Note: This copy is for your personal non-commercial use only.

To order presentation-ready
copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.

ORIGINAL RESEARCH
Roux-en-Y Gastric Bypass
Surgery for Morbid Obesity:

䡲 GASTROINTESTINAL IMAGING
Evaluation of Postoperative
Extraluminal Leaks with Upper
Gastrointestinal Series1
Laura R. Carucci, MD
Purpose: To retrospectively evaluate the radiographic features of
Mary Ann Turner, MD
extraluminal leak after Roux-en-Y gastric bypass (RYGBP)
Robert C. Conklin, MD
surgery at upper gastrointestinal (GI) examinations in a
Eric J. DeMaria, MD large series of patients and to determine morbidity and
John M. Kellum, MD mortality in those patients with leak.
Harvey J. Sugerman, MD
Materials and The investigational review board approved this HIPAA-
Methods: compliant study, and the need for patient informed con-
sent was waived. Radiologic database review revealed
1202 upper GI studies performed over a 4-year period in
906 patients after RYGBP. Extraluminal leak was identified
in 50 patients. Two patients with leaks that occurred be-
fore the study period were excluded. Of the remaining 48
patients, 12 were men and 36 were women (mean age, 45
years; range, 26 – 64 years). Surgery had been laparo-
scopic in 23 patients and open in 25. Upper GI studies
were analyzed by two radiologists in consensus for the
origin, extent, and severity of leaks and associated find-
ings. Chart review was performed to determine clinical
course, treatment, and outcome.

Results: Fifty extraluminal leaks were detected in 48 of 904 patients

(5.3%) at upper GI examinations. All leaks were identified
within 28 days, and, in 37 of 48 patients (77%), leakage
was diagnosed within 1 week of surgery. The majority of
leaks (n ⫽ 37) originated from the gastrojejunal anastomo-
sis. Leaks also occurred at the distal portion of the esoph-
agus (n ⫽ 5), the gastric pouch (n ⫽ 5), the oversewn
jejunum (n ⫽ 2), and the distal anastomosis (n ⫽ 1). Leaks
extended into the left upper quadrant in 30 patients. Ob-
struction or ileus was present in 35 of 48 patients (73%).
Leak into the excluded stomach was observed in 15 of 48
patients. The occurrence of extraluminal leak prolonged
hospital stays; organ failure occurred in 14 (29%) and
death in three (6%) of the 48 patients.

Conclusion: Extraluminal leak was identified on upper GI series in 48 of

904 patients (5.3%) after RYGBP for morbid obesity. Ex-
1
From the Departments of Radiology (L.R.C., M.A.T.,
traluminal leak most commonly arises from the gastrojeju-
R.C.C.) and Surgery (E.J.D., J.M.K., H.J.S.), Virginia Com- nal anastomosis and extends into the left upper quadrant.
monwealth University Medical Center, 1250 E Marshall Extraluminal leak affects morbidity and mortality.
St, Main Hospital, 3rd Floor, Rm 417, PO Box 980615,
Richmond, VA 23298-0615. Received September 8, 2004; 娀 RSNA, 2006
revision requested November 15; revision received Febru-
ary 4, 2005; final version accepted March 1.

姝 RSNA, 2006

Radiology: Volume 238: Number 1—January 2006 119

GASTROINTESTINAL IMAGING: Extraluminal Leak after Roux-en-Y Gastric Bypass Carucci et al

M
orbid obesity is a major health
problem in the United States
and is increasing in epidemic
proportions (1,2). Currently, more than
50% of adults in the United States are
considered to be overweight or obese
(body mass index, ⬎25 kg/m2) (1,3–5).
Furthermore, owing to the associated
Figure 1
increased risks of comorbidity, disabil-
ity, and early mortality, obesity places a
tremendous burden on U.S. health care
(2,4–9).
Nonsurgical approaches to weight
loss have had limited long-term efficacy
for the treatment of morbid obesity (1,9–
11). The most effective treatment of mor-
bid obesity in terms of sustained weight
loss, decreased morbidity, and prolonged
life expectancy is bariatric surgery (1,3,9,
12–14). Accordingly, rates of bariatric
surgery have dramatically increased in re-
cent years. For patients in whom other
methods of weight reduction have failed,
bariatric surgery is considered if the body
mass index is greater than 40 kg/m2 or
greater than 35 kg/m2 with the presence
of associated comorbidities (9,13,15).
Roux-en-Y gastric bypass (RYGBP) sur-
gery is the most successful bariatric pro-
cedure in terms of sustained weight loss
and decreased obesity-related morbidity
(9,16–20). It is currently the procedure of
choice for the management of morbid
obesity in the United States (1,8,10,11,
14,21).
The surgical procedure consists of
forming a small gastric fundal pouch to
exclude the remainder of the stomach
and duodenum. Next, a jejunal Roux
loop is anastomosed to the gastric
pouch with a small stoma (typically
8 –12 mm in diameter). This creates a
blind-ending jejunal limb and an ante-
grade-flowing jejunal limb. Finally, a
distal side-to-side anastomosis of the
excluded jejunal limb and the ante-
grade-flowing jejunal limb is created
(Fig 1a, 1b).
The most serious complication of
bariatric surgery is postoperative ex-

