Chinese Journal of Traumatology 2012;15(5):273-278
Diagnostic accuracy of Focused Abdominal Sonography
for Trauma in blunt abdominal trauma patients in a trauma
centre of Hong Kong
Cheung Kent Shek*, Wong Hay Tai, Leung Ling Pong, Tsang Tat Chi and Leung Gilberto Ka Kit
【Abstract】Objective: Focused Abdominal Sono-
graphy for Trauma (FAST) is widely used for the detection
of intraperitoneal free fluids in patients suffering from blunt
abdominal trauma (BAT). This study aimed at assessing the
diagnostic accuracy of this investigation in a designated
trauma centre.
Methods: This was a retrospective study of BAT pa-
tients over a 6 year period seen in a trauma centre in Hong
Kong. FAST findings were compared with laparotomy, ab-
dominal computed tomography or autopsy findings,
which served as the gold standard for presence of intraperi-
toneal free fluids. The patients who did not have FAST or
gold standard confirmatory test performed, had preexisting
peritoneal fluid, died at resuscitation or had imcomplete docu-
mentation of FAST findings were excluded. The performance
of FAST was expressed as sensitivity, specificity, predic-
F
ocused Abdominal Sonography for Trauma
(FAST) has been used for over 20 years in the
management of blunt abdominal trauma (BAT)
patients.1 Previous studies demonstrated that FAST is
an accurate and expeditious modality for the initial de-
tection of intraperitoneal free fluids in BAT patients.2
Moreover, no significant difference in accuracy is found
for FAST performed by radiologist or non-radiologists.3
In addition, FAST has been shown to play an important
role in emergency departments especially when inputs
from radiologists are not readily available.4 In Hong Kong,
there are 5 designated trauma centres serving its 7
million populations. Each trauma centre has its own
DOI: 10.3760/cma.j.issn.1008-1275.2012.05.003
Department of Accident and Emergency, Queen Mary
Hospital, 102 Pokfulam Road, Pokfulam, Hong Kong,
China (Cheung KS, Wong HT, Leung LP and Tsang TC)
Department of Surgery, Li Ka Shing Faculty of Medicine,
University of Hong Kong, Hong Kong, China (Leung GKK)
*Corresponding author: Tel: 852-22555441, Fax: 852-
22554655, Email: kentshek@graduate.hku.hk
. 273 .
tive values (PV), likelihood ratios (LR) and accuracy.
Results: FAST was performed in 302 patients and 153
of them were included in this analysis. The sensitivity,
specificity, positive PV, negative PV, positive LR, negative
LR and accuracy for FAST were respectively 50.0%, 97.3%,
87.0%, 84.6%, 18.8, 0.5 and 85.0%. FAST was found to be
more sensitive in less severely injured patients and more
specific in more severely injured patients.
Conclusion: FAST is a reliable investigation in the
initial assessment of BAT patients. The diagnostic values
of FAST could be affected by the severity of injury and
staff training is needed to further enhance its effective use.
Key words: Laparotomy; Autopsy; Tomography, X-
ray computed; Tomography, spiral computed; Ultra-
sonography
Chin J Traumatol 2012;15(5):273-278
trauma team structure and not all the medical staff in-
volved in acute trauma care has formal FAST training.
Furthermore, there is no local study investigating its
accuracy. Therefore this study aims at evaluating the
diagnostic accuracy of FAST performed in one of the
designated trauma centres in Hong Kong. The purpose
is to see whether the local performance of FAST is com-
parable to international standards.
METHODS
This was a retrospective observational study done
in the Emergency Department (ED) of Queen Mary
Hospital, which is a designated trauma centre serving
the central, western and southern districts of Hong Kong
Island with a population of around 0.53 million.5 All the
BAT patients who were identified in the hospital trauma
registry and managed by the hospital trauma team in
the ED from 1st January, 2005 to 31st December, 2010
were included. The first tier trauma team members in-
clude doctors from the departments of emergency
. 274 .
