The American Journal of Surgery (2014) -, --

Computed tomographic scan mapping of gastric

wall perfusion and clinical implications
Alan A. Saber, M.D., F.A.C.S.a,*, Nami Azar, M.D.b,
Mahmoud Dekal, M.D.c, Tamer N. Abdelbaki, M.D., M.S., M.R.C.S.a
a

Department of Surgery, bDepartment of Radiology, University Hospitals Case Medical Center, Case
Western Reserve School of Medicine, Cleveland, OH, USA; cDepartment of Radiology, The Brooklyn
Hospital Center, Weill Cornel Medical College, New York, NY, USA
KEYWORDS:
Gastric perfusion;
Leaks;
Sleeve leak

Abstract
BACKGROUND: Several postoperative gastrointestinal complications are attributed to ischemia. We
herein evaluate the gastric wall perfusion using computed tomography (CT) scan perfusion index on
trial to address the etiology of ischemic complication after sleeve gastrectomy.
METHODS: A retrospective study of 205 patients undergoing CT scan of the abdomen to evaluate the
pattern of gastric vascular perfusion was performed. The perfusion index of the gastric mucosa was
measured at 5 gastric points using CT perfusion scanning.
RESULTS: Gastric perfusion at the angle of His (AOH) (53.51 6 14.38) was statistically significantly lower (P , .001) than that at the other gastric points studied: fundus, greater curvature, lesser
curvature, incisura angularis, and mid gastric points (76.16 6 15.21, 73.27 6 16.55, 76.12 6 16.12,
and 75.24 6 14.9, respectively). Gastric perfusion was significantly lower at all the gastric points
(and especially so at the AOH) among obese patients (33 cases) compared with nonobese patients
(18 cases). Gastric perfusion at all the points studied showed a decrease as the body mass index increases. Hypertensive patients had a better gastric perfusion compared with nonhypertensive patients.
CONCLUSIONS: Gastric wall perfusion is statistically significantly decreased at the AOH and gastric
fundus compared with perfusion at other gastric points. Gastric perfusion at all the gastric points studied decreased with the increase in body mass index. Gastric leakage in obese patients following sleeve
gastrectomy could be attributed to a decrease in the blood supply at AOH.
2014 Elsevier Inc. All rights reserved.

Several postoperative gastric complications have been

attributed to ischemia. Recently, there has been increasing
interest in gastric surgery because of the surge in the
number of bariatric surgery. Sleeve gastrectomy has been
The authors declare no conflicts of interest.
* Corresponding author. Tel.: 11-440-991-6765; fax: 11-216-9837230.
E-mail address: saber6231@gmail.com
Manuscript received May 2, 2013; revised manuscript April 26, 2014
0002-9610/$ - see front matter 2014 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.amjsurg.2014.05.023

gaining popularity and has become a standalone weight loss

procedure with an excellent long-term effectiveness in
terms of weight loss and comorbidity resolution.13
However, the dreaded risk of gastric leak through the staple
line is still considered a setback for the sleeve gastrectomy.
Ischemia has been incremented for leak in several gastrointestinal surgeries. There are very few studies analyzing the
gastric blood flow.4 It was found that most leaks after sleeve
gastrectomy occur at the angle of His (AOH).5 The rational
of this study is to evaluate the gastric wall perfusion to

The American Journal of Surgery, Vol -, No -, - 2014

determine the pattern of gastric wall vascularity, utilizing

computed tomography (CT) scan perfusion index (PI).

Evaluation of the gastric perfusion in the

enhanced computed tomography

Principles of computed tomography perfusion

study
The CT perfusion techniques are based on a direct
correlation between enhancement of the soft tissue after
intravenous contrast agent administration and the blood
volumetric flow to this tissue.6
Compartmental and deconvolution analyses are the 2
commonly used analytical methods to quantify vascular
physiology from the data acquired in the dynamic CT. A
general perfusion CT (PCT) technique typically requires a
baseline unenhanced image acquisition, followed by a
series of images acquired over a time period after an
intravenous bolus injection of iodinated contrast media.7
The study and analysis by PCT involve temporal
alterations of tissue attenuation measured in Hounsfield
units following intravenous contrast injection. The tissue
enhancement depends on the iodine concentration, and
indirectly reflects the tissue vascularization and the vascular
physiology.8
After iodinated contrast medium injection, the tissue
enhancement may be divided into 2 phases according to
the contrast agent distribution in the intra- and extravascular compartments.8 At the early phase, the enhancement is purely attributed to the contrast agent
distribution within the intravascular space, and such a
phase usually last 40 to 60 seconds after uptake into
this compartment.

