Roux-en-Y Gastric Bypass for Clinically Severe Obesity

Radiology
Elmar M. Merkle, MD
Peter T. Hallowell, MD
Cathleen Crouse, RN
Dean A. Nakamoto, MD
Thomas A. Stellato, MD
Published online before print
10.1148/radiol.2343030333
Radiology 2005; 234:674 – 683
Abbreviations:
BMI ⫽ body mass index
GI ⫽ gastrointestinal
RYGB ⫽ Roux-en-Y gastric bypass
1
From the Department of Radiology,
Duke University Medical Center, Erwin
Rd, Duke North, Rm 1417, Durham,
NC 27710 (E.M.M.); and Departments of Surgery (P.T.H., C.C., T.A.S.)
and Radiology (D.A.N.), University
Hospitals of Cleveland/Case Western
Reserve University, Cleveland, Ohio.
Received March 3, 2003; revision requested May 23; final revision received January 16, 2004; accepted
February 16. Address correspondence to E.M.M. (e-mail: elmar
.merkle@duke.edu).
Authors stated no financial relationship to disclose.
©
RSNA, 2005
Roux-en-Y Gastric Bypass for
Clinically Severe Obesity:
Normal Appearance and
Spectrum of Complications at
Imaging1
Surgery currently appears to be the most effective method to curtail the effects of
morbid obesity and all of its comorbid conditions. Although the ideal procedure has
yet to be devised, Roux-en-Y gastric bypass has proved to be successful for many
morbidly obese patients pursuing weight loss and increased health. As the technical
aspects of this procedure become less cumbersome and the patient population
increases, it is vital for radiologists to be proficient in the specific evaluation of these
patients, in order to provide optimal care. Complications can be minimized, managed more efficiently, or prevented with prompt evaluation by the radiologist. It is
important to appreciate the patency of both the gastrojejunostomy and the jejunojejunostomy, as well as adequate progression of contrast material before the
patient is discharged (preferably 24 –72 hours after surgery). Follow-up complications include anastomotic leak, staple-line disruption, stomal stenosis, occlusion of
the Roux limb, small-bowel obstruction due to adhesions or internal hernia, and
obstruction of the enteroenterostomy leading to acute gastric distention. These
complications may be life threatening, since clinical symptoms are often inconclusive. To achieve optimal outcome, therefore, conventional radiographic and computed tomographic studies should not be delayed.
©
RSNA, 2005
The World Health Organization has recognized a “global epidemic of obesity,” which is
one of the most alarming health problems not only in the United States but also in the
whole Western world (1,2). Despite extended educational measures and increased public
awareness, the number of obese individuals continues to increase in industrialized countries. Obesity contributes to at least 300 000 deaths annually in the United States, and the
health care costs of obese American adults amount to approximately $100 billion, which
equals approximately 1% of the U.S. gross domestic product. A recent report in the Journal
of the American Medical Association (3) indicated that the prevalence of adult obesity
increased from 22.9% in the period 1988 –1994 to 30.5% in 1999 –2000. Data reflecting the
same trend are reported from Europe, as well: The prevalence of obesity in Denmark has
increased in men from 10% in 1982 to 13% in 1992 and from 9% to 11% in women during
the same period (4).
The body mass index (BMI) is currently used as a mathematic calculation to determine
if a patient is overweight. The BMI is calculated by dividing a person’s body weight in
kilograms by their height in meters squared. While the BMI is currently widely accepted as
the best way to determine obesity, this number can be misleading in very muscular people
or in pregnant or lactating women.
The risk for morbidity and mortality accompanying obesity is proportional to the degree
of excess weight. By definition, adults with a BMI in the range of 18.5–24.9 kg/m2 are
considered to be of normal weight, while those with a BMI of 25 kg/m2 or greater are
considered to be overweight. Adults with a BMI of 30 kg/m2 or greater are classified as
obese class I; those with a BMI 35 kg/m2 or greater, as obese class II; and those with a BMI
40 kg/m2 or greater, as obese class III or extremely obese. The Table shows the continuous
674
Radiology
ESSENTIALS
●
The World Health Organization has
recognized a global epidemic of obesity.
●
The majority of American bariatric surgeons prefer a combination of restrictive and malabsorption procedures
such as the Roux-en-Y gastric bypass.
●
Complications after Roux-en-Y gastric
bypass surgery can be minimized,
managed more efficiently, or even prevented with prompt evaluation by the
radiologist.
increase in the prevalence of overweight,
obese, and severely obese American
adults over the past 25 years. Alarmingly,
the increase in this trend is most pronounced in the younger age groups.
Obesity is associated with more than 30
medical conditions, including sleep apnea,
hypertension, coronary vascular disease,
type 2 diabetes mellitus, degenerative joint
and disk disease, and some types of cancer
(eg, breast cancer, colorectal cancer) (7).
For persons who are overweight or obese,
weight loss of as little as 10% can improve
some obesity-related conditions, including
diabetes and hypertension. Despite important advances in the pharmacologic treatment of obesity, nonsurgical treatment
approaches such as diets, behavioral modification, exercise, and pharmacologic
agents rarely achieve a greater than 10%
weight loss in the majority of obese patients. Low-calorie diets alone have
been unsuccessful in achieving permanent
weight loss. Therefore, in 1991 the National Institutes of Health Consensus Development Conference on Gastrointestinal
Surgery for Severe Obesity stated that gastric restrictive or bypass procedures should
be considered (a) for patients with a BMI of
greater than 35 kg/m2 with high-risk comorbidities such as life-threatening cardiopulmonary problems or severe diabetes
mellitus and (b) for those with a BMI of
greater than 40 kg/m2 (8).