Published online
10.1148/radiol.2381041557

Radiology 2006; 238:119 –127

Abbreviations:
GI ⫽ gastrointestinal
LPO ⫽ left posterior oblique
RYGBP ⫽ Roux-en-Y gastric bypass
SLL ⫽ staple-line leak

Author contributions:
Guarantors of integrity of entire study, L.R.C., M.A.T.;
study concepts/study design or data acquisition or data
analysis/interpretation, all authors; manuscript drafting or
Figure 1: Images depict normal postoperative anatomy after RYGBP. (a) Diagram of surgical procedure manuscript revision for important intellectual content, all
shows creation of a small gastric pouch with a gastrojejunal anastomosis, a blind-ending jejunal limb (ar- authors; approval of final version of submitted manuscript,
rows), and a distal side-to-side jejunojejunal anastomosis. (b) Diagram of postoperative anatomy depicts the all authors; literature research, L.R.C.; clinical studies,
L.R.C., M.A.T., E.J.D., J.M.K., H.J.S.; and manuscript
expected course of food and liquid. G-J ⫽ gastrojejunal, J-J ⫽ jejunojenunal. (c) Diagram and (d) corre-
editing, all authors
sponding fluoroscopic spot image obtained at upper gastrointestinal (GI) examination with the patient in the
supine left posterior oblique (LPO) position show the pouch (P), the gastrojejunal anastomosis (A), the ante- Address correspondence to L.R.C.
grade-flowing jejunal limb (J), and the blind-ending jejunal limb (BL). Region-of-interest circles in b and c (e-mail: lcarucci@vcu.edu).
show sites of possible postoperative leaks.
Authors stated no financial relationship to disclose.

120 Radiology: Volume 238: Number 1—January 2006

GASTROINTESTINAL IMAGING: Extraluminal Leak after Roux-en-Y Gastric Bypass
traluminal leak with resultant peritoni-
tis. Therefore, upper GI examinations
are routinely performed 1–2 days after
surgery to assess for such leaks, as well
as anastomotic narrowing and obstruc-
tion. At upper GI examinations per-
formed after RYGBP, contrast material
is seen to flow from the esophagus into
the gastric pouch and to exit through a
small stoma to enter both the blind-end-
ing and antegrade-flowing jejunal limbs
(Fig 1c, 1d). The presence of extrava-
sated contrast material, which often fills
an extraluminal collection or communi-
cates with a drain, is diagnostic of a free
leak, while filling of the excluded stom-
ach at initial evaluation is indicative of
staple-line leak (SLL).
The radiology literature to date fo-
cuses on general complications after
RYGBP and emphasizes computed to-
mographic (CT) findings. The location
and pattern of postoperative extralumi-
nal leaks on upper GI series have not
been well described in the literature.
Furthermore, other reports describe
complications seen on fluoroscopic
studies with contrast material enhance-
ment obtained after other bariatric pro-
cedures and other types of gastric by-
pass procedures that are no longer
performed (ie, loop gastric bypass)
(15,22–26). Because RYGBP surgery is
increasingly being performed and pa-
tients treated with this procedure are
routinely evaluated with contrast-en-
hanced studies, radiologists must be
aware of the typical postoperative anat-
omy and findings after RYGBP, as well
as the potential complications of this
procedure. Therefore, the purpose of
our study was to retrospectively evalu-
ate the radiographic features of extralu-
minal leak after RYGBP surgery at up-
per GI examinations in a large series of
patients and to determine morbidity
and mortality in those patients with
leak.
Materials and Methods
Review of the radiology database at our
institution revealed that 1202 upper GI
studies were performed in 906 patients
after RYGBP during a 4-year period
(October 1998 to October 2002). Dur-
Radiology: Volume 238: Number 1—January 2006
ing this period, upper GI studies were
performed on a routine basis 1–2 days
postoperatively to assess potential sur-
gical complications. The present study
was conducted in accordance with all
guidelines set forth by the approving in-
stitutional review board. Given the ret-
rospective nature of the study, the re-
quirement for patient informed consent
was waived. Our study was compliant
with the Health Insurance Portability
and Accountability Act.
Patients with Extraluminal Leak
Review of reports of upper GI studies
revealed that postoperative extralumi-
nal leaks occurred in 50 of the 906 pa-
tients. Two of these 50 patients were
excluded from the study because they
had experienced a documented postop-
erative leak before the onset of the
study period (October 1998). There-
fore, our study group consisted of 48
patients (12 men, 36 women; mean age,
45 years; age range, 26 – 64 years) in
whom extraluminal leaks were identi-
fied at upper GI examinations. Upper GI
studies were performed a mean of 2.1
days after surgery (range, 1–18 days).
Three patients underwent initial upper
GI examinations more than 1 week after
surgery (on postoperative days 8, 9, and
18) because their clinical status prohib-
ited earlier examinations. Follow-up up-
per GI or CT studies were performed to
assess any change in clinical status or
symptoms, to monitor treatment of a
leak, and to assess extraluminal collec-
tions.
Patient weight ranged from 230 to
563 lb (103.5–253.4 kg); the mean
weight was 324 lb (145.8 kg). The body
mass index ranged from 40 to 82 kg/m2
(mean, 51 kg/m2). Surgery had been
performed laparoscopically in 23 of the
48 patients (48%) and as an open pro-
cedure in the remaining 25 patients.
One of the 48 patients underwent the
initial RYGBP procedure at an outside
institution. Of the 48 patients, 11 had
previously undergone the following gas-
tric surgical procedures: RYGBP (n ⫽
7), antireflux fundoplication (n ⫽ 2),
laparoscopic gastric banding (n ⫽ 1),
and vertical banded gastroplasty (n ⫽
1). In the nine patients who had previ-
Carucci et al
ously undergone bariatric surgery, sur-
gical revision was performed owing to
failure to maintain weight loss (n ⫽ 7)
or obstruction/anastomotic stenosis
(n ⫽ 2). No patient in our study group
underwent surgical revision owing to a
previous leak.
Upper GI Examination Technique
After a preliminary overhead radio-
graph was acquired with the patient in
the supine position, upper GI examina-
tions were initially performed with the
patient in the supine LPO position so
that the gastrojejunal anastomosis could
be optimally assessed. Examinations
were performed with an EPS 30 digital
fluoroscopy system (Toshiba Medical
Systems, Tokyo, Japan), and additional
views were obtained as necessary. Each
patient was given approximately 50 –100
mL of water-soluble contrast material
(Gastrografin; E-Z-Em, Anjou, Quebec,
Canada), and fluoroscopic spot images
of the distal portion of the esophagus,
the proximal anastomosis, the gastric
pouch, and the proximal small bowel
were obtained. If no leak was identified
with the water-soluble contrast mate-
rial, low-density (60% wt/vol) barium
suspension (Barosperse; Lafayette Phar-
maceuticals, Lafayette, Ind) was adminis-
tered to enable assessment of more sub-
tle leaks. Overhead radiographs were
obtained with the patient in the supine
position immediately after the fluoro-
scopic examination and at 20 –30-
minute intervals until contrast material
was seen to pass the distal small-bowel
anastomosis.
Image Analysis
The radiologic studies performed in the
48 patients were analyzed in consensus
by two abdominal radiologists (L.R.C.,
with 7 years of experience; and M.A.T.,
with more than 25 years of experience
with upper GI and CT studies). The
studies consisted of 308 upper GI series
and 98 CT examinations. Each patient
underwent a mean of 5.4 upper GI stud-
ies and 2.2 CT examinations (ranges,
1–16 and 0 –24, respectively).
Upper GI studies were analyzed to
determine the origin and extent of free
leaks. Leaks were graded as mild, mod-
121
GASTROINTESTINAL IMAGING: Extraluminal Leak after Roux-en-Y Gastric Bypass Carucci et al