Contrast material administration

A commercially available contrast agent (Iodixanol,
Visipaque 320; GE Healthcare, West Milwaukee, WI) was
infused intravenous injection (IV) using an automated
power injector (Stellant Dual Head Injector; Medrad
Healthcare, Warrendale, PA). Eighty milliliter of Visipaque 320 was administrated through a 20-guage IV
catheter placed in a peripheral vein. The infusion rate
was 5 mL/second. Thirty milliliter of normal saline was
infused at 5 mL/second after the infusion of contrast
material.

Imaging acquisition
All studies were performed using CT Phillips ingenuity
software version 4.0 mas 180 kvp 120 idose and 2 Phillips ict
brilliance software version 3.2.1.1 mas 225 kvp 120 idose 3.
All scans were obtained in the craniocaudal direction at end
inspiration. Slice thickness was 5 mm and additional sagittal
and coronal reconstructions were obtained. All scans were
obtained after 70 seconds from the beginning of the IV
contrast injection.

To our knowledge, CT-based assessment of the gastric

perfusion in different parts of the stomach had not been
undertaken before. In our retrospect study, we compared
the gastric perfusion at the AOH with other gastric points:
fundus, greater curvature, lesser curvature, incisura angularis, and mid gastric. In all patients, these gastric points
were determined and an region of interest of chosen size
was placed to obtain the CT Hounsfield units (CT density
measurement), which reflects the amount of the IV contrast
enhancement in these points and therefore the perfusion of
these points separately. By comparing the enhancement of
the AOH with the other gastric points at the same time of
the enhancing study, we evaluate the pattern of gastric
perfusion with these points.

Patients and Methods

A retrospective study of patients who had CT scan of the
abdomen with intravenous contrast during the month of
July 2012 at the Case Western Medical Center was
screened. A total of 205 patients were included. Our
exclusion criteria were any patient with decompress gastric
tube (nasogastric tube, percutaneous endoscopic gastrostomy tube), hiatal hernia, patients younger than 18 years
old, previous upper abdominal surgery, hemodynamically
unstable patients, and any space occupying the mass
compressing the stomach.
PI of the gastric mucosa was measured at 5 points: the
AOH (junction between the esophagus and stomach),
highest point of the gastric fundus, greater and lesser
curvature at the level of incisura angularis, and mid gastric
antrum, respectively. The technique used to measure the
gastric PI is similar to that described by Choi et al.9 All the
images were reviewed by the same radiologist (N.A.).

Statistical analysis
Data were fed to the computer using the Predictive
Analytics Software (PASW Statistics version 18). Quantitative data were described using median, minimum, and
maximum, as well as mean and standard deviation. The
distributions of quantitative variables were tested for
normality using KolmogorovSmirnov test and Shapiro
Wilk test. DAgstino test was used if there was a conflict
between the 2 previous tests. If it reveals normal data
distribution, parametric tests were applied. If the data were
abnormally distributed, nonparametric tests were used.
For normally distributed data, comparison between 2
independent population were done using independent t test
while more than 2 population were analyzed with F test
(analysis of variance) and (Scheffe) test. Significance test
results are quoted as 2-tailed probabilities. Significance of
the obtained results was judged at the 5% level.

A.A. Saber et al.

Table 1

Gastric wall perfusion

Demographic data

Demographic data
Sex
Male
Female
Age (year)
Minimummaximum
Mean 6 standard deviation
Median
Weight (kg)
Minimummaximum
Mean 6 standard deviation
Median
Height (m)
Minimummaximum
Mean 6 standard deviation
Median
BMI
Minimummaximum
Mean 6 standard deviation
Median

Table 2
Number (%)
18 (35.3)
33 (64.7)
18.082.0
46.69 6 17.42
49.0
46.0204.1
83.38 6 29.11
75.7
1.502.0
1.70 6 .10
1.70
16.065.10
29.21 6 8.96
26.80

BMI 5 body mass index.