Patients who may be candidates for
surgical treatment must demonstrate failure of established nonsurgical weight control programs and must be well informed,
motivated, and able to participate in treatment and long-term follow-up.
Teams consisting of abdominal surgeons, psychologists, registered dieticians,
Volume 234
䡠
Number 3
and registered nurses have been formed
throughout the Western world to meet the
rising demand for bariatric surgery. Radiologists also play an important role on these
teams, because morbidly obese patients are
already considered to be high-risk owing to
their comorbidities, and clinical examination in these patients may be difficult.
Therefore, radiologists must be aware of
the common surgical procedures for obesity, the normal imaging appearance after
such procedures, and the spectrum of possible complications.
SURGICAL TECHNIQUES
Surgical procedures for weight reduction
can be divided into two major categories:
gastric restriction and intestinal malabsorption (7). Restrictive bariatric procedures such as vertical gastric banding and
adjustable gastric banding involve the
creation of a small neogastric pouch and
gastric outlet to decrease food intake.
Malabsorption procedures such as jejunoilial bypass, biliopancreatic diversion,
or duodenal switch involve a rearrangement of the small intestine to decrease
the functional length or efficiency of the
intestinal mucosa for nutrient absorption. Roux-en-Y gastric bypass (RYGB) is
a combination of both techniques: A
small gastric pouch with a small outlet is
created, and the functional length of the
small bowel is decreased.
Data from the 1989 and 1998 membership rosters of the American Society for
Bariatric Surgery indicate that 82% of the
North American surgeons prefer a combination restrictive-malabsorption procedure such as RYGB, compared with 15%
who prefer purely restrictive procedures
such as vertical banded gastroplasty, silicone ring gastroplasty, or adjustable
gastric banding (9). European surgeons,
however, seem to prefer purely restrictive procedures over RYGB. Each procedure has its own assets and drawbacks.
While weight loss is, in our opinion,
more persistent with RYGB, the major
drawback of the procedure is its irreversibility. Gastric banding procedures,
on the other hand, are reversible, although in our opinion the weight loss
results are not as consistent.
It is mandatory for the radiologist involved in bariatric surgery cases to be familiar with the surgical technique used
by referring surgeons, since the various
procedures have a different appearance
on imaging studies and a different spectrum of complications. This review will
deal only with the imaging appearance of
RYGB and its complications.
RYGB was first introduced by Mason
and Ito (10) in 1967 for the treatment of
morbid obesity. This surgical procedure
has had many modifications since its original inception (11). The hallmarks of the
procedure are as follows (Fig 1): (a) creation
of a small gastric reservoir of 15–30 mL,
which may be stapled in continuity or divided; (b) creation of a small (⬍12-mmdiameter) gastrojejunostomy that is either
stapled, sutured, or both; and (c) division
of the small bowel distal to the ligament of
Treitz (suspensory muscle of duodenum).
The distal segment is brought up to the
pouch to form the Roux limb. This distal
segment can be brought up “antecolic” (ie,
anterior to the transverse colon) or “retrocolic” (ie, posterior to the transverse colon)
through an incision in the transverse mesocolon. A short-limb RYGB is less than
100 cm in length and a long-limb bypass is
greater than 100 cm. These lengths vary
according to surgeon preference and patient size. The Roux limb and biliopancreatic limb (ie, bypassed or defunctionalized
stomach, native duodenum, and variable
length of in-contiguity proximal jejunum— usually 75 cm) are connected by either a stapled or a sutured jejunojejunostomy. These steps can be accomplished
both laparoscopically and during laparotomy.
PREOPERATIVE IMAGING
BEFORE RYGB
Other than chest radiographs in two views
and an echocardiogram to assess cardiopulmonary function, only one additional
imaging study is widely recommended in
the preoperative work-up for bariatric surgery: An ultrasonographic (US) study of the
right upper quadrant should be performed
to rule out cholelithiasis (12,13). In proved
cases of cholelithiasis, most surgeons remove the gallbladder at the time of bariatric surgery, because gallstone formation
increases during rapid weight loss and endoscopic treatment of choledocholithiasis
is technically extremely challenging after
RYGB.
At present, there is no consensus on
whether an upper gastrointestinal (GI)
radiographic study should be performed
to assess esophageal motility, gastric
emptying, the size of a hiatal hernia if
present, and the anatomy in patients
who have previously undergone gastric
surgery. While some bariatric surgeons
do not believe routine preoperative radiographic studies are necessary, others
consider radiographic studies to be a routine component of the preoperative
Roux-en-Y Gastric Bypass for Severe Obesity
䡠
675
Radiology
Prevalence of Excess Weight among American Adults over Past 25 Years
Years
Overweight (%)
Obese (%)
Severely Obese (%)
1999–2000
1988–1994
1976–1980
64.5
56.0
46.0
30.5
23.0
14.4
4.7
2.9
No data
Sources.—References 3, 5, 6.
Note.—Overweight is BMI ⱖ 25 kg/m2; obese is BMI ⱖ 30 kg/m2; severely obese is BMI ⱖ 40
kg/m2.
Figure 2. Upper GI image obtained 1 day after laparoscopic RYGB with megapouch in a
58-year-old woman shows gastric pouch (ⴱ)
that is explicitly larger than a lumbar vertebral
body.
