Applied Radiology

Over the past decade, computerized tomography (CT) has become the"gold" standard in evaluation of patients with blunt abdominal trauma(BAT).1,2 CT has been shown to be a highly sensitive modality and highlyspecific in evaluating solid organ injury in the abdomen. Furthermore,diagnostic peritoneal lavage (DPL) has been shown to be quite accurate indetection of hemorrhage in the peritoneal cavity of blunt abdominal traumapatients. More recently, especially in Europe, ultrasound has been used for theinitial exam in patients with blunt abdominal trauma.3,4 With two highlysensitive examinations that are widely utilized (CT and DPL), is there a needfor another modality, such as ultrasound, to evaluate patients with BAT? Thereare certain advantages of each modality and several questions regarding the useof ultrasound versus CT or DPL in evaluation of patients with BAT. These issueswill be discussed herein.

Diagnostic peritoneal lavage

In the United States, both diagnostic peritoneal lavage and CT have beenused for assessment of patients with blunt abdominal trauma. DPL has proven tobe an effective tool in detection of intraperitoneal hemorrhage, with a falsepositive and a false negative rate of 1.4% and 1.3%, respectively, in largeseries, a sensitivity of greater than 95%, and a specificity of greater than98%.5 Thus, DPL has become a standard procedure to assess BAT. However,peritoneal lavage can only be performed without patient movement. Also, DPL isan invasive procedure and is not appropriate for the majority of patients withBAT who remain awake and stable. While the complication rates for thisprocedure are low, complications do occur in approximately 1 to 2% ofpatients.6 Furthermore, peritoneal lavage cannot diagnose some injuries such aspancreatic, retroperitoneal, vascular, or solid organ injuries without capsulardisruption unless there is hemorrhage into the peritoneal cavity.

Computerized tomography

CT is currently widely used for evaluation of patients with BAT.1,2 At ourinstitution, if the patient is stable CT is performed for those with anequivocal abdominal examination, persistent abdominal pain, and a decreasinghematocrit. CT may be performed in comatose patients who have no measurableabdominal examination. Patients with significant hematuria and those withfractures requiring operative repair also may undergo CT. Lastly, CT may beused to screen for occult intra-abdominal injury in patients with anunexplained decreased hematocrit. There are, however, some inherentdisadvantages to CT. CT requires patient transport, and it is relativelyexpensive compared to DPL. It also subjects the patient to intravenousiodinated contrast which carries a small risk of acute allergic reaction.Additionally, CT cannot be utilized for unstable BAT patients.
Like DPL, CT can detect free fluid in the abdomen. CT may reliably show thespecific site of organ injury, which cannot be done with DPL. Furthermore, aclassification of the severity of organ injury by CT is important indetermining whether to use operative or nonoperative management.7,8 This typeof determination of the severity of the injury cannot be done with DPL. Forinstance, splenectomy was previously the most common treatment for splenicinjury; more recently, conservative management of splenic injuries on the basisof CT findings is now preferred, if possible.9,l0 Because of the advantages ofCT, it has replaced the use of DPL in many circumstances in detection oftraumatic intra-abdominal injuries. With all the advantages of CT, is there anyroom for the use of ultrasound evaluation of patients with BAT?

Ultrasound vs CT and DPL

While ultrasound has been used routinely for over 20 years in evaluation ofintra-abdominal abnormalities, it is only within the past five years that ithas been widely advocated for the evaluation of patients with BAT. In Europe,ultrasound has been used by both surgeons and emergency department physiciansin the initial examination of patients with BAT.3,4
Use of ultrasound has many advantages. It is relatively inexpensive compared toCT. It can be performed at bedside for unstable BAT patients duringresuscitation, and unlike CT it does not require patient transport. Sonographyis noninvasive, in contrast to DPL. Additionally, no intravenous contrast isnecessary for the examination. In light of all its advantages, the realquestion concerning ultrasound is whether or not it is a reliable modality forevaluation of BAT.

Ultrasound sensitivity

In many reports, sensitivities and specificities of ultrasound have beenequal to the those of CT or DPL in the BAT patient population. Sensitivitiesgreater than 85 to 90% and specificities as high as 99% have been commonlyreported for sonography.3,4,11-13 These reports focus on ultrasound's abilityto detect free fluid in patients with BAT. However, closer scrutiny in themajority of these reports showed several obvious limitations of study design.The most severe limitation was lack of comparison of ultrasound with otherimaging modalities, such as CT, or lack of comparison with operative findings.For instance, in a report by McKenney et al12 of 899 patients, there was nocorrelation with CT, DPL, or laparotomy in the majority (558) of theirpatients. Many of these studies rely on improved physical examination alone astheir comparison to determine sensitivities, specificities, and accuracies.However, there is some fallacy in this logic. It is well known that manypatients with solid organ injuries will show improvement in their physicalexaminations given enough time. Thus, ultrasound could potentially miss solidorgan injuries in patients without free fluid. When correlated with CT, DPL, oroperative findings, McGahan et all4 reported a sensitivity of ultrasound of63%, while the specificity remained high at 95%. There were a number ofinjuries not requiring operation and identified only with the comparison CT.