Results
Among the 205 patients CT scans, 154 were excluded
according to the above exclusion criteria. A total of 51 (18
male/33 female) patients were included in the study.
Demographic data showed that the mean age was 46.69
6 17.42 years (range 1882), mean height 1.7 6 .1 m,
mean weight 83.4 6 29.11 kg (range 46.0204.1), and
mean body mass index (BMI) 29.21 6 8.96 (range 16
65.1) (Table 1).
Table 2 shows the associated comorbidities; hypertension, dyslipidemia, and gastroesophageal reflux disease
were the most common, and 19 patients (37.3%) were in
fact hypertensive.
The mean PI at each of the 5 gastric points studied was
53.51 6 14.38, 76.16 6 15.21, 73.27 6 16.55, 76.12 6 16.12,
and 75.24 6 14.90 at the AOH, fundic dome, greater and lesser
curvature at the incisura angularis, and mid gastric antrum,
respectively. The perfusion at the AOH was significantly lower
(P ,.001) as compared with the other points. The perfusion at
the greater curvature was found, however, to be lower than that
at the lesser curvature but this was not statistically significant
(P , .898) (Table 3).
When the age of the patient was considered and we divided
the patients into 2 groups according to age, younger , 60 (39
patients) and older R 60 (12 patients), the PI was compared for
all 5 points; it was found that there were no statistically
significant differences between the PI at each of the 5 gastric
points in older and younger patients (Table 4). There were no
significant differences in the PI at each of the gastric points
studied in both male and female patients (Table 5).
The mean height of the patients studied was 1.7 6
.1 m (Table 1). We divided the patients studied into 2

Associated comorbidities

Comorbidity

Number (%)

Hypertension
Dyslipidemia
Gastroesophageal reflux
Depression
Coronary heart disease
Asthma
Diabetes mellitus
Chronic renal failure
Hypothyroid
Lymphoma
Lung cancer

19
10
8
6
5
5
5
4
3
3
3

(37.3)
(19.6)
(15.7)
(11.8)
(.1)
(.1)
(.1)
(.08)
(.06)
(.06)
(.06)

groups: taller (n 5 12) and shorter (n 5 39) patients according to whether the height of the patient was equal/
more or less than the mean height for the whole group,
respectively. The perfusion at the AOH was found to be
significantly lower in the taller patients (P , .029)
(Table 6).
We looked at the effect of the associated comorbidities
on the gastric PI. However, because of the sample size, we
only looked at the PI among hypertensive patients (32) and
compared with nonhypertensive patients.19 The PI was
found to be significantly higher among the hypertensive patients (P , .019) at the AOH but not at all the other gastric
points (Table 7).
In this study, we had 33 obese patients (BMI R 30) and
18 nonobese patients (BMI , 30); the PI at each of the 5
locations studied was found to be lower among obese
patients, and this was, however, statistically significant only
at the fundus (P , .02) (Table 8).
The relation between PI and BMI of the patients studied
is illustrated in Table 9. When the patients were divided
into 4 groups, normal weight, overweight, obese, and
morbidly obese, the PI varied significantly at the fundus,
lesser curvature, and greater curvature. It was found that
as the BMI increases the gastric perfusion decreases. It
was also found that the PI at AOH was always lower than
any other measured point when compared across all of
the subdivided BMI groups.