Figure 1. (a) Schematic representation of RYGB with (b) corresponding appearance at upper GI study obtained with water-soluble
contrast material. 1 ⫽ Gastric pouch; 2 ⫽ staple line; 3 ⫽ upper
side-to-side anastomosis between gastric pouch and Roux limb; 4 ⫽
defunctionalized stomach; 5 ⫽ duodenum; 6 ⫽ pancreaticobiliary
limb consisting of the second, third, and fourth portion of the duodenum and variable length of in-contiguity jejunum; 7 ⫽ blind loop,
8 ⫽ Roux limb; 9 ⫽ side-to-side anastomosis between pancreaticobiliary limb and Roux limb.
work-up. In the recent literature, we are
aware of only one article that addresses
this issue: Frigg et al (14) reported their
preoperative radiologic and endoscopic
evaluation performed prior to laparoscopic adjustable gastric banding. Of 147
upper GI series, 74 (50%) showed hiatal
hernias, two (1%) showed motility disorders of the esophagus, and one (⬍1%)
showed incomplete malrotation. In fewer
than 50% of the patients, no pathologic
condition was found. Their results seem
to support the need for a preoperative GI
radiographic study, particularly if a laparoscopic approach is planned, because
anatomic orientation may be confusing
when viewed through a laparoscope (14).
An active peptic ulcer is an absolute contraindication for bariatric surgery (13). Un676
䡠
Radiology
䡠
March 2005
fortunately, preoperative upper GI studies
with contrast material have demonstrated
a high false-negative rate for the diagnosis
of peptic ulcer. Frigg et al (14), using gastroscopy, found gastritis or duodenitis in
10 of 104 patients, an ulcer in one patient,
and erosions in two patients. These findings were not identified on upper GI series.
Many institutions currently screen patients
for Helicobacter pylori (13). If patients test
positive for H pylori, a 2-week course of
antibiotics and histamine-2 receptor blockers is prescribed.
NORMAL IMAGING
APPEARANCE AFTER RYGB
Patients in our bariatric program usually
undergo an upper GI study within the
Figure 3. Upper GI image obtained 1 day after laparoscopic RYGB in a 32-year-old woman
shows gastric pouch (black arrow) that is explicitly smaller than a lumbar vertebral body.
Defunctionalized stomach and duodenum are
not opacified with contrast material, although
stomach is outlined with residual air in the
fundus (ⴱ) and antrum (white arrows). The
blind loop (arrowhead) may be misconstrued
as the gastric pouch.
first 48 hours after RYGB to rule out an
anastomotic leak (Fig 1). After scout image acquisition, the patient is asked to
swallow a small sip of a water-soluble
contrast agent, which may remain in the
distal esophagus and gastric pouch for a
variable period (seconds to minutes) before passing the side-to-side anastomosis
and filling the blind loop and Roux limb
(ie, enteric, or efferent, limb). Delayed
pouch emptying may be misconstrued as
an anastomotic stenosis, but in the majority of cases it simply reflects a swollen
anastomosis after surgery.
Merkle et al
Radiology
Figure 4. Images obtained 2 days after RYGB in a 52-year-old woman. (a) Scout radiograph shows staple line (arrow) and
nasogastric tube. (b) Upper GI image shows nasograstric tube in place. (c) On upper GI image obtained after removal of the
tube, an anastomotic leak (arrow) is evident.
A gastric pouch volume of approximately 30 mL and a 12-mm-diameter
outlet stoma exhibit the best results. On
conventional radiographs, it is hardly
possible for the radiologist to quantify
the volume of the gastric pouch. As a rule
of thumb, however, the gastric pouch
should be of a size similar to that of a
lower thoracic or lumbar vertebral body
(Fig 1). This comparison allows fast and
reliable detection of a pouch that is either
too large or extremely small (Figs 2, 3).
The distal side-to-side anastomosis is
difficult to visualize because contrast material rarely refluxes from the jejunum into
the pancreaticobiliary limb (15). Similarly,
the duodenum and the defunctionalized
stomach are difficult to visualize unless
they are outlined with gas (Fig 3).
In cases of open RYGB surgery, a nasogastric tube is often placed with its distal
end in the Roux limb to temporarily provide a “stent” for the side-to-side anastomosis. In these cases, it is mandatory to
repeat the upper GI study after removal
of the nasogastric tube, because the tube
may obscure an anastomotic leak (Fig 4).
In our current practice, we do not use a
guidewire to maintain easy access to the
jejunum. However, in patients with a
high clinical suspicion for an anastomotic leak (that has not been detected
prior to removal of the nasogastric tube),
Volume 234
䡠
Number 3
the insertion of a guidewire may be helpful before the nasogastric tube is removed.
This approach may avoid potential injury
in these immediately postoperative patients if an anastomotic leak is detected
after removal of the nasogastric tube, necessitating reinsertion of the nasogastric
tube.
In technical terms, examinations of severely obese patients are difficult because
of patient size. Technical problems include difficulty in positioning of patients
for optimal radiographs, inability to place
the image intensifier over the patient,
extreme difficulty during fluoroscopy in
depicting intraabdominal structures, and
suboptimal radiographs caused by markedly scattered radiation (12). If water-soluble contrast material–enhanced radiographs
are inadequate, a repeat study with barium
contrast material may be helpful.
Helical computed tomography (CT) offers a detailed view of the anatomy after
RYGB, with all important structures clearly
depicted (Fig 5). Currently, no data are
available regarding the optimal administration of oral contrast agents. After approval by the referring physician, we
administer a 2% dilution of diatrizoate
meglumine (Gastrografin; Bracco Diagnostics, Princeton, NJ) 1 hour prior to the CT
scan, encouraging the patient to drink as
much of the solution as possible immedi-
ately prior to the CT examination. Since an
anastomotic leak is often a differential consideration, we avoid using barium as an
oral contrast agent and do not use effervescent crystals.
