- Email: [email protected]
The halo sign and peripancreatic fluid: useful CT signs of hypovolaemic shock complex in adults
Clinical Radiology (2005) 60, 599–607
The halo sign and peripancreatic fluid: useful CT signs of hypovolaemic shock complex in adults M.F. Ryana,*, P.A. Hamiltona, J. Sarrazina, P. Chub, O. Benjaminova, K. Lamc Departments of aMedical Imaging, bSurgery, and cBiostatistics, Sunnybrook and Women’s College Health Sciences Centre, University of Toronto, 2075 Bayview Avenue, Toronto, Ont., Canada M4N 3M5 Received 20 October 2003; received in revised form 20 January 2004; accepted 18 February 2004
KEYWORDS Abdomen; Signs in imaging; Trauma; Emergency radiology; Computed tomography
AIM: To report two new, useful computed tomography (CT) signs of the hypovolaemic shock complex (HSC) in adults admitted after blunt abdominal trauma: the halo sign (ring of fluid around a collapsed intra-hepatic inferior vena cava (IVC)), and peripancreatic retroperitoneal fluid. MATERIALS AND METHODS: CT images of 498 consecutive patients admitted after blunt abdominal trauma were reviewed, of which 27 had CT signs of the HSC. The CT images of these 27 patients were analysed. A control group of 101 patients examined using CT for suspected blunt abdominal trauma who did not have the HSC were chosen for comparison. RESULTS: The most common features involved the vascular compartment: diminished IVC diameter ðn ¼ 27Þ; a positive halo sign ðn ¼ 21Þ; diminished anteroposterior diameter of the aorta ðn ¼ 13Þ; and abnormal vascular enhancement ðn ¼ 10Þ: Peripancreatic retroperitoneal fluid in the absence of pancreatic injury, pancreatitis or pancreatic disease was observed in eight patients. Hollow visceral abnormalities included: diffuse increased mucosal enhancement of both the small and large bowel ðn ¼ 19Þ; diffuse thickening of the small bowel wall ðn ¼ 11Þ; and small bowel dilatation ðn ¼ 7Þ: Solid visceral abnormalities included both decreased and or increased enhancement. Several concomitant intra- and extra-abdominal injuries were also identified. CONCLUSION: In the setting of blunt abdominal trauma, early abdominal CT can show diffuse abnormalities due to the HSC, which occasionally may alert clinicians of unsuspected shock. Recognition of these signs as distinguished from injured viscera is important in order to avoid unnecessary laparotomy. Two new signs are described: the halo sign and peripancreatic retroperitoneal fluid. q 2004 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction Hypovolaemic shock complex (HSC) is an infrequently encountered entity found using computed tomography (CT) in victims of severe trauma.1 – 8 Shock is defined as inadequate organ perfusion to maintain oxygen delivery to the capillary exchange *Guarantor and correspondent: Max F. Ryan, Department of Clinical Radiology, Cork University Hospital, Wilton, County Cork, Ireland. Tel.: þ 1-353-21-546400; fax: þ1-353-21-343307. E-mail address: [email protected]
beds along with inadequate removal of the acid metabolites from the tissues, and is thus associated with poor prognosis.9 Clinically, early recognition of shock is mandatory, and treatment should be initiated simultaneously with the identification of a probable cause. The response to initial treatment, coupled with physical assessment, usually provides sufficient information as to the cause of the shock state. Nevertheless, despite close monitoring by the trauma team, a number of patients are clinically silent before systemic manifestations occur. Thus,
0009-9260/$ - see front matter q 2004 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2004.02.012
600
M.F. Ryan et al.
a few patients go undetected before emergent diagnostic imaging.1 The diagnostic accuracy of CT as a primary diagnostic tool in the management of patients admitted after blunt abdominal trauma is well established.10,11 In HSC, a spectrum of vascular and visceral CT signs have been described.3,6,12 – 14 The vascular signs include reduced calibre of major intra-abdominal vessels and increased luminal enhancement. Visceral signs may involve hollow viscera (diffuse small bowel thickening, mucosal enhancement and dilatation) and, or solid viscera (increased or reduced end organ perfusion).3,6,12 – 14 The main importance of recognizing the CT manifestations of HSC is to spare these patients unnecessary surgery. The purpose of this report is to describe and assess the value of a spectrum of specific abdominal CT imaging parameters in identifying patients in shock and in predicting their clinical outcome.
above, at and 2 cm below the origin of the renal arteries.15,16,17 Hollow visceral abnormalities includes: small bowel dilatation (. 2.5 cm); bowel wall thickening (. 3 mm) and increased bowel wall mucosal enhancement (greater than the psoas muscle);4,14 enhancement and dilatation of the colon; and increased gallbladder mucosal enhancement.16 – 18 Solid visceral abnormalities includes: intense adrenal gland enhancement (greater than the IVC)5 and renal cortical enhancement (greater than that of the aorta).1,3 Two features of the hypovolaemic shock complex not previously described were evaluated: the halo sign, a ring of oedema around a collapsed intrahepatic IVC (ratio of the short axis diameter of the opacified intra-hepatic IVC less than half the short axis diameter of the intra-hepatic IVC bed at the same level), and low density extra-peritoneal peripancreatic fluid (lower than 20 HU) around the entire gland (in the absence of biochemical or CT evidence of pancreatitis or pancreatic disease).
Methods and materials
Imaging technique
From our comprehensive trauma database (Microsoft, Excel 2000), a retrospective review of all posttraumatic abdominal CT images and medical records of 498 consecutive patients admitted to our level 1 trauma centre, from January 1998 to June 2000 was undertaken.
The indication for abdominal and pelvic CT examination in all cases was suspected intra-abdominal injury, based on a history of blunt abdominal trauma and an initial assessment including a physical examination. All patients were managed according to standard ATLS protocols. After this initial evaluation, and resuscitation with intravenous fluid support, all patients were thought to be haemodynamically stable enough to undergo abdominal CT examination. A helical CT machine (GE HiSpeed Advantage, WI, USA) was used in all cases: 7 mm collimation, pitch value of 1 and 50% overlapped slices to reconstructions with 3.5 mm inter-scan spacing were obtained. A rapid power-injected bolus of 2 ml/kg body weight, up to a maximum of 150 ml of non-ionic contrast material (Omnipaque 300, Nycomed Amersham plc, Buckinghamshire, England) at a rate of 3 ml/s was used. Twenty-five of 27 (92%) patients received oral or nasogastric contrast medium (100 – 500 ml Hypaque 2%, Amersham Health Inc.) in the trauma room before transfer to CT.
The clinical diagnosis of shock The clinical diagnosis of shock was established by a combination of several established clinical and biochemical findings including: hypotension defined as a systolic blood pressure less than 100 mmHg, tachycardia and reduced urine output (advanced trauma life support (ATLS) manual). In addition, supportive criteria such as blood transfusion requirements within the first 24 h of admission were included.
The CT manifestations of shock According to the literature,1 – 3,15,16,17 various CT signs of HSC have been described. The CT signs of hypovolaemia include diminished calibre of the inferior vena cava (IVC), defined as an anterior– posterior diameter of less than 9 mm at three levels (within the intra-hepatic portion of the IVC, at and 2 cm below the level of the renal arteries);4,7 diminished calibre of the aorta, defined as an anteroposterior diameter less than 1.3 cm, 2 cm
Study group Two authors who were skilled in interpreting abdominal and pelvic trauma CT images identified 27 of 498 patients with CT signs of the HSC based on the combination of two or more vascular or visceral criteria aforementioned above by consensus
The halo sign and peripancreatic fluid: useful CT signs of hypovolaemic shock complex in adults
601
Table 1 Incidence of hypovolaemic shock on 498 post-traumatic abdominal CT images Status
Overall frequency
Survived frequency
Died frequency
No hypovolaemic shock complex Hypovolaemic shock complex
471 (94.5%) 27 (5.5%)
411 (87.3%) 8 (29.6%)
60 (12.7%) 19 (70.4%)
opinion. One of these authors, obtained electronic calliper and density measurements of the relevant vasculature, solid and hollow viscera and fluid collections from all CT examinations on a CT viewing console. Other data collected included: the presence of intra- and extra-peritoneal fluid and density measurements (other than peripancreatic fluid).
Control group A control population examined consisted of 101 patients randomly selected from the remaining group of 498 trauma patients. The data collected in this group included: the presence or absence of the halo sign and the presence of low attenuation (, 20 HU) peripancreatic fluid.
