products given with packed red blood cells (see Chapter 13). Once the patient has been adequately resuscitated, a more thorough exam of the peritoneal cavity must be completed. If indeed the bleeding source is localized to the liver and bleeding continues after manual compression is released, then the portal triad should be identified and a Pringle maneuver performed (Fig. 29-8).
FIGURE 29-7 Manual compression of major liver injury.
FIGURE 29-8 The Pringle maneuver controls arterial and portal vein hemorrhage from the liver. Any hemorrhage that continues must come from the hepatic veins. (Reproduced with permission from Burch JM, Moore EE. Injuries to the liver, biliary tract, spleen, and diaphragm. In: Wilmore DW, ed. ACS Surgery: Principles & Practice. New York: WebMD Corporation; 2002.)
Much controversy has evolved around the normothermic ischemic time produced by the use of the Pringle maneuver. Many authors have advocated clamping for 20 minutes and then allowing reperfusion for 5 minutes.67,68 This practice has not been proven to be beneficial in traumatic liver injury. Multiple studies have emerged indicating that longer portal triad occlusion can be accomplished with similar results. In one study describing the management of 1,000 cases of hepatic trauma, the Pringle maneuver was utilized for between 30 and 60 minutes in many of the high-grade injuries without adverse sequella.69Pachter et al. managed 81 patients with the assistance of the Pringle maneuver for up to 75 minutes without any apparent morbidity from the procedure.70 Therefore, it seems that longer normothermic ischemic time can be used without added morbidity in the severely injured liver.
The Pringle maneuver often does not control all bleeding. It will control the inflow bleeding from the hepatic artery and portal vein but not the retrograde bleeding from the vena cava and hepatic veins.
Hemostatic Maneuvers for Severe Parenchymal Injury
Packing
Perihepatic packing has become the most widely used and successful method for management of severe liver injury. Laparotomy pads are packed around the liver, thus compressing the wound between the anterior chest wall, diaphragm, and retroperitoneum. This “damage control” laparotomy provides hemostasis while the patient is able to be hemodynamically optimized in the intensive care unit (ICU) as well as provides pressure on the wound to achieve hemostasis. Beal reported an 86% survival rate in 35 patients in whom perihepatic packing was used.71 In order to provide the tamponade necessary for effective packing, the surgeon must mobilize the liver by taking down the right and left triangular, coronary, and falciform ligaments. If, however, there is obvious hematoma in a ligament, this area should not be entered. Hematoma in the ligament may indicate a vena caval or hepatic vein injury and mobilization may lead to rapid exsanguination. The decision to pack must be made early in the exploration, in order to provide the best chances for patient survival.72 Indeed, early packing is associated with the increased survival of liver trauma patients. Richardson et al. found that the death rate associated with packing significantly decreased after 1989 and was linked to less packing time, as was demonstrated by a decrease in the average blood loss despite the equal severity of injury.33 One of the difficulties with packing comes with removal. Often, the bare liver area that has become hemostatic is now adherent to the packs. Pulling off the packs can then cause further bleeding. Different solutions to this problem have been described from wetting the gauze with saline on removal to a more innovative technique described by Feliciano and Pachter.73 They suggest placing a nonadherent plastic drape directly on top of the hepatic surface and then placing the laparotomy pads above this plastic interface. An important issue regarding abdominal packing is abdominal closure. These patients will undoubtedly require significant resuscitation. Abdominal compartment syndrome diagnosed with elevated bladder pressure (above 25 mm Hg), increasing peak airway pressures, decreased urine output, and abdominal distention can be a life-threatening consequence of this resuscitation (see Chapter 38). Abdominal compartment syndrome can be avoided in these patients by leaving the fascia and skin edges open and placing a temporary closure device over the open abdomen.
Packing is often useful in blunt, venous injury but cannot be expected to provide hemostasis in major arterial injury. Major arterial injury is often seen with penetrating trauma and therefore packing in penetrating bleeding may not be successful.
The timing of packing removal continues to be the subject of debate. Correction of coagulopathy, acidosis, and hypothermia can almost always be accomplished within 24–48 hours of packing. Intra-abdominal sepsis is a risk of prolonged packing. Krige et al. found that packs that remained for more than 3 days had an 83% incidence of developing perihepatic sepsis, whereas those left less than 3 days had a 27% chance of sepsis.74 A 1986 report found a 10.2% sepsis rate for patients who had packs removed within 24–48 hours along with complete clot evacuation and debridement of devitalized tissue.75 Caruso et al. advocate the removal of packs at 36–72 hours because they have experienced a higher rate of repacking for recurrent hemorrhage in the group of patients who had their packs removed earlier.72 Nicol et al. reported a significantly higher repacking rate in those hemodynamically stable patients whose packs were removed at 24 hours compared to those patients whose packs were removed after 48 hours.76Overall, it seems that pack removal prior to 72 hours, effective residual peritoneal clot evacuation, and excision of devitalized tissue will provide the optimal circumstance for minimizing perihepatic sepsis.
Direct Suture
Grade III and IV liver lacerations often do not respond to the more topical procedures listed for minor injury control. One of the oldest reported techniques to control deep parenchymal bleeding is direct suturing of the tissue with large, blunt-tipped 0-chromic suture. Utilizing a large blunt needle with 0 suture prevents the suture from tearing through Glisson’s capsule when tying. The stitches can be continuous or if a deeper laceration is encountered, a mattress configuration is preferred. This technique is most appropriate for lacerations less than 3 cm in depth. It is best to avoid the direct suture approach as blind passage of these large blunt needles may injure bile ducts and vascular structures thereby leading to possible intrahepatic hematomas or hemobilia.
Finger Fracture
More severe parenchymal laceration may involve larger branches of the hepatic artery or portal system and will not respond to the attempted tamponade with large parenchymal suturing. In these cases some clinicians prefer the technique of finger fracture (Fig. 29-9).70 The utilization of this technique involves careful extension of the laceration using finger fracture until bleeding vessels can be identified and then controlled with clips, ligation, or direct repair. This technique can lead to extensive additional parenchymal bleeding while searching for the initially damaged vasculature.
FIGURE 29-9 Hepatotomy with selective ligation is an important technique for controlling hemorrhage from deep (usually penetrating) lacerations. This technique includes finger fracture to extend the length and depth of the wound (A), division of vessels or ducts encountered (B), and repair of any injuries to major veins (C). (Reproduced with permission from Burch JM, Moore EE. Injuries to the liver, biliary tract, spleen, and diaphragm. In: Wilmore DW, ed. ACS Surgery: Principles & Practice. New York: WebMD Corporation; 2002.)
Omental Packing
Omental packing has been used successfully on its own as well as in conjunction with finger fracturing. It fills the potential dead space with viable tissue that also is a source of macrophage activity. Stone and Lamb original work was reinforced by Fabian and Stone when they managed to stop the venous hemorrhage of severe parenchymal laceration in 95% of patients with an 8% mortality.66,77 The technical aspects of this process include first mobilizing the greater omentum from the transverse mesocolon in the avascular plane. Next, the omentum is mobilized from the greater curvature preserving the usually right gastroepiploic vascular pedicle (Fig. 29-10). The tongue of omentum is then placed into the injury defect. The ability to achieve hemorrhage cessation with this method reiterates that most hepatic bleeding is venous. Tamponade with viable omental packing is superior to most of the direct techniques of hemorrhage control.
FIGURE 29-10 Omental mobilization employed for liver packing.