The gastric pouch may be found by
following the course of the distal esophagus. It is usually positioned right laterally to the right of the proximal anastomosis. A normal gastric pouch is usually
collapsed at the time of CT. The blind
loop and the Roux limb are usually seen, as
well. Both structures should not be larger
than 2.5 cm in diameter. Whereas the
blind loop ends after several centimeters
(depending on the surgeon’s technique),
the Roux limb can usually be followed either retrocolically or antecolically to the
distal anastomosis with the pancreaticobiliary limb. The defunctionalized stomach
may contain a small amount of air and
fluid, but it should never be distended. The
native duodenum and the proximal jejunum (part of the pancreaticobiliary limb)
can also be followed to the lower anastomosis. These bowel loops should not be
larger than 2.5 cm in diameter.
Despite the substantial advantages of
CT over upper GI series, the routine use
of CT in the immediate postoperative
course after RYGB is not recommended
because of the increased radiation exposure to the patient and the increased cost;
Roux-en-Y Gastric Bypass for Severe Obesity
䡠
677
Radiology
ceed the size and weight limits of CT systems. Our current CT system (MX8000;
Philips Medical Systems, Best, the Netherlands) has a manufacturer-recommended
table weight limit of 157.5 kg. A rigid table
weight limit, however, falls short of addressing the problem appropriately, since
the patient’s weight also needs to be spread
over the CT table. It is obvious that a patient with a height of 1.6 m and a weight of
160 kg will strain the table more than will
a 1.9-m-tall patient with the identical
weight. A reasonable approach for the future may be to encourage the manufacturers to shift from a rigid table weight limit to
a BMI limit, which would take patient’s
height and weight into account. The CT
gantry may also be the limiting factor; our
current CT system has a maximum size of
68 cm in the right-to-left dimension and
56 cm in the anteroposterior dimension.
Patient weight also has a profound effect on the magnitude of intravenous
contrast material enhancement both in
the vascular system and in parenchymal
organs such as the liver. Whereas the total dose of iodine (as determined by the
volume and concentration of contrast
material) correlates directly with the magnitude of hepatic enhancement, patient
weight is inversely related (16).
GENERAL AND SPECIFIC
COMPLICATIONS AFTER RYGB
Figure 5. Transverse contrast material– enhanced CT scans in a 51-year-old woman 3 months
after RYGB. Patient presented with crampy abdominal pain after eating, but CT findings are
unremarkable. (a) CT scan through upper abdomen shows staple line (white arrowhead) between
gastric pouch (black arrowhead) and defunctionalized stomach (ⴱ). Side-to-side anastomosis
(black arrow) between Roux limb (white arrow) and gastric pouch is positioned anterior to the staple
line (white arrowhead). Of note is a small air bubble in the defunctionalized stomach; this is a normal
finding after RYGB. Curved arrow ⫽ blind loop. (b) CT scan obtained 2 cm caudal to a shows Roux
limb (straight arrow), end of the blind loop (curved arrow), and defunctionalized stomach (ⴱ). (c) CT
scan at level of the transverse mesocolon shows retrocolic Roux limb (arrow), transverse colon, and
lower side-to-side anastomosis (arrowhead). (d) CT scan obtained 3 cm caudal to c shows decompressed horizontal duodenum (ⴱ), which is part of the pancreaticobiliary limb.
therefore, we believe helical CT should be
used only in ambiguous cases where conventional radiographic studies are indeterminate—for example, as a follow-up
study after a small anastomotic leak.
EQUIPMENT-RELATED
OBSTACLES IN BARIATRIC
PATIENT IMAGING
Our GI radiology suite is currently
equipped with Fluorex Efficiency 450D
systems (Toshiba Medical Systems, Tustin,
Calif). These systems have a weight limit of
150 kg. If bariatric patients exceed this
limit, they are either examined in a stand678
䡠
Radiology
䡠
March 2005
ing position or moved to our angiography
suite, which is equipped with an Axiom
Artis system (Siemens Medical Systems,
Malvern, Pa), which has a table weight
limit of 200 kg.
As mentioned earlier, the distance between the fluoroscopic table and the image intensifier is also of concern. While
our Toshiba systems in the GI radiology
suite allow a maximum patient size (in
the anteroposterior dimension) of 47 cm,
our Siemens angiography system offers a
maximum distance of 60 cm between the
fluoroscopic table and the image intensifier.
Patient dimensions also frequently ex-
Severely obese patients are already considered to be at high risk because of their
comorbidities. Because most of the surgical procedures for obesity are completely
elective, prospective patients need to be
aware of the possible complications, which
can be divided into general and specific
categories.
General Complications
Pulmonary embolism.—Pulmonary embolism is the leading cause of perioperative death in patients who undergo bariatric surgery, but it can occur as many as
6 months after the procedure (17). The
overall incidence of deep vein thrombosis and pulmonary embolism is reported
to be less than 2% (9,11,13,17); however,
nearly one-third of the bariatric surgery
patients who have pulmonary embolism
will die.