medicine, surgery, orthopaedics, anaesthesiology and
intensive care. The trauma team would be activated if
any of the following criteria was present, namely, ab-
normal vital signs (systolic blood pressure <90 mm Hg;
respiratory rate <10 or >29 per minute), abnormal level
of consciousness (Glasgow Coma Scale <13), signifi-
cant blunt or crush injury to chest/abdomen, major pel-
vic fracture, serious mechanism of injury (e.g. ejection
from automobile) or evidence of high-energy impact (e.g.
fall from >6 m height). Patients were excluded if they
died at the ED after initial resuscitation or had known
ascites or peritoneal dialysis before the injury. BAT pa-
tients who did not undergo FAST or standard confirma-
tory test and those with incomplete FAST findings were
also excluded. This study was approved by the local
Institutional Review Board.
During the study period, there was no formal guide-
line in the hospital on who and when to do the FAST.
However, it was a usual practice that FAST would be
done by the emergency physicians or surgeons in
charge of the trauma team in the ED as both of them
must be certified advanced trauma life support provider.
The procedures of doing FAST followed the standard
recommendation. Scanning was done on a supine pa-
tient at 6 sites i.e. the four quadrants of the abdomen,
the suprapubic and subxiphoid region. During the study
period, two models of ultrasound machines were used.
They were ALOKA SSD-500 (probe model UST-934N-3.5,
3.5 MHz convex sector probe) and GE Medical Logiq
P5 Premium Ultrasound Console (probe model 4C 1.8-
4.0/D2.9 MHz, curvilinear probe).
Laparotomy (LAP) started within 4 hours of admis-
sion was used as a gold standard confirmatory test for
presence of intraperitoneal blood in BAT patients for whom
surgical exploration was clinically indicated. If the pa-
tient did not require LAP, abdominal computed tomogra-
phy (CT) scan was taken as the gold standard when the
patient was still treated in the ED. Postmortem exami-
nation (PME), in addition to CT scan if available, was
used as surrogate standard test for patients who died
during hospital stay without LAP performed.
Demographic data were analysed with descriptive
statistics. The performance of FAST was expressed as
sensitivity, specificity, predictive values (PV), likelihood
ratios (LR). In order to assess whether severity of injury
would affect the results of FAST, the performance were
Chinese Journal of Traumatology 2012;15(5):273-278
also analysed separately on patients with injury sever-
ity score (ISS) <15 (less severe) and >15 (more severe),
respectively.
RESULTS
During the study period, a total of 320 BAT patients
were identified. Among them, 302 (94.4%) had FAST
performed and 300 had adequate documentation of
FAST findings. In these 300 patients, only 153 (47.8%)
had confirmatory test(s) and thus they were included in
analysis (Figure 1).
Figure 1. Patient selection.
Among the 153 patients included in the analysis,
there was male predominance (67.3%). The average
age was 48.6 years (range 3-94 years). Six (3.9%) pa-
tients were under the age of 18. Thirty three (21.6%)
patients died up to 90 days after the index presentation.
Gold standard tests included 25 LAP, 130 CT scan
and 9 PME (Table 1).
The overall performance of FAST is shown in Table 2.
The sensitivity, specificity, positive PV, negative PV, posi-
tive LR, negative LR and accuracy were 50.0%, 97.3%,
87.0%, 84.6%, 18.8, 0.51 and 85.0% respectively.
The performance of FAST was also analyzed with
respect to the injury severity on presentation. Detailed
data are listed in Table 3. Among patients with false nega-
tive FAST results, more had ISS >15 (16/20, 80.0%) when
compared to those with true positive or negative results
(72/130, 55.4%). The mortality rate at day 90 was also
higher in the false negative group (40.0% vs 18.5%) in
true positive/negative group. All these revealed that FAST
was more sensitive in less severely injured patients and
Chinese Journal of Traumatology 2012;15(5):273-278 . 275 .