Comments
The worldwide outburst in bariatric surgery was only
dampened by a number of complications and technical
difficulties that is innate to each procedure. These challenges have inspired the search for an ideal surgical
procedure, and explain the dynamic nature and evolution
of the field of bariatric surgery.10,11
Sleeve gastrectomy, a relatively new weight loss procedure, has been recently gaining popularity for its technical
simplicity and excellent weight loss. However, leak after
sleeve gastrectomy can be a catastrophic complication. In

The American Journal of Surgery, Vol -, No -, - 2014

4
Table 3

Comparison between gastric perfusion at the different studied locations

AOH

Fundus

GC

LC

Pylorus

22.0100.0
53.51 6 14.38
50.0
,.001*

41.0110.0
76.16 6 15.21
75.0

34.0111.0
73.27 6 16.55
73.0

40.0105.0
76.12 6 16.12
76.0

37.0110.0
75.24 6 14.90
75.0

,.001*

,.001*
.894

,.001*
1.000
.898

,.001*
.999
.966
.999

r
Minimummaximum
Mean 6 SD
Median
P
P1
P2
P3
P4

P 5 P value for F test (ANOVA); P1 5 P value of Schaffer test between AOH with each other location; P2 5 P value of Schaffer test between fundus with
each other location; P3 5 P value of Schaffer test between GC with LC and pylorus; P4 5 P value of Schaffer test between LC and pylorus.
ANOVA 5 analysis of variance; AOH 5 angle of His; GC 5 greater curvature; LC 5 leaser curvature; SD 5 standard deviation.
*Statistically significant at P % .05.

fact, leak is the second most common cause of death after

bariatric surgery. Leakage after sleeve gastrectomy was
reported to vary between .7% and 7%.1215 Sleeve gastrectomy is a high-pressure system; this may explain the
possible persistence of leak after sleeve gastrectomy for a
couple of months to heal.
The nonspecific clinical presentations and limitations of
most of the radiological studies contribute to the challenges
of early diagnosis of leak in such patient populations with a
low tolerance for complications. The possible progression of
leak to peritonitis, septic shock, multiple organ failure, and
even death further complicates the situation. In those who
survived the event, recovery can be protracted and

Table 4

complicated. Leak may require readmission, prolonged

hospitalization, transfer to the intensive care unit, total
parental nutrition, intravenous antibiotics, frequent images,
image-guided drainage, or even reoperation. Endoluminal
stents have been used in acute leak; however, stent migration
and erosion can be problematic. As a result, a stormy
postoperative course with increase morbidity, mortality,
financial burden, and potential medicolegal action may
follow.
Leak following sleeve gastrectomy (LSG) is a unpredictable complication; however, 75% to 100% of leaks
occur at AOH.13,16 Several underlying mechanisms have
been incriminated in the pathogenesis of leak, including

Relationship between age and gastric perfusion (r)

Age (year)

Gastric perfusion index (r)

r at angle of His
MinimumMaximum
Mean 6 standard deviation
Median
r at fundus
Minimummaximum
Mean 6 standard deviation
Median
r at greater curvature
Minimummaximum
Mean 6 standard deviation
Median
r at LC
Minimummaximum
Mean 6 standard deviation
Median
r at pylorus
Minimummaximum
Mean 6 standard deviation
Median
P value for Student t test.
LC 5 leaser curvature.

,60 (n 5 39)

R60 (n 5 12)

22.0100.0
53.41 6 15.19
50.0

41.086.0
53.83 6 11.92
49.50

53.0110.0
76.26 6 13.59
75.0

41.0105.0
75.83 6 20.35
74.0

46.0111.0
74.38 6 15.16
75.0

34.0109.0
69.67 6 20.80
66.0

50.0105.0
78.08 6 14.05
78.0

40.0105.0
69.75 6 21.01
67.0

58.0110.0
76.41 6 13.34
75.0

37.0102.0
71.42 6 19.32
67.0

P value
.930

.934

.393

.119

.417

A.A. Saber et al.

Table 5

Gastric wall perfusion

Relationship between sex and gastric perfusion (r)

Sex

Gastric perfusion index (r)

r at angle of His
Minimummaximum
Mean 6 standard deviation
Median
r at fundus
Minimummaximum
Mean 6 standard deviation
Median
r at greater curvature
Minimummaximum
Mean 6 standard deviation
Median
r at leaser curvature
Minimummaximum
Mean 6 standard deviation
Median
r at pylorus
Minimummaximum
Mean 6 standard deviation
Median

Male (n 5 18)

Female (n 5 33)