Penetrating Tract
Penetrating tracts through the hepatic tissue provide another challenge for the surgeon. Often these are of great depth and length, therefore making visualization of the entire injury impossible. Management of these injuries has included the packing of omentum into the tract for hemostasis. Also, devices such as the rolled cellulose-covered gelatin sponge can be inserted into the tract for hemostasis. Poggetti et al. advocate the use of balloon tamponade of the tract.78 A Penrose drain is placed over a hollow perforated tube and tied on both ends. The balloon is then placed into the tract and inflated with a contrast agent (Fig. 29-11). If successful tamponade has been achieved, the balloon is left in the abdomen and removed 24–48 hours later at a second laparotomy. A similar technique using a Foley balloon has been described.79 A size 16 Foley is inserted into the tract and inflated. If there is continued active bleeding, the catheter is moved back or forward and inflated again. If bleeding continues through the catheter but not out of tract, the balloon is proximal to the bleeder and needs to be repositioned deeper. If the bleeding continues from the tract orifice, then the balloon must be repositioned further out of the tract. Once the catheter is positioned, drains are placed in the area. The drains and catheter are brought out through the skin. The Foley can be removed after deflation produces no further signs of bleeding in 3–4 days or at the time of the next planned reexploration. Sengstaken–Blakemore tubes have also been used in these situations.80
FIGURE 29-11 A handmade balloon from a Robinson catheter and a Penrose drain may effectively control hemorrhage from a transhepatic penetrating wound. (Reproduced with permission from Burch JM, Moore EE. Injuries to the liver, biliary tract, spleen, and diaphragm. In: Wilmore DW, ed. ACS Surgery: Principles & Practice. New York: WebMD Corporation; 2002.)
Deep, small-diameter penetrating injury may continue to bleed from the depths of the wound. In these instances finger fracture of a significant liver segment may be necessary. Another alternative, considering institutional availability, may be angioembolization for these lesions.
Fibrin Sealants and Hemostatic Devices
Fibrin sealants have been a topic of much interest. Fibrin glue combines fibrinogen with thrombin, calcium chloride, and aprotinin to form a stable clot.81 However, difficulties have been found with the use of fibrin glue. Time required for preparing the glue, inability of the glue to stick to a bleeding surface, and hypotension with injection have led to minimal human use of the preparation. Fibrin sealants are currently undergoing clinical trials. The Modified Rapid Deployment Hemostat has been studied in trauma patients with severe liver injury. This device is made of a 1 cm layer of acetylated poly-N-acetyl glucosamine bonded to a 4 × 4 in gauze pad. Applying the bandage in direct contact with a bleeding vessel and placing direct pressure has been shown to decrease the bleeding substantially.82
Resection
Anatomic resections for severe hepatic trauma were often performed in the late 1960s and early 1970s. However, the 80% survival reported by McClelland and Shires has not been experienced in other reports.83 In fact, most series surmise that when anatomic resection is performed for massive bleeding, the mortality is prohibitively high. It has been stated that major lobar resection may be necessary in 10% of liver injuries but more than half of these have a mortal wound.71 Another report from Australia and Hong Kong gave the results of resection in 37 patients with major liver injury.46 This article suggests that better resectional results occur when trained hepatobiliary surgeons perform the resection. However, the results are less than compelling when actual injury grade, the low number of resections during first laparotomy, and morbidity are taken into account. Better results have been reported by Polanco et al.84 They report on 56 patients who underwent hepatic resection during their initial operation with a morbidity of 30% and a mortality of 17.8%. They recommend hepatic resection in the following scenarios: patients with massive bleeding related to a hepatic venous injury that must be repaired directly, massive destruction of hepatic tissue, and finally patients with a major bile leak from a proximal main intrahepatic bile duct. Successful outcome by crushing the injured liver segment between two aortic clamps has also been reported.85 The hemostatic clamps are left in place and 36 hours later the patient is brought back to the operating room where the now necrotic segment is easily removed and the liver edge oversewn.
Hepatic Artery Ligation
Hepatic arterial ligation can be a useful maneuver either in the operating room or with the aid of angiography. Complete selective hepatic artery ligation is used in only about 1% of patients with severe hepatic trauma.33 If a patient has a noticeable decrease in bleeding after the Pringle maneuver has been performed, hepatic artery ligation should be considered. Such instances occur in a scenario with a knife wound or small caliber missile wound with continued deep parenchymal bleeding. When the portal vein remains patent, the chance for severe hepatic dysfunction after hepatic artery ligation is minimal.86 However, with patients who have undergone significant hypoperfusion due to traumatic shock, hepatic artery ligation may produce enough further ischemia to produce necrosis or sepsis.87 One instance of hepatic failure following hepatic artery embolization resulted in the patient receiving an urgent liver transplant.88
Currently, most centers are advocating a multimodality approach to hepatic arterial bleeding. The role of interventional radiology has gained significant importance in the role of bleeding control after packing. Sclafani et al. in 1984 reported on the successful selective arterial embolization of severely injured liver parenchyma after packing.89 Angiography has become an important step in the management algorithm for severe liver injury. One report stated that an approach for high-grade liver injury includes “immediate surgery for control of life-threatening hemorrhage, the use of complex surgical techniques to address these injuries, the institution of early hepatic packing and immediate postoperative hepatic angiography and angioembolization.” In this report of 22 patients that had sustained grade IV and V injury, 15 underwent angiographic embolization (10 had been previously packed). All of these procedures were successful in arresting the continued bleeding.90 Therefore, it cannot be overemphasized that the care for severe hepatic parenchymal injury requires not only operative skill but also the judgment to determine the time for packing and collaboration with interventional radiology specialists (Fig. 29-12).
FIGURE 29-12 (A) Hepatic pseudoaneurysm. (B) Coiled hepatic pseudoaneurysm.
Hepatic Transplantation
Hepatic transplantation has been successfully reported. This is assuredly a drastic approach to traumatic injury and is an alternative for very few patients. The patient must have an overall excellent chance of survival with minimal concomitant injury, especially other intra-abdominal or neurologic injury. Also, if a trauma patient requires a transplant, it must be completed immediately; waiting for a donor organ to arrive is not an option. Case reports from Philadelphia and Miami describe successful transplantation.91,92 These cases present the requirements for possible transplantation as the patients had a single liver injury, no neurologic compromise, the achievement of hemodynamic stability, corrected coagulopathy, and the ability to obtain a donor organ within 36 hours of an anhepatic state.
Retrohepatic Vena Cava/Hepatic Vein Injury
Severe hepatic trauma can injure the vena cava anywhere along its extraparenchymal course. Also, damage to the hepatic veins can be extraparenchymal or intraparenchymal. Life-threatening bleeding from these injuries occurs if the supporting structures, mainly the suspensory ligaments, diaphragm, or liver parenchyma, are disrupted. Therefore, the exposure of a major venous injury may release the tamponade and result in free bleeding and exsanguination. As Buckman et al. outlined, there are three main strategies described to deal with these mortal injuries. The first is to directly repair the venous injury with or without vascular isolation. The second is with a lobar resection. The third is by using a strategy of tamponade and containment of the venous bleeding.93
Direct Venous Repair
Direct venous repair without shunting has been advocated by Pachter and Feliciano. They describe occlusion of the portal triad for a significant time, mobilization of the liver with medial rotation, and efficient finger fracture to the site of injury.94 With these methods they reported a 43% (6/14) survival. Chen et al. have published similar results of a 50% survival.95
Various shunting maneuvers have been attempted for complete vascular control of the liver. Schrock et al. first introduced the atriocaval shunt in 196896 (Fig. 29-13). The goal is to shunt the blood from the infrahepatic vena cava, bypassing the retrohepatic cava, and directing flow into the atria. This, along with the Pringle maneuver, is theoretically used to create a bloodless field. Unfortunately, of the approximately 200 cases published using atriocaval shunting, only at best 10–30% survive their injury.33 The caveats of this maneuver include the need to plan for the procedure essentially before proceeding with the operation. All the equipment must be ready and a thorocoabdominal exposure is necessary. Shunting a patient cannot be successfully accomplished if the patient has already had major blood loss, becomes coagulopathic, and has inadequate operative incisional exposure. Shunting, in general, is not often used at present. The patients who require shunting often have catastrophic injury in which time is of the essence. Therefore, more often these patients are packed urgently and brought to angiography for embolization for any hope of true survival.
FIGURE 29-13 (A) A hole is cut in the right atrial appendage above a 2-0 silk purse-string suture. A Satinsky clamp maintains vascular control. (B) Final position of No. 36 chest tube acting as an atriocaval shunt. Note the extra hole cut in the chest tube at the level of the right atrium. All holes in the chest tube are outside the umbilical tapes, thereby forcing blood from the lower half of the body and the kidneys through the shunt. (Reproduced with permission from Feliciano DV, Pachter HL. Hepatic trauma revisited. Curr Probl Surg. 1989;26:499.)