more specific in more severely injured patients. Pelvic fracture (6/20, 30.0%) was the most com-
monly associated diagnosis involved, followed by splenic
The 3 patients with false positive FAST results were injury (5/20, 25.0%), liver injury (4/20, 20.0%) and me-
all seen in the afternoon. Their absence of free intra- senteric injury (4/20, 20.0%). The clinical details of 20
peritoneal fluid was confirmed by PME in an 8-year-old patients with false negative FAST results were summa-
child and by CT scan in 2 middle-aged male patients. rized in Table 5.
Among the 20 patients with false negative FAST results,
6 (30.0%), 8 (40.0%) and 6 (30.0%) attended in morn- Table 3. Comparison of FAST performance with respect
ing (07:00-13:59), afternoon/evening (14:00-21:59) and to ISS
night shift (22:00-06:59) respectively. In comparison to ISS<15 ISS>15
130 patients with true positive/negative FAST, a greater No. of patient 64 (100%) 89 (100%)
proportion of mistakes occurred in night shift (30.0% vs True positive 6 (9.4%) 14 (15.7%)
18.5%, Table 4). False positive 2 (3.1%) 1 (1.1%)

True negative 52 (81.3%) 58 (65.2%)

Table 1. Demographic characteristics of included
False negative 4 (6.3%) 16 (18.0%)
patients (n=153)
Demographic variables n (%) Sensitivity 60.0% 46.7%

Gender Specificity 96.3% 98.3%

Male 103 (67.3) Positive PV 75.0% 93.3%

Female 50 (32.7) Negative PV 92.9% 78.4%

Average age (yr) 48.6 Positive LR 16.20 27.50

<18 years 6 (3.9) Negative LR 0.42 0.54

Day 90 mortality 33 (21.6) Ac curacy 90.6% 80.9%

Gold standard test

LAP 25 (16.3) Table 4. Demographic characteristics of patients with

CT scan 130 (85.0) false negative FAST versus true positive/negative results
F al s e n e g at i v e True positive/nega-
PME 9 (5.9)
(n=20) tive (n=130)
Gender
Table 2. Overall performance of FAST (n=153) Male 12 89
Variables
Female 8 41
True positive 20 (13.1%)
Average age (yr) 47.2 (22-80) 49.1 (3-94)
False positive 3 (2.0%)
<18 years old 0 (0.0%) 5 (3.8%)
True negative 110 (71.9%)
Day 90 mortality 8 (40.0%) 24 (18.5%)
False negative 20 (13.1%)
Attendance
Sensitivity 50.0%
Morning 6 (30.0%) 46 (35.4%)
Specificity 97.3%
Afternoon/evening 8 (40.0%) 60 (46.2%)
Positive PV 87.0%
Night 6 (30.0%) 24 (18.5%)
Negative PV 84.6%
ISS>15 4 (20.0%) 58 (44.6%)
Positive LR 18.80
ISS<15 16 (80.0%) 72 (55.4%)
Negative LR 0.51

Ac curacy 85.0%
. 276 . Chinese Journal of Traumatology 2012;15(5):273-278

Table 5. Clinical characteristics of 20 patients with false negative FAST results

Age(yr) Sex Injury mechanism ISS Peritonism Associated injury Gold standard test
51 M RTA >15 Yes Mesenteric tear, splenic injury CT: hemoperitoneum;

LAP: 2 L intraperitoneal blood

27 M RTA >15 Yes Mesenteric tear CT: pelvic blood

80 F Compression >15 No Limb injury CT: small amount perihepatic, perisplenic,

pelvic fluid

22 M Crush <15 Yes Pancreatic laceration CT: lesser sac free fluid
77 M Fall >15 No Tension hemopneumothorax CT: slight subhepatic free fluid

32 M RTA >15 No Liver laceration LAP: 2 L hemoperitoneum

28 F Fall >15 No Hemothorax PME: 150 ml hemoperitoneum

78 F RTA >15 No Traumatic subarachnoid CT: perisplenic blood

haemorrhage

23 M RTA >15 Yes Liver laceration, mesenteric tear LAP: 2.5 L intraperitoneal blood
28 F RTA >15 Yes Macerated spleen, liver laceration CT: ascites; LAP: gross hemoperitoneum

53 F RTA >15 No Hemothorax pelvic fracture CT: left subphrenic, perisplenic and para-
colic gutter fluid

26 F Fall <15 No Pelvic fracture CT: Morrison’s pouch fluid

53 M Fall >15 No Pelvic fracture, splenic laceration CT: trace pelvic blood
69 M RTA >15 No Splenic laceration CT: perisplenic and perihepatic free fluid;
LAP: 5 L intraperitoneal blood

