Clinical policy: Critical issues in the evaluation of adult patients presenting to the emergency department with acute blunt abdominal trauma☆

I. What is the diagnostic performance of CT in diagnosing significant intra-abdominal injuries requiring intervention in blunt abdominal trauma? 

Background: CT has been used to evaluate the injured abdomen since the early 1980s. Although CT is widely accepted as a primary modality to evaluate blunt abdominal trauma, its overall accuracy in specific situations is a subject of ongoing study. The difficulty in studying CT is compounded by the fact that technology is improving quickly, as are the criteria for determining when to operate in the setting of a positive result. Whereas the liver and spleen are relatively easy to evaluate by CT, the bowel, pancreas, and diaphragm are more difficult. It also is well recognized that CT interpretation is, to an extent, reader dependent. This element of subjectivity places a premium on interpreter experience. The vast majority of CT studies are retrospective reviews at individual trauma centers. Given the acceptance of CT as nearly a criterion standard for evaluating the abdomen, future prospective randomized studies may be very difficult to perform.

Findings: Liu et al8 completed a prospective, comparative study of CT, diagnostic peritoneal lavage, and ultrasonography. Each of 55 stable patients suspected of blunt abdominal injuries underwent all 3 studies. If the results of any study were positive, the patients underwent laparotomies. Thirty-nine of these patients underwent laparotomies, a relatively high positive test rate for trauma. Results were as follows: CT sensitivity 97% and specificity 95%, diagnostic peritoneal lavage sensitivity 100% and specificity 84%, ultrasonography sensitivity 92% and specificity 95%. The only false negative CT was a case of intestinal perforation. Although it is difficult to draw conclusions from a study of this size published more than a decade ago, it bears inclusion as the only study of its kind available for review. In a separate study, Stafford et al9 studied 394 trauma patients prospectively and reported that for solid organ injuries, CT with oral contrast was 84.2% sensitive and 94% specific.

Specific intra-abdominal injuries that are difficult to diagnose have been the subjects of specific retrospective case reviews. These injuries include those to the bowel, diaphragm, and pancreas. Sherck et al5 studied 26 patients in their institution (3% of the 883 receiving CT scans) who had proven small bowel injuries. These authors found that, for small bowel perforation, CT scan had a sensitivity of 92% and a specificity of 94%. Butela et al12 retrospectively reviewed 50 cases of proven small bowel injuries in which CT was available for review. Sixty-two patients were chosen randomly as a comparison group to create a case-control study. On the basis of the original CT interpretation, the study revealed a sensitivity of 64% and a specificity of 97%.

Killeen et al13 performed a retrospective study of 150 patients with either CT or surgical diagnoses of blunt abdominal trauma or of bowel or mesenteric injuries. CT had a sensitivity of 94% in detecting bowel injury and 96% in detecting mesenteric injury. The lack of a control group and lack of surgical confirmation of bowel injury in 44% of patients were recognized weaknesses of the study. The authors noted that many of these patients had evidence of bowel wall contusions that resolved on follow-up CT. Janzen et al14 retrospectively examined 19 patients with surgically proven bowel or mesenteric injuries and compared them with 12 patients with no bowel or mesenteric injuries at surgery and calculated a sensitivity of 83% and specificity of 84% for CT in surgically proven bowel or mesenteric injuries. The positive predictive value was 77%, and the negative predictive value was 89%, although the control group size was relatively small. In a large, retrospective, case-controlled study of hollow viscus injury associated with blunt abdominal trauma, Fakhry et al15 found that CT was fair at identifying hollow viscus injury and even worse at identifying perforated hollow viscus injury. Although 84.2% of patients with free fluid and no solid organ injury had small bowel injury, only 30.5% of these patients suffered small bowel perforation. Hollow viscus injury was present in 91.5% of patients with pneumoperitoneum. Other radiologic findings on CT (eg, bowel wall thickening, stranding, contrast extravasation, retroperitoneal blood) were less effective in identifying hollow viscus injury.