Other shunting procedures have been utilized as well. Pilcher et al., in 1977, reported on a balloon shunt introduced through the saphenofemoral junction.97 This occlusive method has had some anecdotal success and avoids emergent thoracotomy without destruction of the surrounding ligamentous tamponade.98 The multi-institutional trial results in 1988 however did not show any survival benefit of the balloon shunt versus the atriocaval shunt.99 Venovenous bypass has been used in some institutions as well.100 Again this method requires considerable planning but obviates the hemodynamic instability of caval occlusion and ligamentous disruption. Direct clamping techniques have also been used in a small number of patients. Carrillo et al. had success with vascular clamps placed on hepatic vein injured ends, filling the laceration with viable omentum, and packing with planned reoperation.101 Nicoluzzi et al. report using hepatic vascular exclusion without aortic cross-clamping.102 After fluid loading, vascular clamps are placed on the porta, infrahepatic suprarenal inferior vena cava, and the suprahepatic inferior vena cava. Once clamping is tolerated, a direct vessel repair is accomplished. In general, direct approaches to vein repair are difficult and can often result in a profuse uncontrolled bleeding situation, especially since even the most veteran surgeon has little experience in these uncommon injuries.
Anatomic Resection
As mentioned earlier, anatomic resection has resulted in a high mortality when carried out for traumatic bleeding. In certain circumstances when the dissection has already been done by the injury itself, resection for debridement may be indicated. However, current data do not promote anatomic resection for major venous injury unless direct repair is necessary.
Vena Cava Stenting
Endoluminal stent grafts are now available for many uses. Reports of using the fenestrated graft in blunt trauma have been reported.103 The graft used by Watarida et al. was homemade and stayed patent at the 16-month follow-up. More successful reports of commercially available fenestrated grafts used in retrohepatic vena caval injuries are surfacing. These grafts are being placed both after “damage control” laparotomy and prior to laparotomy when the lesion is seen on CT.104,105 Hommes et al. report the survival of a patient with intraoperative placement of an endovascular stent graft into the IVC for a juxtahepatic IVC injury with parenchymal and packing.106 Though these grafting procedures are not yet commonplace, a significant future for their use is apparent.
Tamponade with Containment
With the high mortality of direct venous repair and anatomic resection evident, the focus on severe vascular injury management has shifted to methods of tamponading and containing venous injury in addition to embolization of arterial bleeding. Many of the methods utilized for severe parenchymal injury are also effective for large venous injury. In Memphis, the mortality of patients with juxtahepatic venous injuries who were treated with omental packing was a low 20.5%.107 Another article emphasizing packing included 14 patients with hepatic vein injury and 6 patients with retrohepatic vena caval injury with an overall mortality of only 14%.71 Cue et al. depict four patients with retrohepatic vena cava, hepatic vein injury, or both who underwent initial packing and survived.108
At this time it seems that the most successful method of managing severe retrohepatic or hepatic venous injury is by using tamponade and containment. Direct repair of damaged vessels continues to have a very high morbidity even in the most experienced hands. Resection also has shown itself to be a morbid alternative with the survival data primarily being in the hands of experienced hepatobiliary surgeons in somewhat stable patients. Overall, the best approach to severe liver injury includes (a) expedient recognition and operative intervention of unstable hemorrhaging patients, (b) mobilization of the liver ligaments not directly involved with hematoma to better visualize the injury, (c) placement of a viable omental tongue into parenchymal defects, (d) rapid determination of the need for gauze packing when direct surgical maneuvers fail, and (e) angiographic embolization of hepatic arterial injured branches when ongoing hemorrhage or CT blush is seen.
Drains
The use of closed-suction drains has clearly been proven to be superior over Penrose drain use in a number of publications. A 1991 study reported a perihepatic abscess rate of 6.7% with no drain, 3.5% with closed suction, and 13% with Penrose drainage.107 A study from Charity Hospital found an abscess rate of 1.8% in those with no drainage, 0% abscess rate in those with closed suction, and 8.3% abscess rate in those with open drains.109 Examination of these figures, however, indicates no significant difference in abscess rate between the no drainage group versus the closed-suction cohort. Indeed, in a review of 161 significant liver injuries, 78 patients underwent closed-suction drainage and 83 were left without a drain.110 The injury grade, blood loss, shock, specific injuries severity, and associated injuries were similar in the two groups. There was no difference in mortality, abscess formation, or biliary fistula between the two groups. Thereby, the study concluded that drainage should be done only in injuries with obvious bile leaks noted at the time of laparotomy. This viewpoint is reiterated in a 1988 article stating that the presence of hypotension and multiple transfusions are more predictive of abscess formation than drain placement.109 With the current use of interventional radiology techniques, routine drainage has become less of an issue. Most centers will treat patients expectantly and only place drains in patients with obvious bile leaks. If indeed a collection or abscess develops, many can be dealt with by percutaneous tube placement under radiologic guidance.
Complications of Operative Management
Bleeding
Postoperative hemorrhage is not a common occurrence. Most series quote a 2–7% hemorrhagic complication rate.73,99 Falling serial hematocrits, increasing abdominal distention, and episodes of hypotension or tachycardia mark continued bleeding. The inability to operatively control the bleeding is often confounded by hypothermia and coagulopathy. In the past these patients were urgently returned to the operating room after correction of their coagulopathy and rewarming. Currently, after resuscitation these patients can often be managed with angiographic localization of the bleeding source and embolization. Of course, unstable patients do need operative intervention and may in fact need reexploration of previously packed areas in order to specifically identify the bleeding source. The areas of bleeding must then be addressed by utilizing the previously discussed maneuvers of severe injury control.
Abdominal Compartment Syndrome
Abdominal compartment syndrome may develop with packing and continued fluid requirements in these severely ill patients. Packing has been shown to even cause pulmonary embolism from venous stasis caused by infrahepatic vena caval compression.111 The physician must be vigilant in his or her care and resuscitation of these patients.
Hemobilia
Immediate attention should be given to a patient who develops a significant upper gastrointestinal bleed following liver repair. Many times this is the only symptom that can point to the development of hemobilia. The often-mentioned signs and symptoms of hemobilia—jaundice, right upper quadrant pain, and falling hematocrit—are common occurrences in most patients after severe liver injury and therefore make the diagnosis of hemobilia difficult. The incidence of hemobilia ranges anywhere from 0.3% to 1.2%.49,112 The presentation may be days to weeks postinjury. Upper endoscopy and bleeding scans are generally unable to locate the source of bleeding. Angiography will frequently delineate a pseudoaneurysm and accomplish embolization of the damaged vessel.113,114 Operative debridement and drainage may be necessary if a large cavity has formed or sepsis is apparent.49
Bilhemia
Biliovenous fistulas have also been described by Clemens and Wittrin in the literature but are quite rare.115 This entity occurs as the bilious venous blood dissolves in the bloodstream and is carried directly to the right heart. Therefore, one sees a patient with a drastically rising bilirubin with relatively normal liver function tests. Glaser et al. discussed three cases of bilhemia, which were identified by ERC.116 The management of these cases involved a left hemihepatectomy in the first, spontaneous resolution in the second, and controlled biliary fistula in the last. Another method of control included placement of a constant suction T-tube with subsequent resolution.117 Although spontaneous resolution has occurred, this entity can have a high mortality if left unaddressed.
Biliary Fistulae
Biliary fistulae are one of the complications that a surgeon is likely to encounter. Biliary fistula can account for up to 22.5% of traumatic liver management complications.90 Overall biliary fistulae seem to occur in about 4–6% of patients who undergo operative management of severe liver injury.118,119 Some bile duct injuries are obvious intraoperatively with significant bile staining and a visible disrupted bile duct. Many persistent fistulae may, however, manifest from smaller radicals, which retract into the liver parenchyma and are not visualized. Operative drain placement is advocated in liver injury with obvious bile staining. It is common for liver injuries to have transient early postoperative serosanguinous and bilious drainage. Bilious drainage of at least 50 mL per day that continues after 2 weeks is considered a biliary fistula.99 Also, persistent earlier drainage of over 300–400 mL of bile a day should be cause for further evaluation.