28.086.0
49.11 6 12.25
49.0

22.0100.0
55.91 6 15.05
56.0

53.0110.0
80.17 6 17.28
79.50

41.0105.0
73.97 6 13.75
74.0

48.0111.0
72.78 6 19.09
68.50

34.0100.0
73.55 6 15.30
73.0

40.0105.0
73.22 6 17.71
71.0

46.0105.0
77.70 6 15.24
80.0

60.0105.0
74.06 6 13.57
69.50

37.0110.0
75.88 6 15.74
78.0

P value
.107

.167

.876

.349

.681

P value for Student t test.

technical errors, staple line crossing, poor vascularization,

and gastric inflammation. However, extensive devascularization of the AOH has been postulated as the incriminating
factor for leak in susceptible patients.15

Table 6

Unfortunately, there is no clear conclusion on the

exact cause of leaks in such a procedure and consequently
no definitive preventive measures have been clearly
identified.15

Relationship between patients height and gastric perfusion (r)

Height (m)

Gastric perfusion index (r)

r at angle of His
Minimummaximum
Mean 6 standard deviation
Median
r at fundus
Minimummaximum
Mean 6 standard deviation
Median
r at greater curvature
Minimummaximum
Mean 6 standard deviation
Median
r at leaser curvature
Minimummaximum
Mean 6 standard deviation
Median
r at pylorus
Minimummaximum
Mean 6 standard deviation
Median
P value for Student t test.
*Statistically significant at P % .05.

,1.75 (n 5 39)

R1.75 (n 5 12)

39.0100.0
55.92 6 13.69
52.0

22.074.0
45.67 6 14.28
44.50

41.0105.0
76.08 6 14.15
75.0

53.0110.0
76.42 6 18.98
71.0

34.0100.0
73.72 6 14.46
73.0

46.0111.0
71.83 6 22.79
68.50

46.0105.0
76.74 6 14.46
76.0

40.0105.0
74.08 6 21.29
74.50

37.0110.0
74.72 6 15.38
75.0

60.0105.0
76.92 6 13.69
77.0

P value
.029*

.947

.791

.622

.659

The American Journal of Surgery, Vol -, No -, - 2014

6
Table 7

Relation between comorbidities and gastric perfusion (r)

Comorbidity

Gastric perfusion index (r)

r at angle of His
Minimummaximum
Mean 6 standard deviation
Median
r at fundus
Minimummaximum
Mean 6 standard deviation
Median
r at greater curvature
Minimummaximum
Mean 6 standard deviation
Median
r at leaser curvature
Minimummaximum
Mean 6 standard deviation
Median
r at pylorus
Minimummaximum
Mean 6 standard deviation
Median

No HTN (n 5 32)

HTN (n 5 19)

22.094.0
49.91 6 12.81
49.0

40.0100.0
59.58 6 15.14
57.0

53.0110.0
77.38 6 15.14
75.50

41.0105.0
74.11 6 15.53
73.0

34.0111.0
73.22 6 18.61
72.50

52.0100.0
73.37 6 12.81
73.0

40.0105.0
76.38 6 17.98
78.0

49.095.0
75.68 6 12.84
75.0

55.0110.0
74.38 6 15.26
70.0

37.097.0
76.68 6 14.57
80.0

P value
.019*

.464

.975

.884

.598

P value for Student t test.

HTN 5 Hypertension.
*Statistically significant at P % .05.

Few studies on the gastric vascular anatomy were done

on cadavers.4,8 The rational of our study is to evaluate any
possible vascular pattern of the stomach of living human
Table 8

through the abdominal CT scan and to look for any correlation that may explain the susceptibility of some patients
and certain gastric area for leak compared with others.

Relationship between BMI (nonobese vs obese) and gastric wall perfusion (r)
BMI

Gastric wall perfusion index (r)

r at AOH
Minimummaximum
Mean 6 standard deviation
Median
r at fundus
Minimummaximum
Mean 6 standard deviation
Median
r at greater curvature
Minimummaximum
Mean 6 standard deviation
Median
r at leaser curvature
Minimummaximum
Mean 6 standard deviation
Median
r at pylorus
Minimummaximum
Mean 6 standard deviation
Median
P value for Student t test.
AOH 5 angle of His; BMI 5 body mass index.
*Statistically significant at P % .05.