Statistical methods SAS Version 8 and Microsoft Excel were used to conduct all the statistical analysis. General descriptive statistics were first created for the 27 HSC patients. Ninety-five percent confidence intervals (CI) were constructed for both the HSC and the control group with reference to the mean maximum anteroposterior calibre of the aorta and IVC, as well as the proportion having the halo sign, and peripancreatic oedema; among other signs that were subsequently compared using two-sided Z-test. The significance level was originally set at 5% level, but adjusted by Bonferroni method to account for the case of multiple testing. As a result, any test that gave a p-value of less than 0.005 indicates a statistically significant result.
Results Demographics Of 498 patients presenting with blunt abdominal trauma, 27 adults were identified as having abdominal CT signs of the HSC (Table 1). There were 17 males and 10 females, all between the ages of 18 – 84 years with mean age of 39.6 years (Table 2). The control group was age matched (95% CI (Control—HSC) 2 6.2272, 10.3145). The indication for CT examination was blunt abdominal trauma in all 27 patients and included: motor vehicle accidents as a passenger or driver ðn ¼ 24Þ; pedestrians involved in a motor vehicle accident ðn ¼ 2Þ; and a fall from a height ðn ¼ 1Þ: The demographics of our patients are summarized in Table 3. All patients were thought to be haemodynamically stable after initial resuscitation therapy that included intravenous volume replacement in the trauma room. Eleven patients that were thought to be haemodynamically stable before emergent abdominal CT image became hypotensive (, 100 mmHg) after the CT study. The CT findings were confirmed in all patients by surgery or autopsy. Twenty-six patients underwent laparotomy for associated visceral injuries. Nineteen patients (70.4%) died as a result of their injuries of which 12 (63.2%) died within 24 h.
CT findings The CT examinations were performed within a mean time of 2.7 h, ranging from 0.5 – 18.6 h of the injury. The CT results observed in our series are summarized in Tables 4 and 5. The most common features involved the vascular compartment: a
Table 2 Age and gender distribution of 27 patients with the hypovolaemic shock complex Item
Overall mean ^ SD ðn ¼ 27Þ
Survived mean ^ SD ðn ¼ 8Þ
Died mean ^ SD ðn ¼ 19Þ
Age (years) Sex (proportion of males) Sex (proportion of females)
39.6 ^ 4.2 17 (63.0%) 10 (37.0%)
25.8 ^ 3.1 6 (75.0%) 2 (25.0%)
45.5 ^ 5.3 11 (57.9%) 8 (42.1%)
602
M.F. Ryan et al.
Table 3 General demographics of 27 patients with the hypovolaemic shock complex Demographics Initial BP (mmHg) Injury severity score Glasgow coma scale Blood units required in 24 h
diminished IVC diameter at multiple levels was observed in all 27 patients (100%; Figs. 1 – 7), mean value 3.9 mm as opposed to 14.5 mm for the controls (p , 0:0001; Table 6). Twenty-one (77.8%) patients demonstrated a positive halo sign (Figs. 2 – 5; Table 7). The halo sign was not observed in the control group ðp , 0:0001Þ: The halo sign was more prevalent in the deceased group (n ¼ 15; Table 4). The sensitivity and specificity of the halo sign for HSC in this series were 77.8 and 100%, respectively. Diminished anteroposterior diameter of the aorta was observed in 13 patients (Figs. 1 – 4 and 7), with a mean value of 12.5 mm recorded over three different levels (Tables 1 –3, 6 and 7). The mean value in the control group was 17.1 mm. Extra-peritoneal fluid was observed in 21 (77.8%) patients. The presence of extra-peritoneal fluid was in some cases attributable to interstitial oedema due to the HSC alone, and in other cases blood as a result of injured viscera. Low-density (, 20 HU) peripancreatic extra-peritoneal fluid was observed in 8 patients (29.6%), p ¼ 0:002; (Figs. 1, 6 and 7; Tables 4 and 7). In no case was pancreatitis or pancreatic injury recorded on
Figure 1 CT image of the upper abdomen in a 34-yearold male involved in a motor vehicle accident. This shows several features of the HSC: a dense small calibre aorta (9 mm) and flat IVC (2 mm), peripancreatic oedema, splenic hypoperfusion and dense enhancement of the right adrenal gland (black arrow). Note a large layering intra-abdominal haematoma is evident in the right subphrenic space.
serum biochemistry, CT, surgery or autopsy. Peripancreatic fluid was observed more frequently in the deceased group (Table 4) and 7 of 8 (87.5%) patients with peripancreatic oedema died.
Discussion Hypovolaemic shock is a state of circulatory dysfunction resulting in a sudden decrease in the intra-vascular blood volume relative to the vascular
Table 4 Vascular and visceral CT manifestations of 27 patients with the hypovolemic shock complex Overall frequency ðn ¼ 27Þ
Survived frequency ðn ¼ 8Þ
Died frequency ðn ¼ 19Þ
Vascular manifestations Diminished aortic diameter (mm) Collapsed inferior vena cava (mm) Halo signa Flat renal veins
13 (48.0%) 27 (100.0%) 21 (77.8%) 27 (100.0%)
3 (37.5%) 8 (100.0%) 6 (75.0%) 8 (100.0%)
10 (52.6%) 19 (100.0%) 15 (78.9%) 19 (100.0%)
Visceral manifestations Increased bowel wall enhancement Bowel wall thickening Small bowel dilatation Increased adrenal gland enhancement Gallbladder wall enhancement Increased renal cortical enhancement Diminished splenic enhancement Peripancreatic oedema Increased pancreatic enhancement Decreased hepatic enhancement Gastric dilatation
19 (70.4%) 11 (40.7%) 7 (25.9%) 14 (51.9%) 9 (33.3%) 15 (55.6%) 8 (29.6%) 8 (44.4%) 1 (3.7%) 3 (11.1%) 19 (70.4%)
5 (62.5%) 2 (25.0%) 3 (37.5%) 4 (50.0%) 3 (37.5%) 7 (87.5%) 1 (12.5%) 1 (12.5%) 0 (0.0%) 0 (0.0%) 7 (87.5%)
14(73.7%) 9 (47.4%) 4 (21.1%) 10(52.6%) 6 (31.6%) 8 (42.1%) 7 (36.8%) 7 (87.5%) 1 (100%) 3 (15.8%) 12(63.2%)
a
Halo sign ¼ rim of fluid around a collapsed intra-hepatic inferior vena cava
The halo sign and peripancreatic fluid: useful CT signs of hypovolaemic shock complex in adults
603
Table 5 Other injuries in 27 patients with the hypovolaemic shock complex Injury
Overall frequency ðn ¼ 27Þ
Survivors ðn ¼ 8Þ
Deceased ðn ¼ 19Þ
Intra-peritoneal fluid Extra-peritoneal fluid Active arterial extravasations Pelvic fractures Bladder rupture Liver laceration Splenic injury Renal injury Adrenal haematoma Urethral tear Ureteral tear Pancreatic injury Head injurya Head injury—contusion Head injury—traumatic subarachnoid haemorrhage Subdural haemorrhage Spinal injury
21 (77.8%) 21 (77.8%) 13 (48.1%) 25 (92.6%) 5 (18.5%) 8 (29.6%) 14 (51.8%) 9 (33.3%) 2 (7.4%) 5 (18.5%) 1 (3.7%) 0 (0.0%) 23 (85.2%) 10 (37.0%) 9 (33.3%) 5 (18.5%) 3 (11.1%)
5 (62.5%) 6 (75.0%) 3 (37.5%) 7 (87.5%) 3 (37.5%) 5 (62.5%) 4 (50.0%) 3 (37.5%) 0 (0.0%) 3 (37.5%) 1 (12.5%) 0 (0.0%) 7 (87.5%) 2 (25.0%) 2 (25.0%) 2 (25.0%) 0 (0.0%)
16 (84.2%) 15 (78.9%) 10 (52.6%) 18 (94.7%) 2 (10.5%) 3 (15.8%) 10 (52.6%) 6 (31.6%) 2 (10.5%) 2 (10.5%) 0 (0.0%) 0 (0.0%) 16 (84.2%) 8 (42.1%) 7 (36.8%) 3 (15.8%) 3 (15.8%)
a
Head injury ¼ all head injuries including concussion, contusion, cerebral oedema and subarachnoid haemorrhage.
