Operative Techniques in
General Surgery
Preface
s the editors of Operative Techniques in General Sur-
A gery, we consider our lives both privileged and
blessed. The opportunity to serve in the capacity as edi-
tors for a number of years has given us a unique insight
into the highly competitive publishing industry and a
more personal look into aspects of the daily lives of the
contributors to this journal and their unique secretaries
(a.k.a. “babysitters/protectors”).
Since this journal’s inception and including prepro-
grammed, yet to be published issues (yes, this has to be
done years ahead of the actual publication date!), we have
been fortunate to entice 22 surgeons of renown to serve as
individual issue guest editors and to receive scientific
contributions from a total of 286 surgeons, radiologists,
and endoscopists, aided by an immensely talented team of
artists, of whose expertise we are in awe of with each
completed issue.
Based on this experience, we would like to offer to you,
our valued readers, some personal observations. These
observations are nonscientific, obviously biased, but are
based on our honest experience with this educational
adventure—some observations are highly complimentary
and, unfortunately, some are not. As always, we welcome
and encourage your input regarding these thoughts—we
are readily available, we promise not to play elevator
music if you are put on hold, and we remain proud that
we do not have voice mail (hard to believe, but true!).
1. Although the two of us obtain a degree of publicity
and notoriety via the journal, kudos belongs, in truth, to
a person behind the scenes who does not have her picture
on this editorial page. We are indebted to this talented
lady, whose name is Marilyn Churchward, and who not
only keeps us on track but who tirelessly reminds and
cajoles the contributors to do the same—not an easy task,
particularly when the latter is done personally and not by
the “easier” e-mail route. All of us have unsung heroines
(heroes) such as Mrs. Churchward in our respective lives,
as do all surgical and medical journals. Let’s tell them how
valuable they are on a regular basis and remind ourselves
Operative Techniques in General Surgery, Vol 5, No 2 (June), 2003: pp 55-56
to whom the true kudos belong.
2. An amazingly large team is required to publish a
quarterly journal—we were previously naı̈ve to this fact.
Our admiration for the executive publisher, who has been
a true coach and a cheerleader to us, as editorial neo-
phytes, continues to grow, as has our respect for the
assistant publisher, the issue managers, the proofreaders,
the artists, and the circulation department. To make this
team efficient and successful, the keyword has been com-
munication on a regular and repetitive basis.
3. An incredible amount of talent has been displayed
by our contributors and our guest editors. Busy surgeons
have a hard time saying no—this is a well-known surgical
truism. The contributors have enthusiastically accepted
the invitation to contribute and have given to you, and to
us, their most precious commodity—their time. Our “fail-
ure to keep a promise rate” (the initial “acceptance” to
contribute changes to “can’t do” six months later) by our
authors has been 2%. Although this is low, it is by right
never acceptable, remembering that many others could
have been asked in the first place. Many contributors have
been well ahead of the deadline (let us do so regularly); in
one instance, a contributor beat the deadline by a full
year. Bravo! This was an endocrine contribution, for
which we offered up the “golden rhinoceros award.”
4. Surgeons are very difficult to get hold of by tele-
phone. This has been frustrating indeed. In a higher per-
centage of instances than we would have imagined, com-
puterized voices were the only “voices” to be found
during a regular business day. The many, many secretar-
ies that we have spoken to were, in most instances, cour-
teous and helpful. Unfortunately, this was not uniformly
so. Shouldn’t it be? We think so! We cannot help wonder
if it is so difficult for surgeons to find other surgeons on
the telephone, how difficult is it for our collective patients
to do the same! Let’s think about this for a moment, shall
we? Are we making ourselves too inaccessible to those
who need us the most? Are we being overly protected
from any intrusion into our busy daily lives? This has
55
56 van Heerden and Farley

certainly been our experience—an experience that war- tive Techniques in General Surgery team, of whom we are
rants sober reflection. so proud, and each of whom fully deserves the “golden
Thank you for allowing us the opportunity to reflect rhinoceros award” for excellence.
and to comment, as well as to thank you for allowing us to Jon A. van Heerden, MD
serve as your editors. Our chosen editorial quote is rele- David R. Farley, MD
vant since it does not apply to any members of the Opera Editors

“Whenever I get the urge to work, I sit down until

the urge passes.”
—Mark Twain
Introduction
one are the days of elective gastric surgery for duo-
G denal ulcer and gastric ulcer disease. Similarly, gas-
trectomy is less frequently used with the noted decrease
in incidence of gastric cancer in the Western world. Cur-
rently, gastric surgery is directed either at the complica-
tions of peptic ulcer disease (perforation, bleeding, and
rarely obstruction), the myriad of sequelae of previously
performed gastric resection(s), and now the relative new-
comer on the block— bariatric surgery.
This edition of Operative Techniques in General Surgery
is devoted to gastric surgery for benign disease. The indi-
vidual authors were carefully chosen not only for their
reputations and that of their respective institutions, but
also because of their interest in the field. Each article
provides a brief, pertinent discussion of historical de-
scriptions, but probably, more importantly, brings to us
the authors’ experience with several helpful technical tips
that facilitate the conduct of the specific operation.
Probably of most use to the more junior surgeons are
© 2003 Elsevier Inc. All rights reserved.
1524-153X/03/0502-0009$30.00/0
doi:10.1053/otgn.2003.35359
Operative Techniques in General Surgery, Vol 5, No 2 (June), 2003: pp 57-58
the separate articles on some of the basic techniques of
vagotomy, drainage procedures, and gastrectomy. With
the infrequency of formal vagotomy and/or gastrectomy
currently in an elective setting, the need for these proce-
dures becomes more common in an emergency setting
and brings with them other considerations of sepsis (du-
odenal or gastric perforation), stress/hypovolemia (ulcer
hemorrhage unresponsive to interventional techniques),
or metabolic/nutritional concerns (gastric outlet obstruc-
tion). These emergency procedures with associated in-
flammatory changes are always a bit more difficult and
often (usually?) lack the nice clean landmarks available in
elective gastric operations. Hopefully some of the tips
provided by the experts, as well as the alternative ap-
proaches, will broaden the available armamentarium of
the less experienced surgeons; indeed, we all learn from
other surgeon’s tips and tricks.
Concerning bariatric surgery, the field has been subject
to constant change and improvement in the last three
decades. Most current bariatric operations are based ini-
tially on some form of gastroplasty/gastric restriction to
limit the size of the receiving gastric reservoir. Minimally
invasive techniques represent the future in this field. The
57
58
article on laparoscopic gastric bypass comes from one of
the world’s experts in the development of the technique—
Dr. Philip Schauer. Although many different laparoscopic
approaches toward the mobilization and position of the
Roux-en-Y limb and cardiojejunostomy exist, this article
provides a foundation of the basic principle, as well as one
expert’s approach.
Michael G. Sarr
Hopefully, this edition of Gastric Surgery for Benign
Disease will offer something of use for young and older
(hopefully no active surgeon becomes “old” and fixed in
his or her ways) surgeons alike.
Michael G. Sarr, MD
Guest Editor
Closure of Duodenal Perforations
Richard C. Thirlby, MD
he incidence of peptic ulcer and the numbers of elec-
T tive operations for intractability in patients with pep-
tic ulceration have been declining in North America for
over two decades. The identification and eradication of
Helicobacter pylori (H. pylori) has also impacted the num-
ber of elective operations for peptic ulcer. Infection of
gastric mucosa by H. pylori is pathogenic in about 90% of
patients less than 60 years old with chronic or intractable
ulcers. However, the rate of H. pylori infection may be less
common in patients with acute complications of peptic
ulcer (eg, perforation and bleeding) and, despite the over-
all decrease in incidence of duodenal ulcers, emergency
operations for bleeding or perforation have not decreased
in incidence. Indeed, with our increasingly aged popula-
tion managed with nonsteroidal anti-inflammatory
agents (NSAIDs) for the various arthropathies, operations
required for acute duodenal perforations are still com-
monly required. Therefore, surgeons treating patients
with peptic ulcer must be knowledgeable not only of
operative techniques but also the relevance of H. pylori
infection in their patients. This chapter will describe the
operative technique for closure of duodenal perforations
as well as the current diagnostic techniques that should
be incorporated into the evaluation and treatment of pa-
tients with duodenal perforations.
PREOPERATIVE DIAGNOSIS
The clinical suspicion and diagnosis of perforated peptic
ulcer is usually straightforward. A history of sudden onset
of diffuse abdominal pain associated with diffuse perito-
nitis and the typical board-like abdomen are virtually
pathognomonic. Free air under the diaphragm or pneu-
moperitoneum is apparent on upright chest radiographs
in about 90% of such patients. When the diagnosis is in
question, or when nonoperative treatment is being con-
sidered, additional testing is warranted. Computed to-
mography (CT) is more sensitive than plain radiographs
in detecting pneumoperitoneum; in addition, CT with
oral or contrast upper gastrointestinal series (UGI) using
water-soluble contrast will typically demonstrate perfora-
From the Virginia Mason Medical Center, Seattle, Washington 98101.
Address reprint requests to Dr. Richard C. Thirlby, Virginia Mason Medical
Center, 1100 Ninth Avenue, C6-SUR, Seattle, Washington 98101.
© 2003 Elsevier Inc. All rights reserved.
1524-153X/03/0502-0013$30.00/0
doi:10.1053/otgn.2003.35363
Operative Techniques in General Surgery, Vol 5, No 2 (June), 2003: pp 59-64
tions and also assess the presence of free versus contained
perforations. As discussed below, in patients in whom
radiographic studies show contained perforations, a pri-
mary, nonoperative approach may be an option.
PREOPERATIVE PREPARATION AND
POSSIBLE NONOPERATIVE
MANAGEMENT
Several studies have demonstrated a role for nonoperative
management of selected patients with duodenal perfora-
tions.1-4 A randomized trial comparing nonoperative ver-
sus operative treatment in patients with perforated peptic
ulcers showed that nonoperative treatment was success-
ful in about 70% of patients.1 Of note, this study did not
include upper GI contrast studies in the entry criteria.
The patient population least likely to be managed suc-
cessfully by nonoperative treatment was the elderly (fail-
ure in 6/9 patients). The conclusion of these investiga-
tions was that a 12-hour trial of nonoperative
management was safe in patients less than 70 years old. In
a retrospective series, Berne and Donovan2 reported that
12% of 294 patients with duodenal or prepyloric perfora-
tions treated at their institution during an 11-year period
were managed nonoperatively.2 Their algorithm proposes
that if an UGI series shows that the ulcer is sealed, the
success of nonoperative management approaches 100%.
Their original recommendations, published before the
current recognition of the importance of H. pylori infec-
tion, also stressed the importance of chronicity of the
ulcer diathesis in the management of patients with perfo-
rated peptic ulcers.2 Patients with chronic ulcers contin-
ued to experience symptoms and complications of peptic
ulcer, whereas patients with acute ulcers frequently did
not experience symptoms in the future. Berne and co-
workers suggested, therefore, that the threshold to oper-
ate on patients with a history or signs of chronic ulcer
should be lower, because definitive ulcer operations can
be combined with ulcer closure. As discussed below, the
appreciation of H. pylori has changed the importance of
this distinction.4
Nonoperative treatment may be appropriate in selected
patients with perforated duodenal ulcers and even in
those younger patients with a typical history of chronic
ulcers. Because current thinking is that most ulcers are
associated with H. pylori and can be cured with antibiot-
ics, the relevance of chronicity may be minimal. At pre-
sentation, patients in whom nonoperative management is
59
60 Richard C. Thirlby

being entertained should probably have UGI studies with
water-soluble contrast to document that the ulcer perfo-
ration is contained. Most elderly patients will not demon-
strate spontaneous sealing of their ulcers. If there is no
free leak of contrast, nonoperative treatment may be
appropriate regardless of patient age. Nonoperative
management thereafter consists of nasogastric suction,
broad-spectrum antibiotics, and high-dose intravenous
antisecretory agents (eg, histamine H2-receptor antago-
nists, proton-pump inhibitors). Gastric pH testing should
confirm achlorhydria (ie, gastric pH ⬎4 to 5). The role of
antifungal treatment (eg, fluconazole) in these patients is
unproven; many series have suggested a pathogenic role
of Candida species in the infective complications in pa-
tients with duodenal perforations and thus support the
use of prophylactic fluconazole.5 Clinical improvement
during a primary nonoperative approach should be ap-
parent within 12 hours.
SURGICAL TREATMENT
The surgical treatment of perforated duodenal ulcers can
be divided into three approaches: simple closure, laparo-
scopic techniques, and closure of complex or giant duo-
denal ulcers. As shown in Figs 1 and 2, the techniques of
simple closure (Graham patch) have not changed signif-
icantly since the classic report by Roscoe Graham in
1937.6 If the omentum is not available or not sufficient to
close the perforation, the falciform ligament may be di-
vided from the anterior abdominal wall and a flap mobi-
lized based on its attachment to the liver (Fig 3). It is
likely that laparoscopic techniques are safe (Figs 4, 5, and
6) and include standard suturing of an omental plug,
stapled omental plugs, use of gelatin sponge as a plug with
fibrin glue, and a “single-suture closure” technique.7-9 All
these techniques should be accompanied by some attempt
at peritoneal toilet by aspirating the free intraperitoneal
fluid with additional irrigation. Larger ulcers may require
complex repairs (Fig 7). In the era before the recognition
of the importance of H. pylori in the pathogenesis of
duodenal ulcer, randomized studies and large case series
have demonstrated better long-term outcomes when de-
finitive operations were added to closure of the ulcer
perforation.10,11 However, the need to perform a defini-
tive procedure (eg, proximal gastric vagotomy) at the

1 The original figure used by Roscoe Graham of Toronto, Canada in ‘The Treatment of Perforated Duodenal Ulcers“ in 1937.
Graham stated, ”Three interrupted catgut sutures are used. . . a piece of omentum, either free or attached, is laid over these sutures,
which are then tied just sufficiently tight to hold the omental graft in situ, but not with sufficient force to cause the sutures to cut
out. . . No attempt is made actually to close the perforation.“ The concept for the ”Graham Patch“ has not changed in 66 years.
However, other suture material may be preferable to catgut. Full-thickness silk sutures are believed to be ulcerogenic by some
surgeons and are not recommended. Absorbable sutures such as polyglactin (eg, VicrylR, Ethicon, Somerville, NJ) or polydiaxone
(eg, PDSR, Ethicon, Somerville, NJ) are preferable (reprinted with permission, ref 6)
Closure of Duodenal Perforations 61

2 Alternative techniques of patch

closure. Two to four seromuscular su-
tures are placed as shown, with bites
proximal and distal to the perforation.
After placement of all sutures, the
omentum is drawn beneath the tagged
sutures and then secured. This tech-
nique is simple, avoids placement of
suture in inflamed tissue and is un-
likely to compromise the duodenal lu-
men.

time of ulcer closure is unclear, currently, especially in
younger patients in whom infection by H. pylori is usually
involved.
The true incidence of H. pylori infection in patients
with perforated duodenal ulcers is controversial.12-14
Most studies have concluded that the incidence of infec-
tion is similar to that in patients with chronic peptic
ulcer.13,14 Others have found lower rates of H. pylori
infection in perforated ulcers.12 It would appear that if
one excludes patients (usually the elderly) with concur-
rent use of high doses of NSAIDs, most patients with
perforated peptic ulcers are infected with H. pylori. A

3 When omentum is not available, the falciform ligament may be used as a patch. After division of the falciform at the anterior
abdominal wall, the vascularized falciform graft based on its attachment to the liver easily reaches the duodenum, Three sutures of
PDS secure the falciform patch over the ulcer.
62 Richard C. Thirlby

4 Laparoscopic repair of perforated duodenal ulcer. Suggested port sites and positions of surgeon and assistants are shown in the
inset. Alternatively, the patient can be placed in a Lloyd-Davis position with reverse Trendelenburg tilt, with the operating surgeon
positioned between the patient’s thighs. Recent studies have suggested that perforations should be less than 10 mm to be repaired
laparoscopically. The first step of the procedure is to tack the apex of the omental patch to the superior aspect of the duodenum,
either with an endoscopic stapler or with a polyglactin suture full thickness through the duodenum and the apex of the omentum.

recent study from Hong Kong reported that about 80% of
all patients with perforated duodenal ulcers were infected
by H. pylori.15 After simple omental patch closure, these
patients were then randomized to receive either a 4-week
course of antisecretory therapy (omeprazole) alone or
antibiotic eradication in conjunction with omeprazole.
After 1 year, ulcer relapse was significantly less common
in patients treated with the added antibiotic eradication
compared with those who received antisecretory therapy
alone (5 versus 38%, respectively).15 Therefore, ideally
before choosing a surgical technique, one should assess
the presence of H. pylori infection before completion of

5 Laparoscopic repair of perforated duodenal ulcer. After securing the omentum to the apex of the perforation, two to three
additional sutures are placed on the periphery of the omental patch and anchored with double clips (Endo Clip: U.S. Surgical,
Norwalk, CT). When the integrity of the repair is questionable, application of commercially available fibrin glue preparations
should be considered (Tisseal, Baxter International Inc. Glendale CA, or Hemaseel APR, Hemacure Corp, Sarasota, FL). This
technique is applicable to virtually all open and laparoscopic repairs described in the previous figures. Catheters are available which
permit application through laparoscopic trocars.
Closure of Duodenal Perforations 63

6 A “plug and glue” technique is

effective in closing duodenal perfora-
tions. Gelatin sponge is rolled into a
plug or a plug of omentum is inserted
into the perforation, secured in place,
and covered with fibrin glue (modi-
fied from ref. 19).

