The current practice of managing splenic injury without surgery grew out of a desire to protect children from postsplenectomy sepsis. It was discovered that most children’s injured spleens stop bleeding without surgery. This practice was gradually extended into the adult population where the results are not as good but still approach 80% among stable patients. Multiple studies have been conducted in an attempt to more accurately predict which patients will succeed and which will fail attempts at nonoperative management. They have focused on combinations of patient factors such as age and vital signs and CT factors such as contrast blush and depth of laceration [3,4,6]. Failure of nonoperative management not only delays effective therapy and consumes resources, but patients who fail attempted nonoperative management have greater morbidity and mortality [8]. Advanced age, portal hypertension, and coagulopathy increase the probability of the failure of nonoperative management.

The nonoperative management of a ruptured spleen must be a joint effort between the surgical team and the ICU team. The parameters that will default the patient to the operative pathway should be agreed upon in advance between those who will be monitoring and supporting and those who will operate. In general, any indication of hemodynamic instability should lead to immediate surgery and splenectomy. If the patient experiences a steadily falling hemoglobin level but never manifests any change in vital signs, there should be prior agreement regarding the number of units of packed red blood cells (PRBCs) to be transfused prior to resorting to surgery. The absolute number will vary with the estimated operative risk, other factors predicting success or failure, and the patient’s preference but should rarely exceed four units of PRBCs for an isolated splenic injury.

Splenic embolization may be an option in some facilities for those patients whose CT demonstrates a contrast blush within the spleen. If embolization is to be utilized, it should
be performed by a team that is readily available and has demonstrated success with the procedure.

Patients admitted to the ICU for nonoperative management of an isolated splenic injury should receive their planned immunizations including pneumococcal, meningococcal, and Hemophilus influenza vaccine since there is evidence that these vaccines are more effective with the spleen in situ [9].

When the splenic injury is successfully managed nonoperatively, there are still potential complications. Delayed bleeding of a lacerated spleen is a well-recognized complication of splenic injury. Many programs will follow elaborate algorithms specifying when patients may increase physical activity and participate in activities such as physical therapy since such activity is perceived to play a role in delayed rupture. However, there is no convincing evidence that, short of avoiding a blow to the flank, one regimen is superior to another. Pain associated with either capsular distention or infracted splenic tissue may eventually necessitate splenectomy, particularly if the spleen is embolized. The other major complication is an infection involving the injured splenic parenchyma or the perisplenic hematoma resulting in either splenic or subphrenic abscess [10]. Unexplained fever, leukocytosis, pleural effusion, or hiccoughs should necessitate an abdominal CT scan looking for evidence of infection. Most such infections can be effectively treated with antibiotics and percutaneous drainage but failure to respond promptly should result in exploration, evacuation of the infected hematoma, and splenectomy.

The other commonly injured organ in blunt abdominal trauma is the liver. The injured liver differs from the injured spleen in two significant ways. First, removal of the injured organ is not a treatment option. Second, the liver secretes bile directly into the GI tract so that liver injuries have a more complex range of complications including bile leak, hemobilia, obstructive jaundice, and so on. While the surgical options differ from the spleen, the decision to operate should be based on similar considerations. The first criterion for successful nonoperative management is hemodynamic stability. A patient who does not meet this condition should be taken to the OR, explored, and if necessary, packed, since the organ cannot be totally removed. Experienced trauma or hepatic surgeons will more often be able to perform a definitive procedure initially but the lack of such surgeons should rarely lead to an attempt to manage an unstable patient nonoperatively. Perihepatic packing followed by either angiography with embolization, reexploration when more experienced personnel are available, or transfer to a more capable facility are all preferable to attempting to manage an unstable patient nonoperatively. Conversely, surgical exploration in the face of hemodynamic stability by an inexperienced team is a recipe for disaster and should be avoided.

Patients with solitary liver injuries admitted to the ICU for nonoperative management should first be evaluated for hemodynamic stability and if stable should next be evaluated to determine whether they are likely to benefit from angiography and embolization. Patients with contrast extravasation or severe lacerations extending deep into the hepatic parenchyma are candidates for angiography with embolization Liver injuries in the face of cirrhosis, portal hypertension, or coagulopathy are much more likely to fail nonoperative management than comparable injuries lacking these comorbidities.

Complications of nonoperative management are primarily the result of bleeding, infection, bile leak, hepatic necrosis, and jaundice. Delayed bleeding from a liver laceration may occur but sudden unrelenting hemorrhage from the liver necessitating emergency surgery is rare beyond 24 hours postinjury. Steadily falling hemoglobin levels in an otherwise stable patient are an indication for either repeat CT scanning to verify that the bleeding is coming from the liver or angiography in an attempt to identify a vessel suitable for embolization.