62 M Hit by fallen object >15 Yes Pelvic fracture CT: trace pelvic fluid
58 M RTA >15 No Pelvic fracture, retroperitoneal LAP: <20 ml blood

haematoma

45 M Fall >15 No Bladder CT: trace intraperitoneal fluid

37 M RTA <15 No Pelvic fracture CT: trace pelvic ascites

52 M RTA >15 No Liver, spleen and LAP: 500 ml blood in right side

mesenteric laceration abdomen and left subphrenic space

43 F Fall <15 No Spinal fracture CT: trace ascites
RTA: road traffic accident.

DISCUSSION can be concluded that FAST performed in this trauma

This investigation is unique in that it is the first local
study of its kind. It reveals that FAST performed in this
trauma centre had a high specificity (97.3%) and accu-
racy (85.0%) while the sensitivity (50.0%) was only
borderline. However, the sensitivity and specificity of a
diagnostic test are often influenced by the prevalence
of the target disease. In this study, almost half of the
BAT patients did not have a confirmatory test for intra-
abdominal fluid post-injury. Thus the true prevalence of
intraabdominal fluid post-injury is unknown. On the other
hand, this study did show a high positive LR. As the LR
is less affected by prevalence of the target disease, it
centre is useful to rule in intraabdominal bleeding if
positive result is shown. Worldwide, most studies on
the performance of FAST only report the sensitivity and
specificity, and the results are variable. When compar-
ing the results between different institutions in the United
States6,7, Canada8, Australia2,9, Turkey10, Israel11, South
Africa4 and Taiwan12, there is a wide range of sensitivi-
ties (43.0%-91.7%), specificities (91.6%-100%) and
accuracies (9%-96%). The overall results of FAST in
the trauma centre we investigated (sensitivity 50.0%;
specificity 97.3%; accuracy 85.0%) were compatible
with the reported findings of similar institutions in other
parts of the world.
Chinese Journal of Traumatology 2012;15(5):273-278
Given the high positive LR of FAST in this study,
further surgical intervention is indicated if a BAT patient
has a positive FAST result. On the contrary, clinical
judgment should be exercised in the face of a negative
FAST result. Further imaging studies should be con-
sidered as it is well known that ultrasonography is not
reliable to detect solid organ or bony pelvic injuries.13
This is reflected by the finding that a relatively high per-
centage of BAT patients with false negative FAST re-
sults have pelvis, spleen, mesentery or liver injuries
undetected in the first instance.
In this study, FAST seems to perform less well in
patients with more severe injuries (ISS>15). The higher
false negative rate and negative LR suggest that a nega-
tive FAST result in a severely injured BAT patient should
be dealt with caution. This may be because that the
trauma team is distracted to manage other associated
injuries or their performance is adversely affected by
the complexity of injuries. This, however, requires fur-
ther investigation.
An ideal investigation for the detection of intraperi-
toneal free fluids in blunt trauma patient should be
sensitive, specific, noninvasive or radioactive, quick
and easy to perform, nonoperator dependent, and suit-
able for hemodynamically unstable patients. However,
none of FAST, LAP or CT scan can satisfy all these
criteria. CT scan and/or LAP are well-established gold
standard confirmatory test for FAST in many previous
studies.2-4,6-12 PME result was less commonly used as
gold standard test. It remains controversial whether a
benign course of clinical observation could be regarded
as a true negative examination. Strictly speaking, this
is not an acceptable methodology. Since without a gold
standard confirmatory test, it would not be possible to
ascertain the incidence of misclassification. Therefore,
only patients having gold standard tests were included
in our study.
Previous studies have demonstrated that hands-on
experience of emergency physicians reduces the inci-
dence of specific technical errors when performing
FAST.15 Also, it has been shown that dedicated emer-
gency department ultrasound training rotation improves
residents’ interpretation accuracy.16 At present, not all
doctors in Hong Kong participating in trauma team have
attended training course of FAST. It is suggested that
enhancement in formal training would lead to further
. 277 .
improvement in the overall performance of this
noninvasive diagnostic investigation.17
However, this study still has several limitations. First,