erate, or severe on the basis of the size
of the extraluminal collection of con-
trast material (⬍3 cm for mild leaks,
3– 6 cm for moderate leaks, and ⬎6 cm
for severe leaks). The sizes of the gas-
tric pouch and stoma were measured in
the LPO view. The upper GI studies
were also assessed for the presence of a
fistula and/or SLL. Associated findings
on the upper GI study, including pneu-
moperitoneum, obstruction, and severe
Table 1
Location of Leak Relative to
Gastrojejunal Anastomosis in 48
Patients
Leak Location
Left
Anterior
Right or medial
Posterior
Inferior
Above gastroesophageal junction
Lower abdomen or pelvis
* Data in parentheses are percentages. Leaks occurred
in more than one location in some patients.
ileus, were documented. Obstruction bowel, proximal stasis, and prolonged
was diagnosed when upper GI studies transit time. Ileus was diagnosed when
revealed dilated proximal bowel with a diffusely dilated bowel without a transi-
transition point to decompressed distal tion point was present. Pleural effusion
and parenchymal changes in the visual-
Figure 3 ized lung bases were assessed. Correla-
tive cross-sectional CT and follow-up
fluoroscopy studies were reviewed after
initial interpretation of the upper GI
study was performed.
Clinical Chart Review
Clinical chart review was performed by
two authors (L.R.C., R.C.C.) to deter-
mine if there were any complications
that occurred during the surgical proce-
dure, the clinical presentation of leaks,
No. of the time of onset of leaks, treatment,
Patients* and patient outcome. The effect of a
36 (75) leak on hospital resources, including the
9 (19) number of procedures performed, criti-
8 (17) cal care requirements, the length of the
4 (8)
Figure 3: Supine frontal overhead radiograph hospital stay, and the number of hospi-
3 (6)
from upper GI examination in 48-year-old woman tal admissions, was assessed. Addi-
shows a large free leak (L) extending to the left of
2 (4) tional complications that occurred after
1 (2)
the gastrojejunal anastomosis and into the left
a leak, such as infection, thromboem-
upper quadrant. Contrast material extends into the
bolic events, bleeding, organ failure,
subphrenic space and left paracolic gutter. P ⫽
pouch.
and death, were documented.