7 Giant perforated duodenal ulcers may require adjuvant measures. One technique is a tube duodenostomy. A 10-French Foley
catheter, inserted into the uninvolved lateral aspect of the second portion of the duodenum, is left to dependent drainage for 14
days. After confirmation of an intact repair, the balloon of the Foley catheter is deflated, and the catheter is withdrawn 2 cm (just
outside the duodenum) and placed to dependent drainage for 24 to 48 hours. An upper GI study should confirm the lack of
extravasation of duodenal contents before removing the catheter completely. Some surgeons prefer using a Malecot catheter for
duodenal decompression, leaving the catheter in place for at least six weeks before removing it. Another technique in patients with
giant ulcers and/or compromised healing, illustrated here, may be to place “feed me, drain me” catheters. Approximately 15 cm
distal to the ligament of Treitz, a 10 to 12 French red rubber catheter is positioned retrograde into the duodenum for decompression.
Multiple side holes are added to the intraluminal portion of the catheter to enhance catheter function. A second feeding catheter
jejunostomy is placed antegrade to provide for postoperative enteral nutrition. A third technique that may rarely be required for a
very large duodenal perforation not amenable to primary closure is to perform a duodenojejunostomy to a Roux-en-Y limb; this
creates a mucosa-to-mucosa internal drainage.
64
operation.16 Serologic testing or histologic evaluation
that become available postoperatively can be of no value
to the surgeon intraoperatively.
The most rapid test available to determine the presence
of H. pylori infection is a recently developed HpSA antigen
test (Meridian Bioscience, Cincinnati, OH) which is per-
formed on stool. Whenever possible, a stool specimen
should be assayed for H. pylori antigen levels before op-
eration. Assay results should be available in about 90
minutes. In our opinion, the addition of a definitive ulcer
operation (eg, proximal gastric vagotomy) is not indi-
cated in H. pylori positive patients at the time of ulcer
closure, because postoperative antibiotic therapy can
eradicate the likely etiology of the ulcer diathesis. The
most widely available method to confirm H. pylori infec-
tion is the CLO test, which is performed on antral mucosa
obtained by endoscopic biopsy. However, the time re-
quired for this assay to turn positive is variable and de-
pends on the concentration of the organism in the biopsy
specimen. In our experience, the assay can take as long as
2 to 3 hours to turn positive, making its reliability for
diagnosing the presence of H. pylori infection at the time
of operation of little value. If there is no evidence of H.
pylori infection as determined by HpSA antigen testing
preoperatively in the emergency room or in the elderly
patient who requires chronic antiinflammatory (NSAIDs)
therapy, addition of a definitive ulcer procedure (eg,
proximal gastric vagotomy) should at least be consid-
ered.4 In contrast, if the HpSA antigen test or CLO testing
confirms H. pylori infection, then simple ulcer closure
with postoperative antibiotic eradication of infection in
conjunction with a short course of acid suppression
should cure the ulcer diathesis in the majority of patients.
In patients with perforations of prepyloric ulcers, the
possibility of cocaine-related ulcers should be consid-
ered.17 This condition occurs almost exclusively in males
who smoke crack cocaine. The perforations are typically
only 3 to 5 mm in diameter and are located in the prepy-
loric gastric antrum. In these patients, simple omental
closure seems appropriate.
FOLLOW-UP AND RECURRENCE
Ulcer recurrence in patients with perforated duodenal
ulcers is affected by many factors. As mentioned previ-
ously, patients with a history of chronic ulcer before the
perforation are much more likely to develop or continue
to have persistent signs, symptoms and complications of
peptic ulcer. Continued use of NSAIDs is associated with
recurrent ulcers. Infection with H. pylori will result in
nonhealing or recurrent ulceration. Finally, gastrinomas
or the Zollinger-Ellison syndrome can be another cause of
recurrent ulcers.
The postoperative follow-up of patients with perfo-
rated duodenal ulcers, therefore, should include the fol-
lowing: first, antral mucosal biopsies and/or serologies to
Richard C. Thirlby
rule out H. pylori should be performed in all patients;18
second, combination antibacterial and antisecretory ther-
apy to eradicate H. pylori should be administered to in-
fected (or perhaps all) patients; third, all patients should
have fasting serum gastrin levels measured and, in select
patients, it may be appropriate to obtain salicylate aspirin
and ibuprofen levels; and fourth, after six weeks postop-
eratively, endoscopy with confirmation of both ulcer
healing, and eradication of H. pylori should be completed
in all patients.
REFERENCES
1. Crofts TJ, Park KGM, Steele RJC, et al: A randomized trial of
nonoperative treatment for perforated peptic ulcer. NEJM 320:
970-973, 1989
2. Berne TV, Donovan AJ: Nonoperative treatment of perforated
duodenal ulcer. Arch Surg 124:830-832, 1989
3. Donovan AJ, Vinson TL, Maulsey GO, et al: Selective treatment of
duodenal ulcer with perforation. Ann Surg 189:627-636, 1979
4. Donovan AJ, Berne TV, Donovan AJ: Perforated duodenal ulcer.
Arch Surg 133:1166-1171, 1998
5. Peoples JB: Candida and perforated peptic ulcers. Surgery 100:
758-764, 1986
6. Graham RR: The treatment of perforated duodenal ulcers. Surg
Gynec Obstet 64:235-238, 1937
7. Thompson AR, Hall TJ, Anglin BA, Scott-Conner CE: Laparo-
scopic plication of perforated ulcer. South Med J 88:185-189,
1995
8. Matsuda M, Nishiyama M, Hanai T, Saeki S: Laparoscopic omental
patch repair of perforated peptic ulcer. Ann Surg 221:236-240, 1995
9. Siu WT, Leong HT, Law BK, et al: Laparoscopic repair for perfo-
rated peptic ulcer: A randomized controlled trial. Ann Surg 235:
313-319, 2002
10. Jordan PH Jr: Proximal gastric vagotomy without drainage for
treatment of perforated duodenal ulcer. Gastroenterology 83:179-
183, 1982
11. Hay JM, Lacaine F, Kohlmann G, et al: Immediate definitive sur-
gery for perforated ulcer does not increase operative mortality: A
prospective controlled trial. World J Surg 12:705-709, 1988
12. Reinbach DH, Cruickshank G, McColl KEL: Acute perforated
duodenal ulcer is not associated with Helicobacter pylori infec-
tion. Gut 34:1344-1347, 1993
13. Sebastian M, Chandran VPP, El Ashaal, et al: Helicobacter pylori
infection in perforated peptic ulcer disease. B J Surg 82:360-362, 1995
14. Matsukura N, Onda M, Tokunaga A, et al: Role of Helicobacter pylori
infection in perforation of peptic ulcer: An age-and gender-matched
case-control study. J Clin Gastroenterol 25:S235-S239, 1997
15. Enders KW, Lam YH, Sung JJY, et al: Eradication of Helicobacter
pylori prevents recurrence of ulcer after simple closure of duode-
nal ulcer perforation. Ann Surg 231:153-158, 2000
16. Stabile BE: Redefining the role of surgery for perforated duodenal
ulcer in the Helicobacter pylori era. Ann Surg 231:159-160, 2000
17. Feliciano DV, Ojukwu JC, Rozycki GS, et al: The epidemic of
cocaine-related juxtapyloric perforations: With a comment on the
importance of testing for helicobacter pylori. Ann Surg 229:801-
806, 1999
18. Greenberg PD, Koch J, Cello JP: Clinical utility and cost effective-
ness of Helicobacter pylori testing for patients with duodenal and
gastric ulcers. Am J Gastroenterol 91:228-232, 1996
19. Lau WY, Leung KL, Kwong KH et al: A randomized study com-
paring laparoscopic versus open repair of perforated peptic ulcer
using suture or sutureless technique. Ann Surg 224:131-138,
1996
Pyloroplasty
Jon Arne Söreide, MD, PhD, FACS and Kjetil Söreide, MD
y incision of the pyloric muscle and “plastic” recon-
B struction of the pyloric channel, pyloroplasty facili-
tates gastric emptying when the gastric outlet is ob-
structed either mechanically by ulcer, stenosis, or
hypertrophy, or functionally by antropyloric vagotomy.
First described by Heineke in 1886 and by Mikulicz in
1888, the Heineke-Mikulicz technique is still considered
the principal pyloroplasty operation favored by most sur-
geons. Nevertheless, several modifications and some al-
ternative techniques have been introduced during the last
century (Fig 1).
INDICATIONS
Traditionally, pyloroplasty was performed following se-
lective or truncal vagotomy in the surgical treatment of
peptic ulcer disease with pyloric stenosis.1,2 The need for
pyloroplasty in the absence of mechanical pyloric ob-
struction has been a matter of discussion for many
years.3-5 During emergency surgery for a bleeding duo-
denal ulcer, clear exposure of the offending ulcer is im-
portant, and access to control the bleeding site is manda-
tory. Closure of the duodenotomy is completed by per-
forming a pyloroplasty to prevent stenosis at the level of
the duodenotomy.6 Caution should be undertaken to per-
form a pyloroplasty when the pyloric area is severely
distorted or inflamed as a result of the ulcer disease. With
the marked decrease in gastric surgery for peptic ulcer
disease, pyloroplasty has become a much less common
procedure.
Although most commonly related to peptic ulcer dis-
ease and its complications (ie, obstruction, bleeding or
perforation), a pyloroplasty may be considered in other
rare situations, including Crohns disease,7 progressive
systemic sclerosis,8 and during operations when gastric
replacement of the esophagus is necessary and a vagot-
omy (vagectomy) is necessitated by the resective proce-
dure.9,10
From the Division of Gastroenterologic Surgery, Dept. of Surgery, Rogaland
Central Hospital, University of Bergen, Institute of Surgical Sciences, Stavanger,
Norway.
Address reprint requests to Professor Jon Arne Söreide, MD, PhD, FACS, Dept.
of Surgery, Rogaland Central Hospital, POB 8100, N-4068 Stavanger, Norway;
e-mail: jon.soreide@kir.lib.no.
© 2003 Elsevier Inc. All rights reserved.
1524-153X/03/0502-0016$30.00/0
doi:10.1053/otgn.2003.35366
Operative Techniques in General Surgery, Vol 5, No 2 (June), 2003: pp 65-72
GENERAL CONDITIONS
To make the operation technically feasible, the anterior
surface of the pylorus should be minimally involved and
the duodenum sufficiently mobile. Occasionally, a pylo-
roplasty can also be performed in the presence of an
anterior ulcer if the tissue surrounding the ulcer is mini-
mally involved.
There are two basic types of pyloroplasty: the Heineke-
Mikulicz and the Finney procedures. The Jaboulay pylo-
roplasty is really a gastroduodenostomy; the incision does
not extend through the pylorus. In addition, several im-
portant modifications of the standard procedures have
been described; we will address these as well.
Heineke-Mikulicz Pyloroplasty
A longitudinal incision is placed through the pylorus,
extending from the distal antrum to the proximal duode-
num. By closing this incision transversely, the outlet di-
ameter of the pylorus is increased.
The duodenum is mobilized by a Kocher maneuver,
and the pylorus identified (Fig 2). The pyloric veins of
Mayo may aid its identification, as does its palpable thick-
ening appreciated by transmural inspection. Scarring on
the anterior surface of the pylorus should be minimal, and
mobilization of the duodenum is necessary to facilitate
the pyloroplasty. Between two traction sutures placed
about 1 cm apart on the anterior surface of the pylorus, a
longitudinal incision is made, extending about 3 cm onto
the antrum and a similar distance onto the duodenum
(Fig 2). The total length should not exceed 5 to 7 cm. The
incision is created by diathermy, which also affords he-
mostasis. Careful inspection of the stomach and the duo-
denum for bleeding ulcer or point of obstruction is man-
datory. If pyloric stenosis is present, the initial entry into
the gut lumen could be made either in the duodenum or
in the stomach, because the obstructed lumen in the py-
loric region may be eccentrically placed. Use of a grooved
director or narrow tipped clamp passed into the pylorus
from the initial distal antrotomy will facilitate the appro-
priate placement of the pylorotomy. If bleeding from an
associated gastric or duodenal ulcer is encountered, a
transfixion suture (polydioxone, PDS™ 3– 0 or 2– 0 silk)
may be necessary.
Various techniques have been designed to accomplish
closure of the incision. The longitudinal incision is closed
transversely by rostral and caudal distraction of the re-
65
66 Söreide and Söreide

1 Pyloroplasty incisions: (A) Heineke-Mikulicz, (B) Finney, (C) Moschel, and (D) Jaboulay
Pyloroplasty 67

2 Gastroduodenotomy for Heineke-Mikulicz pyloroplasty

68
3 Finney pyloroplasty
traction sutures. Of importance is that the outlet diameter
is kept sufficient. A one-layer closure (Weinberg modifi-
cation of the Heineke-Mikulicz pyloroplasty) is most fre-
quently employed (Fig 2C and 2D), although a two-layer
closure may be preferred by some surgeons (Fig 2E).
Whichever method is used, the suture should pass
through all layers with meticulous approximation of the
separate serosal and mucosal layers. Our personal prefer-
ence is a single layer with continuous absorbable suture
(poliglecaprone, Monocryl™ 3– 0 or 4 – 0).
Finney Pyloroplasty
By extending the incision of the pyloric area onto the
stomach and first portion of the duodenum, a Finney
Söreide and Söreide
pyloroplasty (ie, a gastroduodenostomy with incision of
the pylorus) can be completed (Fig 3). This technique is
especially well suited for a J-shaped stomach in which the
pylorus may be retracted and fixed rostrally, making a
Heineke-Mikulicz pyloroplasty tenuous. The duodenum
should be widely Kocherized and partially detached from
the gastro-hepatic ligament to facilitate the descending
duodenum to be laid alongside the greater curvature of
the distal antrum. Adjacent gastric and duodenal walls are
first united by means of a seromuscular suture (poligle-
caprone Monocryl™, 3– 0 or 4 – 0), from above down-
wards, closing the angle between the pylorus. A traction
suture is placed in the superior margin of the pyloric ring,
a second one placed on the duodenal wall about 10 cm
Pyloroplasty 69

4 Moschel pyloroplasty

distal to the pyloric ring, and, a third placed on the greater
curvature of the stomach 10 cm proximal to the pylorus.
With diathermy, a full thickness incision is made along
the inverted horseshoe-shaped line which runs from the
gastric antrum 4 to 5 cm proximal to the pylorus curving
through the duodenal bulb and down the descending
duodenum. Care should be taken to stop any bleeding.
Again, initial entry into the lumen could be either in the
antrum or the duodenum.
To close the incision, the posterior adjacent walls are
approximated by means of a continuous seromuscular
suture (poliglecaprone Monocryl™, 3– 0), starting at the
superior end of the pylorotomy. This suture is continued
through the inferior gastroduodenal angle, to proceed
70 Söreide and Söreide

onto the anterior wall of the gastroduodenostomy until a

safe and well-approximated closure forms the anterior
part of the pyloroplasty. From a personal point of view,
one or two additional single seromuscular sutures to sup-
port the anastomosis at the inferior gastroduodenal angle
may ease the surgeon’s mind that night.
Jaboulay pyloroplasty. This procedure is not a true
pyloroplasty, because the pylorus is not incised (see Fig
1). Indeed, separate antrotomy and duodenotomy inci-
sions are necessary. However, the aim of, and the indica-
tion for, this operation is to increase the luminal size of
the gastric outlet, and the Jaboulay modification will in
principal be performed by following the steps described
in the Finney procedure but without necessarily opening
the pyloric ring.

Modifications of Standard Procedures

Judd pyloroplasty. This operation includes excision
of an anterior ulcer with removal of the anterior two-
thirds of the pyloric sphincter. This and other modifica-
tions of the two standard pyloroplasties (Heineke-Miku-
licz and Finney pyloroplasty) are generally considered
more of historic interest, since our present understanding
of duodenal ulcer disease does not support the need for
excision of the ulcer. Recurrent stenosis tended to occur
more frequently after the Judd pyloroplasty, due to the
larger amount of tissue removed with the failure to assure
and maintain an adequate pyloroplasty.
Moschel pyloroplasty. Using a Y-shaped incision,
with the base of the Y extending through the pylorus onto
the duodenum and the arms of the Y onto the antrum, this
technique was designed to maintain an adequate blood
supply to the advancement flap of the antrum. The antral
flap is then sutured to the duodenum, generally using a
one-layer closure (Fig 4).

STAPLING TECHNIQUES
Surgical stapling techniques have also been employed to
perform a modified Heineke-Mikulicz pyloplasty (Fig 5)
and a Jaboulay pyloroplasty (Fig 6). Transverse closure of
a longitudinal pylorotomy can be accomplished using a
linear stapler. Similarly, an antroduodenostomy can be
accomplished using a gastrointestinal stapler, the arms of
which are introduced through separate antrotomy and
duodenotomy creating a stapled side-to-side anastomosis
(Fig 6). The possible benefits in comparison to standard
techniques, however, remain to be shown. Recently, fur-
ther technical developments, including laparoscopic ap-
proaches have been introduced.11,12 Given that general
laparoscopic skills and experience are present, this ap-
proach may have its place in some patients.

