In the United States, trauma accounts for 37 million emergency room visits annually and is the leading cause of death in persons younger than 45 years old [1]. Five million cases of blunt abdominal trauma (BAT) occur each year, and 50,000 deaths occur annually because of blunt injury [2]. Early recognition of severe abdominal injury is paramount, as the rapid identification of hemoperitoneum expedites patient care to the operating room for exploratory laparotomy and definitive management. The traditional diagnostic tool for identification of hemoperitoneum has been the diagnostic peritoneal lavage (DPL). Over the last 30 years, bedside ultrasound has been used with increasing frequency to assess for intra-peritoneal hemorrhage [3-7]. Focused Assessment with Sonography for Trauma (FAST) utilizes clinician-performed ultrasonography at the patient’s bedside to rapidly identify the presence of hemoperitoneum or hemopericardium associated with acute life threatening injuries. Current literature has demonstrated improved outcomes and shortened time to operative intervention when this examination is performed by trained personnel and included in the initial evaluation of the unstable trauma patient [8,9]. Point of care ultrasound (POCUS) is defined as a focused, goaloriented bedside ultrasound examination performed by the treating physician to answer a specific question or guide an invasive procedure. POCUS is performed by physicians of various specialties such as emergency medicine, intensive care, and various surgical specialties. The archetypal POCUS examination, the FAST, was initially developed by emergency physicians (EP) and surgeons caring for traumatically injured patients, and over the past 20 years the number of applications a nd types of specialists using bedside ultrasound has grown [10,11]. On August 1st 2012, Harborview Medical Center, a large inner-city hospital and training site for the University of Washington, formally incorporated EP-performed FAST exams for the initial evaluation of severely injured BAT patients. Prior to this date, unstable BAT patients received either a FAST exam or a DPL as the initial diagnostic evaluation for hemo-peritoneum. The FAST exam was performed either by a radiologist or emergency physician and interpreted in real time by the performing physician, while the DPL was performed by a general surgeon. After the protocol change, the FAST exam became the first method of investigation, with the DPL reserved for patients with a negative FAST and ongoing hemodynamic instability. We sought to evaluate the accuracy of EP FAST exams for unstable BAT patients. Secondary analysis included determination of the trends in use for DPL for unstable BAT patients.

Study design

We conducted a retrospective analysis of all Emergency Severity Index (ESI) level 1 trauma patients at Haroborview Medical Center using the Emergency Department electronic tracking system over a 26-month period (July 1, 2011 to August 31, 2013), thirteen months before and after the protocol change. Patient inclusion criteria were: blunt thoracoabdominal injury, full trauma team activation, and a FAST examination performed as part of the initial diagnostic evaluation. The criterion for full trauma team activation is hemodynamic instability, as demonstrated by hypotension (systolic blood pressure <90 mm Hg), need for cardiopulmonary resuscitation, obvious major vascular injury or need for ongoing transfusion of blood products to maintain blood pressure, as determined by the trauma team. Patients with penetrating injury, no FAST exam, no confirmatory computed tomography (CT) or laparotomy, or who were transferred from an outside hospital with CT imaging of the abdomen already performed, were excluded from analysis. Data was collected using a closed-ended abstraction form for demographics, results of the FAST exam, peritoneal fluid aspirate or lavage, CT, and operative findings when available. FAST exam results were noted to be positive, negative, or indeterminate per the medical record. Indeterminate FAST exams were excluded from determining test characteristics. DPL results were positive if there were greater than 100,000 cells/mm3 red blood cells, 500 white blood cells/mm, elevated amylase, bile or the presence of fibrous material. The primary endpoint was the presence of any amount of intra-peritoneal hemorrhage on CT, or intra-peritoneal hemorrhage identified in the operating room. Inter- rater reliability was determined by comparing FAST results to CT or laparotomy findings using a weighted kappa score. The study was approved by the University of Washington institutional review board.

Setting

Haroborview Medical Center is the only designated level 1 trauma center in a five-state region in the Pacific Northwest that includes Washington, Wyoming, Idaho, Alaska, and Montana, with an annual Emergency Department census of over 75,000 patients per year. It serves as the regional referral hospital for a large portion of the Pacific Northwest, and is the primary training site for physicians of various specialties including radiology, trauma surgery and emergency medicine.

A total of 667 ESI level 1 BAT patients were identified using the ED electronic tracking system (Figure 1). 249 patients had full trauma activation, and 185 (74%) unstable BAT patients had a FAST examination performed as part of their initial evaluation with a confirmatory test for comparison. A total of 71/185 (38%) patients also had DPL performed in conjunction with the FAST exam to identify or confirm hemoperitoneum in critically unstable BAT patients. Most patients had full trauma team activation for hypotension and were injured in a motor vehicle collision (Table 1).

Figure 1: ESI = Emergency Severity Index; BAT = Blunt Abdominal Trauma; FAST = Focused Assessment with Sonography in Trauma.

Table 1: Patient Demographics.

Emergency physician FAST

90 FAST examinations were performed by EPs for unstable BAT patients in whom a confirmatory CT or operative report was available. There were 23/90 positive, 53/90 negative, and 13/90 indeterminate EP FAST exams. There were 22 true positive, 46 true negative, 1 false positive, and 7 false negative examinations for a sensitivity, specificity and accuracy of 76% (95% CI 56 to 89 %), 98% (95% CI 87 to 99 %), and 89% for any amount of free intra-peritoneal hemorrhage (Table 2). The one false positive EP FAST had a normal DPL, no CT findings of intraperitoneal hemorrhage, and did not require operative management. Of the seven false negative exams, four had trace or small amounts of intra-peritoneal hemorrhage. Three had moderate to large amounts of hemorrhage found on laparotomy, two of which were post-angiography with embolization (Table 3). 6/7 false negatives had associated pelvic fractures. Excluding cases of trace or small fluid collections would have increased the EP FAST sensitivity to 88% (95% CI 67 to 96 %). Of the indeterminate FAST exams, one case was notable for moderate amount of hemo-peritoneum diagnosed by CT not requiring operative management. All other indeterminate EP FAST exams had no intraperitoneal hemorrhage.