Bile leaks from the injured liver may result in either contained collections known as bilomas or more diffuse biliary ascites. Bilomas may cause compression of adjacent structures producing jaundice or gastric outlet obstruction in the subhepatic location but the more common problem resulting from bile leak is secondary infection. Small bile leaks occur commonly after liver injury but most are of no clinical significance. Elevated liver function tests after liver injury are an indication for hepatobiliary imaging, or hepatobiliary iminodiacetic acid (HIDA) scan to evaluate for a bile leak. Signs and symptoms of infection are usually better evaluated with a CT scan. Patients in whom a fluid collection is identified should undergo percutaneous drainage if they show evidence of infection. If the drained fluid shows a bilirubin level significantly above that of serum, the patient should then undergo HIDA scanning. Most such bile leaks will seal with adequate drainage of the fluid collection. If bilious drainage persists, they should be evaluated for endoscopic retrograde cholangiopancreatogram (ERCP) with stent placement.

High fevers, often exceeding 39°C, may be seen in patients with liver injury typically beginning 48 to 72 hours postinjury. These fevers have been blamed on atelectatic lung immediately above the diaphragm or on areas of hepatic necrosis. Solid evidence to firmly establish the cause of such fevers is not available. Patients who sustain severe liver injuries but remain hemodynamically stable may nonetheless harbor significant areas of devitalized liver. In the vast majority, this necrotic liver does not require resection. However, if the necrotic liver becomes infected or if the patient deteriorates, resectional debridement of the necrotic material may be necessary.

Hemobilia is a rare complication of hepatic injury. The classic triad of gastrointestinal hemorrhage, jaundice, and right upper quadrant pain should suggest the diagnosis. It may present anytime from the first few days postinjury to months later. Diagnosis is often difficult and delayed. The bleeding is usually intermittent so that diagnostic endoscopy may demonstrate no source for the bleeding. Any patient with a history of hepatic trauma, either immediate or more remote, who has evidence of unexplained gastrointestinal hemorrhage, should undergo diagnostic angiography coupled with therapeutic embolization if a hepatic pseudoaneurysm is identified [11].

Renal injury is most often the result of blunt trauma and frequently occurs in conjunction with other injuries. Right renal injury most frequently occurs in conjunction with hepatic injury and left renal injury in conjunction with splenic injury. Renal injury is almost always associated with hematuria but the severity of the hematuria and the degree of the renal injury are often discordant. Gross hematuria may appear dramatic but most renal bleeding diminishes spontaneously within a few hours of injury. Even impressive perinephric hematomas on CT often have little impact on management decisions [12,13].

The kidney has two possible responses to injury that may require monitoring and or intervention, contrast extravasation from bleeding or a urine leak Rarely will the hemodynamically stable patient continue to bleed from a lacerated kidney. In such cases, the management is similar to the other solid organs with appropriate imaging to confirm the source of bleeding followed either by embolization or surgical exploration. Usually, extravasation of urine from an injured kidney will resolve spontaneously [12,13]. Extravasated contrast that is confined within Gerota’s fascia does not mandate immediate intervention since it will frequently resolve spontaneously or respond to minimally invasive methods. Leakage of urine as demonstrated by delayed contrast extravasation outside of Gerota’s
fascia may still resolve but is more likely to benefit from percutaneous drainage of the renal collecting structures. Persistent urine leakage often indicates ureteral obstruction from either urinoma or retroperitoneal hematoma and may benefit from ureteral stenting.

Renal vascular injury is most often recognized on CT with intravenous contrast as an area of renal parenchyma that does not enhance. This injury may involve a single segment of the kidney or the entire kidney. Although gross hematuria may occur, it is typically of very short duration and may be absent altogether. Microscopic hematuria is virtually always present. The arterial injury may be either complete disruption or thrombosis. However, even with complete disruption, significant hemorrhage into the retroperitoneum is rare. Revascularization is rarely of benefit since in most cases, the time required for diagnosis, surgical exposure, and repair is beyond the warm ischemia tolerance of the kidney. Segmental infarction or even infarction of one entire kidney is usually well tolerated if the other kidney is healthy. Sequelae such as pain, abscess, bleeding, or hypertension are rare. Compression of the kidney by either hematoma or urinoma with subsequent renovascular hypertension (Page kidney) is extremely rare.