The meaningful study of pancreatic and diaphragmatic injuries is difficult because these injuries occur in small numbers. Akhrass et al6 examined 72 patients with pancreatic injuries, only 10 of whom had CT scans for blunt mechanisms. Of these 10, only 3 had initially positive results on CT scan for pancreatic injuries (sensitivity 30%). Murray et al7 identified 11 patients with diaphragmatic rupture proven at surgery and reviewed their CT scans retrospectively using 3 radiologists. The investigators chose a control group of 21 patients who also had surgery but did not have injured diaphragms. The authors reported a sensitivity of 61% (95% confidence interval [CI] 41% to 81%) and a specificity of 87% (95% CI 76% to 99%) for the CT diagnosis of diaphragmatic rupture. Similar studies with smaller numbers yielded conflicting results.

Patient Evaluation Recommendations: What is the diagnostic performance of CT in diagnosing significant intra-abdominal injuries requiring intervention in blunt abdominal trauma?

Level A recommendations. None specified.

Level B recommendations. When either liver or spleen injury is suspected, CT can reliably exclude injuries that require emergent operative intervention. CT alone cannot be used to exclude either bowel, diaphragm, or pancreas injury.

Abdominal CT accurately identifies hemoperitoneum among patients with blunt abdominal trauma.

Level C recommendations. None specified.

II. Does oral contrast improve the diagnostic performance of CT in blunt abdominal trauma? 

Background: Trauma centers frequently administer oral contrast before each CT scan. Patients either drink contrast or have it administered through nasogastric tubes. Authors report contrast volumes in the range of 450 to 1,000 mL. Multiple authors have written that oral contrast improves the diagnostic accuracy of abdominal CT by identifying extravasation of bowel contents, delineating mesentery, and setting opacified bowel apart from hematomas and pancreatic injuries.16 Some authors have become concerned, however, about the possible risks of oral contrast administration: vomiting, aspiration, and delayed diagnosis related to bowel transit time of the contrast.17

Findings: Few studies examine the utility of using oral contrast during abdominal CT when evaluating blunt abdominal trauma. Stafford et al9 performed the only prospective randomized study of this technique currently in the literature. The authors randomized 394 injured patients to receive either oral contrast through a nasogastric tube or no contrast before CT evaluation of the abdomen. When analyzing only small bowel injuries, the authors reported a sensitivity of 86% (6 out of 7 injuries discovered) for CT with oral contrast and a sensitivity of 100% (3 out of 3 injuries discovered) for CT without oral contrast. The authors reported that for solid organ injuries, the oral contrast and no contrast groups demonstrated sensitivities of 84.2% and 88.9%, respectively, and specificities of 94% and 57.1%, respectively. The average time from nasogastric tube placement to CT scan was 39±18 minutes for the no contrast group and 46±24 minutes for the contrast group. These time intervals suggested that contrast was not allowed to transit for an optimal amount of time before scanning, although the authors did not assess the adequacy of contrast administration. The authors concluded that oral contrast did not provide additional benefit in CT scanning of the abdomen after blunt abdominal trauma. However, the authors also admitted that the small numbers of bowel injuries in their study suggested the need for additional studies with larger sample sizes.

Fakhry et al15 performed a retrospective, case-control study of 275,557 trauma center admissions from 95 participating institutions. This study included 408 patients with perforated small bowel injuries identified at laparotomy who also received abdominal CT with oral contrast; 2.9% exhibited extravasation of oral contrast. Tsang et al17 performed a retrospective review of contrast CT scans between 1988 and 1993 to determine whether oral contrast was essential to these scans in the opinion of a radiologist blinded to the clinical outcomes. Thirty-one liver and spleen injury cases were chosen randomly for review, and contrast was not judged to be essential in any of these cases. Contrast was not helpful among 20 randomly chosen patients who had no injuries. Contrast was not essential to the diagnosis of 22 intestinal injuries, but was essential in the diagnosis of 2 of 6 pancreatic injuries. Among the 22 intestinal injuries, the initial CT scan diagnosed only 1. The authors pointed out that multiple intestinal injuries were diagnosed in their institution by laparotomy without CT scan. The authors concluded that contrast was not helpful in the diagnosis of most solid organ injuries and did not improve sensitivity for intestinal injuries, but may help improve the sensitivity of CT for pancreatic injuries. The authors also acknowledged the significant weaknesses of their study. The study was retrospective, relied on subjective reviews of CT for contrast utility, and contained few intestinal and pancreatic injuries.