Table 2: Emergency Physician FAST.

Table 3: False Negative EP FAST

Radiology FAST

Of the 95 FAST exams performed by the Department of Radiology, there were 12 positive, 68 negative, and 16 indeterminate examinations. There were 11 true positive, 63 true negative, 1 false positive, and 5 false negative FAST exams for a sensitivity, specificity and accuracy of 69% (95% CI 41 to 87 %), 98% (95% CI 90 to 99 %), and 93% for any amount of free intra-peritoneal hemorrhage (Table 4). The lone false positive case was followed by a negative DPL, a CT without evidence of free intra-peritoneal hemorrhage and no operative management. All 5 false negative exams were associated with trace or small amount of hemorrhage, and none required an exploratory laparotomy. 3/5 false negative Radiology FAST exams were associated with pelvic fractures, and two required angiography guided embolization (Table 5). Excluding cases of trace or small hemorrhage would have increased the sensitivity of the radiology FAST exam to 100%. Of the indeterminate Radiology FAST examinations, four were ultimately positive for internal hemorrhage, three of which had moderate to large amounts of blood on laparotomy.

Table 4: Radiology FAST.

Table 5: False Negative Radiology FAST

Accuracy of FAST

Of the 185 FAST exams performed collectively, there were 33 true positive, 109 true negative, 2 false positive, 12 false negative, and 29 indeterminate FAST examinations. The sensitivity, specificity and accuracy were 73% (95% CI 57 to 84%), 98% (95% CI 93 to 99%), and 91% respectively. Inter-rater reliability between two study investigators (AA and BB) was good (κ=0.78). Test characteristics for hemodynamically unstable BAT patients with moderate to large amounts of hemo-peritoneum, were sensitivity 92%, specificity 98%, and accuracy 97%. The difference in sensitivity between EP FAST exam (76% [95% CI 56 to 89 %]) and Radiology FAST exam (69% [95% CI 41 to 87 %]) was statistically insignificant (p=0.7284), and both exams had near identical accuracy rates (89% v 93%) making neither exam better or worse than the other.

Diagnostic peritoneal lavage

A total of 71 patients underwent a diagnostic peritoneal lavage to evaluate for hemoperitoneum. Of the 71 DPLs performed, 53 (75%) were performed after a negative FAST, 15 (21%) after an indeterminate FAST, and 3 (4%) after a positive FAST exam. Five patients had grossly positive peritoneal aspirates, four of which had a final operative diagnosis of traumatic hemoperitoneum. 34 DPLs were performed in 79 BAT patients before the new protocol and 38 DPLs were performed in 106 patients after the protocol change. While the overall use of DPL in critically injured BAT patients decreased with the implementation of the new protocol, this trend was not statistically significant (p=0.17).

One challenge of trauma care is the rapid identification of significant abdominal injuries in hemodynamically unstable patients. Highly accurate, the FAST exam is also repeatable, noninvasive, easily learned and can be performed at the bedside, and thus plays a vital role in the evaluation of trauma patients [12-15]. While computed tomography is the gold standard for diagnosing solid organ injury, the FAST exam is associated with expedited care in BAT, shorter hospital stay, and lower overall cost of care [16-18]. The FAST exam has the potential to mobilize resources quickly for critically injured patients, rapidly identifying those who would benefit from laparotomy

Our findings demonstrate that EP FAST exams in unstable BAT patients had a sensitivity, specificity and overall accuracy of 76%, 98%, and 89% for any amount of free intra-peritoneal hemorrhage. The sensitivity for FAST examinations in our study was somewhat lower than previously reported because our study was designed, a priori, to determine test characteristics for any amount of free hemorrhage, regardless of its clinical significance. It is generally appreciated that the sensitivity of the FAST exam is highest for hemodynamically unstable patients with moderate to large amounts of intra-peritoneal hemorrhage. The FAST exam is limited in its ability to detect trace and small amounts of hemorrhage 18. With this in mind, EP FAST exams in our study would have demonstrated a sensitivity, specificity and accuracy of 88%, 98%, and 94% respectively for moderate to large amounts of blood in unstable BAT patients, results comparable to published findings in similar patients [19-24]. We did not find a significant difference in test characteristics for FAST performed by Radiology versus EPs.

The main limitations to our study were small sample size, single center, and retrospective design. Because we searched for ESI level 1 trauma patients only, patients initially triaged at a lower acuity level, yet subsequently became unstable, were not identified, and could not be included in this study. Patients with indeterminate FAST exams, no documented FAST exam results, or without confirmatory findings were not included in the analysis, all of which could have significantly changed accuracy rates. Lastly, length of time between the FAST exam and confirmatory testing was not abstracted, thus fluid accumulations that occurred after a negative FAST could not be accounted for.

Overall, the FAST exam demonstrated modest sensitivity and high specificity and accuracy for the detection of any amount of intraperitoneal hemorrhage in hemo-dynamically unstable BAT patients. EP performed FAST showed a sensitivity of 76%, specificity of 98%, and accuracy of 89% for any amount of hemorrhage, test characteristics similar to Radiology performed FAST. Utilization of diagnostic peritoneal aspirate and/or lavage in critically injured BAT patients decreased, although the change was not significant.