Results

Extraluminal Leaks
Figure 2
In two of the 48 patients with extralumi-
nal leakage, leaks were identified at two
sites simultaneously. Hence, there were
50 leaks in 48 patients. The mean gas-
tric pouch size was 3.5 ⫻ 5.0 cm (range,
0 –7 ⫻ 0 – 8 cm), and the mean stomal
size was 6 mm (range, 0 –15 mm). All
leaks were identified within 28 days af-
ter surgery (mean, 5.5 days; range,
1–28 days), and, in 37 of 48 patients
(77%), the leak was diagnosed within 1
week after surgery. In 12 patients, the
initial upper GI series did not reveal a
leak but a leak was diagnosed with fol-
low-up upper GI studies performed
4 –28 days (mean, 11.9 days) after the
initial surgery. In 11 of 12 patients, leak-
Figure 2: Extent of leak on supine LPO spot images from upper GI examinations performed on postopera- age occurred after bowel obstruction.
tive day 1 after RYGBP in two patients. (a) Image in 35-year-old woman shows a small leak (L) and a 1.5-cm Six of these 12 patients underwent sur-
collection of contrast material (arrow). Follow-up studies revealed that the leak healed without intervention. gical exploration between the time of
(b) Image in 56-year-old woman shows a large leak (L) with an 8-cm collection of contrast material (arrows) in the initial upper GI study and the diag-
the left upper quadrant. This large leak required immediate surgical repair. A ⫽ gastrojejunal anastomosis,
nosis of leak. In one patient, the leak
J ⫽ jejunum, P ⫽ pouch.
was diagnosed after the patient was ini-

122 Radiology: Volume 238: Number 1—January 2006

GASTROINTESTINAL IMAGING: Extraluminal Leak after Roux-en-Y Gastric Bypass Carucci et al

Table 2 Figure 4
Anatomic Origin of Leak in 48
Patients
No. of
Leak Origin Patients

Gastrojejunal anastomosis 37 (77)

Distal esophagus* 5 (10)
Gastric pouch 5 (10)
Blind jejunal limb 2 (4)
Jejunojejunal anastomosis 1 (2)

Note.—Data in parentheses are percentages.

* In two of these five patients, leaks from both the
gastrojejunal anastomosis and the distal esophagus
were identified.

tially discharged from the hospital, and Figure 4: Leak from gastrojejunal anastomosis extending to the left with resultant left upper quadrant col-
there were corresponding acute symp- lection. (a) Supine frontal spot image from upper GI examination in 55-year-old woman shows a leak from the
toms. gastrojejunal anastomosis (arrow) filling a large collection (L). (b) Transverse CT image in same patient
Leak severity was assessed on the shows the leak (arrow) and a left upper quadrant collection of extraluminal contrast material (L). Also note the
basis of the size of the extraluminal col- associated left pleural effusion. P ⫽ pouch.
lection of contrast material. In 11 pa-
tients, leak was classified as mild—that
is, resulting in a collection smaller than
3 cm (Fig 2a). The leak was severe (ie, Figure 5
resulted in a collection larger than 6 cm)
in 24 patients (Fig 2b). The remaining
13 leaks resulted in 3– 6-cm collections
and were classified as moderate.
The majority (75%) of extraluminal
leaks (Table 1) extended to the left of
the stoma (Fig 3), and left upper quad-
rant collections (Fig 4) occurred in 30 of
48 patients (62%). Leaks that extended
anteriorly, medially, or posteriorly (Fig
5) relative to the stoma occurred less
often. Rarely, collections of contrast
material extended inferior to the stoma,
above the gastroesophageal junction, or
in the pelvis.
The leak (Table 2) originated from
the gastrojejunal anastomosis in the ma- Figure 5: Posterior leak. (a) Left lateral upper GI spot image in 39-year-old woman shows a leak (arrow)
jority of patients (n ⫽ 37, 77%) (Fig 6). that extends posteriorly from the gastrojejunal anastomosis. (b) Transverse CT image in same patient shows a
Leaks from the distal esophagus oc- left upper quadrant collection of fluid and gas (arrow).
curred in five of the 48 patients (Fig 7).
In two of these patients, stomal leaks
were also present and were confirmed jejunojejunal anastomosis on a 30- ileus was diagnosed at upper GI exami-
surgically at the time of leak repair. minute delayed overhead radiograph nations in 35 of 48 patients (73%).
Leaks originated from the gastric pouch (Fig 10). Leak from the distal anastomo- Moderate to severe ileus was identified
in five patients (Fig 8) and from the sis was confirmed surgically. in seven patients. Partial postoperative
blind-ending jejunal limb in two patients obstruction was identified in 28 patients
(Fig 9). In one patient, no leak was Upper GI Examination Findings in the following locations: jejunojejunal
present at initial fluoroscopic evalua- Associated with Free Leak anastomosis (n ⫽ 17), gastrojejunal
tion; however, a large collection was Extraluminal leak was identified at the anastomosis (n ⫽ 8), and at the site
seen in the inferior abdomen near the same time as or after obstruction or where the surgically mobilized jejunum