The diagnosis of a biliary injury can be done by a fistulogram if a drain is in place, HIDA scan (though not anatomically exact), MRCP, or ERC. Major left or right bile duct injury often requires further intervention for closure. In the past the surgical approach was recommended with resection or Roux-en-Y procedures predominating. More recently, nonoperative approaches have proven successful. Percutaneous stenting of injuries and drainage of biloma collections has been utilized.120 Also, many reports are surfacing of management using ERC sphincterotomy with stenting and percutaneous drainage of biloma. One study described five patients with intrahepatic bile duct injuries.121 The injuries included left main hepatic duct, right second-order bile duct, and more peripheral lesions. All were successfully managed nonoperatively. Repeat ERC of these patients led the authors to conclude that “therapeutic ERC and percutaneous interventional radiology can both treat the complication of the ductal injury and allow healing of the ductal disruption.” Confirmation of healing of major ductal injury after ERC stenting and percutaneous drainage has been documented.90 For bile fistulae that do not involve a main bile duct, drainage alone will provide adequate treatment and other maneuvers are rarely necessary.
Hepatic Necrosis
Major hepatic necrosis can be a complication of the multimodality management of severe liver injury. Dabbs et al. found that 29 of 30 patients that they encountered with major hepatic necrosis underwent initial operative intervention.122 Many of the patients then had embolizations performed making their risk of major hepatic necrosis between 65% and 68%. A large number of these patients then required resection of their necrotic hepatic parenchyma.
Other Fistulae Problems
Thoracobiliary fistulae are also encountered with traumatic liver injury. Though it is a rare complication, identification and management can prevent morbidity of progression to bronchobiliary fistula. Many of these injuries occur after thorocoabdominal penetrating injury. Often the patient does well initially with resolution of hemothorax, no evidence of jaundice, and stabilization of liver injury only to become significantly tachypneic a week or more later. One report described the treatment of a thoracobiliary fistula with chest tube drainage and ERC.123 One patient returning for routine follow-up was operatively managed with thoracic and abdominal drains and diagnosed from an abnormal chest x-ray and subsequent CT. Rothberg et al. promote operative intervention in order to evaluate for significant diaphragmatic injury, liver necrosis, or lung necrosis with possible bronchial involvement.124
Penetrating injury can potentially provide a means for many severe fistula communications. Pleurocaval fistula may result from thorocoabdominal injury. This fistula may be the source of life-threatening air embolism.125Arterioportal fistula are associated with initial hemorrhage and subsequent portal hypertension.126 One case report described a GSW that formed a left hepatic artery to portal vein fistula. This fistula was able to be successfully managed by interventional radiology embolization. Portosystemic venous shunts have also been reported in severe blunt liver injury.127
Traumatic Extrahepatic Biliary Tract Injury
Overview
Extrahepatic biliary and portal triad injuries make up only about 0.07–0.21% of all trauma admissions at Level I trauma centers.128,129 Though these injuries are rare, their evaluation and management prove difficult. Technical problems including continued hemorrhage, adjacent organ injury, and small duct size can prove insurmountable. A timely diagnosis and treatment method may prove to be the survival difference in patients with these severe injuries. In a Seattle paper, 38% were a result of blunt mechanisms, similar to the 31% with blunt mechanism quoted in a 1995 multi-institutional trial.128,129 Injury to this area carries an overall 50% mortality, with vascular injury (portal vein or hepatic artery) being the most morbid. When examining those with both portal vein and hepatic artery injury, the mortality is 99%. It is evident that the management of these injuries is a significant challenge (see Chapter 34). Most street weapons are now of high caliber and medium to high velocity. These weapons usually do not result in simple, single injury. Instead, multiple injuries to the liver, porta, vena cava, and surrounding viscera most often occur. Not only are these portal triad injuries difficult to manage, but also the specific injury cannot be identified preoperatively and, therefore, intraoperative decision making is crucial.
Injury Types and Diagnostics
Gallbladder
Gallbladder injury accounts for up to 66% of extrahepatic biliary tract injuries.128 Injury can be from either blunt or penetrating mechanisms. Blunt injury often involves avulsion, either partial or complete, contusion, or perforation. Penetrating injury has been seen involving everything from the body of the gallbladder down to the cystic duct. A review from 1995 warned that 100% of 22 cases of blunt gallbladder injuries were associated with other intra-abdominal trauma; however, this is not uniformly reported and isolated gallbladder injury is encountered.130 Therefore, though a patient may present with an isolated gallbladder injury, the surgeon must carefully rule out further intra-abdominal trauma. A trauma patient may also manifest a gallbladder injury as a result of a significant contusion. Blood in the gallbladder can cause stasis and blockage of the cystic duct, which may present as acute cholecystitis.131
Gallbladder injury is successfully evaluated by CT (Fig. 29-14). The findings of an ill-defined contour of the wall, collapse of the lumen, or intraluminal hemorrhage highly suggest blunt gallbladder injury.132Patients may also present with bile peritonitis and right upper quadrant pain. Ultrasound examination in gallbladder injury has not been formally evaluated but intuitively should provide useful information about this injury. Despite these diagnostic methods, the diagnosis of gallbladder injury is most often secured at laparotomy.
FIGURE 29-14 CT scan revealing a distended gallbladder filled with blood (dark arrow) in a patient with blunt abdominal trauma and virtually no peritoneal signs.
Bile Duct
Bile duct injury is most often encountered in penetrating injury.128 Blunt ductal injury is most likely to happen where the bile duct is fixed to its surroundings, for example, the pancreaticoduodenal junction.133In a multi-institutional trial it was found that blunt injuries were predominately a complete transaction, whereas penetrating injuries were partial 75% of the time.128
Extrahepatic bile duct injuries are evident in two distinct settings: first, at the time of laparotomy for a patient in shock with other severe liver, vascular, pancreatic, or duodenum injury; second, in a late presentation often more than 24 hours and up to 6 weeks after the original injury time. The patients with late presentation may develop jaundice, abdominal distention and pain, intolerance to enteral feeding, fever, or worsening base deficit due to bilious ascitis or infection.129
Evaluation of the stable patient with CT scan or ultrasound in the acute setting will not be able to differentiate abdominal blood with biliary leak. There may be some indication of pancreatic head fullness, duodenal thickening, or portal edema but these are nonspecific findings. In the presence of bile staining during an operative procedure and no obvious injury, a cholangiogram through the gallbladder can be helpful.133 DPL has also shown a lack of specificity for biliary injury as duodenal, small bowel, and liver injuries may produce bile.134 Also, the small amount of bile may be obscured by the presence of blood in the peritoneum with the DPL. Late presenters of bile duct injury cannot be recognized until symptoms are apparent. At that time CT, ultrasound, or ERC can be used to visualize bile collections and localize injury.135
Management of Extrahepatic Biliary Injuries
General Considerations
Patients with portal triad and extrahepatic biliary injuries usually arrive in shock. Therefore, the tenets described for major liver injury apply to portal injury as well. A midline incision should be made. Evacuation of blood clots and hemoperitoneum with urgent packing of the bleeding portions should be completed. The patient should be resuscitated and coagulopathy correction initiated by the anesthesia team. Hematoma or bleeding around or within the hepatoduodenal ligament or severe parenchymal injury leading to the porta hepatis should raise suspicion of a portal triad injury. Bile staining should also be fully investigated as 12% of bile duct injuries may be missed at the initial operation.133 The Pringle maneuver may be helpful in decreasing the inflow to a portal injury. In order to obtain adequate examination and exposure for repair, a wide Catell maneuver should be performed, which includes mobilizing the ascending and hepatic flexure areas of the colon, thus exposing the duodenum completely to the head of the pancreas and inferior vena cava.