Several factors increase the possibility
of deep vein thrombosis and/or pulmonary embolism in these patients. Morbidly obese patients are less active and
may be confined to a bed or chair. Venous stasis disease and polycythemia due
to underlying respiratory insufficiency
Merkle et al
Radiology
Figure 6. Transverse contrast-enhanced CT
scan obtained 20 days after revision of previous RYGB in a 40-year-old-woman shows large
fluid collection in the left upper quadrant (arrowheads). Note also small seroma (arrow) in
the subcutaneous fatty layer, which had resolved at follow-up imaging (not shown). CTguided biopsy revealed an abscess, which was
subsequently drained.
are also common in these patients. The
combination of some or all of these factors, in addition to the patient undergoing an abdominal surgical procedure in
the supine position for several hours,
leads to the development of deep vein
thrombosis and pulmonary embolism
(9). Various prophylactic methods have
been employed to prevent deep vein
thrombosis and pulmonary embolism,
including subcutaneous low-dose anticoagulant injections, pneumatic compression stockings, elastic stockings or bandages, intravenous low-molecular-weight
dextran injections, and use of the Trendelenburg position during surgery (13).
Unfortunately, none of these methods
has been proved to decrease the incidence of postoperative venous thromboembolism in bariatric patients. In our
opinion, early postoperative ambulation
may be the most important method to
prevent postoperative deep vein thrombosis and pulmonary embolism.
The diagnosis of deep vein thrombosis
and pulmonary embolism at clinical examination can be difficult. In our opinion, color Doppler US is the imaging modality of choice for the work-up of
bariatric surgery patients suspected of
having lower-extremity deep vein thrombosis, and contrast-enhanced helical CT
is the imaging modality of choice for the
work-up of those suspected of having
pulmonary embolism. If pulmonary CT
cannot be performed because of weight
Volume 234
䡠
Number 3
Figure 7. Transverse unenhanced CT scans obtained 3 days after RYGB in a 51-year-old woman
with stenosis of the enteroenteric anastomosis leading to acute gastric distention. (a) Staple line
(arrow), blind loop (ⴱ), and Roux limb (arrowhead) are shown. Note also marked distention of
fundus of the defunctionalized stomach. (b) Antegastric Roux limb (arrowhead) and distention of
defunctionalized stomach and duodenal bulb (ⴱ) as part of the pancreatobiliary limb are shown.
(c) Descending part of the duodenum (ⴱ) and proximal jejunum (straight arrow) as parts of the
pancreatobiliary limb, both markedly dilated, are shown. Note also side-to-side enteroenteric
anastomosis (curved arrows) and Roux limb (arrowhead). (d) The horizontal part of the duodenum (ⴱ) as part of the pancreatobiliary limb is markedly dilated.
constraints, ventilation-perfusion scintigraphy may serve as an alternative imaging modality. To the best of our knowledge, no studies have been performed to
compare the diagnostic utility of helical
pulmonary CT and ventilation-perfusion
scintigraphy in bariatric surgery patients
suspected of having pulmonary embolism.
Incisional hernia.—One of the major advantages of laparoscopic procedures for
obesity is the lack of associated incisional
hernias, which occur in approximately
5% of cases after open RYGB surgery (11).
In general, the diagnosis of an incisional
hernia at clinical examination is easy,
and no additional imaging studies are required.
Gallstone formation.—Gallstone formation is increased during rapid weight loss
caused by either dietary methods or bariatric surgery. Because most surgeons remove the gallbladder during bariatric surgery in proved cases of cholelithiasis, the
incidence of cholelithiasis is approximately 1%–3% after RYGB (11,17,18). US
is the imaging modality of choice for patients suspected of having cholelithiasis.
Since endoscopic retrograde cholangiography is technically challenging after
RYGB, magnetic resonance (MR) cholangiography may be performed in patients
suspected of having choledocholithiasis. However, as mentioned earlier, the
weight limit of the MR imager table may
be an insurmountable obstacle that could
prevent bariatric surgery patients from
undergoing MR cholangiography.
Major wound infection and seroma.—Major wound infection after RYGB is seen in
Roux-en-Y Gastric Bypass for Severe Obesity
䡠
679
Radiology
examinations on our interventional CT
scanner in stand-by mode, which means
that we perform CT-guided aspiration or
drainage in the same imaging session if a
suspicious abdominal fluid collection has
been detected.
Procedure-specific Complications
Figure 8. Upper GI studies obtained 18 months after RYGB in a
53-year-old woman in whom gastric staple-line disruption occurred.
After initial weight loss, the patient lost the sensation of early satiety
and regained weight. (a) Gastric pouch (white arrows), Roux limb (arrowhead), previously excluded stomach (ⴱ), and disrupted staple line
(black arrow) are visible. (b) Subsequent image obtained within minutes
of a shows patency of the Roux limb (arrowheads). Note rugal folds in
previously excluded stomach (ⴱ). White arrow ⫽ gastric pouch.
Figure 9. (a– c) Images from upper GI series obtained 5 weeks after RYGB in a 44-year-old
woman who presented with multiple episodes of nausea and vomiting show distention of gastric
pouch (ⴱ) and wedge-shaped tight anastomosis (arrow) that never changes shape or fully distends. Patient subsequently underwent successful endoscopic dilation.
up to 10% of patients, while seromas in
the subcutaneous fatty layer are more
common, with an incidence of up to 40%
(13). Although diagnosis of a superficial
wound infection at clinical examination
is easy, CT is often required to exclude
the presence of intraabdominal abscess
in patients with fever and wound infection.