Clancy et al18 reviewed 492 patients who underwent CT scan over a period of 4 years, only 8 of whom received oral contrast. They reported an overall sensitivity of 98.4% and a specificity of 99.8% for intra-abdominal injuries. The 1 missed injury in this series was cecal ischemia with intestinal perforation. The 2 false positive CT scans were a suspected bowel injury and a suspected splenic rupture. The authors concluded that contrast was not necessary, but may prove helpful on follow-up CT scans for patients with findings suggestive of bowel injury on the first CT.

Federle et al19 retrospectively reviewed the records of 510 patients who received contrast and found that no patients had evidence of aspiration pneumonitis. The authors concluded that oral contrast was safe for this population.

Patient Evaluation Recommendations: Does oral contrast improve the diagnostic performance of CT in blunt abdominal trauma?

Level A recommendations. None specified.

Level B recommendations. Oral contrast is not essential to the evaluation of blunt abdominal trauma.

Level C recommendations. None specified.

III. What is the diagnostic performance of FAST in diagnosing hemoperitoneum in blunt abdominal trauma? 

Background: Rapidly identifying hemoperitoneum among unstable blunt trauma victims is a linchpin of early intervention and effective management. Emergency physicians and surgeons have used diagnostic peritoneal lavage in this role, but diagnostic peritoneal lavage is invasive and associated with procedural complications.8, 20 Many trauma centers now use abdominal ultrasonography instead of diagnostic peritoneal lavage for the rapid identification of hemoperitoneum.

FAST is an ultrasonographic technique used to visualize, at a minimum, Morrison's pouch (ie, the right upper quadrant), the splenorenal recess (ie, the left upper quadrant), and the pouch of Douglas in the pelvis. The presence of intra-peritoneal fluid (ie, blood) in the setting of blunt abdominal trauma facilitates decisions regarding laparotomy and further diagnostic studies.

Findings: The performance of FAST in identifying hemoperitoneum in blunt abdominal trauma victims is summarized in the Table.

Table. The performance of FAST in identifying hemoperitoneum in blunt abdominal trauma victims.

	Study	Study Class	Study Size	No. of Patients With Hemoperitoneum	Sensitivity, %	Specificity, %	Positive Likelihood Ratio	Negative Likelihood Ratio
	Ma et al 21	I	245 patients	64	90	99	90	0.1
	Rozycki et al22	II	1,227 patients	96	83.3	99.7	278	0.17
	Shackford et al23	I	241 patients	51	68	98	34	0.33
	Smith et al24	III	841 patients	45	73	98	36.5	0.28
	Tso et al10	II	163 patients	11	91

FAST detects intra-peritoneal fluid with 68% to 91% sensitivity and excellent specificity.22 In a prospective study of 1,540 blunt abdominal trauma victims, FAST was 100% sensitive and 100% specific in identifying hemoperitoneum among hypotensive patients.22

Tso et al10 calculated a sensitivity of 69% for FAST in identifying all intra-abdominal injuries (combined), including solid organ injuries, viscus trauma, and hemoperitoneum. Similarly, Richards et al25 found that, among 30 patients with false negative FAST examinations, 16 had either bowel or mesenteric injuries.

Patient Evaluation Recommendations: What is the diagnostic performance of FAST in diagnosing hemoperitoneum in blunt abdominal trauma?

Level A recommendations. None specified.

Level B recommendations. FAST is useful as an initial screening examination to detect hemoperitoneum in blunt abdominal trauma patients.

Level C recommendations. None specified.