Gallbladder
Isolated gallbladder injury is most often managed with open cholecystectomy. However, there have been reports of laparoscopic cholecystectomy in penetrating trauma.136 This procedure should be done with great reserve since many gallbladder injuries are associated with other intra-abdominal injury in both penetrating and blunt trauma. Though the laparoscope can give a good superficial exam of the peritoneal cavity, visualization of the duodenum, pancreas, and porta is in most hands not sufficient. Minor gallbladder contusions can often be managed nonoperatively.137 This may lead to cholecystitis or delayed rupture if hematoma is present. Historically, it had been suggested that simple lacerations should undergo cholecystorraphy with absorbable suture.138 Cholecystorraphy, however, remains a rare procedure, as small gallbladder lacerations are rarely encountered and cholecystectomy can be rapidly performed. Cholecystectomy should also be performed on all patients with injury to the cystic duct or right hepatic artery that would eliminate the blood supply to the gallbladder.
Bile Duct
Bile duct injury should be addressed after hemorrhage has been controlled. In the patient who remains in shock and coagulopathic, packing and placement of a Jackson-Pratt drain in the area of biliary injury is adequate until reexploration is performed. In the somewhat more stable patient who is becoming coagulopathic, a small T-tube placed in the injured duct will provide adequate drainage until a formal repair can be accomplished.139 With a partial transection of a right or left hepatic duct, insertion of a small T-tube into the common hepatic duct with a long limb traversing the partially transected area even without suturing may provide enough support for full healing.140
For the stable patient definitive repair is preferred at the first operation. Four broad categories of biliary duct injury have been described: (a) avulsion of cystic duct or small laceration, (b) transection without loss of tissue, (c) extensive defect in the wall, and (d) segmental loss of ductal tissue.140
Avulsions and small lacerations in the duct can be repaired primarily with 6-0 polyglycolic suture making sure not to narrow the lumen. A T-tube with a limb under the repair can be used; however, this may be difficult to insert in a patient with a normally small duct. The techniques used to place a T-tube may also devascularize an already compromised duct. Therefore, the authors will not place a T-tube in primary repair. For avulsions in which primary repair may narrow the lumen, a piece of the cystic duct or proximal gallbladder wall can be used for the repair.141
Penetrating injury very occasionally results in a transection of the bile duct without significant tissue loss. In these instances an end-to-end anastomosis can be performed. One must be sure to perform minimal dissection around the duct or the lacerated ends in order to maintain adequate blood supply for healing. Tension on the anastomosis will most certainly lead to stricture. Ivatury et al. reported a 55% stricture rate in the end-to-end anastomosis that then required enteric conversion.142 Similarly, Stewart and Way had initial success in 67% of patients initially managed with Roux-en-Y for complete laceration following laparoscopic cholecystectomy with failure in all lacerations treated with end-to-end anastomosis.143
Extensive wall defects and segmental tissue loss require biliary-enteric anastomosis (Fig. 29-15). In the past many methods of “patching” were attempted. Saphenous vein grafts have had difficulties with shrinking and fibrosis, which then required stenting.144 Prosthetic patches and jejunal mucosal patches have also been tried with anecdotal success only.145
FIGURE 29-15 Roux-en-Y choledochojejunostomy. Anastomosis is performed in a one-layer fashion. The T-tube is brought out through a separate proximal stab wound. The gallbladder has been removed.
Deciding which type of biliary-enteric anastomosis to perform depends on the injury location, access, and size. Roux-en-Y hepaticojejunostomy with cholecystectomy and T-tube drainage is the most utilized approach to complex injury. The retrocolic Roux limb is at least 40 cm long and can be brought up to the common hepatic duct or even to the hilar plate similar to the Kasai procedure. An avulsion of the hepatic ducts at the bifurcation can be managed by suturing the ducts together medially before the end-to-side hepaticojejunostomy.146 If the distal common duct is not found due to its retraction behind the pancreas, drainage of the area may be all that is necessary.134 Roux-en-Y choledochojejunostomy with cholecystectomy and T-tube drainage is also useful for the management of common bile duct injury. However, the vascularity in this anastomosis is crucial and any sign of common bile duct vascular injury would lead the surgeon to construct an anastomosis closer to the common hepatic duct. Cholecystojejunostomy and ligation of the very distal common bile duct is a possibility if intraoperative cholangiography reveals a patent cystic duct. This is a viable option especially in patients with small caliber ducts or instability.
Blunt distal hepatic duct injury is rare. However, the surgical treatment of these injuries must be individualized to each situation. Both the right and left hepatic duct injuries have been reported.147,148 Biliary-enteric anastomosis are sometimes possible right at the hilar plate; however, if the repair is difficult, ligation of a left or right duct has been reported to lead to atrophy of the involved lobe, not biliary cirrhosis.149
Stenting in biliary anastomosis is a controversial topic. Surgeons in favor of stenting report that stenting allows for decompression, when edema post-trauma may be significant, as well as allows access for cholangiography. T-tubes must exit the duct outside of the repair area or stricture will result. Enteric stents are not necessary and some surgeons feel comfortable without their use, stating that a foreign body in an already small duct may promote stricture or obstruction.150 Morbidity data cannot support a definitive answer for or against stenting and therefore a stent must be used at the discretion of the surgeon, taking each situation separately.
When ampullary or intrapancreatic bile duct injury is discovered, a pancreaticoduodenectomy may be appropriate if duodenal and pancreatic injury is also seen. An isolated ampullary primary repair or reimplantation may be possible. The authors have repaired an ampullary injury by performing a transduodenal sphincteroplasty and primary repair of the ductal injury. Hepatic resection is necessary only in the case of combination injury to the liver parenchyma and hepatic duct traversing that segment.140
The major complications associated with biliary duct injury are fistula and stricture. A fistula may be able to be nonoperatively managed with drainage. Persistent fistula may require reexploration. Strictures may present with recurrent cholangitis or biliary cirrhosis. Stenting by endoscopists has become frequent; however, long-term results are not conclusive. A recent publication used an aggressive technique of placing an increasing number of stents until complete disappearance of the biliary stricture occurred. Though the authors did have a complication rate of 9%, their mean duration of treatment was 12 months with a 48.8-month stricture-free interval posttreatment thus far.151Conversely, Johns Hopkins reported their experience with operative management of all postoperative bile duct strictures and had a 98% success rate.152
REFERENCES
1. Croce MA, Fabian TC, Menke PG. Nonoperative management of blunt hepatic trauma is the treatment of choice for hemodynamically stable patients. Results of a prospective trial. Ann Surg. 1995;221:744.
2. Moynihan BG. Abdominal Operations. Philadelphia: WB Saunders; 1905:473.
3. Tilton B. Some considerations regarding wounds of the liver. Ann Surg. 1905;41(1):27.
4. Pringle JH. Notes on the arrest of hepatic hemorrhage due to trauma. Ann Surg. 1908;48:541.
5. Madding GF, Lawrence KB, Kenney PA. Forward surgery of the severely injured. Second Aux Surg Group. 1942;1:307.
6. Couinaud C. Les envelopes vasculo-biliares de foie ou capsule de Glisson: leur interet dons la chirurgie vesiculaire, les resections hepatiqes et l’abord du hile du foie. Lyon Chir. 1954;49:589.
7. Moore EE, Cogbill TH, Jurkovitch GJ, et al. Organ injury scaling—spleen, liver (1994 revision). J Trauma. 1995;38:323.
8. Olsen WR, Redman HC, Hildreth DH. Quantitative peritoneal lavage in blunt abdominal trauma. Arch Surg. 1972;104:536.
9. Root HD, Hauser CW, McKinley CR, et al. Diagnostic peritoneal lavage. Surgery. 1965;57:633.
10. Richards JR, McGahan JP, Pali MJ, et al. Sonographic detection of blunt hepatic trauma: hemoperitoneum and parenchymal patterns of injury. J Trauma. 1999;47:1092.
11. Rozycki GS, Ochsner MG, Feliciano DV, et al. Early detection of hemoperitoneum by ultrasound examination of the right upper quadrant: a multicenter study. J Trauma. 1998;45:878.
