Keywords
blunt - embolization - penetrating - trauma
Introduction
Historically intra-abdominal arterial bleeding has been an indication for laparotomy,
accounting for 20 to 35% of patients who fail nonoperative treatments. This approach
is, however, associated with additional complications and hospital resources.[1] More recent advances in emergency interventional radiology have resulted in utilization
of angiographic techniques for embolization of involved arteries. This includes coiling
and proximal gelfoam embolization techniques, which can be used as a stand-alone or
adjunct procedure to surgical intervention to return patients to hemodynamic stability.[2]
[3] Patients who are less hemodynamically stable may need additional treatment such
as laparotomy.[4] Nonoperative management of blunt trauma (BT) has been highlighted extensively in
the literature.[4]
[5]
[6] As for penetrating trauma (PT), there is a greater paucity of literature regarding
optimal endovascular management. Biagioni et al demonstrated the successful use of
stent grafts in PT, although embolization was not a focus of this study.[7] It is difficult to ascertain the type of treatment that would most likely result
in the success of bleeding control, especially for retroperitoneal organs.[8]
There have been numerous studies that highlight the different types of operative and
nonoperative management strategies, and recent literature has demonstrated the efficacy
of embolization for BT.[4]
[5]
[6]
[9] However, there is a lack of data regarding endovascular management for PT. Herein,
this study analyzed data regarding patients who underwent embolization for penetrating
abdominal trauma at a single institution to assess differences in outcomes compared
to BT. Additionally, factors that may be predictors of mortality and successful embolization
were assessed.
Methods
Institutional Review Board approval was obtained to perform this retrospective study.
Between January 2018 and December 2020, 43 patients (38 male and 5 female, average
age: 36.6 years old, standard deviation [SD]: 17.0, range: 4–73) received arterial
embolization for either BT or PT injury in a tertiary care academic center in an urban
setting with preponderance of PT.
Data Collection
Electronic medical chart review was performed to document demographic information
(age and gender), clinical history, location of trauma, procedural course (OR prior
to embolization, computed tomography angiography [CTA], bleeding arterial vessels,
number of vessels embolized, embolics used, fluoroscopy time, lab values pre- and
24 hours postprocedure), and hospital course (date of admission and discharge, days
of survival).
Embolization Technique
All embolization procedures were performed by a board-eligible or board-certified
fellowship-trained interventional radiologist. Patients were brought to interventional
radiology from the emergency room or operating room following BT or PT. Routine protocol
prior to intervention was to obtain preintervention CTA for all patients deemed hemodynamically
stable to document active bleeding and procedure planning.
In the angiography suite, angiography was performed from a femoral approach and angiography
of the affected body was performed to identify the site of vascular injury. Subsequently,
selective catheterization of the bleeding vessel was performed with the use of a microcatheter
(2.4–2.8 Fr). Embolization was performed with a combination of microcoils and/or gelfoam
slurry based on operator preference and clinical scenario. Technical success was defined
as the arrest of any active sites of extravasation seen on pre-embolization angiography.
Statistical Analysis
All statistical analysis was performed using JMP software (JMP, Version 16, SAS Institute
Inc., Cary, North Carolina, United States, 1989–2021). Collected patient data were
analyzed to compare profiles of BT and PT. Additionally, factors associated with mortality
was assessed. Number of vessels was analyzed both as continuous and categorical (0,
1, and ≥ 2) variables. Categorical variables were assessed using Fischer's exact and
Pearson's chi-square test and are represented as n (%). Continuous variables were evaluated using Student's t-test and Wilcoxon rank sum test and are represented as mean (SD).
Results
A total of 23 patients with BT and 20 patients with PT received arterial embolization.