Abdominal fluid collection, subphrenic
abscess, and peritonitis.—These conditions
occur in fewer than 2% of all bariatric
surgery patients. They are considered to
be extremely serious complications and
may be life threatening in morbidly
680
䡠
Radiology
䡠
March 2005
obese patients (9,18). Clinical signs may
be subtle, and diagnosis is often difficult. CT has a critical role in the
work-up of bariatric surgery patients
suspected of having an abdominal fluid
collection, and it should be performed
without hesitation since failure to immediately diagnose and treat an abdominal catastrophe almost always
leads to severe systemic sepsis, multiple
organ failure, and death (9). This is even
more true for patients with a history of
bariatric surgery because the overall complication rate in these patients is markedly
higher (Fig 6). We usually perform these
Anastomotic or staple-line leak.—The incidence of anastomotic or staple-line
leaks is reported to be in the range of
1%– 4% for patients who have undergone
RYGB (Fig 4) (11,13,17–19). Early postoperative leaks are sometimes difficult to
recognize because fever and abdominal
tenderness are frequently present due to
respiratory insufficiency, atelectasis, and
postsurgical abdominal pain. Early diagnosis is important, however, owing to
the resultant peritonitis and high mortality. Left shoulder pain, tachycardia, and
anxiety are early symptoms. Upper GI
studies have an important role in the
identification of these leaks. Oral contrast material extravasation is usually
seen without difficulty. However, it may
be challenging to detect the specific site
of the leak, since image quality may be
limited owing to the patient’s body habitus. Notably, the incidence of anastomotic leaks is five to 10 times higher after
revision procedures (11,20).
Acute gastric distention.—Afferent gastric distention (or afferent or closed-loop
obstruction) occasionally occurs after
open or laparoscopic RYGB because of
obstruction or edema at the enteroenterostomy site (9). The distention may be
severe enough to cause a staple-line disruption or, if the Roux limb is brought
antegastric, put enough tension on the
gastrojejunostomy to cause a leak. Patients usually report a bloated feeling and
may often hiccup. Abdominal radiographic studies may demonstrate a distended excluded stomach and often
show an air-fluid level. In our experience,
CT is the imaging modality of choice for
patients who may have obstruction of
the enteroenterostomy, because all important structures (eg, the defunctionalized stomach, duodenum, pancreaticobiliary limb, and lower side-to-side
anastomosis) are well depicted (Fig 7).
Typical imaging findings include severe
distention of the pancreaticobiliary limb
adjacent to the lower anastomosis and
variable distention of the duodenum and
defunctionalized stomach. Gastric and
duodenal distention will become progressively worse over time because pancreaticobiliary fluid will accumulate. Percutaneous fine-needle decompression may be
Merkle et al
Radiology
temporarily successful; however, if distention recurs, emergent surgery may be necessary.
Staple-line disruption.—Staple-line disruption is a complication that can occur in
both divided and undivided RYGB procedures. Many cases may be subclinical, and
patients may not lose the sensation of early
satiety. Surgical correction involves restapling and dividing the stomach. In our
opinion, an upper GI series is the imaging
modality of choice in cases suspicious for
staple-line disruption (Fig 8). The primary
imaging finding is detection of oral contrast material that outlines rugal folds in
the defunctionalized stomach. As mentioned earlier, it may be challenging to detect the exact location of the staple-line
disruption, because image quality may be
limited owing to the patient’s body habitus.
Stomal stenosis.—Stomal stenosis is another complication of RYGB (18) (Fig 9).
The gastrojejunal stoma is created to be
approximately 12 mm in diameter. Patients with stenosis of the gastrojejunal
stoma typically present with postprandial epigastric pain and vomiting.
While diagnosis and treatment are usually accomplished by means of endoscopy, fluoroscopically guided balloon
dilation may serve as an alternative treatment modality. Vance et al (21) performed 41 fluoroscopically guided balloon dilations in 28 patients with
symptoms of gastric outlet obstruction
after gastric restrictive surgery for morbid
obesity. After initial fluoroscopically
guided balloon dilation, 11 (39%) of 28
patients remained asymptomatic during
at least 3 months of follow-up. Recurrent
dysphagia occurred in 17 (61%) patients,
who were treated either surgically (n ⫽
3), with endoscopically guided balloon
dilation (n ⫽ 4), with fluoroscopically
guided balloon dilation (n ⫽ 9), or medically (n ⫽ 1). Of the nine patients who
underwent a second fluoroscopically
guided balloon dilation, two became
asymptomatic. The seven patients who
remained symptomatic were treated with
surgery (n ⫽ 3), endoscopically guided
balloon dilation (n ⫽ 1), or fluoroscopically guided balloon dilation (n ⫽ 3). Of
the three patients treated with a third
fluoroscopically guided balloon dilation,
one became asymptomatic. Overall, fluoroscopically guided balloon dilation was
effective in 14 (50%) of 28 patients.
Stomal ulceration.—Patients with stomal
ulceration usually present with severe
dyspepsia, burning retrosternal pain, and
vomiting. Diagnosis is usually made with
the aid of endoscopy, because detection
Volume 234
䡠
Number 3
Figure 10. Transverse unenhanced CT scans obtained 19 months after RYGB in a 40-year-old
woman with internal hernia through the transverse mesocolon. (a) Scan at level of the stomach
shows massive gastric distention. Note normal caliber of blind loop (straight arrow) and Roux
limb (arrowhead). Curved arrow ⫽ splenic colonic flexure. (b) Scan above level of the transverse
mesocolon shows massive dilatation of duodenum and distention of retrocolic jejunum (large
black ⴱ) as parts of the pancreaticobiliary limb. Roux limb (arrowhead) is compressed. Note
proximal transverse colon (straight arrows), dilated jejunal loop posterior to transverse colon
(white ⴱ), distal transverse colon (small black ⴱ), and descending colon (curved arrow). (c) Scan
at level of the transverse mesocolon shows compression of jejunum (straight arrows) as part of the
pancreaticobiliary limb. Roux limb (small arrowhead) is barely visible. Note also dilated jejunal
loop (white ⴱ), hepatic colonic flexure (large arrowhead), distal transverse colon (black ⴱ), and
descending colon (curved arrow). (d) Scan below level of the transverse mesocolon shows dilated
jejunum (large white ⴱ) as part of the pancreaticobiliary limb. Note also dilated jejunal loop (small
white ⴱ), ascending colon (arrowhead), distal transverse colon (black ⴱ), and descending colon
(curved arrow).