12. Sirlin CB, Casola G, Brown MA, et al. Patterns of fluid accumulation on screening ultrasonography for blunt abdominal trauma. J Ultrasound Med. 2001;20:351.
13. Branney SW, Moore EE, Cantrill SV, et al. Ultrasound based key clinical pathway reduces the use of hospital resources for the evaluation of blunt abdominal trauma. J Trauma. 1997;42:1086.
14. Rozycki GS, Ballard RB, Feliciano DV, et al. Surgeon-performed ultrasound for the assessment of truncal injuries: lessons learned from 1540 patients. Ann Surg. 1998;228:557.
15. Chiu WC, Cushing BM, Rodriguez A, et al. Abdominal injuries without hemoperitoneum: a potential limitation of focused abdominal sonography for trauma. J Trauma. 1997;42:617.
16. Ballard RB, Rozycki GS, Newman PG, et al. An algorithm to reduce the incidence of false-negative FAST examinations in patients at high risk for occult injury. Focused Assessment for the Sonographic Examination of the Trauma patient. J Am Coll Surg. 1999;189:145.
17. Valentino M, Serra C, Zironi G, et al. Blunt abdominal trauma: emergency contrast-enhanced sonography for detection of solid organ injuries. Am J Roentgenol. 2006;186:1361.
18. McGahan JP, Horton S, Gerscovich EO, et al. Appearance of solid organ injury with contrast-enhanced sonography in blunt abdominal trauma: preliminary experience. Am J Roentgenol. 2006;187:658.
19. Catalano O, Sandomenico F, Raso MM, et al. Real-time, contrast-enhanced sonography: a new tool for detecting active bleeding. J Trauma. 2005;59:933.
20. Udobe KF, Rodriguez A, Chiu WC, et al. Role of ultrasonography in penetrating abdominal trauma: a prospective clinical study. J Trauma. 2001;50:475.
21. Soffer D, McKenney MG, Cohn S, et al. A prospective evaluation of ultra-sonography for the diagnosis of penetrating torso injury. J Trauma. 2004;56:953.
22. Murphy JT, Hall J, Provost D. Fascial ultrasound for evaluation of anterior abdominal stab wound injury. J Trauma. 2005;59:843.
23. Bokhari F, Nagy K, Roberts R, et al. The ultrasound screen for penetrating truncal trauma. Am Surg. 2004;70:316.
24. Federle MP, Goldberg HI, Kaiser JA, et al. Evaluation of abdominal trauma by computed tomography. Radiology. 1981;138:637.
25. Poletti PA, Mirvis SE, Shanmuganathan K, et al. Blunt abdominal trauma patients: can organ injury be excluded without performing computed tomography? J Trauma. 2004;57:1072.
26. Phillips T, Sclafani SJ, Goldstein A, et al. Use of the contrast-enhanced CT enema in the management of penetrating trauma to the flank and back. J Trauma. 1986;26:593.
27. Fabian TC, Croce MA, Stewart RM, et al. A prospective analysis of diagnostic laparoscopy in trauma. Ann Surg. 1993;217:557.
28. Zantut LF, Ivatury RR, Smith RS, et al. Diagnostic and therapeutic laparoscopy for penetrating abdominal trauma: a multicenter experience. J Trauma. 1997;42:825.
29. Fabiani P, Iannelli A, Mazza D, et al. Diagnostic and therapeutic laparoscopy for stab wounds of the anterior abdomen. J Laparoendosc Adv Surg Tech A. 2003;13(5):309.
30. Hurtuk M, Reed RL, Esposito TJ, et al. Trauma surgeons practice what they preach: the NTDB story on solid organ injury management. J Trauma. 2006;61:243.
31. Coimbra R, Hoyt DB, Engelhart S, et al. Nonoperative management reduces the overall mortality of grades 3 and 4 blunt liver injuries. Int Surg. 2006;91(5):251.
32. Malhotra AK, Fabian TC, Croce MA, et al. Blunt hepatic injury: a paradigm shift from operative to nonoperative management in the 1990s. Ann Surg. 2000;231:804.
33. Richardson JD, Franklin GA, Lukan JK, et al. Evolution in the management of hepatic trauma: a 25-year perspective. Ann Surg. 2000;232:324.
34. Pachter HL, Knudson MM, Esrig B, et al. Status of nonoperative management of blunt hepatic injuries in 1995: a multicenter experience with 404 patients. J Trauma. 1996;40:31.
35. Ciraulo DL, Luk S, Palter M, et al. Selective hepatic arterial embolization of grade IV and V blunt hepatic injuries: an extension of resuscitation in the non-operative management of traumatic hepatic injuries. J Trauma. 1998;45:353.
36. Misselbeck TS, Teicher EJ, Cipolle MD, et al. Hepatic angioembolization in trauma patients: indications and complications. J Trauma. 2009;67:769.
37. Kozar RA, Moore JB, Niles SE, et al. Complications of nonoperative management of high-grade blunt hepatic injuries. J Trauma. 2005;59:1066.
38. Kozar RA, Moore FA, Cothren CC, et al. Hepatic-related morbidity associated with nonoperative management of complex blunt hepatic injuries: AAST multicenter trial. In: Sixty-Fourth Meeting of the American Association for the Surgery of Trauma; 2005; Atlanta, GA.
39. Kelly MD, Armstrong CP, Longstuff A. Characterization of biliary injury from blunt liver trauma by MRCP: case report. J Trauma. 2008;64:1363.
40. Bridges A, Wilcox MC, Varadarajulu S. Endoscopic management of traumatic bile leaks. Gastrointest Endosc. 2007;65:1081.
41. Marks JM, Ponsky JL, Shillingstad RB, et al. Biliary stenting is more effective than sphincterotomy in the resolution of biliary leaks. Surg Endosc. 1998;12:291.
42. Miyayama S, Matsui O, Taki K, et al. Bile duct disruption after blunt hepatic trauma: treatment with percutaneous repair. J Trauma. 2006;60:640.
43. Griffen M, Ochoa J, Boulanger BR. A minimally invasive approach to bile peritonitis after blunt liver injury. Am Surg. 2000;66:309.
44. Gates JD. Delayed hemorrhage with free rupture complicating the nonsurgical management of blunt hepatic trauma: a case report and review of the literature. J Trauma. 1994;36:572.
45. Anderson IB, Al Saghier M, Kneteman NM, et al. Liver trauma: management of devascularization injuries. J Trauma. 2004;57:1099.
46. Strong RW, Lynch SV, Wall DR, et al. Anatomic resection for severe liver trauma. Surgery. 1998;123:251.
47. Smadja C, Traynor O, Blumgart LH. Delayed hepatic resection for major liver injury. Br J Surg. 1982;69:361.
48. Pollack CV. Hemobilia presenting as lower gastrointestinal hemorrhage without pain or jaundice: a case report. J Miss State Med Assoc. 1990; 31:1.
49. Croce MA, Fabian TC, Spiers JP, et al. Traumatic hepatic artery pseudoaneurysm with hemobilia. Am J Surg. 1994;168:235.
50. Franklin GA, Richardson JD, Brown AL, et al. Prevention of bile peritonitis by laparoscopic evacuation and lavage after nonoperative treatment of liver injuries. Am Surg. 2007;73:611.
51. Letoublon C, Chen Y, Arvieux C, et al. Delayed celiotomy or laparoscopy as part of the nonoperative management of blunt hepatic trauma. World J Surg. 2008;32:1189.
52. Pearl LB, Trunkey DD. Compartment syndrome of the liver. J Trauma. 1999;47:796.
53. Markert DJ, Shanmuganathan K, Mirvis SE, et al. Budd Chiari syndrome resulting from intrahepatic IVC compression secondary to blunt hepatic trauma. Clin Radiol. 1997;52:384.
54. Cuff RF, Cogbill TH, Lambert PJ. Nonoperative management of blunt liver trauma: the value of follow-up abdominal computed tomography scans. Am Surg. 2000;66:332.
55. Ciraulo DL, Nikkanen HE, Palter M, et al. Clinical analysis of the utility of repeat computed tomographic scan before discharge in blunt hepatic injury. J Trauma. 1996;41:821.