Patients with BT were older than those with PT (42.3 [17.08] vs. 30.0 [14.5]; p = 0.015). Preprocedural hemoglobin was higher for patients with BT compared to PT
(12.25 [2.28] vs. 10.6 [2.37]; p = 0.028). Postprocedural hemoglobin was similar for patients with BT compared to
PT (10.32 [1.46] vs. 10.82 [1.92]; p = 0.357). Change in hemoglobin was higher for patients with BT compared to PT (–1.93
[2.37] vs. 0.33 [1.54]; p = 0.003). Type of injuries that patients with BT suffered included motor vehicle
collisions (n = 18; 78.26%), falls (n = 4; 17.39%), and a straddle injury (n = 1; 4.35%). Type of injuries that patients with PT suffered included gunshot wounds
(n = 17; 85.0%) and stabs (n = 3; 15.0%). Embolized vessels of those with BT included pelvic (n = 12; 54.55%), abdominal (n = 7; 31.82%), vertebrae (n = 2; 9.09%), and lower extremity (n = 1; 4.55%) vessels. Embolized vessels of those with PT included abdominal (n = 8; 44.44%), pelvic (n = 4; 22.22%), thoracic (n = 3; 16.67%), vertebrae (n = 1, 5.56%), back (n = 1; 5.56%), and lower extremity (n = 1; 5.56%) vessels. Mean number of vessels embolized was higher in the BT group
compared to PT (mean [SD]: 2.26 [1.32] vs. 1.44 [1.03], p = 0.044). Out of all 43 patients who underwent embolization, 40 (93.0%) had technical
success. There was a total of 16 patients (37.2%) who had control for bleeding surgically
prior to embolization and 4 (9.30%) after embolization. Additional results are displayed
in [Table 1].
Table 1
Descriptive analysis for patients with blunt or penetrating trauma
|
Blunt (n = 23)
|
Penetrating (n = 20)
|
p-Value
|
|
Age, mean (SD)
|
42.30
|
17.08
|
29.95
|
14.55
|
0.015
|
|
Fluoroscopy time (in min), mean (SD)
|
25.15
|
13.99
|
20.30
|
9.93
|
0.204
|
|
Radiation dose (mGy), mean (SD)
|
1282.02
|
1199.70
|
.
|
.
|
|
|
Preprocedure hemoglobin, mean (SD)
|
12.25
|
2.28
|
10.59
|
2.37
|
0.028
|
|
Postprocedure hemoglobin (next morning), mean (SD)
|
10.32
|
1.46
|
10.82
|
1.92
|
0.357
|
|
Days of survival, mean (SD)
|
13.70
|
2.60
|
19.10
|
2.79
|
0.160
|
|
Gender
|
|
|
|
|
1.000
|
|
Female
|
3
|
13.04%
|
2
|
10.00%
|
|
|
Male
|
20
|
86.96%
|
18
|
90.00%
|
|
|
Type of injury
|
|
|
|
|
< 0.001
|
|
Fall
|
4
|
17.39%
|
0
|
0%
|
|
|
Gunshot wound
|
0
|
0%
|
17
|
85%
|
|
|
Motor vehicle collision
|
18
|
78.26%
|
0
|
0%
|
|
|
Stab
|
0
|
0%
|
3
|
15%
|
|
|
Straddle
|
1
|
4.35%
|
0
|
0%
|
|
|
OR prior to embolization
|
4
|
17.39%
|
12
|
60.00%
|
0.005
|
|
CTA done before embolization
|
18
|
78.26%
|
15
|
75.00%
|
1.000
|
|
Number of vessels embolized, mean (SD)
|
2.26
|
1.32
|
1.44
|
1.03
|
0.044
|
|
Number of vessels embolized (≥ 2 or < 2)
|
|
|
|
|
0.422
|
|
0
|
1
|
4.35%
|
2
|
10.53%
|
|
|
1
|
6
|
26.09%
|
8
|
42.11%
|
|
|
≥ 2
|
16
|
69.57%
|
9
|
47.37%
|
|
|
Type of vessel embolized
|
|
|
|
|
< 0.001
|
|
Back
|
0
|
0.00%
|
1
|
5.56%
|
|
|
Lower extremity
|
1
|
4.55%
|
1
|
5.56%
|
|
|
Pelvic
|
12
|
54.55%
|
4
|
22.22%
|
|
|
Thorax
|
0
|
0.00%
|
3
|
16.67%
|
|
|
Vertebrae
|
2
|
9.09%
|
1
|
5.56%
|
|
|
Abdominal (breakdown below)
|
7
|
31.82%
|
8
|
44.44%
|
|
|
Renal
|
1
|
|
2
|
|
|
|
Splenic
|
4
|
|
0
|
|
|
|
Gastric
|
1
|
|
0
|
|
|
|
Hepatic
|
1
|
|
6
|
|
|
|
Gelfoam used?
|
17
|
73.91%
|
11
|
55.00%
|
0.219
|
|
Coils used?
|
18
|
78.26%
|
11
|
55.00%
|
0.191
|
|
Embolization targeted or prophylactic or both
|
|
|
|
|
0.654
|
|
None
|
1
|
4.55%
|
0
|
0.00%
|
|
|
Prophylactic
|
0
|
0.00%
|
1
|
5.56%
|
|
|
Targeted
|
17
|
77.27%
|
15
|
83.33%
|
|
|
Both
|
4
|
18.18%
|
2
|
11.11%
|
|
|
Technical success
|
21
|
91.30%
|
19
|
100.00%
|
0.493
|
|
Return for repeat angiography
|
3
|
13.04%
|
3
|
15.00%
|
1.000
|
|
Required OR for bleeding control after embolization
|
2
|
8.70%
|
2
|
10.53%
|
0.841
|
|
Mortality
|
3
|
13.04%
|
2
|
10.53%
|
1.000
|
Abbreviations: CTA, computed tomography angiography; SD, standard deviation; OR, Operation
Room.