,30 (n 5 33)

R30 (n 5 18)

22.0100.0
54.58 6 16.03
50.0

28.075.0
51.56 6 10.86
50.0

53.0110.0
79.73 6 15.09
76.0

41.095.0
69.61 6 13.49
71.0

34.0111.0
76.58 6 17.80
78.0

46.088.0
67.22 6 12.23
68.50

40.0105.0
79.06 6 17.31
83.0

49.094.0
70.72 6 12.35
69.50

55.0110.0
77.36 6 14.78
75.0

37.091.0
71.33 6 14.72
69.0

P value
.479

.022*

.053

.077

.170

A.A. Saber et al.

Table 9

Gastric wall perfusion

Relationship between BMI categories and gastric wall perfusion (r)

BMI

Gastric perfusion index (r)

r at angle of His
Minimummaximum
Mean 6 standard deviation
Median
r at fundus
Minimummaximum
Mean 6 standard deviation
Median
r at greater curvature
Minimummaximum
Mean 6 standard deviation
Median
r at leaser curvature
Minimummaximum
Mean 6 standard deviation
Median
r at pylorus
Minimummaximum
Mean 6 standard deviation
Median

.24.9 (n 5 37)

2529.9 (n 5 16)

3034.9 (n 5 8)

R35 (n 5 10)

39.094.0
54.12 6 13.32
56.0

22.0100.0
55.06 6 18.93
49.50

41.061.0
49.88 6 5.82
49.50

28.075.0
52.90 6 13.85
50.0

69.0110.0
83.88 6 14.11
79.0

53.0105.0
75.31 6 15.27
74.50

53.085.0
70.0 6 11.89
70.0

41.095.0
69.30 6 15.28
71.0

56.0111.0
83.88 6 14.75
82.0

34.0100.0
68.81 6 17.86
65.0

55.088.0
72.50 6 11.70
74.0

46.078.0
63.0 6 11.46
64.0

59.0105.0
85.94 6 15.71
89.0

40.095.0
71.75 6 16.30
73.0

60.094.0
70.38 6 11.27
69.0

49.089.0
71.0 6 13.75
75.0

61.0110.0
81.71 6 14.13
80.0

55.099.0
72.75 6 14.47
68.0

60.091.0
73.75 6 12.08
73.0

37.091.0
69.40 6 16.92
66.50

P value
.870

.046*

.005*

.018*

.152

P value for F test (ANOVA).

ANOVA 5 analysis of variance; BMI 5 body mass index.
*Statistically significant at P % .05.

CT perfusion has been recently described in different parts

of gastrointestinal tract. CT perfusion imaging has been
described for evaluation of gastric perfusion,16,17 as well as
small bowel perfusion.7 CT perfusion measurements have
been reported to differentiate between colon cancer and
diverticulitis with a sensitivity of 80% and specificity of
70%.18 In a prospective study, 52 patients with treatment of
HCV infection underwent PCTand percutaneous liver biopsy
on the same day. Liver samples were scored for fibrosis. CT
perfusion detected that perfusion changes occur early in the
liver during fibrosis in chronic hepatitis C virus infection
with a sensitivity of 71% and a specificity of 65%.19
Gastric PI was measured, in this study, at 5 gastric points
and demonstrated a statistically significant decrease in the
mean PI at the AOH (53.51 6 14.38; P , .001) compared
with PI measured at the other gastric point (4 gastric
points). It is to be noted that perfusion at the gastric fundus
was also low although not statistically significant. These
findings when considered with the increased incidence of
post sleeve gastrectomy leak at the upper third of the stomach may point to an underlying impaired vascular perfusion
as a possible cause of leakage.
In addition, gastric leakage occurs commonly in the first
2 postoperative weeks. In a multicenter study13 on 2,834
patients, 73.2% of leakage cases occurred between days 3
and 14 postoperatively, 20% occurred between days 0 and
2 postoperatively, and 7% after day 14. The fact that
leakage occurs during the active healing phase support
the assumption that leak could be attributed to alteration
in the normal acute healing process because of local risk