Table 6 Comparison of mean aorta and IVC diameters of 27 HSC patients and 101 controls Mean and standard error
Aorta IVC
Hypovolaemic shock complex patients ðn ¼ 27Þ
Controls ðn ¼ 101Þ
95% CI for difference (control—HSC)
p-Valuea
12.5 ^ 0.5 mm 3.9 ^ 0.3 mm
17.3 ^ 0.3 mm 14.5 ^ 0.4 mm
[3.497, 5.948] [9.125, 12.1303]
,0.0001 ,0.0001
Aorta ¼ mean anterioposterior diameter of aorta at three locations: supra-renal, renal and infra-renal. IVC ¼ mean diameter of inferior vena cava at three locations: supra renal, renal and infra-renal levels. CI, confidence interval. a Wilcoxon rank sum test p-value.
Figure 2 CT image of the upper abdomen in a 45-yearold male involved in a motor vehicle accident. This shows the “halo sign”. This refers to a rim of oedema surrounding a collapsed intra-hepatic IVC.
Figure 3 CT image of the upper abdomen at the level of the aortic hiatus in a 55-year-old man involved in a motor vehicle accident. This shows the “halo sign”, a rim of oedema surrounding a collapsed intra-hepatic IVC (3 mm). This finding was present in 21 (78%) of cases in our series.
604
Figure 4 CT image of the upper abdomen in a 37-yearold man with HSC. The “halo sign”, a rim of oedema surrounding a collapsed intra-hepatic IVC, is evident. Note a large liver laceration that is of higher attenuation (35 HU) than the fluid surrounding the collapsed IVC (12 HU).
capacity, to the extent that effective tissue perfusion cannot be maintained. It is thus a failure to meet the metabolic demands of the organs, initially leading to reversible and then to irreversible cellular injury.9 Several theories about the pathogenesis of HSC have been reported. These include the activation of complement, coagulation, cytokine and kallikrein – kinin systems pathways.9 The end result is hypovolaemia, vasoconstriction, and replacement of the intravascular volume with intravenous contrast agent.12 The incidence of hypovolaemic shock is variable
Figure 5 CT image of the mid-abdomen in a 25-year-old woman involved in a motor vehicle accident. This shows low attenuation peripancreatic oedema (9 HU) in the anterior para-renal extra-peritoneal space surrounding a normal pancreas. Note increased gallbladder mucosal enhancement.
M.F. Ryan et al.
Figure 6 CT image of the upper abdomen at the level of the coeliac axis, showing retroperitoneal peripancreatic oedema (11 HU) in a 47-year-old woman. A normal pancreatic body and tail is noted at this level. Features of shock bowel are also noted including: increased small bowel thickening and dilatation. There is flattening on the IVC that was shown on several continuous images. Note increased renal parenchymal enhancement, which was a common, but unreliable, finding in our series.
and largely depends on the referral centre. In our experience, HSC occurs at a frequency of about 5% in patients admitted after multiple trauma. Haemorrhage is the most common cause of shock and is due to a combination of blood and fluid losses due to tissue damage and release of vasoactive substances.9,20 Despite recent advances in the treatment and thus increase in survival of critically
Figure 7 HSC in a 29-year-old male motor vehicle accident victim. Intravenously administered contrast material-enhanced CT image obtained at the level of the pancreas. This shows several features of the HSC: small calibre vasculature, extra-peritoneal peripancreatic oedema, diffuse mural thickening and enhancement of thickened small bowel loops. Note the oral contrast agent was not administered to this patient.
The halo sign and peripancreatic fluid: useful CT signs of hypovolaemic shock complex in adults
605
Table 7 Ninety-five percent confidence intervals (CI) for various signs of HSC HSC patients ðn ¼ 27Þ
Observed percentage/SD
95% CI
Halo sign Peripancreatic oedema Decreased splenic enhancement Decreased hepatic enhancement Adrenal enhancement Gallbladder wall enhancement
77.8% (8.0%) 29.6% (9.6%) 29.6% (8.8%) 11.1% (6.0%) 51.9% (9.6%) 33.3% (9.1%)
(59.2%, 89.4%) (12.4%, 46.9%) (12.4%, 46.9%) (0.0%, 23.0%) (34.0, 69.3%) (15.6%, 51.1%)
Controls ðn ¼ 101Þ Halo sign Decreased splenic enhancement Decreased hepatic enhancement Adrenal enhancement Gall bladder wall enhancement
Observed Percentage/ proportion/SD 0.0% 1.0% (1.0%) 8.0% (2.7%) 2.0% (1.4%) 10.9% (1.9%)
95% CI (0, 3.7%) (0.0%, 2.9%) (2.7%, 13.2%) (0.0%, 3.7%) (6.19%, 18.5%)
a
p-Valuea ,0.0001 0.005 0.0229 ,0.0001 0.0134
Two-sided Z-test (comparing proportions of cases and controls).
injured patients, the majority of patients with multiple injuries will have an element of hypovolaemia.20 To correct this hypovolaemia, preliminary resuscitation with fluids is instituted to promote and maintain adequate tissue perfusion and to prevent further metabolic decline. Volume replacement is directed by the response to this therapy rather than by reliance on the initial classification of blood loss. Blood loss is further complicated by shifts of fluids among the fluid compartments, in particular in the extravascular, extracellular fluid compartment.20 These fluid changes may be directly due to major soft-tissue and bone injuries, or secondary to activation of a systemic inflammatory response. Termed the systemic inflammatory response syndrome, this is directly related to the extent of the soft-tissue injury and is associated with the production and release of multiple cytokines.9 These hormones affect vascular endothelium causing increased permeability and thus tissue oedema. Increased permeability may accentuate hypovolaemia by shifts of fluid from the intravascular into the extravascular, extracellular space. The treatment of hypovolaemia may be further complicated by a combination of coexisting cardiogenic, neurogenic and even septic shock. Furthermore, the effect of the vasopressors and fluids used for resuscitation may further accentuate the features.20 Patients that have responded to initial treatment in the emergency room, and who are thought to be haemodynamically stable enough to undergo CT examination, may nonetheless have a poorly compensated shock state, not recognizable with blood pressure and pulse readings alone. These patients may become hypotensive during or immediately after the CT study.12,21,22 On CT, the findings observed in the HSC may be classified as vascular and non-vascular.1,3,7,15 – 17
The potential CT findings are numerous, with variable association of one to another. In addition, when haemodynamic instability is corrected, only end-organ anomalies may be observed. In the HSC, these end organ findings relate to hypovolaemia and not to structural lesions of the involved viscera, and thus require supportive therapy rather than a surgical procedure per se. In our series, collapse of the IVC at multiple levels associated with flattening of the renal veins was most common (Fig. 7). This is due to decreased circulating blood volume and thus reduced venous return in patients with systemic hypotension.7 As a normal IVC may appear flattened on a single abdominal section, the diagnosis of a collapsed IVC was made when flattening was observed at multiple levels within the abdomen. The proposed halo sign refers to a collapsed intra-hepatic IVC, as a function of hypovolaemia, surrounded by a circumferential zone of low attenuation, probably representing extracellular fluid (Figs. 2 – 5). This sign was most frequently observed at a level corresponding with the superior segments of the liver, below the confluence of the hepatic veins. This finding was observed in 21 of 27 (77.8%) patients ðp , 0:0001Þ and in the setting of trauma is considered a specific sign of HSC and thus probably a poor prognostic indicator. For example, 15 of 21 (71.4%) of patients with the halo sign died. The halo sign was not observed in any patient in the control group nor in any of the 471 non-HSC trauma CT cases initially reviewed. The halo sign is not specific to HSC. The authors have observed similar findings in patients with underlying liver disease and liver congestion. Other reports have also observed similar findings around portal vasculature in patients with biliary cirrhosis, hepatitis, and tumours that obstruct lymphatic drainage at the porta hepatis.23,24
606
In our series, low attenuation (, 20 HU) extracellular fluid in the anterior extra-peritoneal compartment around the pancreas and base of the mesentery is reported (Figs. 1, 6 and 7). This manifestation has not been previously described, but was observed in eight cases, of which seven (87.5%) died. In no case was a pancreatic injury observed on CT, or recorded at surgery or autopsy. In all cases, the pancreas appeared normal in size and uniform in attenuation. Extracellular fluid was originally reported in the HSC3,19 but was not included in subsequent reports.2,12 In the setting of shock, extracellular fluid is thought to be related to loss of pre-capillary arteriolar sphincter tone that causes arteriolar constriction, and thus blood flows into the ischaemic microcirculatory capillary bed, but venous constriction hinders its return. Blood that pools in the peripheral circulation causes increased hydrostatic pressure, which in the presence of increased capillary membrane permeability, forces fluid and protein out of the vascular space into the interstitial space further decreasing the circulating blood volume.9 As a result blood volume expanders are only temporarily retained in the circulation and are only of transient therapeutic benefit.9 Small bowel manifestations of the hypoperfusion shock complex are better known than the halo sign and peripancreatic extracellular fluid signs.3,4,12,14 Termed shock bowel, this refers to increased small bowel wall mucosal enhancement, mural thickening and luminal dilatation. This is thought to be due to hypovolaemia induced reduced bowel perfusion and subsequent venous vasoconstriction, in response to sympathetic stimulation and possibly vasopressors used in resuscitation.6 Replacement of a depleted intravascular volume with contrast agent along with increased bowel wall permeability results in wall thickening and intense bowel wall enhancement due to diminished perfusion and interstitial leak of molecules of the contrast agent. Accumulation of third space fluid and decreased fluid resorption causes luminal dilatation.14 The small bowel is typically diffusely involved, whereas the colon is sporadically involved. The aforementioned changes do not necessarily lead to irreversible bowel ischaemia and infarction if a minimal threshold of blood flow is maintained through adequate resuscitation measures.4,25 Reported solid visceral signs of the HSC includes a prolonged intense nephrogram. This may be due a combination of hypoperfusion induced renal failure and or contrast induced nephrotoxicity and renal impairment1 (Fig. 7). The present authors find hyperintense renal parenchyma enhancement as a sign of HSC unreliable. Hyperdense renal parench-
M.F. Ryan et al.