Death postembolization was not associated with age (p = 0.50), type of trauma (p = 1.00), OR prior to embolization (dead: 40.0% vs. alive: 35.1%, p = 1.00), multiple (≥ 2) vessels embolized (100% vs. 54.1%, p = 0.50), mean number of vessels embolized (mean [SD]: 0.81 [SD: 1.26] vs. 0.82 [0.41];
p = 0.072), use of gelfoam embolic (80% vs. 64.9%, p = 0.65), use of coils (40.0% vs. 70.3%, p = 0.31), prophylactic or targeted embolization (targeted: 82.9% vs. 50.0%; p = 0.29), and use of repeat angiography (40.0% vs. 10.8%, p = 0.14) and radiation dose (mean [SD]: 1276.13 [1630.75] vs. 1282.9 [1177.54] mGy,
p = 0.99). Patients with more severe cases requiring OR for bleeding control after
embolization were associated with a higher likelihood of failure and death (40.0%
vs. 5.56%, p = 0.038). Technical success of embolization was not associated with death (80.0%
vs. 97.2%, p = 0.23). Additional results are displayed in [Table 2].
Table 2
Descriptive analysis for mortality
|
Alive (n = 37)
|
Dead (n = 5)
|
p-Value
|
|
Age, mean (SD)
|
36.00
|
16.93
|
41.60
|
19.96
|
0.500
|
|
Fluoroscopy time (in min), mean (SD)
|
21.87
|
11.65
|
29.64
|
17.74
|
0.196
|
|
Radiation dose (mGy), mean (SD)
|
1282.90
|
1177.54
|
1276.13
|
1630.75
|
0.990
|
|
Preprocedure hemoglobin, mean (SD)
|
11.33
|
2.48
|
12.74
|
2.22
|
0.238
|
|
Postprocedure hemoglobin (next morning), mean (SD)
|
10.47
|
1.57
|
10.60
|
2.40
|
0.873
|
|
Days of survival, mean (SD)
|
14.60
|
9.76
|
25.40
|
25.36
|
0.072
|
|
Gender
|
|
|
|
|
0.099
|
|
Female
|
3
|
8.11%
|
2
|
40.00%
|
|
|
Male
|
34
|
91.89%
|
3
|
60.00%
|
|
|
Type of trauma
|
|
|
|
|
1.000
|
|
Blunt
|
20
|
54.05%
|
3
|
60.00%
|
|
|
Penetrating
|
17
|
45.95%
|
2
|
40.00%
|
|
|
OR prior to embolization
|
13
|
35.14%
|
2
|
40.00%
|
1.000
|
|
CTA done before embolization
|
28
|
75.68%
|
4
|
80.00%
|
1.000
|
|
Number of vessels embolized, mean (SD)
|
0.81
|
1.26
|
0.82
|
0.41
|
0.072
|
|
Number of vessels embolized (≥ 2 or < 2)
|
|
|
|
|
0.358
|
|
0
|
3
|
8.11%
|
0
|
0.00%
|
|
|
1
|
14
|
37.84%
|
0
|
0.00%
|
|
|
≥ 2
|
20
|
54.05%
|
4
|
100.00%
|
|
|
Gelfoam used
|
24
|
64.86%
|
4
|
80.00%
|
0.650
|
|
Coils used
|
26
|
70.27%
|
2
|
40.00%
|
0.313
|
|
Embolization targeted or prophylactic or both
|
|
|
|
|
0.290
|
|
None
|
1
|
2.86%
|
0
|
0.00%
|
|
|
Prophylactic
|
1
|
2.86%
|
0
|
0.00%
|
|
|
Targeted
|
29
|
82.86%
|
2
|
50.00%
|
|
|
Both
|
4
|
11.43%
|
2
|
50.00%
|
|
|
Return for repeat angiography
|
4
|
10.81%
|
2
|
40.00%
|
0.141
|
|
Required OR for bleeding control after embolization
|
2
|
5.56%
|
2
|
40.00%
|
0.038
|
|
Technical success
|
35
|
97.22%
|
4
|
80.00%
|
0.232
|
Abbreviations: CTA, computed tomography angiography; SD, standard deviation.