IV. What is the diagnostic performance of diagnostic peritoneal lavage in diagnosing significant intra-abdominal injuries requiring intervention in blunt abdominal trauma? 

Background: Clinical examination alone does not identify all intra-abdominal injuries, especially in blunt trauma patients with multiple injuries.26 Diagnostic peritoneal lavage was introduced in 1965 as a rapid method to identify hemoperitoneum.3 Multiple studies documented that diagnostic peritoneal lavage is a sensitive test for intra-peritoneal blood.20, 27, 28, 29 Since 1965, diagnostic peritoneal lavage decreased the number of trauma deaths and nontherapeutic laparotomies in severely injured patients.30

Diagnostic peritoneal lavage is performed using either an open or closed (ie, percutaneous over a guide wire) technique. Sensitivity, specificity, and complication rates of the 2 techniques are similar.31, 32, 33, 34 The closed technique, however, is completed more quickly.31, 32, 33

Results of diagnostic peritoneal lavage are positive for hemoperitoneum in blunt abdominal trauma in the presence of the following: (1) the aspiration of 5 to 10 mL of frank blood, (2) an RBC count of 100,000/mL in the effluent after lavaging the abdomen with 1 L of isotonic fluid, or (3) a WBC count greater than 500/mL in the effluent.35 Lavage fluid Gram's stain and biochemical markers such as alkaline phosphatase and amylase are nonspecific and insensitive for intra-abdominal injury.29, 36, 37

Findings: The sensitivity of diagnostic peritoneal lavage for hemoperitoneum approaches 98%.8, 20, 28, 38, 39 A mean sensitivity of 95% for diagnostic peritoneal lavage was determined in a review of 58 studies in a decision analysis of optimal modalities evaluating blunt trauma in hemodynamically stable adults.40 Diagnostic peritoneal lavage can detect as little as 20 mL of intra-peritoneal blood.35

Diagnostic peritoneal lavage has practical limitations in identifying retroperitoneal, diaphragmatic, and enteric injuries because intra-abdominal bleeding usually is limited.30, 41 For example, in 1 study where diagnostic peritoneal lavage was performed early after blunt abdominal trauma, the authors found that lavage was falsely negative in 18% of patients with bowel injuries.42

Several prospective studies compared diagnostic peritoneal lavage to ultrasonography and abdominal CT in blunt abdominal trauma. Diagnostic peritoneal lavage was at least as sensitive (ie, 83% to 100%) as CT and ultrasonography in detecting hemoperitoneum.8, 27, 39, 43

Although diagnostic peritoneal lavage efficiently identifies hemoperitoneum resulting from intra-abdominal injuries, many injuries are self-limited and do not require laparotomy. The false-positive rate for diagnostic peritoneal lavage is between 13% and 54%.28, 38, 39, 43, 44, 45 Nontherapeutic laparotomies can have significant complications; 1 prospective study had a complication rate of up to 20%.46 Excluding nontherapeutic laparotomies, the major complication rate for diagnostic peritoneal lavage is 1% to 2%.8, 20

Mele et al47 used diagnostic peritoneal lavage as a screening test. These authors followed each diagnostic peritoneal lavage with a positive result with an abdominal CT to identify those patients requiring laparotomies for surgical repair. The authors experienced no nontherapeutic laparotomies, no missed injuries, and an overall decreased utilization of CT in this small series. Besides having a small sample size, this study was not randomized and suffered from selection bias.

Patient Evaluation Recommendations: What is the diagnostic performance of diagnostic peritoneal lavage in diagnosing significant intra-abdominal injuries requiring intervention in blunt abdominal trauma?

Level A recommendations. None specified.

Level B recommendations. Diagnostic peritoneal lavage can be used to exclude hemoperitoneum in blunt abdominal trauma patients. Diagnostic peritoneal lavage does not define the extent of injury, has a 1% to 2% complication rate, and may lead to nontherapeutic laparotomies.

Level C recommendations. On the basis of consensus and current practice patterns, the initial choices for the evaluation of blunt abdominal trauma are CT and FAST, depending on the patient's hemodynamic stability.