COMMENTS AND CONCLUSION

Providing that the duodenal area is not severely inflamed
5 Stapled Heineke-Mikulicz pyloroplasty and scarred, pyloroplasty is usually a safe and technically
Pyloroplasty 71

6 Stapled Jaboulay pyloroplasty

72
easy operation. However, the advantage of facilitating
gastric emptying is also accompanied by some disadvan-
tages. The rapid emptying of liquids and solids due to
destruction of the pyloric sphincter in association with a
vagotomy may cause symptoms such as dumping and
diarrhea. In addition, some patients may suffer from re-
flux alkaline gastritis due to reflux of duodenal fluid into
the stomach. With the decreasing incidence of peptic
ulcer disease as well as the virtual disappearance of trun-
cal vagotomy, the need for a pyloroplasty is vanishing.
REFERENCES
1. Rachlin L: Vagotomy and Heineke-Mikulicz pyloroplasty in the
treatment of pyloric stenosis. Am Surg 36(4):251–253, 1970
2. Cade D, Allan D: Long term follow-up of patients with gastric
ulcers treated by vagotomy, pyloroplasty and ulcerectomy. Br J
Surg 66(1):46 – 47, 1979
3. Aeberhard P, Walther M: Results of a controlled randomized trial
of proximal gastric vagotomy with and without pyloroplasty. Br J
Surg 65(9):634 – 636, 1978
4. Gutierrez de la Pena C, Marquez R, Fakih F, Dominguez-Adame E,
Medina J: Simple closure or vagotomy and pyloroplasty for the
Söreide and Söreide
treatment of a perforated duodenal ulcer: comparison of results.
Dig Surg 17(3):225–228, 2000
5. Chan V, Reznick R, O’Rourke K, Kitchens J, Lossing A, Detsky A:
Meta-analysis of highly selective vagotomy versus truncal vagot-
omy and pyloroplasty in the surgical treatment of uncomplicated
duodenal ulcer. Can J Surg 37(6):457– 464, 1994
6. Brolin R, Stremple J: Emergency operation for upper gastrointes-
tinal hemorrhage. Am Surg 48(7):302–308, 1982
7. Taor RE: Adult hypertrophic pyloric stenosis and Crohn’s disease.
Proc R Soc Med 69(3):228, 1976
8. Hirakata M, Akizuki M, Okano Y, et al: Pyloric stenosis in a patient
with progressive systemic sclerosis. Clin Rheumatol 7(3):394 –
397, 1988
9. Fok M, Cheng SW, Wong J: Pyloroplasty versus no drainage in
gastric replacement of the esophagus. Am J Surg 162(5):447– 452,
1991
10. Law S, Cheung MC, Fok M, Chu KM, Wong J: Pyloroplasty and
pyloromyotomy in gastric replacement of the esophagus after
esophagectomy: a randomized controlled trial. J Am Coll Surg
184(6):630 – 636, 1997
11. Ng JW, Yeung GH: Laparoscopic vagotomy and open pyloroplasty
for bleeding duodenal ulcer not controlled endoscopically. Surg
Laparosc Endosc 8(2):127–131, 1998
12. Danikas D, Geis W, Ginalis E, Gorcey S, Stratoulias C: Laparo-
scopic pyloroplasty in idiopathic hypertrophic pyloric stenosis in
an adult. [In Process Citation]. JSLS 4(2):173–175, 2000
Open Gastrojejunostomy
Christopher J. Sonnenday, MD, and Charles J. Yeo, MD
hile the volume of gastric surgery performed in
W surgical training programs in the United States has
decreased markedly in recent years with the virtual dis-
appearance of elective ulcer operations,1 open gastrojeju-
nostomy remains an important tool in the armamentar-
ium of any general surgeon. Gastrojejunostomy is often
performed to restore gastro-enteric continuity after resec-
tion or palliative bypass of malignant disease (gastric or
periampullary), or as part of a gastric bypass procedure
for morbid obesity. Further, gastrojejunostomy serves as
a treatment option for benign obstruction to the gastric
outlet or duodenum. This chapter will concentrate on the
role of open gastrojejunostomy in the management of
patients with benign disease.
INDICATIONS
Open gastrojejunostomy may be indicated for any condi-
tion that functionally or mechanically impairs normal
gastric emptying. Historically, the most common cause of
gastric outlet obstruction requiring operative treatment
has been peptic ulcer disease. Despite the recognition of
Helicobacter pylori as a causative agent in peptic ulcer
disease and the establishment of effective medical thera-
pies, gastric outlet obstruction may occur in 6% to 8% of
patients with chronic duodenal ulcer.2 Many of these
patients may be managed initially with nonoperative ther-
apy consisting of nasogastric suction, gastric acid inhibi-
tion, eradication of H. pylori, endoscopic dilation, and
total parenteral nutrition. Those patients that fail nonop-
erative intervention require surgical therapy for persis-
tent or recurrent obstruction.3,4 Endoscopic balloon dila-
tion coupled with aggressive H. pylori therapy and high
dose proton pump inhibition has been proposed as an
alternative to surgical therapy, but recent trials report a
symptomatic recurrence rate as high as 50%,5,6 with those
patients experiencing a recurrence typically requiring
surgical intervention.
Gastrojejunostomy, accompanied by highly selective
From the The Johns Hopkins Medical Institutions, Baltimore,MD.
*Note: All illustrations have been modified from Atlas of Surgery -Volume 2.
Editor John L. Cameron; Illustrator Corinne Sandone; Mosby; St. Louis, 1994.
Address reprint requests to Dr. Charles J. Yeo, Professor of Surgery and Oncol-
ogy, The Johns Hopkins Hospital, Department of Surgery, Blalock 606, 600 N.
Wolfe Street, Baltimore, MD 21287-4606.
© 2003 Elsevier Inc. All rights reserved.
1524-153X/03/0502-0029$30.00/0
doi:10.1053/otgn.2003.35367
Operative Techniques in General Surgery, Vol 5, No 2 (June), 2003: pp 73–79
vagotomy (HSV), has been advocated as the primary non-
resectional treatment of obstructing duodenal ulcer, be-
cause it appears to be a more effective gastric drainage
procedure than pyloroplasty and carries a lower risk of
perioperative morbidity and mortality than antrec-
tomy.2,7 Csendes and colleagues2 compared gastrojeju-
nostomy plus HSV, gastroduodenostomy plus HSV, or
antrectomy plus HSV in a prospective trial and found
both the gastrojejunostomy and antrectomy groups to
have excellent and equivalent long-term outcomes. It is
important to emphasize that gastrojejunostomy alone is
not believed to be an adequate therapy for obstructing
peptic ulcer disease and should be accompanied by vagot-
omy to prevent marginal ulceration at the gastrojejunal
anastomosis.
Chronic pancreatitis is another important cause of gas-
tric outlet obstruction. While acute pancreatitis is often
associated with transient gastroduodenal dysmotility,
these patients rarely require operative intervention.
While less than 1% of all patients with chronic pancreati-
tis will develop fixed duodenal stenosis, up to 25% of
patients with complicated chronic pancreatitis will re-
quire surgery.8,9 Thorough evaluation of these patients is
necessary to rule out other associated conditions which
would mandate different therapy, such as peptic ulcer
disease, pancreatic pseudocyst, or pancreatic ductal ade-
nocarcinoma.
Gastroduodenal obstruction secondary to duodenal
stricture may occur in 1% to 3% of patients with Crohn’s
disease.10 Approximately one-third of these patients may
eventually need surgical treatment, and gastrojejunos-
tomy rather than duodenoplasty is the procedure of
choice in these patients.11 Again, vagotomy is recom-
mended at the time of this procedure to reduce the inci-
dence of marginal ulceration.10
Other rare causes of mechanical gastroduodenal ob-
struction that may be treated effectively with gastrojeju-
nostomy include congenital duodenal obstruction (due to
atresia, webs, or duplication), annular pancreas, and
posttraumatic strictures of the pylorus or duodenum.
Patients with severe functional impairment to gastric
emptying, in the absence of a mechanical obstruction,
present a tremendous challenge in management. Systemic
disorders such as diabetes mellitus, various collagen-vas-
cular diseases, and autonomic neuropathies may be asso-
ciated with severe gastroparesis. Idiopathic gastric atony
has also been reported.12 Patients with functional gastric
73
74
outlet obstruction may be referred to surgeons after failed
trials of the prokinetic agents. In some instances, attempts
at gastric pacing may have been attempted and failed.13
Although no large series or prospective trials exist for this
small group of patients, gastrojejunostomy has been re-
ported to provide benefit but only in highly selected pa-
tients.12,14,15 However, partial or complete gastric resec-
tion may eventually be necessary in some patients.
Patients with gastroparesis are at extremely high risk of
persistent gastric emptying problems after primary drain-
age procedures (pyloroplasty, gastrojejunostomy, and
even antrectomy), and therefore procedures such as trun-
cal vagotomy (which may worsen gastric motility) should
be avoided.16
PREOPERATIVE EVALUATION
The goals of the preoperative evaluation for a patient with
gastroduodenal obstruction are: (1) appropriate resusci-
tation; (2) nutritional supplementation; and (3) to defin-
itively exclude causes of obstruction that would require
other medical or surgical therapy, particularly malig-
nancy. Patients with chronic gastroduodenal obstruction
can present with impressive dehydration and malnutri-
tion due to protracted decreased oral intake and vomiting.
Electrolyte and acid-base disturbances, classically a con-
traction hypochloremic metabolic alkalosis with hypoka-
lemia, should be promptly identified and corrected. Na-
sogastric decompression may cease the vomiting and
allow relief of gastric distention. Total parenteral nutri-
tion is often necessary to begin the correction of protein-
calorie malnutrition. Operative intervention is usually
delayed until adequate hydration and electrolyte correc-
tion are achieved and until nutritional repletion is begun.
The diagnosis of gastroduodenal obstruction is often
suspected based on the findings at history and physical
examination, and it may be further supported by the ap-
pearance of a large gastric air bubble on plain radiography
of the abdomen. Confirmatory tests such as an upper
gastrointestinal (GI) series or computed tomography
(CT) with oral contrast may further support the diagnosis
and isolate the location of the obstruction. In patients
without a clear mechanical obstruction, scintigraphic im-
aging using both solid and liquid phases (dual phase nu-
clear medicine gastric emptying scan) may document an
abnormality in gastric emptying.
Before proceeding to the operating room for surgical
therapy, it is usually essential that the mucosa of the
stomach and duodenum be examined endoscopically in
an effort to define the cause (and possibly the site) of the
obstruction. Any thickened, inflamed, or ulcerated re-
gions of the mucosa should be sampled, with the goal
being to assess for malignancy, Crohn’s disease and H.
pylori status. Other imaging studies such as CT, magnetic
resonance, and endoscopic ultrasonography may be help-
Sonnenday and Yeo
ful in defining the underlying cause of the gastroduodenal
obstruction.
TECHNICAL CONSIDERATIONS
Antecolic Gastrojejunostomy. Open antecolic gas-
trojejunostomy is typically performed via an upper mid-
line incision. If a nasogastric tube was not placed preop-
eratively, then such a tube is placed either immediately
before or after the induction of general anesthesia to de-
compress the stomach. As always, careful inspection of
the abdominal contents is performed, paying particular
attention to the stomach, duodenum, and upper abdomen
for any pathology not suspected in the preoperative eval-
uation. Any suspicious masses should be sent for frozen
section analysis. The dependent portion of the greater
curvature of the stomach is identified and cleared of the
gastrocolic omentum. Typically the gastroepiploic arcade
is divided along the gastric greater curvature (Fig 1).
Some surgeons prefer to create the anastomosis equidis-
tantly between the lesser and greater curvatures of the
stomach, thereby avoiding the need to ligate any vessels
along the greater curvature. The transverse colon with its
attached greater omentum is then retracted cephalad, and
the ligament of Treitz and proximal jejunum are identi-
fied (Fig 2). A gentle loop of proximal jejunum is brought
up in an antecolic fashion without any associated tension
and approximated to the greater curvature of the stomach
where the greater curvature has been exposed. It is appro-
priate to create a relatively short afferent loop, however, it
is important not to make the length of the afferent loop so
short that it compresses the transverse colon. Our pre-
ferred technique of gastrojejunal anastomosis is per-
formed by initially placing a posterior outer layer of 3– 0
silk seromuscular sutures (Fig 3). The gastrojejunal anas-
tomosis is usually made 3 to 4 cm in length. Linear open-
ings are then created within the stomach (as depicted in
Fig 3) and subsequently in the jejunum adjacent to the
posterior outer layer of 3– 0 silk sutures. Next, the inner
layer of the posterior row is performed using a continuous
locking suture of 3– 0 absorbable synthetic material (Fig
4). This suture is then brought anterior as the inner layer
of the anterior row using a Connell technique. After the
inner layer of the anterior row is completed, an outer layer
of the anterior row is placed using interrupted 3– 0 silk
seromuscular sutures. The final configuration of a typical
antecolic gastrojejunostomy is shown in Fig 5. There are
many alternative suturing techniques for performing a
gastrojejunostomy. Additionally, stapling devices such as
a linear stapler or a circular stapler can be used. In a
considerable proportion of patients, an antecolic gastroje-
junostomy does not appear to empty nearly as effectively
as a retrocolic gastrojejunostomy. For this reason, our
preference is to perform a retrocolic gastrojejunostomy in
most patients who require open gastrojejunostomy.
1 The abdomen is entered through an upper midline incision and thoroughly explored. If an antecolic gastrojejunostomy is to be
performed, then the most dependent portion of the gastric greater curvature is identified and cleared of omentum over a distance
of 6 to 8 cm. The gastroepiploic arcade is typically left intact, caudal to the opening in the gastrocolic omentum.

2 The transverse colon is elevated with its omentum at-

tached, and the ligament of Treitz and proximal jejunum are
identified. A mobile loop of proximal jejunum is then
brought up in an antecolic fashion, insuring that the afferent
limb of the gastrojejunostomy does not compress the trans-
verse colon.
76
3 After approximating the gastric greater curvature and
the mobile loop of proximal jejunum, the anastomosis is
commenced. The anastomosis is performed by initially plac-
ing a posterior outer layer of 3– 0 silk Lembert sutures be-
tween the stomach and the jejunum. After insuring that the
stomach has been decompressed via a nasogastric tube, a
linear 3 to 4 cm gastrotomy is created. A parallel opening is
then made in the jejunum, adjacent to the anterior gastrot-
omy.
4 The afferent and efferent limbs of jejunum are controlled
with noncrushing clamps. The inner layer of the posterior row
of the gastrojejunal anastomosis is placed using a continuous
locking suture of 3– 0 synthetic absorbable material. This su-
ture is continued anteriorly as the inner layer of the anterior
row using a Connell stitch. The inner layer of the anterior row
is then completed, and the noncrushing intestinal clamps are
removed.
Sonnenday and Yeo
Retrocolic Gastrojejunostomy. After initial ab-
dominal exploration, the most dependent portion of the
greater curvature of the stomach is cleared of omentum,
as was shown in the initial step of an antecolic gastroje-
junostomy (Fig 1). An opening is then made bluntly in
the transverse mesocolon, avoiding the mesocolic vessels
(which run vertically), and typically a bit to the left of the
midline and to the left of the middle colic vessels. A
proximal loop of jejunum is brought through this rent in
the transverse mesocolon (Fig 6), and the anastomosis is
performed above the mesocolon in a fashion similar to
that of an antecolic gastrojejunostomy (Fig 7). The com-
pleted anastomosis, which lies above the transverse me-
socolon should not be left in this position. The gastroje-
junostomy is then pulled down through the transverse
mesocolon, and the transverse mesocolon is tacked to the
stomach, circumferentially, at least 1 cm above the gas-
trojejunal anastomosis (Fig 8), thereby fixing the jejunal
loop below the mesocolon. This will prevent an obstruc-
tion of the jejunum by the mesocolon.
POSTOPERATIVE CARE
The primary principles of management of patients after
gastrojejunostomy are: (1) ensuring continued adequate
nutrition and hydration; (2) slow resumption of oral in-
take; and (3) anticipation of possible complications in-
 5 The outer layer of the anterior row is com-
pleted using 3– 0 silk Lembert sutures. The fi-
nal anatomy of an antecolic gastrojejunal anas-
tomosis is shown here. There are no mesenteric
defects to close, and no rents in the gastrocolic
omentum or transverse mesocolon to repair.