Radiology: Volume 238: Number 1—January 2006 123

GASTROINTESTINAL IMAGING: Extraluminal Leak after Roux-en-Y Gastric Bypass
crosses the transverse mesocolon (n ⫽
3). Left pleural effusion was observed in
22 of 48 patients (46%). The effusion
required drainage and/or chest tube
placement in 11 patients (23% of the 48
patients with free leak). Free air was
noted in 24 patients (50%). Left lower
Figure 6
Figure 6: Leak from gastrojejunal anastomo-
sis. Supine LPO spot image from upper GI exami-
nation in 48-year-old man shows a leak from the
gastrojejunal anastomosis (arrow) with several
collections of contrast material in the left upper
quadrant. A ⫽ gastrojejunal anastomosis, J ⫽
jejunal limb.
Figure 7
Figure 7: Leak from distal esophagus. Supine
LPO spot image from upper GI examination in
49-year-old woman shows a small leak (arrows)
from the distal esophagus. A ⫽ gastrojejunal
anastomosis, P ⫽ pouch.
124
lobe atelectasis was present in 44 pa-
tients (92%).
Chronic cutaneous fistulas were di-
agnosed after extraluminal leak at fol-
low-up fluoroscopic contrast studies in
five of 48 patients (10%). Four of these
patients had enterocutaneous fistulas;
the remaining patient had a colocutane-
ous fistula. SLL was diagnosed at upper
GI examinations in 15 of the 48 patients
with extraluminal leak (31%) (Fig 11).
In six of these 15 patients, SLL occurred
simultaneously with a free leak at the
time of initial diagnosis of the leak; in
the remaining nine patients, SLL was
diagnosed at follow-up upper GI exami-
nations a mean of 21 days (range, 4 – 45
days) after the diagnosis of free leak. Of
the 15 patients with leak into the ex-
cluded stomach, 10 had undergone an
open surgical procedure and five had
undergone a laparoscopic procedure. In
four of the five patients with SLL after
laparoscopic surgery, an extraluminal
leak preceded the development of leak
into the excluded stomach. Because the
stomach is transected in the laparo-
scopic technique, communication with
the excluded stomach presumably oc-
curs via a fistula. Follow-up upper GI
studies in the 15 patients with SLL re-
vealed chronic SLLs in seven patients
and healing of the SLL without inter-
vention in four patients. Three patients
ultimately underwent reversal of the
gastrojejunostomy because of complica-
tions caused by the extraluminal leak,
Figure 8
Figure 8: Leak from gastric pouch. Two consecutive supine LPO spot images from upper GI examination in
33-year-old woman show a postoperative leak from the gastric pouch (P). (a) Early filling of the pouch reveals
a small leak (arrow). (b) Moments later, the leak (arrow) appears slightly larger, and contrast material has
exited the pouch (P) and entered the jejunum through the gastrojejunal anastomosis (A).
Carucci et al
and one patient underwent successful
surgical revision. According to findings
at follow-up upper GI examination, this
latter patient had a good result.
Review of Initial Surgical Reports
Review of surgical reports revealed
complications at the time of initial sur-
gery in 23 of 48 patients (48%). Anasto-
moses were tested intraoperatively in
all patients with either methylene blue
dye or endoscopy. Extraluminal leaks
were found and repaired intraopera-
tively in 12 of the 48 patients (25%). A
single leak site was found at the time of
initial surgery in seven patients; in six
patients, the leak occurred at the gas-
trojejunal anastomosis, and in one pa-
tient, it arose from the distal esophagus.
In five patients, leaks were found to
originate from more than one site, in-
cluding the stoma, esophagus, gastric
pouch, and/or the excluded stomach.
Other problems occurred at the time of
initial surgery in 11 of the 48 patients
(23%) and included the following: mild
solid organ injury (n ⫽ 6), abnormal
bowel requiring resection (n ⫽ 3), ex-
tensive adhesions requiring lysis (n ⫽
2), prolonged bleeding (n ⫽ 2), difficulty
creating the gastrojejunal anastomosis
(n ⫽ 2), small-bowel ischemia (n ⫽ 1),
hypotensive shock (n ⫽ 1), and equip-
ment failure that inadvertently resulted
in massive dilatation of the excluded
stomach (n ⫽ 1). (More than one com-
Radiology: Volume 238: Number 1—January 2006
GASTROINTESTINAL IMAGING: Extraluminal Leak after Roux-en-Y Gastric Bypass
Table 3
Effect of Extraluminal Leak on Health
Care Resources Used for 48 Patients
Parameter Mean
No. of upper GI studies
performed 5.4 (1–16)
No. of CT studies performed 2.2 (0–24)
No. of fluoroscopic procedures
performed 1.1 (0–6)
No. of admissions 2.3 (1–9)
No. of surgical procedures
performed 3.1 (1–10)
Length of hospital stay (d) 42.4 (6–142)
Note.—Data in parentheses are ranges.
plication occurred in five of the 11 pa-
tients.)
Clinical Presentation of Extraluminal Leak
Clinical findings of postoperative leak
included elevated white blood cell
count, fever, and/or tachycardia in 44 of
48 patients (92%). Severe hypotension