yma enhancement is occasionally seen on normal CT examinations and depends on several other factors, including cardiac output and scan acquisition parameters. In the HSC, intense symmetric enhancement of the adrenal glands is reported5 (Fig. 1). The adrenal glands are usually normal in configuration and size. This finding is likely related to a sympathetic response to hypovolaemic shock along with preservation of perfusion to the adrenal glands as a vital organ. Although useful, the authors have also observed this sign in patients after severe burns or surgery. In HSC, low attenuation values of the spleen, is related to severe hypoperfusion rather than an injury to the splenic artery or parenchyma (Fig. 1).13 Splenic arterial flow has no autoregulatory mechanism and is highly sensitive to sympathetic stimulation and thus vasoconstriction.26 Other post-traumatic causes for splenic hypoperfusion include vessel injury, splenic rupture and infarction, and inadequate volume and timing of intravenous contrast agent.7,13,27 In the absence of hepatic injury, abnormal hepatic arterial enhancement is infrequent, as arterial flow is autoregulated and mixed with hepatopetal portal venous inflow. A generalized marked reduction of hepatic enhancement is abnormal and occurs in the most severe form of HSC. In our series, hepatic enhancement was typically heterogeneous (Fig. 1). This feature is less common than other abnormalities, such as the halo sign and peripancreatic extracellular fluid.
Outcome In our series, the HSC was a marker of severe injury and poor outcome. The mortality for adults with the hypovolaemic shock complex was 70%. Mortality was affected by associated injuries, such as active extravasation of arterial contrast and intra-cranial injuries. This reflects the complexity of injuries that occur to victims of severe trauma and the multifactorial aetiologies involved. In conclusion, in the setting of blunt abdominal trauma, attention to the features that may suggest poorly compensated shock is advocated. The CT findings are due to shock rather than structural lesions of the affected viscera per se, and thus require supportive therapy rather than laparotomy.
Acknowledgements The authors thank Michael Schull, MD, Department
The halo sign and peripancreatic fluid: useful CT signs of hypovolaemic shock complex in adults
of Medicine, University of Toronto and William Sharkey, trauma statistics office for their assistance.
13.
14.
References 1. Sivit CJ, Taylor GA, Bulas DI, Kushner DC, Potter BM, Eichelberger MR. Posttraumatic shock in children: CT findings associated with hemodynamic instability. Radiology 1992;182:723—6. 2. Rotondo A, Catalano O, Grassi R, Scialpi M, Angelelli G. Thoracic CT findings at hypovolemic shock. Acta Radiol 1998;39:400—4. 3. Taylor GA, Fallat ME, Eichelberger MR. Hypovolemic shock in children: abdominal CT manifestations. Radiology 1987;164: 479—81. 4. Mirvis SE, Shanmuganathan K, Erb R. Diffuse small-bowel ischemia in hypotensive adults after blunt trauma (shock bowel): CT findings and clinical significance. AJR Am J Roentgenol 1994;163:1375—9. 5. O’Hara SM, Donnelly LF. Intense contrast enhancement of the adrenal glands: another abdominal CT finding associated with hypoperfusion complex in children. AJR Am J Roentgenol 1999;173:995—7. 6. Kaufman RA, Stalker HP. Hypovolemic shock in children. Radiology 1988;166:579—80. 7. Jeffrey Jr RB, Federle MP. The collapsed inferior vena cava: CT evidence of hypovolemia. AJR Am J Roentgenol 1988; 150:431—2. 8. Guyton AC. In: Guyton AC, editor. Circulatory shock and physiology of its treatment. Philadelphia: WB Saunders; 1971. p. 263—71. 9. Shamji FM, Todd TRJ. Hypovolemic shock. Symposium on cardiovascular crises in the critically ill. Crit Care Clin 1985; 1. 10. Amoroso TA. Evaluation of the patient with blunt abdominal trauma: an evidence based approach. Emerg Med Clin North Am 1999;17:63—75. 11. 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;35:267—70. 12. Sivit CJ, Eichelberger MR, Taylor GA. CT in children with rupture of the bowel caused by blunt trauma: diagnostic
15.
16.
17.
18.
19. 20.
21.
22.
23.
24.
25.
26. 27.
607
efficacy and comparison with hypoperfusion complex. AJR Am J Roentgenol 1994;163:1195—8. Berland LL, Van Dyke JA. Decreased splenic enhancement on CT in traumatized hypotensive patients. Radiology 1985; 156:469—71. Hara H, Babyn PS. Significance of bowel wall enhancement on CT following blunt abdominal trauma in childhood. J Comput Assist Tomogr 1992;16:94—8. Shin MS, Berland LL, Ho KJ. Small aorta: CT detection and clinical significance. J Comput Assist Tomogr 1990;14: 102—3. Taber P, Korobkin MT, Gooding CA, Palubinskas AJ, Neuhauser EB. Growth of the abdominal aorta and renal arteries in childhood. Radiology 1972;102:129—34. Calabuig R, Seggerman RE, Weems WA, Weisbrodt NW, Moody FG. Gallbladder and gastrointestinal motility after hemorrhagic shock. Surgery 1990;107:568—73. Moody FG, Calabuig R, Li YF, Harari Y, Rodriguez LF, Weisbrodt NW. Biliary and gut function following shock. J Trauma 1990;30(Suppl. 12):S179—84. Cox TD, Kuhn JP. CT scan of bowel trauma in the pediatric patient. Radiol Clin N Am 1996;34:807—18. Chapter 3: Shock. In Advanced Trauma Life Support for Doctors. Student Course Manual. 6th ed. American College of Surgeons 1997; 88—107. Brody JM, Leighton DB, Murphy BL, et al. CT of blunt trauma bowel and mesenteric injury: typical findings and pitfalls in diagnosis. RadioGraphics 2000;20:1525—36. Radin DR, Boswell Jr WD, Ralls PW, Halls JM. Recognition of hypotensive shock on abdominal computed tomography. J Comput Tomogr 1986;10:299—301. Wenzel JS, Donohoe A, Ford III KL, Glastad K, Watkins D, Molmenti E. Primary biliary cirrhosis: MR imaging findings and description of MR imaging periportal halo sign. AJR Am J Roentgenol 2001;176:885—9. Lawson TL, Thorsen MK, Erickson SJ, Perret RS, Quiroz FA, Foley WD. Periportal halo: a CT sign of liver disease. Abdom Imaging 1993;18:42—6. Bulkley GB, Kvietys PR, Parks DA, Perry MA, Granger DN. Relationship of blood flow and oxygen consumption to ischemic injury in the canine small intestine. Gastroenterology 1985;89:852—7. Goodman LR, Aprahamian C. Changes in splenic size after abdominal trauma. Radiology 1990;176:629—32. Federle MP, Griffiths B, Minagi H, Jeffrey Jr RB. Splenic trauma: evaluation with CT. Radiology 1987;162:69—71.