6 The performance of a retrocolic gastrojejunostomy com-

mences with similar maneuvers as for the antecolic gastrojeju-
nal anastomosis. As shown in Fig 1, the greater omentum is
cleared from the most dependent portion of the gastric greater
curvature for a distance of 6 to 8 cm. Next, an opening is then
made in the transverse mesocolon, usually to the left of the
middle colic vessels, avoiding injury to major vessels within the 7 The completed retrocolic gastrojejunal anastomosis using
mesocolon. A proximal loop of jejunum is brought through the an outer layer of interrupted 3– 0 silk Lembert sutures and an
transverse mesocolon, and the gastrojejunal anastomosis is per- inner layer of running 3– 0 synthetic absorbable suture. It is
formed above the transverse mesocolon, with the stomach and important that the anastomosis not be left in this suprameso-
the jejunum approximated side-by-side. Note that for a retro- colic position, as the narrow opening in the transverse mesoco-
colic gastrojejunostomy, the length of the afferent limb can be lon can serve as an obstruction to the afferent and efferent
shorter than for an antecolic gastrojejunostomy. jejunal limbs.
78
8 The gastrojejunal anastomosis is pulled down through the
transverse mesocolon, and the transverse mesocolon is tacked
to the stomach approximately 1 cm cephalad to the gastrojeju-
nal anastomosis. Using this maneuver, the most dependent
portion of the gastric greater curvature remains dependent, and
the afferent and efferent limbs of the gastrojejunostomy reside
entirely beneath the transverse mesocolon without the risk of
obstruction by the mesocolon itself.
cluding delayed gastric emptying and marginal ulcer for-
mation.
Fluid and electrolyte replacement are standard, with
attention to any preoperative deficits that may require
further supplementation. Parenteral or enteral nutrition
initiated in the preoperative period should be continued
into the postoperative period until adequate oral intake
can be established. Typically, we remove the nasogastric
tube on the morning of the first postoperative day, as long
as the output has not been excessive (greater than 500 to
1000 mL per 24 hours) and as long as the gastric pH is
neutral (ie, acid secretion is suppressed pharmacologi-
cally). One exception to this practice of early nasogastric
tube removal is the patient with a chronically dilated
stomach that might be anticipated to be slow to regain
motility, where high nasogastric outputs are identified; an
option to be considered in these patients is to place a tube
gastrostomy at the time of operation. We do not routinely
administer postoperative antiemetics, preferring to eval-
uate complaints of nausea or emesis at the bedside to
ensure there is no significant gastric dilation. Prokinetic
agents such as metoclopramide or the motilin agonist
erythromycin may be used in these patients to promote
gastric emptying,17 but no prospective randomized trials
are available to prove their effectiveness after gastrojeju-
nostomy alone. We do not routinely administer these
agents.
Sonnenday and Yeo
Our current practice is to offer limited sips of water and
ice to the patient on the first postoperative day, with clear
liquids ad libitum on the second postoperative day, and a
regular diet on the third postoperative day. Many patients
seem to best tolerate several small meals or snacks per day
rather than three large meals. Delayed gastric emptying
(DGE) is the most common complication after gastroje-
junostomy and is of particularly high risk in patients with
chronic gastric outlet obstruction. Patients with advanced
age, diabetes mellitus, hypothyroidism, and autonomic
disorders appear also to be at higher risk for DGE.18,19 A
unique group of patients who appear to be at risk for DGE
include those with an unresectable pancreatic cancer that
is not mechanically obstructing the duodenum but who
manifest symptoms of DGE preoperatively.20 While there
is no definitive therapy for DGE, most patients have a
self-limited course and are slowly able to initiate oral
intake over time. Adjunctive strategies such as prokinetic
agents, limitation of narcotics, and early return of physi-
cal activity may play some role in promoting normal mo-
tility. In patients unable to tolerate oral intake after 7 to 10
days, an upper GI series with diluted barium can evaluate
the anatomy and establish that the anastomosis is patent.
If the gastrojejunal anastomosis is tight or cannot be vi-
sualized by the upper GI series, upper endoscopy by a
skilled endoscopist to evaluate the anastomosis is appro-
priate. In selected patients, it may be appropriate to intu-
bate and gently balloon dilate the anastomosis. However,
these measures are necessary only in the minority of
cases, and patience will allow most patients to resolve
early DGE without invasive intervention.
Gastrojejunostomy has long been recognized as an ul-
cerogenic procedure. The incidence of marginal ulcer-
ation can be reduced by performance of vagotomy (if
appropriate) and possibly by evaluation for and treatment
of H. pylori infection. When vagotomy is not performed,
we believe that most patients should be maintained on
gastric acid suppression therapy, using either histamine
H2-receptor antagonists or proton-pump inhibitors.
REFERENCES
1. Parsa CJ, Organ CH, Barkan H: Changing patterns of resident
operative experience from 1990 to 1997. Arch Surg 135:573–575,
2000
2. Csendes A, Maluenda F, Braghetto I, et al: Prospective randomized
study comparing three surgical techniques for the treatment of
gastric outlet obstruction secondary to duodenal ulcer. Am J Surg
166:45– 49, 1993
3. Jaffin BW, Kaye MD: The prognosis of gastric outlet obstruction.
Ann Surg 201:176 –179, 1985
4. Millat B, Fingerhut A, Borie F: Surgical treatment of complicated
duodenal ulcers: controlled trials. World J Surg 24:299 –306, 2000
5. DiSario JA, Fennerty MB, Tietze CC, et al: Endoscopic balloon
dilatation for ulcer induced gastric outlet obstruction. Am J Gas-
troenterol 89:868 – 871, 1994
6. Lau JYW, Chung SCS, Sung JJY, et al: Through the scope balloon
dilatation for pyloric stenosis: long term results. Gastrointest En-
dosc 43:98 –102, 1996
Open Gastrojejunostomy
7. Kennedy T, Johnston GW, Love AH, et al: Pyloroplasty versus
gastrojejunostomy. Results of a double-blind, randomized trial.
Br J Surg 60:949 –953, 1973
8. Aranha GV, Prinz RA, Greenlee HB, et al: Gastric outlet and
duodenal obstruction from inflammatory pancreatic disease. Arch
Surg 119:833– 835, 1984
9. Prinz RA, Aranha GV, Greenlee HB: Combined pancreatic duct
and upper gastrointestinal and biliary tract drainage in chronic
pancreatitis. Arch Surg 120:361–366, 1985
10. Delaney CP, Fazio VW: Crohn’s disease of the small bowel. Surg
Clin North Am 81:137–158, 2001
11. Nugent FW, Roy MA: Duodenal Crohn’s disease: an analysis of 89
cases. Am J Gastroenterol 84:249 –254, 1989
12. Shellito PC, Warshaw AL: Idiopathic intermittent gastroparesis
and its surgical alleviation. Am J Surg 148:408 – 412, 1984
13. Schirmer BD: Mechanical and motility disorders of the stomach
and duodenum. In GD Zuidema, CJ Yeo (Eds): Shackelford’s Sur-
gery of the Alimentary Tract (5th ed). Philadelphia: WB Saunders,
2002, pp 178 –184
79
14. Behrns KE, Sarr MG: Diagnosis and management of gastric emp-
tying disorders. Adv Surg 27:233–255, 1994
15. Guy RJ, Dawson JL, Garrett JR, et al: Diabetic gastroparesis from
autonomic neuropathy: surgical considerations and changes in
vagus nerve morphology. J Neurol Neurosurg Psychiatry 47:686 –
691, 1984
16. Cullen JJ, Kelly KA: Gastric motor physiology and pathophysiol-
ogy. Surg Clin North Am 73:1145–1160, 1993
17. Yeo CJ, Barry MK, Sauter PK, et al: Erythromycin accelerates gastric
emptying following pancreaticoduodenectomy: a prospective, ran-
domized placebo controlled trial. Ann Surg 218:229 –238, 1993
18. Kung SP, Lui WY, P’Eng FK: An analysis of the possible factors
contributing to the delayed return of gastric emptying after gas-
trojejunostomy. Surg Today 25:911–915, 1995
19. Woods SD, Mitchell GJ: Delayed return of gastric emptying after
gastroenterostomy. Br J Surg 76:145–148, 1989
20. Sarr MG, Gladen HE, Beart RW Jr., van Heerden JA: Role of
gastroenterostomy in patients with unresectable carcinoma of the
pancreas. Surg Gynecol Obstet 152:597– 600, 1981
Open Truncal Vagotomy
David W. Mercer, MD, and Terrence H. Liu, MD
espite advances in medical therapy to inhibit acid
D secretion and eradicate Helicobacter pylori, surgery
remains important in managing patients suffering from
the complications of peptic ulcer disease. Over the last
two decades there has been an increase in emergency
operations performed for complications of peptic ulcers
while the number of operations for elective indications
has decreased markedly.1,2 Moreover, Taylor3 reported
that despite the introduction of effective antiulcer medi-
cations, the mortality of peptic ulcer disease has remained
stable. Furthermore, peptic ulcers have a high recurrence
rate after discontinuing medical therapy.4 Thus, there has
been a renewed interest in operative treatment of peptic
ulcer disease. However, the type of operation performed
for peptic ulcer disease has changed in the Helicobacter
pylori era.5,6 Specifically, gastric resections are less fre-
quently used, and some type of vagotomy procedure with
or without drainage seems most effective. Consequently,
the purpose of this chapter is to describe techniques for
performance of truncal vagotomy (total abdominal vagot-
omy), review its indications, and discuss recurrences and
side effects after vagotomy. However, before reviewing
truncal vagotomy, the understanding of the pathophysi-
ology of peptic ulcer disease will be discussed briefly.
PATHOPHYSIOLOGY OF PEPTIC ULCER
DISEASE
Benign Gastric Ulcer. In contrast to duodenal ulcer
disease, there has been no reduction in hospitalization
rates for patients with benign gastric ulcer over the last
several decades.7 Operations for both elective and urgent
indications have likewise remained unchanged. However,
hospitalizations and operations may be increasing
slightly as a result of greater use and abuse of nonsteroidal
anti-inflammatory drugs (NSAIDS), particularly in
women in their later years.8,9 Over 20% of patients will
develop a gastric ulcer within three months of beginning
therapeutic doses of NSAIDS and, of these, approximately
From the Department of Surgery, University of Texas Houston Medical School,
HoustonTX.
Address reprint requests to Dr. David W. Mercer, Department of Surgery,
University of Texas Houston Medical School, 5656 Kelley St, Houston, TX 77026.
© 2003 Elsevier Inc. All rights reserved.
1524-153X/03/0502-0012$30.00/0
doi:10.1053/otgn.2003.35362
80 Operative Techniques in General Surgery, Vol 5, No 2 (June), 2003: pp 80-85
30% will present with an acute complication, principally
bleeding.10
Type 1 gastric ulcers, the most common form, are lo-
cated on the greater curvature at or proximal to the inci-
sura. Hypersecretion of gastric acid is rarely if ever en-
countered with this type of ulcer, but acid appears to play
a permissive role in ulcer development and accentuates
progression once it occurs. Evidence also suggests that
Helicobacter pylori plays a role in its pathogenesis, al-
though the etiologic role is not as strong as with duodenal
ulcer. In contrast, the Type 2 ulcer occurs in the same
location as the Type 1 lesion but is associated with duo-
denal ulcer disease and excessive acid secretion. Simi-
larly, a Type 3 gastric ulcer is located within 2 cm of the
pylorus (ie, prepyloric) and is also associated with excess
acid secretion. The fourth type of gastric ulcer is located
within 2 cm of the gastroesophageal junction and tends to
be associated with hypochlorhydria; these ulcers are rare
but are associated with considerable operative mortality.
Duodenal Ulcer. While hospitalizations and espe-
cially elective operations for duodenal ulcer disease have
decreased dramatically over the past three decades,7 ur-
gent operations appear to be increasing. Because patients
operated currently are older than before, morbidity and
mortality are increased with such operations.1,11 Duode-
nal ulcer disease is a disease of multiple etiologies.12 The
only absolute requirements are secretion of acid and pep-
sin in conjunction with either Helicobacter pylori infec-
tion or ingestion of NSAIDS. Gastric acid secretory rates
are usually increased in patients with duodenal ulcer dis-
ease,13 in large part related directly to gastric antral infes-
tation with Helicobacter pylori, particularly in patients
resistant to or developing recurrence after standard anti-
secretory therapy.12 Interestingly, complete eradication
of the organism results in recurrence rates of only 2%.4 In
contrast, the pathogenesis of duodenal ulcer disease asso-
ciated with NSAID use is less well understood but appears
to involve an interplay between acid and mucosal injury
secondary to the NSAID use.
ANATOMY AND PHYSIOLOGY OF THE
VAGUS NERVE
The vagus nerve arises from the vagal nucleus in the floor
of the fourth ventricle, traverses the neck in the carotid
sheath, and enters the mediastinum where it divides into
several branches around the esophagus. These branches
coalesce above the esophageal hiatus to form separate left
Open Truncal Vagotomy
1 As shown, truncal vagotomy requires division rostral to the
celiac and hepatic branches of the vagus nerve, which arise
from the right (posterior) and left (anterior) divisions of the
vagus nerve, respectively.
and right vagus nerves (see Fig 1). Quite commonly, there
are more than two vagal trunks in the region of the distal
esophagus.14 At the level of the gastroesophageal (GE)
junction, the left vagus nerve lies anterior and the right
vagus posterior. The left vagus gives off a branch to the
liver and then continues along the lesser curvature as the
anterior nerve of Latarget. “The criminal” nerve of Grassi,
a branch of the posterior vagus nerve, has been recog-
nized in the etiology of recurrent ulcer when left undi-
vided (see Fig 8).15 The right vagus nerve also gives off a
branch to the celiac plexus and then continues posteriorly
along the lesser curvature.
Vagotomy effectively inhibits the cephalic phase of gas-
tric acid secretion and reduces the sensitivity of the pari-
etal cells to gastrin and histamine by 65 to 70%.16 But, in
addition to inhibiting acid secretion, vagotomy also
causes loss of receptive relaxation of the proximal stom-
ach which speeds gastric emptying of liquids, and by
causing loss of antral pump function, slows gastric emp-
tying of solids. The effect of vagotomy on dividing the
90% of vagal fibers that are afferent remains unclear.
INDICATIONS FOR TRUNCAL
VAGOTOMY
Dragstedt introduced gastric vagotomy by a transthoracic
approach in 1943 for the treatment of duodenal ulcer
81
disease.17 In 1945, he developed the first surgical treat-
ment for peptic ulcers, ie, truncal vagotomy and pyloro-
plasty.18 Today, truncal vagotomy is probably the most
common operation performed for duodenal ulcer disease.
The principal indications are as follows: (1) uncompli-
cated duodenal and prepyloric ulcers; (2) emergency pro-
cedures for bleeding or perforated peptic ulcers; (3) py-
loric stenosis requiring a drainage operation; (4)
recurrent ulceration after gastrectomy or proximal gastric
vagotomy; (5) some gastric ulcers; and (6) possibly hem-
orrhagic gastritis. Elective surgical intervention may be
indicated for: (1) the rare patient with disease resistant to
medical therapy, despite compliance for two years, or
patients who have had two or more proven recurrences
after medical therapy; (2) patients who cannot be fol-
lowed regularly because of geographic or socioeconomic
reasons or who cannot afford antiulcer medication; and
(3) patients who have had complications such as perfora-
tions or hemorrhage.
PREOPERATIVE PREPARATION
Preoperative evaluation includes assessment of comorbid
conditions and risk factors as well as evaluation of the
underlying peptic ulcer disease with endoscopy and, in
selected patients, secretory analysis and serum gastrin
levels (elective procedures). After endoscopy to localize
the ulcer, secretory tests can be done to evaluate basal
acid output as well as peak acid output after stimulation
with pentagastrin. More precise results regarding acid
secretion can be obtained by utilizing 24 hour pH stud-
ies.19 These tests enable one to assess the degree of hy-
peracidity and to identify deviations in the normal circa-
dian rhythm of acid secretion in patients refractory to
medical therapy. These tests can also be used to confirm a
decrease in acid secretion postoperatively after truncal
vagotomy. Depending on clinical presentation, a serum
gastrin level probably should be obtained to exclude a
gastrinoma.
INTRAOPERATIVE TESTING
While we do not perform intraoperative testing to assess
for completeness of vagotomy, some surgeons have advo-
cated intraoperative testing. The Grassi or pH test is based
on the premise that denervated parietal cells are less sen-
sitive to circulating pentagastrin than innervated cells.16
The stomach is washed through a small gastrotomy, and a
probe topically maps out the local mucosal pH during
intravenous infusion of pentagastrin (6 ␮g/kg/hr). A pH
less than 5.0 suggests residual vagal innervation.
DRAINAGE
Most surgeons employ some form of a drainage procedure
in association with truncal vagotomy. Drainage proce-
82 Mercer and Liu

TRUNCAL VAGOTOMY TECHNIQUES

2 Exposure to esophageal hiatus: A

midline incision extending cephalad
along the side of the xiphisternum is
used. To expose the esophageal hiatus
and to facilitate dissection under direct
vision, we utilize a Thompson Retractor
to elevate the sternocostal margin. The
left triangular ligament of the liver is
divided, as is the falciform ligament, to
diminish the likelihood of the retractor
tearing the liver.