as a sign of leak occurred in seven pa-
tients (15%). Substantial pain at one or
more sites occurred in 26 of 48 patients
(54%) and included left shoulder pain
(n ⫽ 9), generalized abdominal pain
(n ⫽ 5), back pain (n ⫽ 4), epigastric
pain (n ⫽ 4), left flank pain (n ⫽ 4), and
retrosternal chest pain (n ⫽ 3). Addi-
tional clinical signs associated with leak
included the following: shortness of
breath (n ⫽ 8), hiccups (n ⫽ 2), in-
creased drain output (n ⫽ 3), and syn-
cope (n ⫽ 1). Nausea and vomiting oc-
curred in 39 of 48 patients (81%).
Clinical Consequences of
Extraluminal Leak
As a consequence of postoperative leak,
39 of 48 patients (81%) required repeat
surgery. The number of surgical proce-
dures per patient in our study group
ranged from one to 10; the mean num-
ber of surgical procedures was 3.1 (Ta-
ble 3). Four patients ultimately required
reversal of the gastrojejunostomy owing
to extensive complications. The seven
patients who did not require repeat sur-
gery were treated with a combination of
bowel rest, the administration of intra-
venous fluids and antibiotics, and percu-
taneous drain placement. Two of the 48
Radiology: Volume 238: Number 1—January 2006
Figure 9
Figure 9: Leak from blind-ending jejunal limb.
Supine LPO fluoroscopic spot image from upper
GI examination in 38-year-old woman shows a
large leak (arrow) from the oversewn, blind-ending
jejunal limb (BL). Note the large collection of con-
trast material in the left upper quadrant (L). A ⫽
gastrojejunal anastomosis, J ⫽ antegrade jejunal
limb, P ⫽ pouch.
patients had very small leaks that were
seen at upper GI examinations but were
believed to be clinically unimportant.
Postoperative leaks had a substan-
tial effect on the duration of the hospital
stay and the use of services (Table 3).
The majority (n ⫽ 45, 94%) of the 48
patients with leak required long-term
tube feeding or total parenteral nutrition.
Of the 48 patients, 22 (46%) required at
least one fluoroscopic procedure, includ-
ing feeding tube manipulations and drain-
age catheter treatment of abscesses or
distention of the excluded stomach. Pa-
tients with leaks after RYGBP had a pro-
longed hospital stay (mean duration of
stay, 42.4 days). Multiple complications
occurred as a consequence of the long
hospital stay. Leaks resulted in inten-
sive care unit stays of longer than 1
week and/or ventilator dependence in
30 patients (62%). Severe wound infec-
tion occurred in 15 patients (31%). In-
fection with multidrug-resistant organ-
isms developed in 15 patients (31%).
Thromboembolic events, including pul-
monary embolism or deep venous
thrombosis, occurred in six patients
(12%). Gastrointestinal bleeding oc-
curred in five patients (10%). One or
more kinds of organ failure, including
Carucci et al
Figure 10
Figure 10: Leak from jejunojejunal anastomo-
sis. Thirty-minute delayed supine frontal overhead
radiograph from upper GI examination in 64-year-
old woman shows a large collection of contrast
material (arrows) in the lower abdomen in the
region of the jejunojejunal anastomosis. Leak from
the distal small-bowel anastomosis was con-
firmed at surgery. P ⫽ pouch.
acute respiratory distress syndrome
(n ⫽ 14) and acute renal failure (n ⫽ 5),
occurred in 14 patients (29%). Cardio-
pulmonary arrest occurred in four pa-
tients (8%) who subsequently survived;
death occurred in an additional three
patients (6%).
Discussion
Extraluminal leak is the most common
serious early complication of RYGBP
and bariatric surgery in general
(11,19,24,25,27–29). Extraluminal leak
was identified in 5.3% of patients in our
series. The incidence of leak in our
study included the occurrence of leak in
some patients who underwent revision
surgery and is similar to the incidence
rates reported in the literature, which
range from 1.0%–5.6% (1,5,20,21,27–
29). If not recognized early and treated
promptly, postoperative leak is a poten-
tially lethal complication of gastric by-
pass surgery (19,28). In our series,
125
GASTROINTESTINAL IMAGING: Extraluminal Leak after Roux-en-Y Gastric Bypass
death after extraluminal leak occurred
in 6% of the 48 patients with postoper-
ative leak and in 0.3% of all patients
who underwent RYGBP in the study pe-
riod. Therefore, early detection of post-
operative leak is critical for minimizing
morbidity and preventing mortality
(24,25,27).
Evidence has shown that radiologic
evaluation is important in the early de-
tection of leak in the postoperative
setting (25,29). Clinical symptoms are
often nonspecific, and physical exami-
nation is difficult owing to the large size
of the patient; therefore, routine early
postoperative upper GI examinations
are advocated for the detection of leaks
(20,25,26,29,30). The use of planned
early upper GI examinations may mini-
mize the morbidity caused by a leak af-
ter bariatric surgery (30). However, if
there is strong clinical evidence of a