56. Cox JC, Fabian TC, Maish GO III, et al. Routine follow-up imaging is unnecessary in the management of blunt hepatic injury. J Trauma. 2005;59:1175.
57. Dulchavsky SA, Lucas CE, Ledgerwood AM, et al. Efficacy of liver wound healing by secondary intent. J Trauma. 1990;30:44.
58. Nance FC, Cohn I. Surgical judgement in the management of stab wounds of the abdomen: a retrospective and prospective analysis based on a study of 600 stabbed patients. Ann Surg. 1969;170:569.
59. Renz RM, Feliciano DV. Gunshot wounds to the liver. A prospective study of selective nonoperative management. J Med Assoc Ga. 1995;84:275.
60. Demetriades D, Gomez H, Chahwan S, et al. Gunshot injuries to the liver: the role of selective nonoperative management. J Am Coll Surg. 1998;188:343.
61. DuBose J, Inaba K, Teixeira PG, et al. Selective non-operative management of solid organ injury following abdominal gunshot wounds. Injury. 2007;38:1084.
62. Demetriades D, Hadjizacharia P, Constantinou C, et al. Selective nonoperative management of penetrating abdominal solid organ injuries. Ann Surg. 2006;244:620.
63. Shanmuganathan K, Mirvis SE, Chiu WC, et al. Penetrating torso trauma: triple-contrast helical CT in peritoneal violation and organ injury: a prospective study in 200 patients. Radiology. 2004;231:775.
64. Velmahos GC, Constantinou C, Tillou A, et al. Abdominal computed tomographic scan for patients with gunshot wounds to the abdomen selected for nonoperative management. J Trauma. 2005;59:1155.
65. Omoshoro-Jones JA, Nicol AJ, Navsaria PH, et al. Selective nonoperative management of liver gunshot injuries. Br J Surg. 2005;92:890.
66. Stone HH, Lamb JM. Use of pedicled omentum as an autogenous pack for control of hemorrhage in major injuries of the liver. Surg Gyn Obstet. 1975;141:92.
67. Sheldon G, Rutledge R. Hepatic trauma. Adv Surg. 1989;22:179.
68. Man K, Fan ST, Ng IO, et al. Prospective evaluation of Pringle maneuver in hepatectomy for liver tumors by a randomized study. Ann Surg. 1997;226:704.
69. Feliciano DV, Mattox KL, Jordan GL, et al. Management of 1000 consecutive cases of hepatic trauma (1979–1984). Ann Surg. 1986;204:438.
70. Pachter HL, Spencer FC, Hofstetter SR, et al. Significant trends in the treatment of hepatic trauma: experience with 411 injuries. Ann Surg. 1992;215:492.
71. Beal SL. Fatal hepatic hemorrhage: an unresolved problem in the management of complex liver injuries. J Trauma. 1990;30:163.
72. Caruso SM, Battistella FD, Owings JT, et al. Perihepatic packing of major liver injuries. Arch Surg. 1999;134:958.
73. Feliciano DV, Pachter HL. Hepatic trauma revisited. Curr Probl Surg. 1989;26:453.
74. Krige JE, Bornman PC, Terblanche J. Therapeutic perihepatic packing in complex liver trauma. Br J Surg. 1992;79:43.
75. Feliciano DV, Mattox KL, Birch JM. Packing for control of hepatic hemorrhage: 58 consecutive patients. J Trauma. 1986;26:738.
76. Nicol AJ, Hommes M, Primrose R, et al. Packing for control of hemorrhage in major liver trauma. World J Surg. 2007;31:569.
77. Fabian TC, Stone HH. Arrest of severe liver hemorrhage by an omental pack. South Med J. 1980;73:1487.
78. Poggetti RS, Moore EE, Moore FA, et al. Balloon tamponade for bilobar transfixing hepatic gunshot wounds. J Trauma. 1992;33:694.
79. Demetriades D. Balloon tamponade for bleeding control in penetrating liver injuries. J Trauma. 1998;44:538.
80. Ozdogan M, Ozdogan H. Balloon tamponade with Sengstaken–Blakemore tube for penetrating liver injury: case report. J Trauma. 2006;60:1122.
81. Kram HB, Reuben BI, Fleming AW, et al. Use of fibrin glue in hepatic trauma. J Trauma. 1988;28:1195.
82. King DR, Cohn SM, Proctor KG, et al. Modified rapid deployment hemostat bandage terminates bleeding in coagulopathic patients with severe visceral injuries. J Trauma. 2004;57:756.
83. McClelland RN, Shires T. Management of liver trauma in 259 consecutive patients. Ann Surg. 1965;161:248.
84. Polanco P, Leon S, Pineda J, et al. Hepatic resection in the management of complex injury to the liver. J Trauma. 2008;65:1264.
85. Ginzburg E, Klein Y, Sutherland M, et al. Prolonged clamping of the liver parenchyma: a salvage maneuver in exsanguinating liver injury. J Trauma. 2004;56:922.
86. Aaron WS, Fulton RL, Mays ET. Selective ligation of the hepatic artery for trauma of the liver. Surg Gynecol Obstet. 1975;141:187.
87. Lucas CE, Ledgerwood AM. Liver necrosis following hepatic artery transection due to trauma. Arch Surg. 1978;113:1107.
88. Anderson IB, Kortbeek JB, Al-Saghier M, et al. Liver transplantation in severe hepatic trauma after hepatic artery embolization. J Trauma. 2005;58:848.
89. Sclafani SJ, Shaftan GW, McAuley J, et al. Interventional radiology in the management of hepatic trauma. J Trauma. 1984;24:256.
90. Asensio JA, Demetriades D, Chahwan S, et al. Approach to the management of complex hepatic injuries. J Trauma. 2000;48:66.
91. Angstadt J, Jarrell B, Moritz M, et al. Surgical management of severe liver trauma: a role for liver transplantation. J Trauma. 1989;29:606.
92. Delis SG, Bakoyiannis A, Selvaggi G, et al. Liver transplantation for severe hepatic trauma: experience from a single center. World J Gastroenterol. 2009;15(13):1641.
93. Buckman RF, Miraliakbari R, Badellino MM. Juxtahepatic venous injuries: a critical review of reported management strategies. J Trauma. 2000;48:978.
94. Pachter HL, Feliciano DV. Complex hepatic trauma. Surg Clin North Am. 1996;76:763.
95. Chen RJ, Fang JF, Lin BC, et al. Surgical management of juxtahepatic venous injuries in blunt hepatic trauma. J Trauma. 1995;38:886.
96. Schrock T, Blaisdell W, Mathewson C. Management of blunt trauma to the liver and hepatic veins. Arch Surg. 1968;96:698.
97. Pilcher DB, Harman PK, Moore EE. Retrohepatic vena cava balloon shunt introduced via the sapheno-femoral junction. J Trauma. 1977;17:837.
98. McAnena OJ, Moore EE, Moore FA. Insertion of a retrohepatic vena cava balloon shunt through the saphenofemoral junction. Am J Surg. 1989;158:463.
99. Cogbill TH, Moore EE, Jurkovich GJ, et al. Severe hepatic trauma: a multicenter experience with 1,335 liver injuries. J Trauma. 1988;28:1433.
100. Baumgartner F, Scudamore C, Nair C, et al. Venovenous bypass for major hepatic and caval trauma. J Trauma. 1995;39:671.
101. Carrillo EH, Spain DA, Miller FB, et al. Intrahepatic vascular clamping in complex hepatic vein injuries. J Trauma. 1997;43:131.
102. Nicoluzzi JE, Von Bahten LC, Laux G. Hepatic vascular isolation in treatment of a complex hepatic vein injury. J Trauma. 2007;63:684.
103. Watarida S, Nishi T, Furukawa A, et al. Fenestrated stent-graft for traumatic juxtahepatic inferior vena cava injury. J Endovasc Ther. 2002;9:134.
104. Castelli P, Caronno R, Piffaretti G, et al. Emergency endovascular repair for traumatic injury of the inferior vena cava. Eur J Cardiothorac Surg. 2005;28:906.