Ultrasound findings-free fluid

The focus of the ultrasound exam is on detection of free fluid. Free fluidusually will localize to either the perisplenic or perihepatic areas, theparacolic gutters, and the pelvis. We have found that free fluid is mostcommonly missed by ultrasound in the pelvis, as patients who undergo pelvicultrasound have an empty bladder. We encourage a full bladder technique, withplacement of a Foley catheter if necessary (figure 1). Free fluid is typicallyhypoechoic (figures 2,3). However, with acute hemorrhage, it usually is moreechogenic. With massive hemorrhage the fluid may be observed to be swirlingwith debris (figure 4).
Ultrasound detection of free fluid will vary with the type of organ injury. Forinstance, free fluid is more frequently detected with splenic injuries thanhepatic injuries.l4 In series directed to detection of free fluid for specificorgan injury ultrasound detection of free fluid is somewhat disappointing. Forinstance, at our institution we have identified free fluid with ultrasound inonly 35% of patients with isolated renal injuries that have been documented byeither CT or laparotomy.15 This is due to the fact that many renal injuries areminor and not associated with renal capsular disruption. Also, as the kidney isa retroperitoneal structure, hemorrhage may be confined to this space and maynot be detected in the peritoneal cavity.
We have had similar disappointing results in detecting free fluid in patientswith isolated bowel or mesenteric injuries.l6

Ultrasound findings-organ injury

Another potential limitation of the ultrasound examination is that it hasshown limited success in predicting the exact organ injured. Most ultrasoundexams focus on the presence of free fluid, but often the exam is unsuccessfulin predicting the exact site of injury. Except for splenic injuries, ultrasoundsensitivity in detecting parenchymal organ injury has been quite low.Ultrasound has been shown to detect from approximately 69 to 90% of patientswith splenic injuries (figures 2-4).14,17
Splenic injuries are more commonly visualized than injuries to the liver orkidneys. Hypoechoic and somewhat cystic-appearing areas may be identified inthe spleen parenchyma. More commonly, the spleen has a relatively disorganizedappearance, with mixed echogenic regions. Hypoechoic rims corresponding tosubcapsular hematomas may be observed with sonography. The spleen is oftenfound to be enlarged with acute splenic trauma. Additionally, Siniluoto and hiscoworkers found that repeat or delayed ultrasound performed 24 to 72 hoursafter the initial exam showed a higher detection rate of splenic injuries.17
Other injuries, such as injuries to the liver and kidney, are less reliablydetected with sonography. Liver injuries probably are poorly visualized becausecapsular disruption is less frequent. Thus, an intraparenchymal hematoma and/orsmall liver laceration may be easily missed with ultrasound. The appearance ofliver lacerations is variable. In our experience these may appear eitherhypoechoic (figure 5), isoechoic, or echogenic. They usually are welllocalized. Hyperechoic liver lesions were quite common in one series18 whichidentified 33 patients of 831 trauma patients studied with a geographichyperechoic pattern in the liver. All these hyperechoic areas corresponded tominor hepatic lacerations. The hyperechoic pattern faded to, most commonly, anisoechoic pattern, but occasionally a hypoechoic pattern appeared 7 to 19 daysafter injury.l9 Subcapsular hematomas of the liver are recognized as ahypoechoic rim surrounding the liver (figure 5).
With minor renal injuries such as renal contusions, small renal lacerationscould be easily overlooked by ultrasound. The appearance of the kidney andultrasound may be normal in such injuries. However, with more severe renalinjuries, such as complete disruption of the renal pedicle or a shatteredkidney, the normal renal architecture will be lost (figure 6). Theretroperitoneum will instead have a very disorganized appearance, and there maybe an accompanying hypoechoic or anechoic hematoma surrounding the kidney.15
Finally, other injuries, such as bowel injures, to this date have not beendirectly detected with sonography. Instead, bowel injuries usually areidentified by the presence of accompanying free fluid in the abdomen (figure7). On the initial ultrasound examination only about one-third of individualswith an isolated bowel on mesenteric injuries will have free fluid in theabdomen.16 Given these limitations of ultrasound, we must question the rolethat ultrasound can or should play in the evaluation of patients with BAT.