factor including inadequate blood supply rather than failure

of staple line.
In this large multicenter study,13 sleeve gastrectomy leak
occurred in 1.5% of the cases and intraoperative leak tests
and postoperative swallow test failed to detect leakage in
97.3% of the cases that developed leakage. This could
minimize the possibility of technical errors with stapling
as a cause of post LSG leak.
Prevention is the best treatment for such a complication,
so every effort must be made to create a reliable staple line
at the initial operation. An appropriate height intact staple
line with well-perfused tissue, good local hemostasis, and
intraoperative evaluation of staple line integrity must be
achieved to minimize the risk for leak. The above information emphasizes the importance of developing a new
strategy to avoid leak during sleeve gastrectomy particularly at the high-risk area, that is, AOH.
In our study, age and sex of the patient did not impact
the gastric perfusion; however, BMI did. Gastric perfusion
was significantly lower at all the gastric points (particularly
at the AOH) among obese patients (33 cases) compared
with nonobese patients (18 cases), and this was statistically
significant only at the fundus. Gastric perfusion at all the
points studied showed a decrease as the BMI increases.
These findings partially agree with a recent review of 4,888
laparoscopic sleeve gastrectomies; significantly high leak
rates were found in heavier patients (BMI .50 vs ,50).16
Our finding of decrease in gastric PI at the AOH and
fundus correlate with the literature that most gastric leakage
following sleeve gastrectomy occurs at AOH. This

The American Journal of Surgery, Vol -, No -, - 2014

correlation supports the fact that local ischemia can be a

major factor for leak at the AOH.
A short height staple may over compress the gastric wall
and augment ischemia in an inherited vascular compromised area of the stomach; AOH, gastric fundus. The
resulting ischemia can explain relatively high incidence of
leak at those susceptible gastric locations, for example,
AOH in sleeve gastrectomy and gastric fundus in esophagogastric anastomosis following esophageal resection.
Also, extensive dissection and devascularization at the
AOH may augment the inherited local ischemia at the
AOH.
The vascular anatomy of the stomach is also particularly
relevant to other surgical procedures, such as esophageal
resections.20 Esophagogastric anastomotic leak has been
associated with leak-related mortality up to 5% in cervical
anastomoses and as high as 60% for thoracic anastomoses.21,22 The most important predisposing factors are attributed to ischemia of the gastric conduit.23 Our findings of
decrease in the gastric mucosal perfusion at the dome of
the gastric fundus may explain the relatively high risk for
esophagogastric anastomotic leak following esophageal
resection.
During laparoscopic sleeve gastrectomy procedures, we
have used several technical maneuver at the AOH to avoid
such catastrophic but potentially preventable complications. These include the following: upgrading the height
of staple line from 3.5 mm to 4.2 mm, avoiding excessive
gastric devascularization at the AOH, and avoiding lateral
thermal gastric injury at the AOH during mobilization with
energy source. If we suspect over-compression of the
stomach at the staple line, we extraluminally support the
staple line with omentum and glue. In addition, in high-risk
patients placing drains, postoperative swallow study may
help in early detection of leak. Because we used these
preventive maneuvers, we did not have any case of leak.
However, prospective randomized studies with large number of procedures comparing these maneuvers are required
to evaluate the utility of such maneuvers.
In conclusion, our study of CT scan mapping of gastric wall
perfusion on 51 cases showed a statistically significant
decrease in vascular perfusion at the AOH and gastric fundus
compared with perfusion at other gastric points. Gastric
perfusion at all the gastric points studied decreased with the
increase in BMI. Gastric leakage in obese patients following
LSG could be attributed to a decrease in the blood supply at the
high-risk gastric area, rather than because of technical error.
Our study has several limitations including the retrospective
design of our study, small sample number, and no comparison
of gastric perfusion before and after such procedures.
A prospective study of CT gastric perfusion scan mapping
of a large number of patients undergoing sleeve gastrectomy
comparing preoperative and postoperative gastric perfusion
in the patients developing this complication compared with
those who did not get postoperative leakage will be helpful to
clarify our findings.