The halo sign and peripancreatic fluid: useful CT signs of hypovolaemic shock complex in adults M.F. Ryana,*, P.A. Hamiltona, J. Sarrazina, P. Chub, O. Benjaminova, K. Lamc Departments of aMedical Imaging, bSurgery, and cBiostatistics, Sunnybrook and Women’s College Health Sciences Centre, University of Toronto, 2075 Bayview Avenue, Toronto, Ont., Canada M4N 3M5 Received 20 October 2003; received in revised form 20 January 2004; accepted 18 February 2004
KEYWORDS Abdomen; Signs in imaging; Trauma; Emergency radiology; Computed tomography
AIM: To report two new, useful computed tomography (CT) signs of the hypovolaemic shock complex (HSC) in adults admitted after blunt abdominal trauma: the halo sign (ring of fluid around a collapsed intra-hepatic inferior vena cava (IVC)), and peripancreatic retroperitoneal fluid. MATERIALS AND METHODS: CT images of 498 consecutive patients admitted after blunt abdominal trauma were reviewed, of which 27 had CT signs of the HSC. The CT images of these 27 patients were analysed. A control group of 101 patients examined using CT for suspected blunt abdominal trauma who did not have the HSC were chosen for comparison. RESULTS: The most common features involved the vascular compartment: diminished IVC diameter ðn ¼ 27Þ; a positive halo sign ðn ¼ 21Þ; diminished anteroposterior diameter of the aorta ðn ¼ 13Þ; and abnormal vascular enhancement ðn ¼ 10Þ: Peripancreatic retroperitoneal fluid in the absence of pancreatic injury, pancreatitis or pancreatic disease was observed in eight patients. Hollow visceral abnormalities included: diffuse increased mucosal enhancement of both the small and large bowel ðn ¼ 19Þ; diffuse thickening of the small bowel wall ðn ¼ 11Þ; and small bowel dilatation ðn ¼ 7Þ: Solid visceral abnormalities included both decreased and or increased enhancement. Several concomitant intra- and extra-abdominal injuries were also identified. CONCLUSION: In the setting of blunt abdominal trauma, early abdominal CT can show diffuse abnormalities due to the HSC, which occasionally may alert clinicians of unsuspected shock. Recognition of these signs as distinguished from injured viscera is important in order to avoid unnecessary laparotomy. Two new signs are described: the halo sign and peripancreatic retroperitoneal fluid. q 2004 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction Hypovolaemic shock complex (HSC) is an infrequently encountered entity found using computed tomography (CT) in victims of severe trauma.1 – 8 Shock is defined as inadequate organ perfusion to maintain oxygen delivery to the capillary exchange *Guarantor and correspondent: Max F. Ryan, Department of Clinical Radiology, Cork University Hospital, Wilton, County Cork, Ireland. Tel.: þ 1-353-21-546400; fax: þ1-353-21-343307. E-mail address: [email protected]
beds along with inadequate removal of the acid metabolites from the tissues, and is thus associated with poor prognosis.9 Clinically, early recognition of shock is mandatory, and treatment should be initiated simultaneously with the identification of a probable cause. The response to initial treatment, coupled with physical assessment, usually provides sufficient information as to the cause of the shock state. Nevertheless, despite close monitoring by the trauma team, a number of patients are clinically silent before systemic manifestations occur. Thus,
0009-9260/$ - see front matter q 2004 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2004.02.012
600
M.F. Ryan et al.
a few patients go undetected before emergent diagnostic imaging.1 The diagnostic accuracy of CT as a primary diagnostic tool in the management of patients admitted after blunt abdominal trauma is well established.10,11 In HSC, a spectrum of vascular and visceral CT signs have been described.3,6,12 – 14 The vascular signs include reduced calibre of major intra-abdominal vessels and increased luminal enhancement. Visceral signs may involve hollow viscera (diffuse small bowel thickening, mucosal enhancement and dilatation) and, or solid viscera (increased or reduced end organ perfusion).3,6,12 – 14 The main importance of recognizing the CT manifestations of HSC is to spare these patients unnecessary surgery. The purpose of this report is to describe and assess the value of a spectrum of specific abdominal CT imaging parameters in identifying patients in shock and in predicting their clinical outcome.
above, at and 2 cm below the origin of the renal arteries.15,16,17 Hollow visceral abnormalities includes: small bowel dilatation (. 2.5 cm); bowel wall thickening (. 3 mm) and increased bowel wall mucosal enhancement (greater than the psoas muscle);4,14 enhancement and dilatation of the colon; and increased gallbladder mucosal enhancement.16 – 18 Solid visceral abnormalities includes: intense adrenal gland enhancement (greater than the IVC)5 and renal cortical enhancement (greater than that of the aorta).1,3 Two features of the hypovolaemic shock complex not previously described were evaluated: the halo sign, a ring of oedema around a collapsed intrahepatic IVC (ratio of the short axis diameter of the opacified intra-hepatic IVC less than half the short axis diameter of the intra-hepatic IVC bed at the same level), and low density extra-peritoneal peripancreatic fluid (lower than 20 HU) around the entire gland (in the absence of biochemical or CT evidence of pancreatitis or pancreatic disease).
Methods and materials
Imaging technique
From our comprehensive trauma database (Microsoft, Excel 2000), a retrospective review of all posttraumatic abdominal CT images and medical records of 498 consecutive patients admitted to our level 1 trauma centre, from January 1998 to June 2000 was undertaken.
The indication for abdominal and pelvic CT examination in all cases was suspected intra-abdominal injury, based on a history of blunt abdominal trauma and an initial assessment including a physical examination. All patients were managed according to standard ATLS protocols. After this initial evaluation, and resuscitation with intravenous fluid support, all patients were thought to be haemodynamically stable enough to undergo abdominal CT examination. A helical CT machine (GE HiSpeed Advantage, WI, USA) was used in all cases: 7 mm collimation, pitch value of 1 and 50% overlapped slices to reconstructions with 3.5 mm inter-scan spacing were obtained. A rapid power-injected bolus of 2 ml/kg body weight, up to a maximum of 150 ml of non-ionic contrast material (Omnipaque 300, Nycomed Amersham plc, Buckinghamshire, England) at a rate of 3 ml/s was used. Twenty-five of 27 (92%) patients received oral or nasogastric contrast medium (100 – 500 ml Hypaque 2%, Amersham Health Inc.) in the trauma room before transfer to CT.
The clinical diagnosis of shock The clinical diagnosis of shock was established by a combination of several established clinical and biochemical findings including: hypotension defined as a systolic blood pressure less than 100 mmHg, tachycardia and reduced urine output (advanced trauma life support (ATLS) manual). In addition, supportive criteria such as blood transfusion requirements within the first 24 h of admission were included.
The CT manifestations of shock According to the literature,1 – 3,15,16,17 various CT signs of HSC have been described. The CT signs of hypovolaemia include diminished calibre of the inferior vena cava (IVC), defined as an anterior– posterior diameter of less than 9 mm at three levels (within the intra-hepatic portion of the IVC, at and 2 cm below the level of the renal arteries);4,7 diminished calibre of the aorta, defined as an anteroposterior diameter less than 1.3 cm, 2 cm
Study group Two authors who were skilled in interpreting abdominal and pelvic trauma CT images identified 27 of 498 patients with CT signs of the HSC based on the combination of two or more vascular or visceral criteria aforementioned above by consensus
The halo sign and peripancreatic fluid: useful CT signs of hypovolaemic shock complex in adults
601
Table 1 Incidence of hypovolaemic shock on 498 post-traumatic abdominal CT images Status
Overall frequency
Survived frequency
Died frequency
No hypovolaemic shock complex Hypovolaemic shock complex
471 (94.5%) 27 (5.5%)
411 (87.3%) 8 (29.6%)
60 (12.7%) 19 (70.4%)
opinion. One of these authors, obtained electronic calliper and density measurements of the relevant vasculature, solid and hollow viscera and fluid collections from all CT examinations on a CT viewing console. Other data collected included: the presence of intra- and extra-peritoneal fluid and density measurements (other than peripancreatic fluid).
Control group A control population examined consisted of 101 patients randomly selected from the remaining group of 498 trauma patients. The data collected in this group included: the presence or absence of the halo sign and the presence of low attenuation (, 20 HU) peripancreatic fluid.