of stomal ulceration may be difficult on
upper GI radiographic studies and nearly
impossible on CT scans (1). Treatment is
with proton pump inhibitors and antibiotics if the patient is positive for H pylori.
Small-bowel obstruction.—This complication occurs within the first several
postoperative weeks in 1%–5% of all patients (13,18,22) and is usually caused by
postoperative adhesions after open RYGB.
While the prevalence of these adhesions
has been reduced by using a laparoscopic
approach, the incidence of internal hernias
has increased—a complication that appears to be more common with a laparoscopic approach. Patients may present
with nonspecific symptoms of crampy abdominal pain. Vomiting is frequently ab-
sent, because few gastric or enteral secretions accumulate from the gastric pouch or
Roux limb. Internal hernias can develop in
the transverse mesocolon, where the limb
passes through (most common type) or
around (Petersen hernia) the mesentery of
the Roux limb, or in the mesenteric defect
at the enteroenterostomy site (15,18,23).
The herniated bowel is usually the
Roux limb or the pancreaticobiliary limb
itself (Figs 10, 11). It is difficult to distinguish small-bowel obstruction caused by
adhesions from that caused by internal
hernia on the basis of findings from CT,
upper GI series, or both. Blachar et al (18)
reported on 15 patients with bowel obstruction following laparoscopic RYGB
surgery. Nine patients had an internal
Roux-en-Y Gastric Bypass for Severe Obesity
䡠
681
Radiology
Figure 11. Images obtained 2 months after RYGB in a 58-year-old woman with internal hernia through a small-bowel
mesenteric defect at jejunojejunostomy site. (a) Upper GI image shows normal-sized small gastric pouch (straight arrow) and
distended blind loop (ⴱ) and Roux limb (arrowheads). (b) Transverse contrast-enhanced CT scan obtained 2 cm caudal to
gastrojejunostomy shows collapsed defunctionalized stomach (ⴱ), dilated blind loop (straight arrow), and dilated Roux limb
(arrowhead). (c) Transverse contrast-enhanced CT scan at level of the jejunojenunostomy shows dilated Roux limb (large
arrowheads) and hernia (straight arrow) at jejunojejunostomy. Note also decompressed ileal loop (small arrowheads) and
transverse and descending colon (curved arrows).
Figure 12. Radiographs in a 22-year-old woman who presented 4 days after RYGB with nausea and vomiting caused by
small-bowel obstruction at level of the transverse mesocolon due to tight mesocolic incision. (a) Unremarkable scout
radiograph shows staple line and nasogastric tube. (b) Contrast-enhanced image from upper GI series demonstrates filling of
pouch (arrow), blind loop (arrowhead), and Roux limb (ⴱ). (c) Delayed image acquired 2 days after contrast material ingestion
shows marked dilatation of pouch (arrow), blind loop (arrowhead), and Roux limb (ⴱ), with no contrast material passing the
level of the transverse mesocolon. Surgery revealed small-bowel obstruction at level of the transverse mesocolon due to tight
mesocolic incision.
hernia, while the remaining six had bowel
obstruction caused by adhesions. At upper
GI series, six (67%) of nine internal hernias
and two (33%) of six adhesions were correctly diagnosed (18). Two internal hernias
were misdiagnosed as adhesions, and one
adhesion was misdiagnosed as an internal
hernia. The remaining four cases were considered equivocal. An abnormal cluster of
dilated bowel segments in the left upper
quadrant or midline was the best imaging
finding to help distinguish between internal hernia and adhesions. This specific
finding was present in 89% of patients
682
䡠
Radiology
䡠
March 2005
with internal hernia and in 33% of those
with adhesions.
At present, knowledge regarding the
CT appearance of internal hernia following
RYGB is limited. Blachar et al (18) reported
the largest series of which we are aware—a
total of three cases: one through a rent in
the transverse mesocolon, one through a
rent in the small-bowel mesentery, and
one Petersen type. In the same series, one
case of small-bowel obstruction due to adhesions was identified. While distended
small-bowel segments (⬎2.5-cm diameter)
with a transition point from dilated to
nondilated small bowel were seen in all
four cases, the mesenteric vessels were engorged and stretched only in the cases of
internal hernia. In that series, the Petersen
hernia was misdiagnosed at CT as adhesions, while the other two internal hernias
were correctly diagnosed. Regardless of the
underlying cause of the small-bowel obstruction, expedient surgical treatment is
mandatory.
Occlusion of Roux limb.—Another type of
small-bowel obstruction is the occlusion
of the Roux limb at the level of the transverse mesocolon, because in RYGB most
Merkle et al
Radiology
of the Roux limbs are positioned posterior to the transverse colon (Fig 5). Occlusion of the Roux limb occurs in 1% of
patients (17) and is easily diagnosed on
an upper GI series (Fig 12).