3 Incision of the phrenoesophageal ligament: After exposure

of the esophageal hiatus, the phrenoesophageal ligament is
incised transversely to expose the esophagus. The anterior va-
gus nerve will usually be seen as a single trunk and is easily
identified. To insure complete division of the anterior vagus
nerves and its associated branches, the periesophageal space
must be entered and all periesophageal neural and adventitial
tissue divided for at least 5 cm on the esophagus.
4 Development of periesophageal space: To allow for the safe
mobilization of the esophagus as well as the precise identifica-
tion of both anterior and posterior vagal nerves, the periesoph-
ageal space must be developed. The left index finger develops
this space which can be exposed by gentle digital dissection
beginning along the right anterolateral aspect of the esophagus.
Open Truncal Vagotomy
5 Division of the anterior vagus nerve: all anterior vagal
nerves are clipped above and below and the intervening seg-
ment excised; we send all specimens for pathologic confirma-
tion. Clips are used to identify the performance of a vagotomy.
6 Esophageal mobilization and division of posterior vagus
nerve: after division of the anterior vagal nerves, the esophagus
is completely mobilized. The right index finger passes into the
space to the left of the esophagus, palpating the aorta posteri-
orly. The index finger passes toward the right elevating the
right posterolateral attachment of the esophagus. This attach-
ment is sharply incised parallel to the esophagus to create a
window. A one-inch Penrose drain is placed around the esoph-
agus to allow for safe traction on the gastroesophageal junction
to facilitate identification of posterior vagal nerve fibers. The
posterior vagus nerve and all branches are clipped above and
below and the intervening segment excised under direct vision.
83
dures can be classified as a form of pyloroplasty or gastro-
enterostomy. In general, there is little difference in the
side effects associated with the various operations. Bile
reflux is more common after gastroenterostomy, diarrhea
is more common after pyloroplasty,22 and the incidence
of dumping is the same for both.
RECURRENT ULCERATION
The rate of recurrent duodenal ulcer in long term (⬎10
years) follow-up varies with the type of operation per-
formed in association with vagotomy. For vagotomy and
gastroenterostomy, the recurrent ulcer rate was 4%; for
vagotomy and antrectomy ⬍1%, for subtotal gastrectomy
(without vagotomy) 2%, and for vagotomy and pyloro-
plasty 10 to 15%.23 Thus, concomitant gastrectomy gave a
lower recurrent ulcer rate. However, although associated
with a lower recurrence rate, gastrectomy also obligates a
greater morbidity, and its role in the Helicobacter pylori
era remains to be determined.5
Recurrent ulcers usually result from a technically inad-
equate operation.24 After truncal (“total abdominal”)va-
gotomy, remnant ulcers usually occur as a result of an
incomplete vagotomy. However, other causes need to be
excluded such as gastrinoma, hypercalcemia, smoking,
abuse of NSAIDS, persistent Helicobacter pylori infection,
and carcinoma. Of all recurrent ulcers, about 95% occur
after surgical intervention for duodenal ulcer, whereas
only 2% to 4% occur after operations for gastric ulcer.24 In
fact, the risk of a recurrent ulcer after gastrectomy for
gastric ulcer proximal to the prepyloric area is only about
2%.
POSTOPERATIVE TESTING OF ACID
SECRETION
Because this chapter focuses on truncal vagotomy, assess-
ment for the completeness of vagotomy postoperatively
will be reviewed briefly. The Hollander test is often used
to assess the completeness of vagotomy.23 This test in-
volves intravenous infusion of insulin to induce hypogly-
cemia, a potent vagal stimulus. Modifications of this test
by Ross and Kay25 allowed classification of individuals
with an incomplete vagotomy as adequate or inadequate.
Individuals with a small but delayed response to insulin
generally have adequate protection against recurrent ul-
ceration; in contrast, a large and early response in acid
secretion suggests inadequate protection against recur-
rence. If the vagotomy is incomplete but has an adequate
response to insulin, a search for an alternative etiology for
the recurrent ulcer should be undertaken. If the patient
has an incomplete vagotomy with an inadequate re-
sponse, operative intervention to “complete the vagot-
omy” is warranted. Although a transabdominal approach
can be undertaken, the transthoracic approach, either
open or thorascopically in a setting of ulcer recurrence, is
easier to complete the vagotomy and probably has less
84 Mercer and Liu

7 If the posterior vagus nerve cannot

be identified, the left index finger
should be placed on the celiac axis with
traction applied caudally. The right
hand is placed behind the esophagus
and an attempt is made to palpate the
posterior vagus nerve with the posterior
aspect of the periesophageal tissues
“stretched.” The posterior vagus can be
identified 1 to 2 cm to the right and
often 1 to 2 cm posterior to the esoph-
agus, on occasion intimately associated
with the right crus of the diaphragm.

morbidity. The patient is placed with the right side down
allowing access to the left chest via a left posterolateral
position. The left lung can be collapsed with a Carlan’s
tube. Four trocars are inserted in a triangular pattern with
the thorascope between the two operating trocars. The
fourth port is used for irrigation and suctioning. The
pleura is incised anterior to the aorta to allow dissection
of the esophagus. Moving the nasogastric tube allows
localization of the esophagus. After dissecting the esoph-
agus circumferentially, an atraumatic hook is utilized to
dissect the vagal nerves and divide them. Use of electro-
cautery near the esophagus should be minimized to avoid
risk of postoperative esophageal necrosis. A complete
search for all other periesophageal nerve fibers is advo-

8 Criminal nerve of Grassi: the most common technical error is to miss the posterior trunk completely by dissecting to close to
the esophagus.20 Nyhus and coworkers have published important details of technique as well as commonly encountered anatomic
variations.21 The branch of the posterior vagus nerve highlighted in this figure is Grassi’s criminal nerve15 which innervates the
proximal greater curvature of the stomach. Although this branch has received the most attention, other branches are equally
important and can arise from the gastric truncal divisions or from the trunks alongside the esophagus.
Open Truncal Vagotomy
cated. Inserting a thoracic drain through one of the ports
and closing the other ports completes the operation. Be-
fore performing this operation, however, the patient
should be tested for Helicobacter pylori, treated if positive,
and its eradication confirmed, especially if it is a recurrent
duodenal ulcer.
POSTOPERATIVE SIDE EFFECTS OF
TRUNCAL VAGOTOMY
After truncal vagotomy and drainage, alcohol is absorbed
more rapidly after ingestion due to more rapid gastric
emptying of liquids. Patients should therefore be warned
to be careful with alcohol intake. Vagal denervation also
changes emptying of solids. Denervation of the antrum
causes loss of coordinated antropyloric pump function
with slowed emptying of solids. Also, by reducing bile
flow, increasing gallbladder volume, delaying gallbladder
emptying, and reducing the response to cholecystoki-
nin,26 vagotomy increases the risk of gallstone develop-
ment. In addition, the usual post gastrectomy syndromes
can exist, including early and late dumping syndrome,
afferent or efferent loop syndrome, alkaline reflux gastri-
tis, gastroparesis, and post vagotomy diarrhea.17 These
are all well described in the references provided.
CONCLUSION
Bilateral truncal (total abdominal) vagotomy combined
with a drainage procedure gives an acceptable incidence
of side effects, a low mortality rate, and an acceptable
recurrent ulceration rate. Other operations either have
higher rates of recurrence with fewer side effects or more
side effects but fewer recurrences. However, the appropri-
ate role of vagotomy in the Helicobacter pylori era contin-
ues to evolve, as less aggressive procedures combined
with eradication of the infection seem to be better for the
patient than the more aggressive operative approaches.
REFERENCES
1. Susser M: Period effects, generation effects and age effects in
peptic ulcer mortality. J Chronic Dis 35:29-40, 1982
2. Sonnenberg A: Costs of medical and surgical treatment of duode-
nal ulcer. Gastroenterology 96:1445-1452, 1989
3. Taylor TV: Deaths from peptic ulceration. Br Med J. 291:653-654,
1985
4. Tytgat GNJ: Treatments that impact favorably upon the eradica-
tion of Helicobacter pylori and ulcer recurrence. Aliment Pharma-
col Ther 8:359-368, 1994
5. Dempsey D, Ashley S, Mercer D, Sillin L: Peptic ulcer surgery in
the H. pylori era: Part 2: indications for operation. Contemporary
Surgery 57:433-441, 2001
6. Ng EK, Lam YH, Sung JJ, et al: Eradication of Helicobacter pylori
prevents recurrence of ulcer after simple closure of duodenal ulcer
85
perforation: randomized controlled trial. Ann Surg. 231:153-158,
2000
7. Sonnenberg A: Changes in physician visits for gastric and duode-
nal ulcer in the United States during 1958-1984 as shown by
National Disease and Therapeutic Index (NDTI). Dig Dis Sci 32:
1-7, 1987
8. Armstrong CP, Blower AL: Non-steroidal anti-inflammatory
drugs and life threatening complications of peptic ulceration. Gut
28:527-532, 1987
9. Soll AH, Kurata J, McGuigan JE: Ulcers, nonsteroidal anti-inflam-
matory drugs and related matters. Gastroenterology 96:561-568,
1989
10. Fries JF, Miller SR, Spitz PW, Williams CA, Hubert HB, Block DA:
Toward an epidemiology of gastropathy associated with nonste-
roidal anti-inflammatory drug use. Gastroenterology 96:647-655,
1989
11. Griffin MR, Ray WA, Schaffner W: Nonsteroidal anti-inflamma-
tory drug use and death from peptic ulcer in elderly persons. Ann
Intern Med 109:359-363, 1988
12. Soll AH: Pathogenesis of peptic ulcer and implications for treat-
ment. New Engl J Med 322:909-916, 1990
13. Cox AJ: Stomach size and its relation to chronic peptic ulcer. Arch
Pathol 54:407-422, 1952
14. Gray SW and Skandalakis DA: Atlas of surgical anatomy for gen-
eral surgeons. Baltimore, MD: Williams and Wilkins, 1985
15. Grassi G: Highly selective vagotomy with intra-operative acid
secretive test of completeness of vagal section. Surg Gynecol Ob-
stet 140:259, 1975
16. Christiansen J, Jensen HE, Ejby-Poulsen P, et al: Prospective con-
trolled vagotomy trial for duodenal ulcer: Primary results, se-
quelae, acid secretion, and recurrence rates two to five years after
operation. Ann Surg 193:49, 1981
17. Dragstedt LR: Gastric vagotomy in the treatment of duodenal
ulcer. In L Nyhus, C Wastell (Eds): Surgery of the stomach and
duodenum (3rd ed). Boston: Little Brown, 1977
18. Dragstedt LR: Section of the vagus nerves to the stomach in the
treatment of peptic ulcer. Ann Surg. 126:687-708, 1947
19. Fuchs KH, Sech A, Frey SM, DeMester TR: Gastric acid secretion
and gastric pH measurement in peptic ulcer disease. Probl Gen
Surg. 9:138-151, 1992
20. Taylor TV, Pearson KW, Torrance HB: Revagotomy for recurrent
peptic ulceration. Br J Surg 64:477, 1977
21. Nyhus LM, Donahue PE, Krystosek RJ, et al: Complete vagotomy:
The evolution of an effective technique. Arch Surg 115:264, 1980
22. Goligher JC, Pulvertaft CN, Watkinson G: Controlled trial of
vagotomy and antrectomy and subtotal gastrectomy in elective
treatment of duodenal ulcer. Br Med J 1:455, 1964
23. Goligher JC, Feather DB, Hall R, et al: Several standard elective
operations for duodenal ulcer: Ten to sixteen year clinical results.
Ann Surg 189:18, 1979
24. Beahrs OH: Surgical management of peptic ulceration recurring
postoperatively. Surg Clin North Am 51:879, 1971
25. Ross B, Kay AW: The insulin test after vagotomy. Gastroenterol-
ogy 46:1379, 1964
26. Torrealba V: Truncal vagotomy and cholelithiasis: Plea for a con-
trolled study. Br Med J 1:1160, 1977
27. Miller TA, Mercer DW: Derangements in gastric function second-
ary to previous surgery. In: Miller TA, (eds). Modern surgical care:
physiologic foundations and clinical applications. St. Louis, MO:
Quality Medical Publishing Inc., 1998, pp 398-409
Antrectomy
Daniel T. Dempsey, MD, and Abhijit Pathak, MD
he antrum serves important hormonal and motor
T functions. The antral mucosa is the primary source of
gastrin, a major stimulus for acid production by the pari-
etal cells. Antral gastrin production is shut off by lumenal
acid. When doing an antrectomy, it is particularly impor-
tant to be certain that all antral tissue is removed distally
when planning a Billroth II or Roux reconstruction. In
these situations, antral tissue left on the duodenal stump
will secrete gastrin unremittingly because it will no longer
be exposed to acid. This syndrome of “retained antrum” is
an important cause of hypergastrinemia and acid hyper-
secretion in the remaining stomach. Recurrent or mar-
ginal ulceration may result.
The antrum also plays a major role in the trituration
and subsequent gastric emptying of solid foods. Antral
resection is generally associated with accelerated gastric
emptying of liquids, and may be associated with delayed
emptying of solids. Symptoms of dumping may or may
not be present. Antrectomy refers to the resection of the
antrum and the pylorus and usually involves removing at
least 30% of the stomach. Distal or hemigastrectomy gen-
erally refers to resection of about 50% of the caudal stom-
ach, including the antrum and pylorus. These operations
are functionally identical, except in the latter in which the
proximal margin of resection is extended cephalad.
PATIENT SELECTION
The most common benign indications for antrectomy or
distal gastrectomy are duodenal ulcer, gastric ulcer, and
large benign gastric tumors (leiomyoma, lipoma). Other
indications include perforation or stricture from trauma
or caustic ingestion. Relative contraindications include
cirrhosis, extensive scarring of the proximal duodenum
(obligating a difficult or tenuous duodenal closure), and
previous operation on the proximal duodenum (eg, cho-
ledochoduodenostomy).
DUODENAL ULCER
In selected patients, antrectomy is still a good operation
for intractable or obstructing duodenal ulcer disease. A
From the Department of Surgery, Temple University, Philadelphia, PA.
Address reprint requests to Dr. Daniel T. Dempsey, 301 South 8th Street,
Philadelphia, PA 19107.
© 2003 Elsevier Inc. All rights reserved.
1524-153X/03/0502-0011$30.00/0
doi:10.1053/otgn.2003.35361
86
truncal or selective vagotomy is added. The advantage of
the latter is that the extragastric visceral innervation is
preserved and long term side effects such as gallstones
and postvagotomy diarrhea may be less.
It must be recognized that operative intervention for
intractable duodenal ulcer disease should be very unusual
nowadays, since over 95% of duodenal ulcers are known
to be associated with Heliobacter pylori infection,
NSAIDs, or gastrinoma. If these causes have been ex-
cluded or treated appropriately, the surgeon should be
wary of the patient referred for “intractable duodenal
ulcer.” This patient may be quite different from the large
number of patients in the surgical literature who are
known to have done well with vagotomy and antrectomy
in the era before effective pharmacologic acid suppression
(eg, cimetidine) and before recognition of the role of H.
pylori. We feel strongly that if vagotomy and antrectomy
are applied indiscriminately today as the operation of
choice for intractable duodenal ulcer, the incidence of
poor surgical outcomes (Visick 3 and 4) will be much
higher than predicted in the existing literature. Medical
therapy has healed all the easy patients, and the ones who
are referred for elective surgery for intractable or non-
healing duodenal ulcer currently are the difficult patients
who may not do well with gastric resection; in this pop-
ulation, nonresective procedures should be strongly con-
sidered.
Most patients hospitalized today for gastric outlet ob-
struction from duodenal ulcer disease have H. pylori in-
fection that should be treated. Nonetheless, most such
patients will require some invasive treatment for their
obstruction. In the low-risk surgical patient with ob-
structing duodenal ulcer disease, antrectomy and selec-
tive vagotomy is a good operation. This is especially true
in the noncompliant patient population in whom eradi-
cation of H. pylori and chronic treatment with antisecre-
tory medication (if necessary for recurrent ulcer) is less
likely to be successful. Results of highly selective vagot-
omy and drainage for obstructing duodenal ulcer have
been good, but recurrent or marginal ulceration is more
common than after antrectomy.
GASTRIC ULCER
Distal gastrectomy (which also includes proximal resec-
tion) to include the gastric ulcer remains the elective
operation of choice for gastric ulcer. The same admoni-
tions discussed above regarding intractability apply. But
Operative Techniques in General Surgery, Vol 5, No 2 (June), 2003: pp 86 –100
Antrectomy
unlike duodenal ulcer, with a nonhealing gastric ulcer,
the specter of malignancy lurks and makes long-term
nonsurgical management less appealing. Resection is also
an attractive option in many patients with a stable gastric
ulcer who require urgent operation for bleeding or perfo-
ration; gastrectomy gets rid of the ulcer, which may be
malignant in 5 to 10% of such patients. In contrast, resec-
tion is less commonly done today in the emergency set-
ting for perforation or bleeding from duodenal ulcer.
87
BENIGN TUMORS
Wedge resection with negative margins is adequate treat-
ment for benign gastric tumors such as leiomyoma (stro-
mal tumors), lipoma, or adenoma. But when these benign
neoplasms are large and located in the antrum or prepy-
loric region, formal distal gastrectomy is often more prac-
tical, because it avoids excessive lumenal narrowing. Fro-
zen section should always be obtained to assure negative
margins and benignity.

OPERATIVE TECHNIQUE

1 (A) We routinely use an upper midline incision and a mechanical retractor. An extended right subcostal incision also gives
good exposure. (B) A minimal access laparoscopic approach usually requires at least 4 ports. Initially these may all be 5-mm ports
with a 5-mm 30 degree scope in the umbilical site (a) and the 5-mm harmonic scalpel and 5 mm grasper in the paramedian ports
(b,c). A 5-mm liver retractor can be placed through the 5-mm subxiphoid port (d) or through a right flank port (e). Once
laparoscopic resection is deemed feasible, usually two sites are changed to 12-mm ports to accommodate the laparoscopic stapler.
One of these sites can be enlarged eventually to remove the specimen.
88 Dempsey and Pathak

2 The antrum includes at least the distal 30% of the stomach. Its proximal border is usually thought of as an imaginary line from
the angularis incisura (a) on the lesser curvature side over to intersect at right angles with the greater curvature side (b); however,
a tongue of antral mucosa usually extends 2 to 3 cm proximally up the lesser curvature. Proximal to this line is the parietal
cell-containing body of the stomach, and distal to the line is the gastrin secreting G-cell containing antral mucosa. Anatomically, the
distal border of the antrum is the pyloric ring usually marked by the vein of Mayo. When done for peptic ulcer disease, the pylorus
should always be resected as part of an antrectomy, defined as resection of the stomach distal to line a-b. Resection from line c-d is
termed distal gastric resection or hemigastrectomy, and from line e-f subtotal gastrectomy.
Antrectomy 89

3 On entering the abdomen, exploration is done. The stomach is decompressed with a nasogastric tube. The nasogastric tube is
then pulled back into the proximal stomach so that it will not be caught in the stapling instruments used to transect the stomach.
It is important to reassess the original plan for antrectomy based on the findings at operation. A high “juxta-cardiac” gastric ulcer
or marked duodenal inflammation might be better treated with nonresective procedures. In these situations, resection of the ulcer
itself, without a vagotomy, is the procedure of choice. Note the presence of a large aberrant “replaced” left hepatic artery originating
from the left gastric artery as occurs in 10% to 20% of patients.
90 Dempsey and Pathak