Statistical methods SAS Version 8 and Microsoft Excel were used to conduct all the statistical analysis. General descriptive statistics were first created for the 27 HSC patients. Ninety-five percent confidence intervals (CI) were constructed for both the HSC and the control group with reference to the mean maximum anteroposterior calibre of the aorta and IVC, as well as the proportion having the halo sign, and peripancreatic oedema; among other signs that were subsequently compared using two-sided Z-test. The significance level was originally set at 5% level, but adjusted by Bonferroni method to account for the case of multiple testing. As a result, any test that gave a p-value of less than 0.005 indicates a statistically significant result.
Results Demographics Of 498 patients presenting with blunt abdominal trauma, 27 adults were identified as having abdominal CT signs of the HSC (Table 1). There were 17 males and 10 females, all between the ages of 18 – 84 years with mean age of 39.6 years (Table 2). The control group was age matched (95% CI (Control—HSC) 2 6.2272, 10.3145). The indication for CT examination was blunt abdominal trauma in all 27 patients and included: motor vehicle accidents as a passenger or driver ðn ¼ 24Þ; pedestrians involved in a motor vehicle accident ðn ¼ 2Þ; and a fall from a height ðn ¼ 1Þ: The demographics of our patients are summarized in Table 3. All patients were thought to be haemodynamically stable after initial resuscitation therapy that included intravenous volume replacement in the trauma room. Eleven patients that were thought to be haemodynamically stable before emergent abdominal CT image became hypotensive (, 100 mmHg) after the CT study. The CT findings were confirmed in all patients by surgery or autopsy. Twenty-six patients underwent laparotomy for associated visceral injuries. Nineteen patients (70.4%) died as a result of their injuries of which 12 (63.2%) died within 24 h.
CT findings The CT examinations were performed within a mean time of 2.7 h, ranging from 0.5 – 18.6 h of the injury. The CT results observed in our series are summarized in Tables 4 and 5. The most common features involved the vascular compartment: a
Table 2 Age and gender distribution of 27 patients with the hypovolaemic shock complex Item
Overall mean ^ SD ðn ¼ 27Þ
Survived mean ^ SD ðn ¼ 8Þ
Died mean ^ SD ðn ¼ 19Þ
Age (years) Sex (proportion of males) Sex (proportion of females)
39.6 ^ 4.2 17 (63.0%) 10 (37.0%)
25.8 ^ 3.1 6 (75.0%) 2 (25.0%)
45.5 ^ 5.3 11 (57.9%) 8 (42.1%)
602
M.F. Ryan et al.
Table 3 General demographics of 27 patients with the hypovolaemic shock complex Demographics Initial BP (mmHg) Injury severity score Glasgow coma scale Blood units required in 24 h
diminished IVC diameter at multiple levels was observed in all 27 patients (100%; Figs. 1 – 7), mean value 3.9 mm as opposed to 14.5 mm for the controls (p , 0:0001; Table 6). Twenty-one (77.8%) patients demonstrated a positive halo sign (Figs. 2 – 5; Table 7). The halo sign was not observed in the control group ðp , 0:0001Þ: The halo sign was more prevalent in the deceased group (n ¼ 15; Table 4). The sensitivity and specificity of the halo sign for HSC in this series were 77.8 and 100%, respectively. Diminished anteroposterior diameter of the aorta was observed in 13 patients (Figs. 1 – 4 and 7), with a mean value of 12.5 mm recorded over three different levels (Tables 1 –3, 6 and 7). The mean value in the control group was 17.1 mm. Extra-peritoneal fluid was observed in 21 (77.8%) patients. The presence of extra-peritoneal fluid was in some cases attributable to interstitial oedema due to the HSC alone, and in other cases blood as a result of injured viscera. Low-density (, 20 HU) peripancreatic extra-peritoneal fluid was observed in 8 patients (29.6%), p ¼ 0:002; (Figs. 1, 6 and 7; Tables 4 and 7). In no case was pancreatitis or pancreatic injury recorded on
Figure 1 CT image of the upper abdomen in a 34-yearold male involved in a motor vehicle accident. This shows several features of the HSC: a dense small calibre aorta (9 mm) and flat IVC (2 mm), peripancreatic oedema, splenic hypoperfusion and dense enhancement of the right adrenal gland (black arrow). Note a large layering intra-abdominal haematoma is evident in the right subphrenic space.
serum biochemistry, CT, surgery or autopsy. Peripancreatic fluid was observed more frequently in the deceased group (Table 4) and 7 of 8 (87.5%) patients with peripancreatic oedema died.
Discussion Hypovolaemic shock is a state of circulatory dysfunction resulting in a sudden decrease in the intra-vascular blood volume relative to the vascular
Table 4 Vascular and visceral CT manifestations of 27 patients with the hypovolemic shock complex Overall frequency ðn ¼ 27Þ
Survived frequency ðn ¼ 8Þ
Died frequency ðn ¼ 19Þ
Vascular manifestations Diminished aortic diameter (mm) Collapsed inferior vena cava (mm) Halo signa Flat renal veins
13 (48.0%) 27 (100.0%) 21 (77.8%) 27 (100.0%)
3 (37.5%) 8 (100.0%) 6 (75.0%) 8 (100.0%)
10 (52.6%) 19 (100.0%) 15 (78.9%) 19 (100.0%)
Visceral manifestations Increased bowel wall enhancement Bowel wall thickening Small bowel dilatation Increased adrenal gland enhancement Gallbladder wall enhancement Increased renal cortical enhancement Diminished splenic enhancement Peripancreatic oedema Increased pancreatic enhancement Decreased hepatic enhancement Gastric dilatation
19 (70.4%) 11 (40.7%) 7 (25.9%) 14 (51.9%) 9 (33.3%) 15 (55.6%) 8 (29.6%) 8 (44.4%) 1 (3.7%) 3 (11.1%) 19 (70.4%)
5 (62.5%) 2 (25.0%) 3 (37.5%) 4 (50.0%) 3 (37.5%) 7 (87.5%) 1 (12.5%) 1 (12.5%) 0 (0.0%) 0 (0.0%) 7 (87.5%)
14(73.7%) 9 (47.4%) 4 (21.1%) 10(52.6%) 6 (31.6%) 8 (42.1%) 7 (36.8%) 7 (87.5%) 1 (100%) 3 (15.8%) 12(63.2%)
a
Halo sign ¼ rim of fluid around a collapsed intra-hepatic inferior vena cava
The halo sign and peripancreatic fluid: useful CT signs of hypovolaemic shock complex in adults
603
Table 5 Other injuries in 27 patients with the hypovolaemic shock complex Injury
Overall frequency ðn ¼ 27Þ
Survivors ðn ¼ 8Þ
Deceased ðn ¼ 19Þ
Intra-peritoneal fluid Extra-peritoneal fluid Active arterial extravasations Pelvic fractures Bladder rupture Liver laceration Splenic injury Renal injury Adrenal haematoma Urethral tear Ureteral tear Pancreatic injury Head injurya Head injury—contusion Head injury—traumatic subarachnoid haemorrhage Subdural haemorrhage Spinal injury
21 (77.8%) 21 (77.8%) 13 (48.1%) 25 (92.6%) 5 (18.5%) 8 (29.6%) 14 (51.8%) 9 (33.3%) 2 (7.4%) 5 (18.5%) 1 (3.7%) 0 (0.0%) 23 (85.2%) 10 (37.0%) 9 (33.3%) 5 (18.5%) 3 (11.1%)
5 (62.5%) 6 (75.0%) 3 (37.5%) 7 (87.5%) 3 (37.5%) 5 (62.5%) 4 (50.0%) 3 (37.5%) 0 (0.0%) 3 (37.5%) 1 (12.5%) 0 (0.0%) 7 (87.5%) 2 (25.0%) 2 (25.0%) 2 (25.0%) 0 (0.0%)
16 (84.2%) 15 (78.9%) 10 (52.6%) 18 (94.7%) 2 (10.5%) 3 (15.8%) 10 (52.6%) 6 (31.6%) 2 (10.5%) 2 (10.5%) 0 (0.0%) 0 (0.0%) 16 (84.2%) 8 (42.1%) 7 (36.8%) 3 (15.8%) 3 (15.8%)
a
Head injury ¼ all head injuries including concussion, contusion, cerebral oedema and subarachnoid haemorrhage.