leak, surgical exploration should not be
delayed until a confirmatory upper GI
study is performed.
The performance and interpretation
of upper GI studies after RYGBP re-
quires knowledge of the surgical proce-
dure, the normal appearance of the up-
per GI tract after RYGBP, the technical
aspects of upper GI studies, and the po-
tential pitfalls of study interpretation.
Figure 11
Figure 11: Extraluminal leak and SLL. (a) Initial supine frontal spot image from upper GI examination in
56-year-old woman shows extraluminal leak (arrows) and a collection of contrast material in the left upper
quadrant (*). (b) A supine right posterior oblique spot image obtained after the patient was rotated to the right
shows that the left upper quadrant collection (*) is located in the excluded gastric fundus. Contrast material
empties into the excluded antrum (A) and the duodenum (D); this finding is consistent with SLL. Continued
extraluminal leakage (arrows) is also identified.
126
We routinely evaluate patients within
1–2 days after RYGBP with upper GI
examinations that are performed ini-
tially with patients in the supine LPO
position. In our experience, this posi-
tion enables optimal visualization of the
proximal anastomosis in the majority of
patients. Other views are obtained as
necessary. Delayed abdominal radio-
graphs are obtained until contrast mate-
rial passes distal to the jejunojejunal
anastomosis because obstruction or
leak may occur at this site.
Radiologists must be aware of the
presence and pattern of leaks on upper
GI studies because postoperative leaks
have drastic implications for patient
morbidity and mortality. In this series,
extraluminal leak most often occurred
at the gastrojejunal anastomosis (77%);
however, leaks may also originate from
the gastric pouch, the distal esophagus,
the blind-ending jejunal limb, or even
the distal anastomosis. Leaks from the
distal anastomosis may be difficult to
recognize on upper GI studies and will
be missed unless overhead radiographs
are obtained until the point at which
contrast material has passed distal to
this site. In addition, leaks may very
rarely arise from the bypassed or ex-
cluded stomach, a finding that would
Carucci et al
not be diagnosed with upper GI studies
unless there was communication with
the excluded stomach (as in SLL).
Therefore, if there is a high clinical in-
dex of suspicion for a leak, surgical re-
exploration should be strongly consid-
ered.
We found that the majority (75%)
of leaks seen on upper GI studies ex-
tended to the left of the gastrojejunal
stoma; resultant left upper quadrant
collections were identified in 62% of all
patients with leak. Leaks are often asso-
ciated with intestinal obstruction or se-
vere ileus that may affect treatment
strategies. The presence or absence of
free air is not a reliable sign of a postop-
erative leak because it was seen on up-
per GI studies in only 50% of patients
and may have been related to the re-
cently performed surgical procedure.
As previously noted, the diagnosis
of postoperative leak in this patient pop-
ulation has important implications for
morbidity and mortality after RYGBP.
The majority of patients require repeat
surgery. In the present study, communi-
cation with the excluded stomach and
chronic cutaneous fistulas occurred as a
consequence of leak in 31% and 10% of
patients, respectively. Chronic SLL may
lead to failed weight loss and subse-
quent failure of the RYGBP procedure.
Extraluminal leak as a complication
of RYGBP has a tremendous effect on
both the patient and on health care re-
sources. Patients with postoperative
leak undergo multiple radiologic studies
(both diagnostic and therapeutic), hos-
pital admissions, and surgeries. In addi-
tion, a lengthy hospital stay is a common
consequence of leak. Whereas the aver-
age hospital stay after routine laparo-
scopic RYGBP is 2 days (18), the length
of the hospital stay increased to an aver-
age of 42.4 days in the setting of an
extraluminal leak.
This study was limited by its retro-
spective nature. In addition, patients
with leaks in this study were identified
through review of a radiology database
of upper GI series. It is therefore proba-
ble that the sickest patients, including
patients who died early in the postoper-
ative period, never underwent upper GI
imaging. Patients with the most severe
Radiology: Volume 238: Number 1—January 2006
GASTROINTESTINAL IMAGING: Extraluminal Leak after Roux-en-Y Gastric Bypass
leaks may therefore have not been in-
cluded in our study, and our calculated
mortality rate may therefore be slightly
low. Future prospective studies could
be performed to adequately address
these issues.
In conclusion, extraluminal leak is
the most common serious complication
of RYGBP surgery for morbid obesity,
resulting in an associated mortality rate
of 6% in our series of patients in whom
leak was diagnosed with upper GI stud-