4 The avascular gastrohepatic ligament (lesser omentum) is opened and the lesser sac entered at about the level of the incisura
angularis. If the posterior stomach is free, the fingers of the operator’s left hand can push up on the greater curvature attachments,
facilitating entry into the lesser sac from this side of the stomach. Alternatively, if the lesser sac is obliterated or if the lesser curvature
is thickened or inflamed, the lesser sac is entered initially from the greater curvature side of the stomach, usually below the spleen
at about the midpoint of the lesser curvature. The right gastroepiploic vessels are divided and ligated here, and a Penrose drain or
catheter passed around the mid-stomach for manipulation. It is often easier to free the greater gastric curvature down toward the
pylorus by staying just outside the right gastroepiploic vessels in the gastrocolic ligament, ligating small omental branches (x).
Often, the lesser curvature dissection can be facilitated by palpating the plane between the lesser curvature and the liver with the aid
of the surgeon’s nondominant hand in the lesser sac, as introduced from the greater curvature side. Occasionally such an approach
will render the omentum ischemic, necessitating resection. The surgeon should pay attention to the location of the middle colic
vessels during this dissection, because inflammation and scarring especially in the obese patient may predispose these vessels to
injury. Alternatively, the distal greater curvature may be mobilized by staying inside the gastroepiploic arcade, ligating the multiple
small branches to the stomach (o).
Antrectomy 91

5 (A) As the distal greater curvature is freed,

the stomach is elevated and the antrum freed
from the pancreas and mesocolon. The right
gastric artery is identified and ligated usually
cephalad and to the right of the pylorus. The
pylorus is gently dissected off the capsule of the
pancreas. Several small vessels may require li-
gation. If the pylorus can be freed only with
great difficulty, and if cancer has clearly been
ruled out, a nonresective operation should be
considered. Generally it is necessary to free 1 or
2 cm of proximal duodenum distal to the mus-
cular pyloric ring to perform an adequate resec-
tion and duodenal closure; this also assures
total resection of antral mucosa distally. The
right gastroepiploic pedicle is doubly ligated
near its origin. We have also used a linear gas-
trointestinal stapler with a vascular load for
this maneuver. Clips are avoided because they
may foil the smooth firing of the stapler used to
divide the duodenum. Losing control of the
right gastroepiploic vein near its origin can re-
sult in troublesome bleeding and/or an impres-
sive mesenteric hematoma, because it usually
drains into the superior mesenteric vein
(SMV). Alternatively the gastroepiploic artery
and vein can be divided on the distal stomach
and dissected away from the pylorus and duo-
denal bulb. (B) The proximal duodenum is
transected with the stapler, and the duodenal
staple line inspected for integrity and hemosta-
sis. At the proposed proximal line of resection,
the neurovascular bundle on the lesser curva-
ture is encircled, divided and ligated. In the
obese patient or if there is inflammation, a sta-
pler with a vascular load may be useful here. A
linear stapler with the 4.8 mm staples (“green
cartridge”) is often most appropriate to
transect the stomach. The specimen is re-
moved, opened and marked, and sent to pa-
thology. We usually request frozen section
confirmation that the distal margin is indeed
duodenum.
92 Dempsey and Pathak

6 If it is easy to free an additional 1 to 2 cm of proximal duodenum from the pancreas, and if a generous Kocher maneuver easily
brings this proximal duodenum to the midline, a Billroth I gastroduodenostomy may be contemplated. We generally do this as a
two-layer, end-to-end anastomosis to the transected stomach on the greater curvature side (A). But if this does not “lay right,” we
do not hesitate to use any part of the transected stomach for the anastomosis. That part of the transected stomach not included in
the anastomosis is oversewn with interrupted silk seromuscular Lembert sutures (B). The tip of the nasogastric tube is positioned
3 to 4 cm proximal to the anastomosis (C).
Antrectomy 93

6 (continued).
94 Dempsey and Pathak

7 More commonly, we perform a 2-layer antecolic isoperistaltic Billroth II gastrojejunal anastomosis to the greater curvature
aspect of the transected stomach (Hofmeister modification) (A). The lesser curvature end of stomach not included in the
gastrojejunostomy is oversewn with Lembert sutures. Alternatively, the entire end of the stomach can be anastomosed to the side
of antecolic jejunum (Polya reconstruction) (B). We usually cut additional holes in the nasogastric tube, and position the tip several
cm past the anastomosis into the afferent limb (ie, toward the duodenum). The additional holes ensure that there is both gastric and
afferent limb decompression.). We do not use a Roux-en-Y limb as primary reconstruction after distal gastrectomy unless the gastric
remnant is 30% or less. However, small gastric remnants predispose to bile reflux esophagitis when reconstructed with a loop-type
gastrojejunostomy, and Roux reconstruction is then most appropriate.
Antrectomy 95

7 (continued)
96 Dempsey and Pathak

8 The duodenal stump is inspected again and the staple line gently wiped with a sponge to ensure hemostasis. If necessary, gentle
cautery or a discreet suture may be placed distal to the staples. We do not routinely oversew the duodenal staple line or turn in the
staple line with a second layer of sutures. The closed end of the transected duodenum is covered with well-vascularized, tension-free
omentum or pericolic fat held in place with two or three carefully placed sutures. A closed suction drain is placed in Morrison’s
pouch (i.e. not near the staple or suture lines).
Antrectomy 97

9 If possible, the surgeon should avoid distal gastric resection if it looks like the duodenum will be difficult to close. But when
faced with a difficult duodenal stump, the surgeon should be guided by the principles of closure, decompression, and drainage.
1. Closure: If the proximal end of the duodenum is problematic, a gastroduodenostomy (Billroth I) is contraindicated. A single
application of a linear stapler is adequate closure of the duodenal stump. This stump is buttressed with a vascularized omental flap.
Inversion or imbrication is unnecessary and in our opinion mettlesome. If it is not possible to free the posterior wall of the
duodenum from the pancreas, then the anterior duodenal wall is carefully preserved and its lip can be sutured to the posterior
duodenal wall in situ (a to b). The location of the common bile duct and its entrance to the duodenum should be ascertained in this
situation. If a posterior ulcer has destroyed the posterior wall of the proximal duodenum and scarified the pancreatic capsule, then
the anterior lip of the duodenal wall can be sutured to this posterior scar with interrupted suture (a and b). Seromuscular bites
anteriorly will “dunk” the mucosa. In our opinion, if the duodenal stump can be sutured to a Roux-en-Y limb, then it can be closed
primarily and this more complex option avoided. Closure of the stump around a large drainage tube is an undesirable last resort that
usually leaks.
2. Decompression of the duodenum should be considered if the duodenal stump closure is worrisome. This maneuver can be
effected by lateral tube duodenostomy, by a transjejunal retrograde tube duodenostomy, by “G-J tube” with passage of the J part into
the afferent limb and into the duodenum, or by similar passage of an NG tube into the afferent limb and into the duodenum. In the
latter instance, a soft tube should be used, and although it can be sutured to the nose, this is less than ideal for long-term
decompression because it is uncomfortable. For a lateral duodenostomy, we prefer a large t-tube or Malecot catheter (18F) placed
through a small duodenotomy and secured with an absorbable pursestring and Lembert sutures. An extensive Kocher maneuver
should be attempted to bring the duodenostomy site to the abdominal wall where it is sutured around the tube. As mentioned, we
avoid placement of a tube through the end of the duodenal stump. The omentum can be wrapped around the tube between the exit
site in the duodenum and the exit site in the peritoneum if the duodenum does not reach the abdominal wall.
3. Drainage: Closed suction drains are placed in the vicinity of (but not in contact with) the duodenal stump closure and tube
duodenostomy and remain for at least 10 days.
98 Dempsey and Pathak

10 While we recognize that an intraluminal circular stapler may be used to fashion a Billroth I anastomosis (A) and that either
a linear or circular stapling technique may be used to create a Billroth II anastomosis (B), we generally do not prefer staplers for
reconstruction after antrectomy for the following reasons:
1. It is our clinical impression that anastomotic bleeding is more common after stapled gastrojejunostomy.
2. A circular stapler technique for Billroth I anastomosis requires more proximal duodenal mobilization than a hand sewn one.
If the first attempt at a stapled anastomosis fails, there may be a critical loss of length in the proximal duodenum.
3. Gastroduodenostomy and gastrojejunostomy are great anastomoses for teaching residents proper surgical technique.
4. There are no data that stapled anastomosis are safer.

SUMMARY loic vessels may be divided with clips or the vascular

stapler, and then dissected away from the pylorus and
Laparoscopic staplers with interchangeable gastrointesti-
proximal duodenum with the harmonic scalpel.
nal and vascular cartridges, and the ultrasonic shears have
4. The proximal duodenum is divided with the lapa-
made laparoscopic antrectomy a more attractive surgical
roscopic stapler, and then the right gastric pedicle is di-
option. A periumbilical camera port and three or four
vided.
working ports are used. A 30 degree angled laparoscope is
5. A double staple technique is used for reconstruc-
essential, and it should be moved to other ports as neces-
tion with an antecolic Billroth II gastrojejunostomy. The
sary for optimal visualization. Although the important
NG tube is gently reintroduced with the tip only a few cm
parts of the operation are the same whether it is done
past the gastroesophageal junction (“3 lines showing at
open or laparoscopically, there are some important differ-
the nose”), and position confirmed with x-ray in the re-
ences in the conduct of the procedure:
covery room. No attempt is made to put the tube through
1. The stomach is decompressed and then the NG
the anastomosis, and the tip should be proximal enough
tube is pulled way back to avoid stapling, since it is vir-
in the stomach that it is not near the anastomosis.
tually impossible to palpate laparoscopically.
2. The proximal gastric margin is usually transected
early. This maneuver facilitates elevation of the antrum Intraoperative Complications
off the pancreas and mesocolon. An endoscopic stapler Antrectomy for benign disease is usually a straightfor-
with a vascular load can be used on the neurovascular ward operation. However intraoperative problems can
bundle on the greater and lesser curvature, and a gastro- arise, and the prudent surgeon will anticipate and avoid
intestinal cartridge with either 3.2 or preferably 4.8 mm these. Significant bleeding is usually the result of venous
staples on the stomach. injury from blunt dissection or excessive traction. Two
3. Just proximal to the pylorus, the right gastroepip- potential trouble spots are the gastroepiploic vein at its
Antrectomy
junction with the superior mesenteric vein, and the left
gastric vein (coronary vein) near its junction usually with
the portal vein. It is important not to make the injury
worse by hasty or rough dissection in an attempt to ex-
pose the source of bleeding. Gentle pressure will control
the hemorrhage indefinitely, and discreet controlled dis-
section will reveal the venous defect that can be repaired
readily with a 5– 0 polypropylene suture. Bleeding from
the splenic capsule can usually be controlled with cautery
and topical hemostatic agents. If staplers with a vascular
load are used during gastrectomy, the staple lines should
always be irrigated and gently abraded to check the ade-
quacy of hemostasis. It is not uncommon that additional
maneuvers are required. We avoid placing hemo-clips in
proximity to where we plan to transect or anastomose
with staplers.
The difficult duodenal stump, more often talked about
than seen nowadays, is best avoided by not doing an
antrectomy in the face of significant scarring and/or in-
flammation in the proximal duodenum (see Fig 9).
Injury to other organs during antrectomy should be
unusual. Dissection is distal to the spleen, making injury
to this organ infrequent. Injudicious traction on the
omentum can occasionally cause a capsular tear, but sple-
nectomy is rarely necessary. If pancreatic injury is sus-
pected, drainage is mandatory. Before drain removal, the
effluent is analyzed for amylase content. Adhesions from
previous operation or gastroduodenal inflammation or
scarring may predispose the colon to injury. Rarely the
transverse colon can be rendered ischemic by injury to
the middle colic artery and partial colectomy may be
necessary. For these reasons, preoperative mechanical
bowel preparation may be prudent when planning gastric
resection in selected patients.
Postoperative management and complications
Early ambulation, incentive spirometry, antisecretory
medications, and prophylactic, perioperative antibiotics
are routine. The choice of prophylactic antibiotic depends
on whether you expect the presence of gastric outlet ob-
struction or achlorhydria; in these situations, coverage of
anaerobic organisms is also wise. Oral liquids are started
when there is clinical evidence that gastrointestinal func-
tion has returned, usually on postoperative day 3 to 5 as
heralded by bowel sounds, decreased NG drainage, and
passage of flatus. First we often obtain a limited upper GI
study with water-soluble contrast to confirm emptying of
the gastric remnant and to rule out leak. Then liquids may
be started with the NG tube in place and residuals
checked, or the tube may be discontinued. Patients are
advised to start oral intake slowly for the first day or two
in an effort to avoid distention of the gastric remnant.
Careful reinsertion of the NG tube should be considered
for persistent vomiting, distention, or hiccups.
Common postoperative complications include atelec-
99
tasis, pneumonia, wound infection, ileus (including gas-
tric stasis), and electrolyte abnormalities. Delayed gastric
emptying during the first postoperative month is by itself
almost never an indication for reoperation. Erythromycin
may be useful in ameliorating this problem; metoclopra-
mide, as a kinetic, is not effective when the antrum has
been removed.
Less common complications include intraabdominal
hemorrhage, suture line bleeding (ie, GI bleed), duodenal
stump leak, anastomotic leak, pancreatitis, intraabdomi-
nal abscess and internal hernia. Duodenal stump leak may
be managed nonoperatively only if the patient continues
to do well and there is radiologic evidence of adequate
drainage and no intraabdominal collection. Otherwise
early reoperation to effect the following is wise: close or
intubate the leak; buttress this closure with omentum;
decompress the duodenum; drain adequately; and place a
feeding jejunostomy to allow long-term postduodenal en-
teral feeding and the ability to reinfuse duodenal drain-
age.
Outcomes
The overall operative mortality rate for antrectomy is
around 2% but is higher in patients with critical organ
dysfunction, insulin-dependent diabetes, or immunosup-
pression. Eighty percent of patients do well (Visick 1 or
2), while up to 20% of patients develop some form of
postgastrectomy and/or postvagotomy complications
(Visick 3 or 4). Problems in this latter group include
recurrent ulcer, dumping, diarrhea, gastric stasis, bile re-
flux gastritis, bowel obstruction and chronic abdominal
pain. If antrectomy is performed judiciously for benign
disease only when necessary, the incidence of poor out-
comes will not increase in our modern era.
SUGGESTED READING
1. Adkins RB Jr., DeLozier JB 3rd, Scott HW Jr, Sawyers JL: The
management of gastric ulcers. A current review. Ann Surg. 201:
741–751, 1985
2. Behrns KE, Sarr MG: Diagnosis and management of gastric emp-
tying disorders. Adv Surg 27:233–255, 1994
3. Csendes A, Maluenda F, Braghetto I, et al: Prospective randomized
study comparing three surgical techniques for the treatment of
gastric outlet obstruction secondary to duodenal ulcer. Am J Surg
166:45– 49, 1993
4. Delcore R, Cheung LY: Surgical options in postgastrectomy syn-
dromes. Surg Clin No Amer 71:57–75, 1991
5. De Vries BC, Schattenkerk ME, Smith EE, et al: Prospective ran-
domized multicentre trial of proximal gastric vagotomy or truncal
vagotomy and antrectomy for chronic duodenal ulcer: results after
5–7 years. Br J Surg 70:701–703, 1983
6. Donahue PE: Parietal cell vagotomy versus vagotomy-antrectomy:
ulcer surgery in the modern era. World J Surg. 24:264 –269, 2000
7. Eagon JC, Miedema BW, Kelly KA: Postgastrectomy syndromes.
Surg Clin No Amer 72:445– 465, 1992
8. Gibril F, Lindeman RJ, Abou-Saif A, et al: Retained gastric antrum
syndrome: a forgotten, treatable cause of refractory peptic ulcer
disease. Dig Dis Sci 46:610 – 617, 2001
100
9. Goligher JC, Feather DB, Hall R, et al: Several standard elective
operations for duodenal ulcer: Ten to 16 year clinical results. Ann
Surg 189:18 –24, 1979
10. Herrington JL Jr: Gastroduodenal ulcer. Overview of 150 papers
presented before the Southern Surgical Association 1888 –1986.
Ann Surg 207:754 –769, 1988
11. Jordan PH Jr, Thornby J: Twenty years after parietal cell vagotomy
or selective vagotomy antrectomy for treatment of duodenal ulcer.
Final report. Ann Surg 220:283–293, 1994
12. Kaneko K, Kondo H, Saito D, et al: Early gastric stump cancer
following distal gastrectomy. Gut 43:342–344, 1998
13. Kendall BJ, Chakravarti A, Kendall E, et al: The effect of intrave-
nous erythromycin on solid meal gastric emptying in patients with
chronic symptomatic post-vagotomy-antrectomy gastroparesis.
Aliment Pharm Therap 11:381–385, 1997
14. Koo J, Lam SK, Chan P, et al: Proximal gastric vagotomy, truncal
vagotomy with drainage, and truncal vagotomy with antrectomy
for chronic duodenal ulcer. A prospective, randomized controlled
trial. Ann Surg. 197:265–271, 1983
15. Mayers TM, Orebaugh MG: Totally laparoscopic Billroth I gastrec-
tomy. J Amer Coll Surg 186:100 –103, 1998
Dempsey and Pathak
16. McConnell DB, Baba GC, Deveney CW: Changes in surgical treat-
ment of peptic ulcer disease within a veterans hospital in the
1970s and the 1980s. Arch Surg 124:1164 –1167, 1989
17. McFadden DW, Zinner MJ: Reoperation for recurrent peptic ulcer
disease. Surg Clin No Amer 71:77–92, 1991
18. Millat B, Fingerhut A, Borie F: Surgical treatment of complicated
duodenal ulcers: controlled trials. World J Surg 24:299 –306, 2000
19. Paimela H, Tuompo PK, Perakyla T, et al: Peptic ulcer surgery
during the H2-receptor antagonist era: a population-based epide-
miological study of ulcer surgery in Helsinki from 1972 to 1987.
Br J Surg 78:28 –31, 1991
20. Ramirez B, Eaker EY, Drane WE, et al: Erythromycin enhances
gastric emptying in patients with gastroparesis after vagotomy and
antrectomy. Dig Dis Sci 39:2295–2300, 1994
21. So JB, Yam A, Cheah WK, et al: Risk factors related to operative
mortality and morbidity in patients undergoing emergency gas-
trectomy. Br J Surg 87:1702–1707, 2000
22. Xynos E, Vassilakis JS, Fountos A, et al: Enterogastric reflux after
various types of antiulcer gastric surgery: Quantitation by
99mTC- HIDA scintigraphy. Gastroenterology 101:991–998,
1991
Highly Selective Vagotomy
Philip E. Donahue, MD, FACS
lceration of the stomach and/or duodenum remains a
U common entity, but because of more effective medi-
cal treatments, the need for performing definitive ulcer
operations has declined markedly. Fifty years ago, before
the advent of histamine receptor antagonists, proton
pump inhibitors, or antimicrobial agents targeted against
Helicobacter species, ulcer operations were very common
in major medical centers. The most effective, most popu-
lar operation for ulcer was a subtotal (⬃70 to 80%) gastric
resection alone until the late 1940s, when a possible role
for vagotomy was recognized. Truncal (or complete ab-
dominal) vagotomy was found to facilitate ulcer healing
by decreasing gastric acid secretion, but it was also found
to have the undesirable side-effect of postoperative gastric
atony in up to 50% of patients. For that reason, surgeons
recognized that a complementary gastric resection or
drainage procedure (pyloroplasty or gastroenterostomy)
had to be performed in conjunction with truncal vagot-
omy. The combination of truncal vagotomy with 25 to
50% distal gastric resection (antrectomy) proved to be an
extremely effective ulcer operation, with an ulcer recur-
rence rate of 0.5 to 1.0%. Importantly, the risk of periop-
erative mortality and morbidity was notably less than
with the subtotal 70 to 80% gastric resection, the previous
state-of-the-art operation for control of ulcer disease be-
fore the advent of vagotomy. Because the risk of postop-
erative morbidity, such as dumping, diarrhea, and weight
loss, was relatively constant after vagotomy and antrec-
tomy (about 20% of patients), vagotomy with drainage
became preferred by many surgeons. This procedure
proved safe, but there was on average a 10% ulcer recur-
rence rate, as well as a constant incidence of undesirable
postvagotomy sequellae such as dumping and diarrhea.
Highly selective vagotomy (HSV), also referrred to as
parietal cell vagotomy or proximal gastric vagotomy, di-
vides only the vagus nerves supplying the gastric corpus
and fundus, the acid-producing portions of the stomach,
while preserving the vagal innervation to the gastric an-
trum. As a result, there is no need for a routine drainage
From the Departments of Surgery, Cook County Hospital and the University of
Illinois at Chicago, Chicago, IL.
Address reprint requests to Dr. Philip E. Donahue, Division of General Surgery,
Rm 6429, Cook County Hospital, 1835 W. Harrison St., Chicago, IL 60612;
e-mail: phil@eolas.com.
© 2003 Elsevier Inc. All rights reserved.
1524-153X/03/0502-0145$30.00/0
doi:10.1053/otgn.2003.35365
Operative Techniques in General Surgery, Vol 5, No 2 (June), 2003: pp 101–105
procedure, and the incidence of postoperative problems is
less. After extensive laboratory investigations, wide appli-
cation of HSV in Europe began after 1970, and HSV be-
came the operation of choice for ulcer disease in Europe
and Asia in the 1980s. However, American surgeons were
not enthusiastic about HSV primarily because concerns
regarding the high reported incidence of ulcer recur-
rence, which was as high as 25% in some reports. This
inordinately high rate of recurrent ulcers was largely a
technical problem. First, the posterior “criminal” nerve of
Grassi was often overlooked because of failure to mobilize
and evaluate the posterior aspect of the gastroesophageal
junction. Second, the branches to the esophagus were not
transected far enough proximally on the esophagus.
Third, the recurrent vagal innervation to the greater cur-
vature of the distal corpus that travels with the right
gastroepiploic vessels was not appreciated and, therefore,
went untransected. At the University of Illinois in Chi-
cago we found that HSV should be modified to include
transection of additional nerves (areas not included in the
original technique of HSV); this “extended HSV” was then
applied in all patients with ulcer. During follow-up, the
noted recurrence rate was low (1.0%), and there were few
postoperative complaints. This article describes the tech-
nical details of HSV and the specific steps to be followed
when performing the procedure.
Although few patients require operations for duodenal
ulcer disease at present, and deaths due to ulcer have
decreased, the incidence of ulcer continues to be quite
high, especially in patients over 65 years of age. The
benefits of an operation such as HSV are many, including
the fact that it can be easily combined with other proce-
dures such as fundoplication or cholecystectomy.
PATIENT PREPARATION
Informed consent. The possibility of symptoms in-
cluding diarrhea, dumping, bloating, early satiety, infec-
tion, bleeding, and the risk of perforation of the esopha-
gus, stomach or contiguous organs should be mentioned.
The patient should be reassured that injuries to the
esophagus or stomach occur rarely, and that the surgeon
will carefully evaluate the tissues directly during the op-
eration.
Retractors and harmonic scalpel. During open
procedures, a fixed mechanical retractor of some type is
invaluable for elevating the rib cage and exposing the
distal esophagus and esophageal hiatus; during laparo-
101
102 Philip E. Donahue