105. Erzurum VZ, Shoup M, Borge M, et al. Inferior vena cava endograft to control surgically inaccessible hemorrhage. J Vasc Surg. 2003;38:1437.
106. Hommes M, Kazemier G, Van Dijk L, et al. Complex liver trauma with bilhemia treated with perihepatic packing and endovascular stent in the vena cava. J Trauma. 2009;67:E51.
107. Fabian TC, Croce MA, Stanford GG, et al. Factors affecting morbidity following hepatic trauma. A prospective analysis of 482 injuries. Ann Surg. 1991;213:540.
108. Cue JI, Cryer HG, Miller FB, et al. Packing and planned reexploration for hepatic and retroperitoneal hemorrhage: critical refinements of a useful technique. J Trauma. 1990;30:1007.
109. Noyes LD, Doyle DJ, McSwain NE. Septic complications associated with the use of peritoneal drains in liver trauma. J Trauma. 1988; 28:337.
110. Mullins RJ, Stone HH, Dunlop WE, et al. Hepatic trauma: evaluation of routine drainage. South Med J. 1985;78:259.
111. Waltensdorfer A, Mahla E, Zink M, et al. Central pulmonary artery embolism after perihepatic packing because of liver trauma. J Trauma. 2007;63:E81.
112. Walt AJ, Wilson RF. Management of Trauma: Pitfalls and Practice. Philadelphia: Lea and Febiger; 1975:348.
113. Cyret P, Baumer R, Roche A. Hepatic hemobilia of traumatic or iatrogenic origin. Recent advances of diagnosis and therapy. Review of the literature from 1976–1981. World J Surg. 1984;8:2.
114. Heimbach OM, Ferguson GS, Harley JD. Treatment of traumatic hemobilia with angiographic embolization. J Trauma. 1978;18:221.
115. Clemens M, Wittrin G. Bilhamie und hamobilie nach reitunfall. Vortrag 166. Hamburg: Tagung Nordwestdeutscher Chirurgen; 1975.
116. Glaser K, Wetscher G, Pointner R, et al. Traumatic bilhemia. Surgery. 1994;116:24.
117. Enneker C, Berens JP. Schwerste Leberruptur mit lebervenenabriss und massive bilhamie. Chirurg. 1978;49:311.
118. Hollands MJ, Little JM. Post-traumatic bile fistulae. J Trauma. 1991;31:117.
119. Howdieshell TR, Purvis J, Bates WB, et al. Biloma and biliary fistula following hepatorraphy for liver trauma: incidence, natural history, and management. Am Surg. 1995;61:165.
120. Dick R, Gilliams A, Dooley JS, et al. Stainless steel mesh stents for biliary stricture. J Intervent Radiol. 1989;4:95.
121. D’Amours SK, Simons RK, Scudamore DH, et al. Major intrahepatic bile duct injuries detected after laparotomy: selective nonoperative management. J Trauma. 2001;50:480.
122. Dabbs DN, Stein DM, Scalea TM. Major hepatic necrosis: a common complication after angioembolization for treatment of high-grade liver injuries. J Trauma. 2009;66:621.
123. Sheik-Gafoor MH, Singh B, Moodley J. Traumatic thoracobiliary fistula: report of a case successfully managed conservatively, with an overview of current diagnostic and therapeutic options. J Trauma. 1998;45:819.
124. Rothberg ML, Kilngman RR, Peetz D, et al. Traumatic thoracobiliary fistula. Ann Thorac Surg. 1994;57:472.
125. Danetz JS, Yelon JA, Fields CE, et al. Traumatic pleurocaval fistula: potential source of air embolism. J Trauma. 2001;50:551.
126. Eastridge BJ, Minei JP. Intrahepatic arterioportal fistula after hepatic gunshot wound: a case report and review of the literature. J Trauma. 1997;43:523.
127. Oppermann TE, Corcos AC, Jones LM, et al. Traumatic intrahepatic portosystemic venous shunt: a rare complication of grade V liver laceration. J Trauma. 2007;63:1230.
128. Jurkovich GJ, Hoyt DB, Moore FA, et al. Portal triad injuries—a multicenter study. J Trauma. 1995;39:426.
129. Dawson DL, Johansen KH, Jurkovich GJ. Injuries to the portal triad. Am J Surg. 1991;161:545.
130. Sharma O. Blunt gallbladder injuries: presentation of twenty-two cases with review of the literature. J Trauma. 1995;39:576.
131. Wilson RF, Walt AJ. Management of Trauma Pitfalls and Practice. 2nd ed. Baltimore: Williams and Wilkins; 1996:476.
132. Erb RE, Mirvis SE, Shanmuganathan K. Gallbladder injury secondary to blunt trauma: CT findings. J Comput Assist Tomogr. 1994;18:778.
133. Michelassi F, Ranson J. Bile duct disruption by blunt trauma. J Trauma. 1985;25:454.
134. Bourque M, Spigland N, Bensoussan A, et al. Isolated complete transection of the common bile duct due to blunt trauma in a child, and review of the literature. J Pediatr Surg. 1989;24:1068.
135. Jones KB, Thomas E. Traumatic rupture of the hepatic duct demonstrated by endoscopic retrograde cholangiography. J Trauma. 1985;25:448.
136. Velez SE, Llaryora RG, Lerda FA. Laparoscopic cholecystectomy in penetrating trauma. J Laparoendosc Adv Surg Tech A. 1999;9:291.
137. Soderstrom CA, Maika K, DuPriest RW. Gallbladder injuries resulting from blunt abdominal trauma. Ann Surg. 1981;193:60.
138. Smith SW, Hastings TN. Traumatic rupture of the gallbladder. Ann Surg. 1954;139:521.
139. Pachter HL, Liang HG, Hofstetter SR. Liver and biliary tract trauma. In: Trauma. 2nd ed. Norwalk, CT: Appleton and Lange; 1991:441.
140. Feliciano DV. Biliary injuries as a result of blunt and penetrating trauma. Surg Clin North Am. 1994;74:897.
141. Sandblom P, Tabrizian M, Rigo M, et al. Repair of common bile duct defects using the gallbladder or cystic duct as a pedicled graft. Surg Gynecol Obstet. 1975;140:425.
142. Ivatury RR, Rohman M, Nallathambi M, et al. The morbidity of injuries of the extrahepatic biliary system. J Trauma. 1985;25:967.
143. Stewart L, Way L. Bile duct injuries during laparoscopic cholecystectomy: factors that influence the results of treatment. In: Proceedings of the Pacific Coast Surgical Association, Seattle, WA, 19–21 February. 1995.
144. Monk JS, Church JS, Agarwal N. Repair of a traumatic noncircumferential hepatic bile duct defect using a vein patch: case report. J Trauma. 1991;31:1555.
145. Thomas JP, Metropol HJ, Myers RT. Teflon patch graft for reconstruction of the extrahepatic bile ducts. Ann Surg. 1964;160:967.
146. Voyles GR, Blumgart LH. A technique for the construction of high biliary-enteric anastomoses. Surg Gynecol Obstet. 1982;154:885.
147. Rodriguez-Montes JA, Rojo E, Martin LG. Complications following repair of extrahepatic bile duct injuries after blunt abdominal trauma. World J Surg. 2001;25:1313.
148. Eid A, Almogy G, Pikarsky AJ, et al. Conservative treatment of a traumatic tear of the left hepatic duct: case report. J Trauma. 1996;41:912.
149. Dawson DL, Jurkovich GJ. Hepatic duct disruption from blunt abdominal trauma: case report and literature review. J Trauma. 1991;31:1698.
150. Innes J, Ferrara J, Carey L. Biliary reconstruction without transanastamotic stent. Am Surg. 1988;54:27.
151. Costamagna G, Pandolfi M, Mutignani M, et al. Long-term results of endoscopic management of postoperative bile duct strictures with increasing numbers of stents. Gastrointest Endosc. 2001;54:162.
152. Lillemoe KD, Meltoon GB, Cameron JL, et al. Postoperative bile duct strictures: management and outcome in the 1990s. Ann Surg. 2000;232:430.