References
1. Bohdjalian A, Langer FB, Shakeri-Leidenmuhler S, et al. Sleeve gastrectomy as sole and definitive bariatric procedure: 5-year results for
weight loss and ghrelin. Obes Surg 2010;20:53540.
2. Tucker ON, Szomstein S, Rosenthal RJ. Indications for sleeve gastrectomy as a primary procedure for weight loss in the morbidly obese.
J Gastrointest Surg 2008;12:6627.
3. Felberbauer FX, Langer F, Shakeri-Manesch S, et al. Laparoscopic
sleeve gastrectomy as an isolated bariatric procedure: intermediateterm results from a large series in three Austrian centers. Obes Surg
2008;18:8148.
4. Gomes M, Ramacciotti E, Miranda Jr F, et al. Vascular flow of the
gastric fundus after arterial devascularization: an experimental study.
J Surg Res 2009;152:12834.
5. Casella G, Soricelli E, Rizzello M, et al. Nonsurgical treatment of staple line leaks after laparoscopic sleeve gastrectomy. Obes Surg 2009;
19:8216.
6. Wu GY, Ghimire P. Perfusion computed tomography in colorectal cancer: protocols, clinical applications and emerging trends. World J Gastroenterol 2009;15:322831.
7. Sitek A, Sheiman RG. Small-bowel perfusion measurement: feasibility
with single-compartment kinetic model applied to dynamic contrastenhanced CT. Radiology 2005;237:6704.
8. Loukas M, Wartmann CT, Louis Jr RG, et al. The clinical anatomy of
the posterior gastric artery revisited. Surg Radiol Anat 2007;29:3616.
9. Choi SH, Chung JW, Kim HC, et al. The role of perfusion CT as a
follow-up modality after transcatheter arterial chemoembolization:
an experimental study in a rabbit model. Invest Radiol 2010;45:
42736.
10. Prentice AM. The emerging epidemic of obesity in developing countries. Int J Epidemiol 2006;35:939.
11. Buchwald H, Avidor Y, Braunwald E, et al. Bariatric surgery: a systematic review and meta-analysis. JAMA 2004;292:172437.
12. Lalor PF, Tucker ON, Szomstein S, et al. Complications after laparoscopic sleeve gastrectomy. Surg Obes Relat Dis 2008;4:338.
13. Sakran N, Goitein D, Raziel A, et al. Gastric leaks after sleeve gastrectomy: a multicenter experience with 2,834 patients. Surg Endosc 2013;
27:2405.
14. Burgos AM, Braghetto I, Csendes A, et al. Gastric leak after laparoscopicsleeve gastrectomy for obesity. Obes Surg 2009;19:16727.
15. Parikh M, Issa R, McCrillis A, et al. Surgical strategies that may
decrease leak after laparoscopic sleeve gastrectomy: a systematic review and meta-analysis of 9991 cases. Ann Surg 2013;257:2317.
16. Aurora AR, Khaitan L, Saber AA. Sleeve gastrectomy and the risk of
leak: a systematic analysis of 4,888 patients. Surg Endosc 2012;26:
150915.
17. Li Zhiyong, Ge Ying, Liu Jinghong, et al. Clinical study of stomach
neoplasm CT perfusion imaging. Chinese-German J Clin Oncol
2009;8:2079.
18. Goh V, Halligan S, Taylor SA, et al. Differentiation between diverticulitis and colorectal cancer: quantitative CT perfusion measurements
versus morphologic criteriadinitial experience. Radiology 2007;242:
45662.
19. Ronot M, Asselah T, Paradis V, et al. Liver fibrosis in chronic hepatitis
C virus infection: differentiating minimal from intermediate fibrosis
with perfusion CT. Radiology 2010;256:13542.
20. Sarli D, Bona D, Abraham M, et al. Conservative and surgical treatment of esophagogastric anastomotic leaks. Ann Ital Chir 2006;77:
3916.
21. Pierie JP, de Graaf PW, van Vroonhoven TJ, et al. Healing of the cervical esophagogastrostomy. J Am Coll Surg 1999;188:44854.
22. Giuli R, Gignoux M. Treatment of carcinoma of the esophagus: retrospective study of 2,400 patients. Ann Surg 1980;192:4452.
23. Urschel JD. Esophagogastrostomy anastomotic leaks complicating
esophagectomy: a review. Am J Surg 1995;169:63440.