1 Vagus nerve branches enter at Sites 1, 2, and 3 (the sites for dissection first described in HSV). Site 4 (the posterior gastric nerves)
and Site 7 (right Gastroepiploic nerve) also contain nerves which require division. Sites 5 and 6 have few branches of the vagus
nerves and are not divided routinely.

scopic procedures, the same exposure is achieved with tures or sutures is recommended (Fig 2). This dissection
reverse Trendelenberg positioning. continues proximally for approximately 10 cm, to where
Operation. During open operations, we begin by the reflected portions of the pancreatico-gastric fold are
opening the lesser sac and identifying the posterior nerve appreciated. The same dissection is thereafter performed
of Latarjet with its distal branches entering the antrum.
Note that the branches of the posterior nerve of Latarjet
enter the antrum approximately 2.0 cm further caudad
than their anterior branch counterparts. Please refer to
the diagram of the areas of vagotomy (Fig 1) as a reference
to the sequential dissection of Areas #2 and #4, #1, #3,
and #7. Area #5, the short gastric area, is not divided, and
#6 (the left gastroepiploic nerve area) is not routinely
divided because it is thought to have very few postgangli-
onic vagus nerves.
During laparoscopic or open procedures, the harmonic
scalpel is the ideal tool for performing dissection and
greatly facilitates the rapidity, ease, and performance of
the operation.
Position for operation. The supine position is rou-
tinely used.
Inspection of the operative field. Expose the duo-
denum to ensure that there is no evidence of pyloric
stenosis; if there is any question, intraoperative gastros-
copy can provide definitive information regarding the
presence or absence of mechanical constriction of the
gastric outlet.
The most proximal branch of the posterior crow’s foot 2 The lesser curvature of the stomach (Area 2) is dissected
(most distal trifurcation of the posterior nerve of Latarjet) first, since the nerves are seen readily, and the chance of dam-
is divided just next to the stomach; the use of fine liga- aging the distal branches of the nerves of Laterjet is slight.
Highly Selective Vagotomy 103

3 At completion of lesser curve dissection, the posterior gastric nerves are divided completely, opening the “window” between the
stomach and the left crus of the diaphragm. Short gastric vessels are not divided because preservation of blood supply to the fundus
is a goal of the operation, and the number of vagal branches which travel with these vessels is small.

4 Periesophageal dissection has been completed in this view;

note the length of esophagus dissected 5 Division of neurovascular bundles in gastropancreatic fold.
104
6 The dissection of the nerves at the distal gastric antrum is
aimed at preserving two discrete branches of both anterior and
posterior Latarjet crow’s foot complex providing innervation to
the gastric antrum.
7 Division of the gastroepiploic nerve is an essential part of the procedure; when performing a laparoscopic procedure the entire
pedicle is divided, whereas at open surgery the nerve alone can be identified and transected.
Philip E. Donahue
anteriorly. As an adjunctive maneuver to facilitate clamp
placement, the surgeon scores the visceral peritoneum at
points between the various neurovascular pedicles that
enter the stomach.
As dissection progresses toward the phrenoesopha-
geal membrane, two layers become apparent. First, the
anterior phrenoesophageal membrane is dissected
from the patient’s right side toward the Angle of His.
Care is taken to demonstrate that the exploratory
clamp, which can be made to slide gracefully to-and-fro
beneath the membrane, is in the right plane and that
the risk of perforation is small. Next, the right edge of
the esophagus and proximal stomach is separated from
the neurovascular connections that tether these struc-
tures (Fig 3). Take care to avoid the anterior vagus and
the posterior vagus nerves (Fig 4). Griffith’s maneuver
(posterior traction on the celiac trunk with the sur-
geon’s left index finger in conjunction with the sur-
geon’s right hand placed behind the esophagus from
the area of the Angle of His) facilitates identification of
the posterior vagus nerve during open procedures.
During laparoscopic procedures, the use of visual clues
makes the posterior nerve more apparent.
The posterior gastric fold (Area #4-gastropancreatic
ligament) (Fig 5). This fold contains the posterior gastric
artery, as well as vagal branches. The esophagus can then
be retracted by means of a Penrose drain or a retraction
device (such as Endo-Retract, Maxi, U.S. Surgical Corpo-
ration) to facilitate identification and division of all aber-
Highly Selective Vagotomy 105
posterior serosal edges over the dissected lesser curvature
of the stomach (Fig 8). This maneuver is believed to
prevent vagal “reinnervation” and to minimize the un-
usual complication of gastric perforation along the lesser
curvature (believed by some surgeons to be secondary to
gastric injury from overjudicious use of electrocautery
during the dissection).

CONCLUSION AND SUMMARY

When operation is required for duodenal ulcer diathesis,
vagotomy of some type will usually be part of the treat-
ment; HSV is arguably the best overall operative treat-
ment and can be performed safely in most patients, ex-
cluding only those who are unstable or have severe
concomitant medical conditions which preclude the extra
time (approximately 30 to 45 minutes) required for its
performance.
While the role of open operations for ulcer complica-
tions is well established, there is little doubt that laparo-
scopic approaches will increase in popularity for appro-
priate candidates. Since practicing surgeons prefer to
perform operations that they have performed for years,
and because there is a diminishing incidence and preva-
lence of ulcer in most areas of the world, surgeons inter-
ested in practicing HSV must make an effort to preserve
the procedure. If the merits of HSV are compared with
those of alternative ulcer procedures, it will always have a
well-deserved place in the surgical armamentarium.
8 Suture approximation of the phrenoesophageal bundles
extends to the vicinity of the gastroesophageal junction, pro-
viding a serosal barrier to nerve regrowth as well as structural
REFERENCES
support to the anti-reflux mechanism.
1. Sonnenberg A: Peptic ulcer. In JE Everhart (Ed): Digestive disease
in the United States: Epidemiology and impact. NIH Publication
No. 94 –1447, 1994
rant fine nerve branches (including the early branching 2. Wangensteen OH, Wangensteen SD, Dennis C: The history of
gastric surgery. In C Wastell, LM Nyhus, PE Donahue (Eds):
posterior vagus nerve branches of Grassi). The goal of the Surgery of the esophagus, stomach and small intestine, 5th ed.
peri-esophageal dissection is to achieve complete division Boston, MA: Little Brown & Co., 1995, pp 354 –385
of all attachments of the neurovascular elements to the 3. Donahue PE, Tsai H-S, Yoshida J, Nyhus LM: Proximal gas-
esophagus (Area #1) for a distance of 5 to 8 cm. tric vagotomy—the first 25 years. Surgery Annual 19:139 –173, 1985
The distal corpus and antrum are then inspected, and 4. Donahue PE, Griffith CA, Richter HM: A 50-year perspective upon
selective gastric vagotomy. Am J Surg 172:9 –12, 1996
neurovascular connections between the stomach and the
5. Donahue PE, Bombeck CT, Condon RE, Nyhus LM: Proximal
hepatogastric ligament are divided (Fig 6). This includes gastric vagotomy versus selective vagotomy with antrectomy: Re-
the tissue deep to the nerves of Latarjet and the respective sults of a prospective, randomized clinical trial after four to twelve
anterior and posterior crow’s feet. Note again, the point of years. Surgery 96:585–590, 1984
entry of the anterior and posterior Latarjet nerves is dif- 6. Donahue PE, Richter HM, Liu K, Anan K, Nyhus LM: Experimental
ferent; reliance on preconceived boundaries (ie, that the basis and clinical application of extended highly selective vagotomy
for duodenal ulcer. Surg Gynecol Obstet 176:39 – 48, 1993
junction of the antrum and corpus is 6.0 cm from the 7. CDC WONDER. Death count, all ages, all races, both genders,
pyloric ring) is less reliable than direct inspection. 1979 –1996 by age-year, the United States. ICD 531.1 to 535.9.
The final area of dissection (#7) is the right gastroepi- http://www.cdc.gov/nchswww/data/gmanual.pdf.
ploic bundle, which can be divided totally (laparoscopic 8. Awad W, Csendes A, Braghetto I, et al: Laparoscopic highly selec-
procedures) or dissected to reveal the contained right tive vagotomy: Technical considerations and preliminary results
in 119 patients with duodenal ulcer or gastroesophageal reflux
gastroepiploic nerve before its division (Fig 7). Both ap-
disease. World J. Surg 21:268 –269, 1997
proaches have been successful in our hands, as verified by 9. Skandelakis LJ, Donahue PE, Skandelakis JE: The vagus nerve and
the intraoperative Congo Red test. The final aspect of the its vagaries. In JE Skandalakis (Ed): Surgical Anatomy and Embry-
procedure involves reapproximating the anterior and ology. Surg Clin North Am 73(4):769 –784, 1993
Roux-en-Y Gastric Bypass by Open Technique
Michael G. Sarr, MD
orbid obesity, probably better termed “medically
M complicated obesity” when talking with patients
(would you like to be called “morbid”?), is a national
health epidemic causing huge economic health care
expenses. The true impact on the American public of
this previously presumed “psychosocial” disease has
become heightened recently not only by its increased
incidence in our population but also because of the
country’s increased awareness of weight consciousness
and health, especially in the younger generations. To-
gether with governmental concerns about health fi-
nancing, the medical community has marketed a grow-
ing campaign against obesity. Numerous weight
reduction programs, dietary approaches, fad diets,
pharmacological dietary suppressants (some proven to
be effective, most worthless; e.g., “melt away fat over-
night”), and surgical (bariatric) programs have ap-
peared with increasing frequency. While this campaign
is appropriate, one must acknowledge the very poor
long-term results obtained with even the best of non-
operative approaches in the patient population with
morbid obesity (i.e., patients ⬎100% above their ideal
body weight). When success is defined as a loss of at
least 50% of the excess body weight and maintenance
thereafter for 5 years, even the highly concentrated
nonoperative programs have successes of ⬍5%—thus,
the need for a more aggressive, permanent means of
weight reduction by anatomical (operative) rearrange-
ment and “forced” behavior modification of eating
habits.1
Surgical history of bariatric surgery is a rather em-
barrassing one for our surgical community. While obe-
sity has undoubtedly always been a severe problem,
much (most?) of the academic community shunned
this problem. Almost all surgical approaches were pur-
sued in the general surgery community by smaller
groups either unwilling or unable to adequately follow
and evaluate the outcomes of many of the bariatric
procedures “introduced” as effective. Two good exam-
ples of this concept of a nonevidence-based approach
From the Mayo Clinic Rochester, Rochester, MN.
Address reprint requests to Dr. Michael G. Sarr, Professor of Surgery, Mayo
Clinic Rochester, 200 First Street S.W., Rochester, MN 55905.
© 2003 Elsevier Inc. All rights reserved.
1524-153X/03/0502-0144$30.00/0
doi:10.1053/otgn.2003.35364
106
include the jejunoileal bypass and its newer relative,
the gastroplasty (stomach “stapling” to the lay public).
Both operations were technically quite easy and theo-
retically attractive in principle. However, far too many
of these operations were performed before a serious,
critical review of their outcomes became available. Je-
junoileal bypass was associated with a very high inci-
dence of acute and chronic liver failure, serious vita-
min, mineral and protein/caloric deficiencies, and the
now better appreciated risk of irreversible oxalate ne-
phropathy. Nonbanded gastroplasty (of which there
were many, many variations) was simply a nondurable
procedure; the channel between the small upper gastric
pouch and the lower stomach rapidly dilated over the
first months postoperatively, thereby effectively de-
stroying the designed efficacy of the operation, and the
staple line partitioning the stomach into a small upper
and large lower pouch often dehisced, effectively abol-
ishing efficacy of the procedure. Introduction of surgi-
cal stapling devices made these gastroplasties techni-
cally quite “easy,” and in many communities, these
operations were even performed by nonboard-eligible
“surgeons,” poorly trained surgeons, or by groups
without a dedicated interest in the need for or impor-
tance of both preoperative counseling or postoperative
surveillance.
Finally, in 1991, the National Institutes of Health
(NIH) convened a consensus conference on bariatric sur-
gery and its role as a medically and cost-effective ap-
proach to the problem.2 Using at that time state-of-the-
art, evidence-based knowledge, this NIH Consensus
Conference finally “legitimized” two bariatric operations
as being proven effective: vertical banded gastroplasty
and Roux-en-Y gastric bypass. Moreover, this conference
also defined medical criteria of direct weight-related mor-
bidity appropriate for consideration of a surgical ap-
proach (i.e., repeated unsuccessful attempts to maintain
satisfactory weight loss in patients with a BMI [body mass
index] of ⬎40 or a BMI ⬎35 with at least one serious
weight-related problem). This NIH Consensus Confer-
ence really began the now-recognized and legitimized era
of bariatric surgery embraced by more academically inter-
ested groups dedicated to an evidence-based approach to
treating patients with the disease of obesity.
While currently both the vertical banded gastroplasty
(VBG) and the Roux-en-Y gastric bypass (RYGB) are the
most commonly performed, medically (and to insurance
Operative Techniques in General Surgery, Vol 5, No 2 (June), 2003: pp 106-113
Roux-en-Y Gastric Bypass
providers) acceptable procedures, new modifications
and concepts are always being introduced (and hope-
fully well studied!). Operations designed to create a
selective maldigestive/malabsorptive anatomy, such as
the biliopancreatic bypass,3 the distal gastric bypass or
very, very long limb gastric bypass,4 and the duodenal
switch with distal bypass5 are more radical approaches
used in the super obese (i.e., patients ⬎225% above
ideal body weight, BMIs ⬎50 to 60). These patients
present special situations, and the operations designed
to deal with such excessive weight would be expected
to have greater side effects related to their maldigestive/
malabsorptive anatomies and should be used selec-
tively. I will not further discuss these procedures in
this section. Interested readers might see reference 4
for an in-depth discussion.
Concerning VBG and RYGB, most studies, at least in
the United States, have questioned the efficacy and dura-
bility of the VBG. The VBG is theoretically attractive be-
cause the operation effectively partitions (using a linear
stapler) the stomach into a small volume (15 to 30 mL)
proximal pouch that communicates with the remain-
der of the stomach by a narrow diameter (9 to 11 mm)
“banded” stoma. Although the operation is technically
easy, has a low operative morbidity/mortality, involves no
anastomoses or gastrointestinal bypass, and the band pre-
vents the stoma connecting the partitioned stomach from
dilating, unfortunately, the short-term (3 years)6 and
107
long-term (10 years)7 efficacy has proven to be poor.
Similarly, a prospective, randomized study comparing
VBG and RYGB showed clearly superior results with
RYGB.8 In addition, in our experience, mechanical or
functional obstruction of the stoma is common, causing
the patients to switch their diets to high-calorie liquids
(with which their weight increases), and vomiting is quite
frequent (so-called maladaptive eating disorder); more-
over, symptomatic gastroesophageal reflux is common,
often requiring conversion to RYGB.9
The basic surgical design of the vertical, disconnected
Roux-en-Y gastric bypass is to separate the cardia of the
stomach (volume ⬍20 mL) from the remainder of the stom-
ach, thereby bypassing the reservoir function of the stom-
ach. By doing so, the operation induces a form of gastric
restriction, preventing ingestion of large meals at any one
sitting. Gastrointestinal continuity is then restored by drain-
ing this small pouch of cardia into a longer than normal
Roux-en-Y limb of jejunum. This anatomy establishes a
dumping-type physiology such that should the patient in-
gest a high caloric density carbohydrate, such as ice cream,
chocolate, etc. (which they should not be eating anyway!),
they develop the symptoms of dumping, i.e., a forced type of
behavior modification. Our results with this approach show
that, at 3 years postoperatively, approximately 75% of pa-
tients have lost and maintained ⬎50% of their preoperative
excess body weight.10
108 Michael G. Sarr