Discussion
Among those who underwent embolization for BT and PT, there was no difference in demographics
(gender) and intraoperative characteristics (coils, gelfoam, CTA use). Additionally,
there was no difference in in-hospital mortality and technical success. Overall, there
were 93.0% (n = 40 out of 43) of patients with successful embolization and a 12% (n = 5 out of 43) rate of mortality.
Cherian et al conducted a multi-institutional study for embolization after blunt abdominal
injury to the spleen, liver, and kidneys. Out of 45 patients, 13.3% died which was
similar to our rate of 13.2% among our cohort of BT. However, their mean length of
stay was 5.2 days while our BT cohort was 13.7 days.[10] Hemodynamic instability is an indication for operative management of BT; Cherian
et al corroborates with our study that shows that a high rate of success for embolization
persists among patients with upper abdominal BT. It is also effective for lower abdominal
BT; Bertelli et al demonstrated this in a case series of mesenteric bleeding among
patients with BT.[11] Additionally, their literature review has indicated a technical success of 96%.
Furthermore, Velez investigated the use of nonoperative management in his cohort of
281 patients. Of them, 183 patients underwent angiography and 166 (91%) underwent
embolization. A total of 7 patients also underwent embolization of multiple vascular
territories, which all achieved technical success. This demonstrates the safety and
efficacy of angioembolization in patients with severe BT.[12] In this study's cohort, 24 patients had more than 2 vessels embolized; 4 of them
had a failure of embolization or mortality. Regarding types of embolization, we found
no difference in the association of gelfoam (p = 0.65) or coil (p = 0.31) with death. Similarly, in a meta-analysis by Rong et al,[13] they found no difference in the success rate of embolization when using coil compared
to gelfoam (odds ratio: 1.41, p = 0.39). However, they demonstrated that coiling was associated with lower odds of
severe complications compared to gelfoam (odds ratio: 0.48, p = 0.02).
O Dell et al described techniques in approaching patients with PT and discussed two
cases that had technical success. They have demonstrated that decisions regarding
which artery to embolize and the embolics used are decided during the procedure; an
interventionist should have gelfoam, coils, and glue ready to be selected and used
for embolization. They also recommend using angiography prior to embolization to plan
access to distal feeding arteries.[14] Meanwhile, in this study's cohort of 20 patients with PT, 75% of them underwent
CTA prior to embolization. Additionally, the presented data indicated similar rates
of technical success and death for those who underwent embolization for PT compared
to BT. As emphasized by O Dell et al, the literature regarding embolization for PT
is scarce.[14] Additionally, a recent position statement from the Society of Interventional Radiology
regarding endovascular treatment of trauma focused on management of specific organs
rather than the type of trauma.[9] This study's data has shown the similarity in preoperative characteristics, intraoperative
aspects, and outcomes of PT compared to BT, indicating that embolization for PT may
have equivalent results as embolization for BT. Both groups in our cohort were similar
with respect to gender, type of embolization (targeted vs. prophylactic), technical
success, and mortality among others. Therefore, embolization for PT may have similar
efficacy as for BT and that it is worth investigating in larger studies.
Although our study holds valuable data regarding novel intervention for PT, there
are several limitations. Patients with PT mostly had single-territory bleeds, lower
mean number of vessels embolized, and higher preprocedural hemoglobin. This may confound
studied outcomes; there is a possibility that PT patients were more hemodynamically
stable compared to BT. However, this may support our hypothesis that PT may have similar
efficacy compared to BT for selected cases. Sixty percent of PT patients went to the
OR prior to embolization; this is due to hospital protocol with severe organ injury
in which patients underwent CTA and immediately were placed on the OR table while
radiology reviews the CT scan.[15] However, it can be inferred that the PT cohort may have been more unstable than
BT yet shows similar outcomes. Additionally, this study had a low sample size of 43
patients with a difference in involved organs, decreasing statistical power and increasing
heterogeneity of presentation among patients.
In conclusion, the rate of mortality, technical success, and requirement of subsequent
OR intervention for hemorrhage control was comparable between BT and PT. BT was associated
with a higher mean number of vessels embolized compared to PT. Our case series may
provide insight in the use of embolization for PT, but further investigation is needed
with larger cohorts.