ies. Although the frequency of leak after
RYGBP was low in this series (5.3%), it
is important for radiologists to recog-
nize the presence and pattern of leaks
on upper GI studies because postopera-
tive extraluminal leaks affect patient
morbidity and mortality.
References
1. Fisher BL, Schauer P. Medical and surgical
options in the treatment of severe obesity.
Am J Surg 2002;184:9S–16S.
2. Allison DB, Fontaine KR, Manson JE,
Stevens J, VanItallie TB. Annual deaths at-
tributable to obesity in the United States.
JAMA 1999;282:1530 –1538.
3. Martin LF, White S, Lindstrom W Jr. Cost-
benefit analysis for the treatment of severe
obesity. World J Surg 1998;22:1008 –1017.
4. Must A, Spadano J, Coakley EH, Field AE,
Colditz G, Dietz WH. The disease burden
associated with overweight and obesity.
JAMA 1999;282:1523–1529.
5. Livingston EH, Ko CY. Assessing the relative
contribution of individual risk factors on sur-
gical outcome for gastric bypass surgery: a
baseline probability analysis. J Surg Res
2002;105:48 –52.
6. Kuczmarski RJ, Flegal KM, Campbell SM,
Johnson CL. Increasing prevalence of over-
weight among US adults: the National Health
and Nutrition Examination Surveys, 1960 to
1991. JAMA 1994;272:205–211.
7. Mokdad AH, Serdula MK, Dietz WH, Bow-
man BA, Marks JS, Koplan JP. The spread of
Radiology: Volume 238: Number 1—January 2006
the obesity epidemic in the United States,
1991–1998. JAMA 1999;282:1519 –1522.
8. Talieh J, Kirgan D, Fisher BL. Gastric bypass
for morbid obesity: a standard surgical tech-
nique by consensus. Obes Surg 1997;7:198 –
202.
9. Gastrointestinal surgery for severe obesity:
National Institutes of Health Consensus De-
velopment Conference Statement. Am J Clin
Nutr 1992;55(2 suppl):615S– 619S.
10. Craig BM, Tseng DS. Cost-effectiveness of
gastric bypass for severe obesity. Am J Med
2002;113:491– 498.
11. Reddy RM, Riker A, Marra D, Thomas R,
Brems JJ. Open Roux-en-Y gastric bypass for
the morbidly obese in the era of laparoscopy.
Am J Surg 2002;184:611– 615.
12. Nguyen NT, Ho HS, Palmer LS, Wolfe BM. A
comparison study of laparoscopic versus
open gastric bypass for morbid obesity. J Am
Coll Surg 2000;191:149 –155.
13. Brolin RE. Update: NIH consensus confer-
ence— gastrointestinal surgery for severe
obesity. Nutrition 1996;12:403– 404.
14. Lonroth H, Dalenback J. Other laparoscopic
bariatric procedures. World J Surg 1998;22:
964 –968.
15. Rahal JP, Lamki N, Raval B, Grace M. Radio-
logical assessment of gastroplasty and gas-
tric bypass surgery. Radiology 1983;147:
274.
16. Sugerman HJ, Starkey JV, Birkenhauer R. A
randomized prospective trial of gastric by-
pass versus vertical banded gastroplasty for
morbid obesity and their effects on sweets
versus non-sweets eaters. Ann Surg 1987;
205:613– 624.
17. Sugerman HJ. Gastric bypass surgery for se-
vere obesity. Semin Laparosc Surg 2002;9:
79 – 85.
18. Higa KD, Boone KB, Ho T. Complications of
the laparoscopic Roux-en-Y gastric bypass:
1,040 patients—what have we learned?
Obes Surg 2000;10:509 –513.
19. Fobi MA, Lee H, Holness R, Cabinda D. Gas-
tric bypass operation for obesity. World
J Surg 1998;22:925–935.
Carucci et al
20. DeMaria EJ, Sugerman HJ, Kellum JM, Mea-
dor JG, Wolfe LG. Results of 281 consecu-
tive total laparoscopic Roux-en-Y gastric by-
passes to treat morbid obesity. Ann Surg
2002;235:640 – 645.
21. Schauer PR, Ikramuddin S, Gourash W, Ra-
manathan R, Luketich J. Outcomes after
laparoscopic Roux-en-Y gastric bypass for
morbid obesity. Ann Surg 2000;232:515–
529.
22. Blachar A, Federle MP. Gastrointestinal
complications of laparoscopic Roux-en-Y
gastric bypass surgery in patients who are
morbidly obese: findings on radiography and
CT. AJR Am J Roentgenol 2002;179:1437–
1442.
23. Blachar A, Federle MP, Pealer KM,
Ikramuddin S, Schauer PR. Gastrointestinal
complications of laparoscopic Roux-en-Y
gastric bypass surgery: clinical and imaging
findings. Radiology 2002;223:625– 632.
24. Koehler RE, Halverson JD. Radiographic ab-
normalities after gastric bypass. AJR Am J
Roentgenol 1982;138:267–270.
25. Moffat RE, Peltier GL, Jewell WR. The ra-
diological spectrum of gastric bypass compli-
cations. Radiology 1979;132:33–36.
26. Goodman P, Halpert RD. Radiological evalu-
ation of gastric stapling procedures for mor-
bid obesity. Crit Rev Diagn Imaging 1991;32:
37– 67.
27. Buckwalter JA, Herbst CA Jr. Complications
of gastric bypass for morbid obesity. Am J
Surg 1980;139:55– 60.
28. Buckwalter JA, Herbst CA Jr. Leaks occur-
ring after gastric bariatric operations. Sur-
gery 1988;103:156 –160.
29. Ganci-Cerrud G, Herrera MF. Role of radio-
logic contrast studies in the early postopera-
tive period after bariatric surgery. Obes
Surg 1999;9:532–534.
30. Serafini F, Anderson W, Ghassemi P, Pokle-
povic J, Murr MM. The utility of contrast
studies and drains in the management of pa-
tients after Roux-en-Y gastric bypass. Obes
Surg 2002;12:34 –38.
127