the investigation is only done in one trauma centre in
Hong Kong, and thus the findings may not be
generalizable to other centres. Further, within the
trauma team, there is a constant rotation of doctors
from various specialties and hospitals. Therefore, no
comments can be made whether the performance is
affected by the clinical background of individual FAST
performers. And subjects are retrieved from the trauma
registry of the hospital, thus case selection bias may
occur since other patients who were not captured in
this registry may have undergone FAST as well.
Moreover, FAST is performed before gold standard con-
firmatory test. The time lag between FAST and con-
firmatory test can lead to bias. Intraabdominal bleeding
may occur after the FAST is done. Lastly, there are
only 6 (4.0%) patients under 18 years old in our series;
therefore our result can not be applied in this paediatric
patient group.
In conclusion, FAST is an accurate investigation in
initial assessment of BAT patients. Its diagnostic val-
ues can be affected by severity of injury. Enhanced
staff training is required to further improve its usefulness.
Acknowledgement
We would like to thank Dr. Lam Wai Man Wendy, Service
Director, Department of Radiology, Queen Mary Hospital,
Hong Kong who has provided valuable radiological expert
opinion on some cases in this study.
REFERENCES
1. Chambers JA, Pilbrow WJ. Ultrasound in abdominal trauma:
an alternative to peritoneal lavage. Arch Emerg Med 1988;5(1):26-33.
2. Hsu JM, Joseph AP, Tarlinton LJ, et al. The accuracy of
focused assessment with sonography in trauma (FAST) in blunt
trauma patients: experience of an Australian major trauma service.
Injury 2007;38(1):71-5.
3. Buzzas GR, Kern SJ, Smith RS, et al. A comparison of
sonographic examinations for trauma performed by surgeons and
radiologists. J Trauma 1998;44(4):604-6.
4. Smith ZA, Postma N, Wood D. FAST scanning in the de-
veloping world emergency department. S Afr Med J 2010;100(2):
105-8.
. 278 .
5. http://www.ha.org.hk/visitor/ha_visitor_index.asp?
Content_ID=10176&Lang=ENG&Dimension=100&parent_ID=
10084&Ver=HTML.
6. Lee BC, Ormsby EL, McGahan JP, et al. The utility of
sonography for the triage of blunt abdominal trauma patients to ex-
ploratory laparotomy. AJR Am J Roentgenol 2007;188(2):415-21.
7. Natarajan B, Gupta PK, Cemaj S, et al. FAST scan: is it
worth doing in hemodynamically stable blunt trauma patients?
Surgery 2010;148(4):695-700.
8. Boulanger BR, Brenneman FD, McLellan BA, et al. A pro-
spective study of emergent abdominal sonography after blunt
trauma. J Trauma 1995;39(2):325-30.
9. Vassiliadis J, Edwards R, Larcos G, et al. Focused assess-
ment with sonography for trauma patients by clinicians: initial
experience and results. Emerg Med (Fremantle) 2003;15(1):42-8.
10. Unlüer EE, Yavasi O, Kara PH, et al. Paramedic-performed
Focused Assessment with Sonography in Trauma (FAST) in the
emergency department. Ulus Travma Acil Cerrahi Derg 2011;17
(2):113-6.
11. Beck-Razi N, Fischer D, Michaelson M, et al. The utility
of focused assessment with sonography for trauma as a triage tool
in multiple-casualty incidents during the second Lebanon war. J
Ultrasound Med 2007;26(9):1149-56.
Chinese Journal of Traumatology 2012;15(5):273-278
12. Liu M, Lee CH, P’eng FK. Prospective comparison of
diagnostic peritoneal lavage, computed tomographic scanning, and
ultrasonography for the diagnosis of blunt abdominal trauma. J
Trauma 1993;35(2):267-70.
13. Kendall JL, Faragher J, Hewitt GJ, et al. Emergency de-
partment ultrasound is not a sensitive detector of solid organ injury.
West J Emerg Med 2009;10(1):1-5.
14. Costa G, Tierno SM, Tomassini F, et al. The epidemiology
and clinical evaluation of abdominal trauma. An analysis of a
multidisciplinary trauma registry. Ann Ital Chir 2010;81(2):95-102.
15. Jang T, Kryder G, Sineff S, et al. The technical errors of
physicians learning to perform focused assessment with
sonography in trauma. Acad Emerg Med 2012;19(1):98-101.
16. Mahler SA, Swoboda TK, Wang H, et al. Dedicated emer-
gency department ultrasound rotation improves residents’ ultra-
sound knowledge and interpretation skills. J Emerg Med 2011;43
(1):129-33.
17. Arafat R, Golea A, Dǎ rǎ mus I, et al. Medical education
for emergency physician focused on basic competence (Focused
Assessment with Sonography in Trauma). Evaluation of the Ro-
manian national program: “Regional Emergency Medical Services
Systems”. Med Ultrason 2011;13(4):283-91.
(Received May 28, 2012)
Edited by LIU Gui-e