INTERVENTIONAL
MANAGEMENT OF
COMPLICATIONS
A number of the possible complications
in bariatric surgery patients can be successfully managed with interventional
radiologic techniques. This spectrum
ranges from simple aspiration and drainage of abdominal fluid collections (Fig
13) to emergent CT-guided gastrostomy
(of the excluded or bypassed stomach)
for temporary decompression in cases of
acute gastric distention caused by obstruction of the enteroenterostomy. Furthermore, duodenal or jejunal feeding
tubes may be placed with CT guidance in
cases of gastrojejunal anastomotic complications, in order to avoid prolonged
intravenous nutrition (Fig 14). Finally,
stenoses of the gastrojejunal anastomosis
can be dilated with the aid of fluoroscopic guidance (21).
Figure 13. Transverse unenhanced CT scans in a 40-year-old woman who presented 1 month
after revision of previous RYGB with persistent fluid collection in left upper quadrant, which
required image-guided drainage. (a) An 18-gauge Yueh needle is visible within fluid collection in
left upper quadrant. Aspiration revealed purulent fluid. (b) A 12-F drainage catheter is visible
within fluid collection, which subsequently resolved.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Stellato TA, Crouse C, Hallowell PT. Bariatric
surgery: creating new challenges for the endoscopist. Gastrointest Endosc 2003; 57:86 –
94.
Martin LF, Hunter SM, Lauve RM, O’Leary JP.
Severe obesity: expensive to society, frustrating to treat, but important to confront.
South Med J 1995; 88:895–902.
Flegal KM, Carroll MD, Ogden CL, Johnson
CL. Prevalence and trends in obesity among
US adults, 1999 –2000. JAMA 2002; 288:1723–
1727.
Heitmann BL. Ten-year trends in overweight
and obesity among Danish men and women
aged 30 – 60 years. Int J Obes Relat Metab
Disord 2000; 24:1347–1352.
Mokdad AH, Serdula MK, Dietz WH, Bowman BA, Marks JS, Koplan JP. The spread of
the obesity epidemic in the United States,
1991–1998. JAMA 1999; 282:1519 –1522.
Popkin BM, Udry JR. Adolescent obesity increases significantly in second and third generation U.S. immigrants: the National Longitudinal Study of Adolescent Health. J Nutr
1998; 128:701–706.
Mun EC, Blackburn GL, Matthews JB. Current status of medical and surgical therapy
for obesity. Gastroenterology 2001; 120:669 –
681.
Gastrointestinal surgery for severe obesity:
National Institutes of Health Consensus Development Conference statement. Am J Clin
Nutr 1992; 55(suppl 2):615S– 619S.
Byrne TK. Complications of surgery for obesity. Surg Clin North Am 2001; 81:1181–
1193, vii–viii.
Mason EE, Ito C. Gastric bypass in obesity.
Surg Clin North Am 1967; 47:1345–1351.
Fobi MA, Lee H, Holness R, Cabinda D. Gastric bypass operation for obesity. World
J Surg 1998; 22:925–935.
Volume 234
䡠
Number 3
Figure 14. Images obtained 6 months after RYGB in a 58-year-old woman who had several
episodes of dehydration and was presenting with nausea and vomiting that required imageguided placement of a feeding tube. (a) Transverse unenhanced CT scan shows tip of 18-gauge
Yueh needle in duodenal bulb. (b) Fluoroscopic image shows diluted contrast material in duodenum (ⴱ), which confirmed correct position of feeding tube.
12.
13.
14.
15.
16.
17.
18.
Bova JG, Robinson JR, McFee AS. Radiologic
evaluation before gastric bypass for morbid
obesity. Am J Surg 1984; 147:372–374.
Brolin RE. Gastric bypass. Surg Clin North
Am 2001; 81:1077–1095.
Frigg A, Peterli R, Zynamon A, Lang C, Tondelli P. Radiologic and endoscopic evaluation for laparoscopic adjustable gastric banding: preoperative and follow-up. Obes Surg
2001; 11:594 –599.
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.
Heiken JP, Brink JA, McClennan BL, Sagel SS,
Crowe TM, Gaines MV. Dynamic incremental CT: effect of volume and concentration of
contrast material and patient weight on hepatic enhancement. Radiology 1995; 195:
353–357.
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.
Blachar A, Federle MP, Pealer KM, Ikramud-
19.
20.
21.
22.
23.
din S, Schauer PR. Gastrointestinal complications of laparoscopic Roux-en-Y gastric bypass surgery: clinical and imaging findings.
Radiology 2002; 223:625– 632.
Toppino M, Cesarini F, Comba A, et al. The
role of early radiological studies after gastric
bariatric surgery. Obes Surg 2001; 11:447–
454.
Cariani S, Nottola D, Grani S, Vittimberga G,
Lucchi A, Amenta E. Complications after gastroplasty and gastric bypass as a primary operation and as a reoperation. Obes Surg 2001;
11:487– 490.
Vance PL, de Lange EE, Shaffer HA Jr,
Schirmer B. Gastric outlet obstruction following surgery for morbid obesity: efficacy of
fluoroscopically guided balloon dilation. Radiology 2002; 222:70 –72.
Filip JE, Mattar SG, Bowers SP, Smith CD.
Internal hernia formation after laparoscopic
Roux-en-Y gastric bypass for morbid obesity.
Am Surg 2002; 68:640 – 643.
Blachar A, Federle MP, Dodson SF. Internal
hernia: clinical and imaging findings in 17
patients with emphasis on CT criteria. Radiology 2001; 218:68 –74.
Roux-en-Y Gastric Bypass for Severe Obesity
䡠
683