Table 6 Comparison of mean aorta and IVC diameters of 27 HSC patients and 101 controls Mean and standard error
Aorta IVC
Hypovolaemic shock complex patients ðn ¼ 27Þ
Controls ðn ¼ 101Þ
95% CI for difference (control—HSC)
p-Valuea
12.5 ^ 0.5 mm 3.9 ^ 0.3 mm
17.3 ^ 0.3 mm 14.5 ^ 0.4 mm
[3.497, 5.948] [9.125, 12.1303]
,0.0001 ,0.0001
Aorta ¼ mean anterioposterior diameter of aorta at three locations: supra-renal, renal and infra-renal. IVC ¼ mean diameter of inferior vena cava at three locations: supra renal, renal and infra-renal levels. CI, confidence interval. a Wilcoxon rank sum test p-value.
Figure 2 CT image of the upper abdomen in a 45-yearold male involved in a motor vehicle accident. This shows the “halo sign”. This refers to a rim of oedema surrounding a collapsed intra-hepatic IVC.
Figure 3 CT image of the upper abdomen at the level of the aortic hiatus in a 55-year-old man involved in a motor vehicle accident. This shows the “halo sign”, a rim of oedema surrounding a collapsed intra-hepatic IVC (3 mm). This finding was present in 21 (78%) of cases in our series.
604
Figure 4 CT image of the upper abdomen in a 37-yearold man with HSC. The “halo sign”, a rim of oedema surrounding a collapsed intra-hepatic IVC, is evident. Note a large liver laceration that is of higher attenuation (35 HU) than the fluid surrounding the collapsed IVC (12 HU).
capacity, to the extent that effective tissue perfusion cannot be maintained. It is thus a failure to meet the metabolic demands of the organs, initially leading to reversible and then to irreversible cellular injury.9 Several theories about the pathogenesis of HSC have been reported. These include the activation of complement, coagulation, cytokine and kallikrein – kinin systems pathways.9 The end result is hypovolaemia, vasoconstriction, and replacement of the intravascular volume with intravenous contrast agent.12 The incidence of hypovolaemic shock is variable
Figure 5 CT image of the mid-abdomen in a 25-year-old woman involved in a motor vehicle accident. This shows low attenuation peripancreatic oedema (9 HU) in the anterior para-renal extra-peritoneal space surrounding a normal pancreas. Note increased gallbladder mucosal enhancement.
M.F. Ryan et al.
Figure 6 CT image of the upper abdomen at the level of the coeliac axis, showing retroperitoneal peripancreatic oedema (11 HU) in a 47-year-old woman. A normal pancreatic body and tail is noted at this level. Features of shock bowel are also noted including: increased small bowel thickening and dilatation. There is flattening on the IVC that was shown on several continuous images. Note increased renal parenchymal enhancement, which was a common, but unreliable, finding in our series.
and largely depends on the referral centre. In our experience, HSC occurs at a frequency of about 5% in patients admitted after multiple trauma. Haemorrhage is the most common cause of shock and is due to a combination of blood and fluid losses due to tissue damage and release of vasoactive substances.9,20 Despite recent advances in the treatment and thus increase in survival of critically
Figure 7 HSC in a 29-year-old male motor vehicle accident victim. Intravenously administered contrast material-enhanced CT image obtained at the level of the pancreas. This shows several features of the HSC: small calibre vasculature, extra-peritoneal peripancreatic oedema, diffuse mural thickening and enhancement of thickened small bowel loops. Note the oral contrast agent was not administered to this patient.
The halo sign and peripancreatic fluid: useful CT signs of hypovolaemic shock complex in adults
605
Table 7 Ninety-five percent confidence intervals (CI) for various signs of HSC HSC patients ðn ¼ 27Þ
Observed percentage/SD
95% CI
Halo sign Peripancreatic oedema Decreased splenic enhancement Decreased hepatic enhancement Adrenal enhancement Gallbladder wall enhancement
77.8% (8.0%) 29.6% (9.6%) 29.6% (8.8%) 11.1% (6.0%) 51.9% (9.6%) 33.3% (9.1%)
(59.2%, 89.4%) (12.4%, 46.9%) (12.4%, 46.9%) (0.0%, 23.0%) (34.0, 69.3%) (15.6%, 51.1%)
Controls ðn ¼ 101Þ Halo sign Decreased splenic enhancement Decreased hepatic enhancement Adrenal enhancement Gall bladder wall enhancement
Observed Percentage/ proportion/SD 0.0% 1.0% (1.0%) 8.0% (2.7%) 2.0% (1.4%) 10.9% (1.9%)
95% CI (0, 3.7%) (0.0%, 2.9%) (2.7%, 13.2%) (0.0%, 3.7%) (6.19%, 18.5%)
a
p-Valuea ,0.0001 0.005 0.0229 ,0.0001 0.0134
Two-sided Z-test (comparing proportions of cases and controls).
injured patients, the majority of patients with multiple injuries will have an element of hypovolaemia.20 To correct this hypovolaemia, preliminary resuscitation with fluids is instituted to promote and maintain adequate tissue perfusion and to prevent further metabolic decline. Volume replacement is directed by the response to this therapy rather than by reliance on the initial classification of blood loss. Blood loss is further complicated by shifts of fluids among the fluid compartments, in particular in the extravascular, extracellular fluid compartment.20 These fluid changes may be directly due to major soft-tissue and bone injuries, or secondary to activation of a systemic inflammatory response. Termed the systemic inflammatory response syndrome, this is directly related to the extent of the soft-tissue injury and is associated with the production and release of multiple cytokines.9 These hormones affect vascular endothelium causing increased permeability and thus tissue oedema. Increased permeability may accentuate hypovolaemia by shifts of fluid from the intravascular into the extravascular, extracellular space. The treatment of hypovolaemia may be further complicated by a combination of coexisting cardiogenic, neurogenic and even septic shock. Furthermore, the effect of the vasopressors and fluids used for resuscitation may further accentuate the features.20 Patients that have responded to initial treatment in the emergency room, and who are thought to be haemodynamically stable enough to undergo CT examination, may nonetheless have a poorly compensated shock state, not recognizable with blood pressure and pulse readings alone. These patients may become hypotensive during or immediately after the CT study.12,21,22 On CT, the findings observed in the HSC may be classified as vascular and non-vascular.1,3,7,15 – 17
The potential CT findings are numerous, with variable association of one to another. In addition, when haemodynamic instability is corrected, only end-organ anomalies may be observed. In the HSC, these end organ findings relate to hypovolaemia and not to structural lesions of the involved viscera, and thus require supportive therapy rather than a surgical procedure per se. In our series, collapse of the IVC at multiple levels associated with flattening of the renal veins was most common (Fig. 7). This is due to decreased circulating blood volume and thus reduced venous return in patients with systemic hypotension.7 As a normal IVC may appear flattened on a single abdominal section, the diagnosis of a collapsed IVC was made when flattening was observed at multiple levels within the abdomen. The proposed halo sign refers to a collapsed intra-hepatic IVC, as a function of hypovolaemia, surrounded by a circumferential zone of low attenuation, probably representing extracellular fluid (Figs. 2 – 5). This sign was most frequently observed at a level corresponding with the superior segments of the liver, below the confluence of the hepatic veins. This finding was observed in 21 of 27 (77.8%) patients ðp , 0:0001Þ and in the setting of trauma is considered a specific sign of HSC and thus probably a poor prognostic indicator. For example, 15 of 21 (71.4%) of patients with the halo sign died. The halo sign was not observed in any patient in the control group nor in any of the 471 non-HSC trauma CT cases initially reviewed. The halo sign is not specific to HSC. The authors have observed similar findings in patients with underlying liver disease and liver congestion. Other reports have also observed similar findings around portal vasculature in patients with biliary cirrhosis, hepatitis, and tumours that obstruct lymphatic drainage at the porta hepatis.23,24
606
In our series, low attenuation (, 20 HU) extracellular fluid in the anterior extra-peritoneal compartment around the pancreas and base of the mesentery is reported (Figs. 1, 6 and 7). This manifestation has not been previously described, but was observed in eight cases, of which seven (87.5%) died. In no case was a pancreatic injury observed on CT, or recorded at surgery or autopsy. In all cases, the pancreas appeared normal in size and uniform in attenuation. Extracellular fluid was originally reported in the HSC3,19 but was not included in subsequent reports.2,12 In the setting of shock, extracellular fluid is thought to be related to loss of pre-capillary arteriolar sphincter tone that causes arteriolar constriction, and thus blood flows into the ischaemic microcirculatory capillary bed, but venous constriction hinders its return. Blood that pools in the peripheral circulation causes increased hydrostatic pressure, which in the presence of increased capillary membrane permeability, forces fluid and protein out of the vascular space into the interstitial space further decreasing the circulating blood volume.9 As a result blood volume expanders are only temporarily retained in the circulation and are only of transient therapeutic benefit.9 Small bowel manifestations of the hypoperfusion shock complex are better known than the halo sign and peripancreatic extracellular fluid signs.3,4,12,14 Termed shock bowel, this refers to increased small bowel wall mucosal enhancement, mural thickening and luminal dilatation. This is thought to be due to hypovolaemia induced reduced bowel perfusion and subsequent venous vasoconstriction, in response to sympathetic stimulation and possibly vasopressors used in resuscitation.6 Replacement of a depleted intravascular volume with contrast agent along with increased bowel wall permeability results in wall thickening and intense bowel wall enhancement due to diminished perfusion and interstitial leak of molecules of the contrast agent. Accumulation of third space fluid and decreased fluid resorption causes luminal dilatation.14 The small bowel is typically diffusely involved, whereas the colon is sporadically involved. The aforementioned changes do not necessarily lead to irreversible bowel ischaemia and infarction if a minimal threshold of blood flow is maintained through adequate resuscitation measures.4,25 Reported solid visceral signs of the HSC includes a prolonged intense nephrogram. This may be due a combination of hypoperfusion induced renal failure and or contrast induced nephrotoxicity and renal impairment1 (Fig. 7). The present authors find hyperintense renal parenchyma enhancement as a sign of HSC unreliable. Hyperdense renal parench-
M.F. Ryan et al.