Ultrasound--present role

At our institution we accept the limitations but also recognize theadvantages of ultrasound in evaluating patients with BAT. In fact, ultrasoundis the initial examination of choice for patients with BAT at our institution.Ultrasound has been extremely helpful in quickly identifying patients withmassive intra-abdominal hemorrhage. The detection of large amounts of freefluid enables patients with massive intra-abdominal hemorrhage to beimmediately taken to the operating room. In patients who have decreasingabdominal tenderness, a stable hematocrit, and a negative ultrasound, nofurther analysis is required. This group comprises the majority of patientswith BAT presenting at our institution.
The two aforementioned groups of patients are easily triaged with ultrasound.However, there is a third group of patients who deserve closer scrutiny. Thisgroup includes patients in whom the physical examination does not correlatewith the ultrasound exam. These are patients with persistent or increasingabdominal tenderness, or those who have a decreasing hematocrit and a negativeultrasound exam. Comatose patients with significant abdominal trauma and anegative ultrasound also deserve closer scrutiny. This group may includepatients with parenchymal organ injuries without free fluid, or those in whichthere is only minimal free fluid with solid or hollow viscus injury. Therefore,reliance on physical examination and laboratory results should dictate the needfor either follow-up ultrasound or CT in this specific group of patients. Inthe initial publication of our study of 500 patients with BAT in whomultrasound was performed, CT was performed in less than 120 of thesepatients.14 Thus, we feel that, in the majority of patients, an ultrasound examwould be adequate in triaging these patients, effectively eliminating the needfor a CT scan.

Ultrasound exam--the technique

Most commonly, ultrasound examination involves examination of the entireperitoneal cavity. The emphasis in the exam is for detection of free fluid. Inour institution we perform exams of the following areas: the right upperquadrant, including the liver, diaphragm, and kidney; the left upper quadrant,including the diaphragm, spleen, and kidney; the right pericolic gutter; theleft pericolic gutter; the pelvis (two views); and the mid-epigastrium (figure8). Additionally, a subxiphoid or transthoracic exam usually was performed todetect pericardial effusion if, in fact, the patient had significant chesttrauma. Such an example is given in figure 9 of an unstable patient after chesttrauma suffered in an automobile accident. This ultrasound image demonstratespericardial hemorrhage. This early sonographic detection of pericardial fluidallows intervention to prevent cardiac tamponade. In our institution we obtainhard-copy imaging of these areas. Additionally, if free fluid is detected, astime permits, a more comprehensive examination of the organs is performed. Forinstance, if fluid is detected in the left upper quadrant, a more comprehensiveexamination of the spleen is performed to detect any splenic abnormality.

Other ultrasound findings

Ultrasound may also be useful in detecting non-traumatic etiologies ofabdominal pain. For instance, we have had patients who have fallen presentingto the emergency department with abdominal pain in which the abdominal pain wasthe cause of the fall and not the fall causing the abdominal pain. We havedetected masses within the abdomen, such as a dissecting aortic aneurysms, asthe etiology of such abdominal pain. We also have seen patients with a numberof other abdominal masses detected incidentally on ultrasound (figure 10).These masses may be confusing as we become focused on the primary purpose ofthe exam, which is to detect free fluid from trauma. For instance, figure 11 isa case of a 24-year-old woman who presented with abdominal pain after an autoaccident. In this individual, a pelvic cystic structure with a solid componentwas noted. This was thought to possibly represent a hematoma in the bladder.However, the patient had no hematuria, and thus a CT scan was performed whichrevealed a cystic mass with a fatty component corresponding to an ovariandermoid.

Ultrasound coverage

There have been a number of different approaches to deciding who, in fact,should perform the ultrasound examination in patients with blunt abdominaltrauma. Prior data has shown that there is some learning curve associated withthe trauma ultrasound exam. For instance, Forster and coworkers19 reported thatsurgeons with less than one year of experience had a positive predictive valueof 60%, surgeons with greater than one year experience but less than threeyears experience had a positive predictive value of 76%, and surgeons withgreater than three years experience in the use of ultrasound had a positivepredictive value of 92% in detection of free fluid in the abdomen.
At many institutions, trauma surgeons perform the trauma ultrasound, while inother institutions, emergency department physicians perform this procedure. Inour institution, we have taken a slightly different approach. Previously, wehad our radiology sonographers in-house from 7:00 a.m. to 11:00 p.m. and theywould be ready for call-back from 11:00 p.m. to 7:00 a.m. We are a major traumacenter, and as the number of exams increased in our emergency department, weprovided 24-hour in-house coverage of sonographers. We found that we were ableto provide this sonographic coverage with minimal cost.20 Prior to introducing24-hour in-house coverage, our sonographers were called back an average of onetime from 11:00 p.m. to 7:00 a.m. to perform an ultrasound examination. After afew months of instituting in-house coverage, our sonographers' case load hadincreased to 5.9 patients per shift.
All exams are reviewed by radiologists on hard-copy with images, as describedabove. In the near future, we will begin soft-copy reading of the images.

Conclusion

Despite the limitations of emergent ultrasound, it has been shown to play auseful role in evaluation of patients with blunt abdominal trauma. Ultrasoundat our institution is used for initial evaluation of patients with BAT. Ifmassive free fluid is detected, these patients are taken directly to theoperating room without further delay. The majority of patients will have anegative ultrasound exam and will show clinical improvement. In individualswith a negative ultrasound exam and continued abdominal pain or abnormal labvalues, such as decreasing hematocrit, a CT scan should be performed. AR