Blunt pancreatic injury is typically the result of high energy impact to the epigastrium. Because the pancreas is well protected by the costal margin and is located deep in the retroperitoneum, isolated pancreatic injury is rare. Physical findings are usually minimal and laboratory and imaging studies are often nondiagnostic. As a result of the difficulty in early diagnosis, isolated pancreatic injuries are rarely the cause of ICU admission. However, patients with injuries to liver, spleen, or kidney may show some abnormality associated with the pancreas during the course of their nonoperative management. Elevations in serum amylase or nonspecific findings on CT scan will not usually change the plan to manage the patient nonoperatively. However, it is important to insure that the duodenum is not injured. Duodenal perforation and pancreatic injury are often difficult to differentiate.

Serum amylase values are commonly relied upon to evaluate the pancreas following injury but the sensitivity and specificity of serum amylase leaves much to be desired in the early postinjury period. Serum amylase values determined within 3 hours of injury appear to be particularly unreliable [14]. A normal serum amylase value later in the patient’s course appears reliable in excluding a significant pancreatic injury. An elevated serum amylase value is much less specific, particularly in the setting of head injury [15]. Certainly, an elevated amylase should raise the level of suspicion sufficiently to pursue further evaluation of the pancreas. CT findings may also be less than diagnostic. Suggestive CT findings include visualization of a fracture of the pancreas, intrapancreatic hematoma, fluid in the lesser sac, retroperitoneal hematoma or fluid, and so on. As with the serum amylase value, CT scans obtained very early postinjury may be falsely negative [16]. These findings should not be interpreted as suggesting that a delayed work up is the preferred method but rather these results emphasize the importance of repeating both the amylase and if necessary the CT scan in cases where suspicion of pancreatic injury remains.

The critical determinant of whether pancreatic injuries can be managed nonoperatively is the integrity of the pancreatic duct. If pancreatic ductal disruption is present, distal resection or internal drainage produces much less morbidity than simple drainage or noninvasive management [17]. If no definitive reason for surgical exploration exists but there is reason to suspect or diagnose a pancreatic injury, it is imperative to evaluate the ductal integrity. If there is any suggestion of instability or peritoneal signs, this should be performed at the time of abdominal exploration. Otherwise, the patient may be a candidate for magnetic resonance cholangiopancreatography (MRCP) or even the more invasive ERCP. Delay in diagnosing and providing definitive therapy for a ductal injury may have devastating consequences.

Pelvic fractures represent the exception to the rule that nonoperative management is only suitable for hemodynamically stable patients. Surgical exploration of the pelvic hematoma is usually not an effective way to control the hemorrhage from a pelvic fracture. Thus, once other sources of bleeding have been excluded, even hemodynamically unstable patients may be managed in the ICU.

Although the focus of pelvic fracture management in the ICU is on dealing with the blood loss into the pelvis, it is important not to lose sight of the abdominal distention, and limitation of diaphragmatic excursion that can occur. Patients with significant bleeding into the pelvis should be monitored very carefully for respiratory compromise. This is particularly true during any transport out of the ICU to sites such as radiology. If there is any doubt of the patient’s ability to maintain adequate spontaneous ventilation, the airway should be secured electively and the patient placed on positive pressure ventilation.

A great deal of force is required to fracture the pelvis. Therefore, it is not surprising that associated injuries are common. Abdominal injuries and lower extremity fractures are both common in patients with pelvic fractures. These associated injuries often make it difficult to ascertain the site of bleeding. It is essential to evaluate the CT scan for the presence of intraperitoneal blood and solid organ injury as well as the size of the pelvic hematoma and the type of pelvic fracture. Lower extremities should be examined and x-rayed if any question exists of fracture. The type and location of pelvic fracture can provide valuable information regarding the likelihood of bleeding. Fractures or ligamentous disruptions of the posterior pelvis are more likely to be associated with severe hemorrhage than anterior fractures, acetabular fractures, or fractures of the iliac wing [18]. So called vertical shear fractures of the pelvis are particularly likely to be associated with arterial bleeding from the superior gluteal artery or other branches of the internal iliac system [19].

It is imperative to carefully examine the perineum for lacerations that may suggest an open pelvic fracture. This includes a careful rectal examination and a vaginal examination for females. If there is any indication of blood in the rectum or vagina, an endoscopic or speculum examination is required. An adequate examination is likely to be extremely painful with the pelvic fracture and often fractured lower extremities that make positioning very difficult. The examination should not be compromised even if it requires airway control and deep sedation. It may also require the assistance of the orthopedist to minimize fracture movement during the examination. The consequences of missing an open pelvic fracture may be disastrous.