yma enhancement is occasionally seen on normal CT examinations and depends on several other factors, including cardiac output and scan acquisition parameters. In the HSC, intense symmetric enhancement of the adrenal glands is reported5 (Fig. 1). The adrenal glands are usually normal in configuration and size. This finding is likely related to a sympathetic response to hypovolaemic shock along with preservation of perfusion to the adrenal glands as a vital organ. Although useful, the authors have also observed this sign in patients after severe burns or surgery. In HSC, low attenuation values of the spleen, is related to severe hypoperfusion rather than an injury to the splenic artery or parenchyma (Fig. 1).13 Splenic arterial flow has no autoregulatory mechanism and is highly sensitive to sympathetic stimulation and thus vasoconstriction.26 Other post-traumatic causes for splenic hypoperfusion include vessel injury, splenic rupture and infarction, and inadequate volume and timing of intravenous contrast agent.7,13,27 In the absence of hepatic injury, abnormal hepatic arterial enhancement is infrequent, as arterial flow is autoregulated and mixed with hepatopetal portal venous inflow. A generalized marked reduction of hepatic enhancement is abnormal and occurs in the most severe form of HSC. In our series, hepatic enhancement was typically heterogeneous (Fig. 1). This feature is less common than other abnormalities, such as the halo sign and peripancreatic extracellular fluid.
Outcome In our series, the HSC was a marker of severe injury and poor outcome. The mortality for adults with the hypovolaemic shock complex was 70%. Mortality was affected by associated injuries, such as active extravasation of arterial contrast and intra-cranial injuries. This reflects the complexity of injuries that occur to victims of severe trauma and the multifactorial aetiologies involved. In conclusion, in the setting of blunt abdominal trauma, attention to the features that may suggest poorly compensated shock is advocated. The CT findings are due to shock rather than structural lesions of the affected viscera per se, and thus require supportive therapy rather than laparotomy.
Acknowledgements The authors thank Michael Schull, MD, Department
The halo sign and peripancreatic fluid: useful CT signs of hypovolaemic shock complex in adults
of Medicine, University of Toronto and William Sharkey, trauma statistics office for their assistance.
13.
14.
References 1. Sivit CJ, Taylor GA, Bulas DI, Kushner DC, Potter BM, Eichelberger MR. Posttraumatic shock in children: CT findings associated with hemodynamic instability. Radiology 1992;182:723—6. 2. Rotondo A, Catalano O, Grassi R, Scialpi M, Angelelli G. Thoracic CT findings at hypovolemic shock. Acta Radiol 1998;39:400—4. 3. Taylor GA, Fallat ME, Eichelberger MR. Hypovolemic shock in children: abdominal CT manifestations. Radiology 1987;164: 479—81. 4. Mirvis SE, Shanmuganathan K, Erb R. Diffuse small-bowel ischemia in hypotensive adults after blunt trauma (shock bowel): CT findings and clinical significance. AJR Am J Roentgenol 1994;163:1375—9. 5. O’Hara SM, Donnelly LF. Intense contrast enhancement of the adrenal glands: another abdominal CT finding associated with hypoperfusion complex in children. AJR Am J Roentgenol 1999;173:995—7. 6. Kaufman RA, Stalker HP. Hypovolemic shock in children. Radiology 1988;166:579—80. 7. Jeffrey Jr RB, Federle MP. The collapsed inferior vena cava: CT evidence of hypovolemia. AJR Am J Roentgenol 1988; 150:431—2. 8. Guyton AC. In: Guyton AC, editor. Circulatory shock and physiology of its treatment. Philadelphia: WB Saunders; 1971. p. 263—71. 9. Shamji FM, Todd TRJ. Hypovolemic shock. Symposium on cardiovascular crises in the critically ill. Crit Care Clin 1985; 1. 10. Amoroso TA. Evaluation of the patient with blunt abdominal trauma: an evidence based approach. Emerg Med Clin North Am 1999;17:63—75. 11. 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;35:267—70. 12. Sivit CJ, Eichelberger MR, Taylor GA. CT in children with rupture of the bowel caused by blunt trauma: diagnostic
15.
16.
17.
18.
19. 20.
21.
22.
23.
24.
25.
26. 27.
607
efficacy and comparison with hypoperfusion complex. AJR Am J Roentgenol 1994;163:1195—8. Berland LL, Van Dyke JA. Decreased splenic enhancement on CT in traumatized hypotensive patients. Radiology 1985; 156:469—71. Hara H, Babyn PS. Significance of bowel wall enhancement on CT following blunt abdominal trauma in childhood. J Comput Assist Tomogr 1992;16:94—8. Shin MS, Berland LL, Ho KJ. Small aorta: CT detection and clinical significance. J Comput Assist Tomogr 1990;14: 102—3. Taber P, Korobkin MT, Gooding CA, Palubinskas AJ, Neuhauser EB. Growth of the abdominal aorta and renal arteries in childhood. Radiology 1972;102:129—34. Calabuig R, Seggerman RE, Weems WA, Weisbrodt NW, Moody FG. Gallbladder and gastrointestinal motility after hemorrhagic shock. Surgery 1990;107:568—73. Moody FG, Calabuig R, Li YF, Harari Y, Rodriguez LF, Weisbrodt NW. Biliary and gut function following shock. J Trauma 1990;30(Suppl. 12):S179—84. Cox TD, Kuhn JP. CT scan of bowel trauma in the pediatric patient. Radiol Clin N Am 1996;34:807—18. Chapter 3: Shock. In Advanced Trauma Life Support for Doctors. Student Course Manual. 6th ed. American College of Surgeons 1997; 88—107. Brody JM, Leighton DB, Murphy BL, et al. CT of blunt trauma bowel and mesenteric injury: typical findings and pitfalls in diagnosis. RadioGraphics 2000;20:1525—36. Radin DR, Boswell Jr WD, Ralls PW, Halls JM. Recognition of hypotensive shock on abdominal computed tomography. J Comput Tomogr 1986;10:299—301. Wenzel JS, Donohoe A, Ford III KL, Glastad K, Watkins D, Molmenti E. Primary biliary cirrhosis: MR imaging findings and description of MR imaging periportal halo sign. AJR Am J Roentgenol 2001;176:885—9. Lawson TL, Thorsen MK, Erickson SJ, Perret RS, Quiroz FA, Foley WD. Periportal halo: a CT sign of liver disease. Abdom Imaging 1993;18:42—6. Bulkley GB, Kvietys PR, Parks DA, Perry MA, Granger DN. Relationship of blood flow and oxygen consumption to ischemic injury in the canine small intestine. Gastroenterology 1985;89:852—7. Goodman LR, Aprahamian C. Changes in splenic size after abdominal trauma. Radiology 1990;176:629—32. Federle MP, Griffiths B, Minagi H, Jeffrey Jr RB. Splenic trauma: evaluation with CT. Radiology 1987;162:69—71.