TECHNIQUE

1 Unlike in the past, when a jejunoileal bypass was performed via a transverse incision, most agree that an upper midline incision
is best. Once the skin incision has been created, rather intense lateral traction on the skin edges will allow the subcutaneous tissue
to “split” down to the midline without much bleeding. After opening the linea alba, there is usually a prominent layer of
preperitoneal fat which can bleed annoyingly if transected in the midline. I prefer entering the peritoneum in the avascular plane
2 to 4 cm to the patient’s left of the midline where this preperitoneal fat/peritoneum fuses with the posterior rectus fascia. One
should always feel for an associated umbilical hernia defect and, when present, the incision can be extended through the defect,
thereby closing it at the end of the operation in conjunction with the fascial closure. These bariatric operations are markedly
facilitated by use of some form of bariatric-designed mechanical retractor. We prefer the Pilling Bariatric retractor (Pilling Co., Ft.
Washington, PA) because the sidebars attached to the table are curved rostrally such that the stem of the retractor is attached to the
table further caudal than the cross bar. In addition, it allows both a stable rostral retraction of the costal margins and lateral
retraction of the incisional edges. Most importantly, the notched liver blade allows rostral retraction of the left lobe of the liver.
Careful placement of the notch of the blade directly over the esophageal hiatus and rather forceful rostral (not anterior) retraction
of the liver provides quite impressive exposure of the fat pad of the gastric cardia.
Roux-en-Y Gastric Bypass 109

2 The first step is to create a retrograde tunnel behind the

cardia extending to the left side of the esophagogastric junc-
tion. This maneuver is markedly facilitated by placing an oro-
gastric 32 Fr tube and mobilizing the fat pad anterorostrally
overlying the cardia of the stomach—the area where the car-
diojejunostomy will be performed. This mobilization should
expose the serosa of the cardia and is bloodless. Patients with
truncal obesity can have a very thick fat pad, which sometimes
requires excision to maintain exposure.

3 A rent in the avascular part of the lesser omentum allows

the surgeon to place his/her left hand into the lesser sac. Several
technical points are important. First, this maneuver is often
easier from the left side of the patient. Second, the easily evident
plane behind the stomach is created caudal to the left gastric
artery, not behind the esophagus. Third, care should be taken
not to transect either the vagal branches to the liver or a re-
placed left hepatic artery. Fourth, the tunnel should be created
at the esophagogastric junction on the patient’s left. The retro-
gastric tunnel is able to be made bluntly by having the index
finger of the left hand meet the index finger of the right hand at
the greater curvature side of the esophagogastric junction.
When the fingers of the left and right hand meet, one can
appreciate the edge of the greater curvature of the stomach by
“rolling” it between the probing fingers; in really obese, tall
male patients, this maneuver is important, because you may not
be able to see this area well. There is a 3- to 4-cm area extending
along the greater curvature from the patient’s left side of the
esophagogastric junction before reaching the first short gastric
vessel; thus, this plane can be developed bluntly.
110 Michael G. Sarr

4 The catheter ends of either two 18 Fr red rubber catheters

or two 18 Fr silicone Foley catheters (with the balloon port cut
off) are passed from the greater curvature side to the lesser
curvature and exiting the rent in the lesser omentum. These
tubes will guide the passage of the stapler. I prefer to guide these
manually rather than using a clamp, because some portions of
passing the clamp are “blind”; using a large clamp can be diffi-
cult and potentially dangerous in the obese patient.

5 A small window is created along the lesser curvature about

2 cm distal to the esophagogastric junction. Again, the edge of
the stomach can be “rolled” between the index finger and the
thumb of the surgeon’s left hand when the fingers are in the
lesser sac through the rent in the lesser omentum and the
thumb passed anteriorly. This window is created at the gastric
margin without injuring the vagal branches. The ends of the
catheters are then brought through this defect.

6 The catheter end of the more proximal tube is passed over the
end of the anvil of a TA90-B linear stapler (U.S. Surgical Corp.,
Norwalk, CT) and the catheter is used to guide the stapler around
the gastric cardia. The second catheter guides a 99-mm linear
stapler just distal to the first one. Both staplers are then angled (see
lateral view) such that there is more anterior wall than posterior
wall, allowing for a small volume punch but with relatively more
anterior wall available for the cardiojejunostomy. Both staplers are
fired (after being certain there are no nasogastric tubes in place)
and the cardia between the staple lines transected. I prefer two
separated staple lines, while others fire two or three staple lines
without formally transecting the cardia.
Roux-en-Y Gastric Bypass
7 This figure shows the “disconnected” gastric anatomy and
the generous anterior wall of cardia. The depiction in the figure
markedly exaggerates the size of the pouch of cardia, which
typically has a volume of ⱕ20 mL.
8 A Roux-en-Y limb is now fashioned from the proximal jeju-
num. By shining the operating room light through the mesentery
from the head of the patient, the vascular supply is well transillu-
minated as seen from below. I prefer to create the Roux limb about
50 to 75 cm from the ligament of Treitz because, as shown, the
primary mesenteric vessels that eventually form the mesenteric
arcade serve to branch off the superior mesenteric artery further
away from the bowel wall, allowing a more mobile Roux limb
based on a relatively larger diameter mesenteric vessel. Transect-
ing the jejunum close to the ligament of Treitz not only mandates
transecting one or two primary mesenteric branches (thus requir-
ing the vascular supply to the end of the Roux limb to rely on the
mesenteric arcade), but it also requires the jejunojejunostomy to
be somewhat tethered to the left retroperitoneal region because the
ligament of Treitz is fixed posteriorly.
111
9 The Roux-en-Y limb is then brought in a retrocolic, ante-
grade position for eventual anastomosis to the cardiac pouch. I
usually enter the lesser sac via the gastrocolic omentum and
make a vertical rent in the transverse mesocolon to the left of
the patient’s middle colic vessels, being careful not to disrupt
the mesenteric arcade along the mesenteric aspect of the bowel
wall. If there is potential tension at the site of the cardiojeju-
nostomy after bringing the Roux limb up to the gastric cardia,
transverse relaxing incisions on both sides of the mesentery of
the Roux limb usually will offer an additional 2 cm of length. In
patients with a shortened mesentery, additional length may be
obtained by mobilizing the entire small bowel from its retro-
peritoneal attachments (usually requiring concomitant mobili-
zation of the ascending colon as well), thereby rotating the
superior mesenteric artery and small bowel mesentery rostrally.
The rent in the transverse mesocolon may have to be enlarged
to allow the base of the small bowel mesentery to enter the
lesser sac through the mesocolon. This maneuver is rarely nec-
essary (once in about 800 such operations – personal series). A
#21 EEA stapler (U.S. Surgical Corp., Norwalk, CT) is then
passed into the lumen through the end of the Roux limb and is
docked with the prong of the anvil. I prefer to position the anvil
through a small cardiotomy and then close the cardiotomy
around the prong of the anvil with a purse string suture. Others
position the anvil via an intraluminal approach by creating a
small gastrotomy in the mid-corpus of the stomach and passing
the prong of the anvil out of the cardia before stapling across the
cardia; I think the latter maneuver is more difficult.
112
10 After firing the EEA stapler, I check the donut of gastric
cardia carefully and, if incomplete, we will place 3 to 5 trans-
mural sutures across the cardiojejunal anastomosis at the area
where the donut was incomplete. I also carefully inspect the
anastomosis for areas of anastomotic disruption created poten-
tially during removal of the anvil. After this, I reinforce the
entire anastomosis with seromuscular Lembert sutures. A na-
sogastric tube is positioned just proximal to the anastomosis for
about 12 or 18 hours to evacuate any intraluminal blood. Fi-
nally, the end of the Roux limb is stapled closed after excising
any redundant blind limb. Intestinal continuity from the by-
passed stomach and duodenum is restored by a jejunojejunos-
tomy. I favor a 150-cm Roux limb as opposed to a 50- to 75-cm
length to prevent any possibility of bile reflux esophagitis.
Michael G. Sarr
11 For patients with super obesity (BMIs ⬎55), I usually
suggest a very, very long limb Roux-en-Y gastric bypass.4 This
procedure has also been called a distal gastric bypass or gastric
bypass with biliopancreatic diversion. The only difference from
the regular gastric bypass is that the enteroenterostomy is a
jejunoileostomy constructed 100 cm proximal to the ileocecal
junction. This maneuver leaves a 300- to 450-cm Roux limb
and a 100-cm common channel where all the digestion and
absorption of ingested complex foodstuffs occurs. I also favor a
temporary gastrostomy in the defunctionalized stomach to al-
low additional nutritional supplements in the first 2 to 3
months (or longer) postoperatively should they be needed. The
operation ends by closing the abdomen with a running fascial
suture; I prefer #2 polypropylene which maintains its strength
past 3 months, unlike all the current absorbable suture material
(polydioxanone, polyglycolic acid, etc.). The subcutaneous tis-
sue is not approximated by suturing, but the skin edges are
sewn together with a dermal suture of 3– 0 polyglycolic acid.
We also do not routinely place a perianastomotic drain at the
cardiojejunostomy unless there is some concern.
Roux-en-Y Gastric Bypass
CONCLUSION
Use of this technique of RYGB has led to at least satisfac-
tory results at the Mayo Clinic.10 By 3 years postopera-
tively, about 75% of patients have lost and maintained at
least 50% of their preoperative excess body weight. Insu-
lin-resistant diabetes mellitus routinely resolves, sleep
apnea improves to the extent that most patients no longer
need or use CPAP (continuous positive airway pressure)
at night, and symptomatically, at least, their degenerative
joint disease improves; the marked weight loss has al-
lowed many patients to undergo total joint arthroplasty
that would otherwise have been either impractical or im-
possible at their preoperative weight.11 Indeed, this latter
topic may serve as an acceptable indication for bariatric
surgery in selected elderly patients.12 Hypertension re-
solves in about 50% of patients, and the associated forms
of hyperlipidemia tend to improve. The ultimate study of
life expectancy, health care expenditure, and quality of
life is not yet completed, but the SOS (Swedish Obesity
Study) currently underway should help to answer some of
these questions, not only for the medical community, but
also for both the general public and the third party payers
in the healthcare industry.
Our technique of vertical, disconnected long limb
Roux-en-Y gastric bypass13 is slightly modified from
some of the earlier reports. The staple line applied verti-
cally rather than horizontally avoids both the need to
ligate any short gastric vessels, keeps the proximal gastric
pouch small, and hopefully uses a part of the stomach that
does not tend to dilate as can occur in the gastric fundus.
Most importantly, there is essentially no acid-producing
capacity such that peptic-type anastomotic ulcers of the
unprotected jejunal mucosal of the Roux-en-Y limb do
not occur. If a stomal ulcer develops, the patient will
undoubtedly be taking aspirin or a nonsteroidal antiin-
flammatory drugs (NSAIDs); their discontinuation will
allow the ulcer to heal. A secondary benefit of this oper-
ation is that anatomically it is the best possible antireflux
procedure. Many patients with morbid obesity (estimated
to be ⬃30%) suffer some element of gastroesophageal
reflux. Not only does this RYGB separate the esophagus
from all acid-peptic secretions,14 but it also prevents any
reflux of bile because of the 150-cm long Roux-en-Y limb.
All open bariatric operations have consequences, both
good (discussed above) and bad. Long-term sequelae of
an open RYGB include an incidence of incisional hernia of
15 to 20%.10 This seemingly high incidence is probably
multifactorial. Obesity predisposes to incisional hernia,
but also, the operation nutritionally challenges these pa-
tients in the first few months postoperatively. Their total
daily caloric intake is probably ⬍500 kcal for the first
postoperative month which establishes a form of planned
113
protein/calorie malnutrition—that is how they lose
weight! Longer term potential complications include iron
deficiency— usually only a problem in menstruating
women—and the potential for vitamin B12 deficiency (al-
beit quite rare).14 For these reasons, we strongly suggest a
daily multivitamin containing iron and monthly paren-
teral vitamin B12 (1000 ␮g IM). Also, because the duo-
denum, the major site of calcium absorption, is by-
passed, we urge daily calcium supplements in the form of
a calcium-containing antacid; these “antacids” are much
easier to take and are more palatable than formal oral
calcium tablets. While we have no good data to support
this approach, it seems to make sense, especially in peri-
and post-menopausal women.
REFERENCES
1. Balsiger BM, Luque-de Leon E, Sarr MG: Concise review for pri-
mary-care physicians—surgical treatment of obesity: who is an
appropriate candidate? Mayo Clin Proc 72:551-558, 1997
2. Consensus Development Conference Panel: NIH conference: Gas-
trointestinal surgery for severe obesity. Ann Intern Med 115:956-
961, 1991
3. Scopinaro N, Gianetta E, Civalleri D, Bonalumi U, Bachi V: Bilio-
pancreatic bypass for obesity: II. Initial experience in man. Brit
J Surg 66:618-620, 1979
4. Murr MM, Balsiger BM, Kennedy FP, Mai JL, Sarr MG: Malabsorp-
tive procedures for severe obesity: comparison of pancreaticobili-
ary bypass and very very long limb Roux-en-Y gastric bypass. J
Gastrointestinal Surg 3:607-612, 1999
5. Marceau P, Hould FS, Simard S, Lebel S, Biron S: Biliopancreatic
diversion with duodenal switch. World J Surg 22:947-954, 1998
6. Nightengale ML, Sarr MG, Kelly KA, Jensen MD, Zinsmeister AR,
Palumbo PJ: A prospective evaluation of vertical banded gastro-
plasty as the primary operation for morbid obesity. Mayo Clin
Proc 66:773-782, 1991
7. Balsiger BM, Poggio JL, Mai J, Kelly KA, Sarr MG: Ten and more
years after vertical banded gastroplasty as primary operation for
morbid obesity. J Gastrointest Surg, in press
8. Sugerman HJ, Starkey JV, Birkenhauer R: A randomized prospec-
tive trial of gastric bypass versus vertical banded gastroplasty for
morbid obesity and their effects on sweets versus non-sweets
eaters. Ann Surg 205:613-624, 1987
9. Balsiger BM, Murr MM, Mai J, Sarr MG: Gastroesophageal reflux
after intact vertical banded gastroplasty: correction by conversion
to Roux-en-Y gastric bypass. J Gastrointest Surg 4:276-281, 2000
10. Balsiger BM, Kennedy, FP, Abu-Lebdeh HS, et al: Prospective
evaluation of Roux-en-Y gastric bypass as primary operation for
morbid obesity. Mayo Clin Proc 75:673-680, 2000
11. Parvizi J, Trousdale RT, Sarr MG: Total joint arthroplasty after
bariatric surgery for morbid obesity. J Arthroplasty, in press
12. Murr MM, Siadati MR, Sarr MG: Results of bariatric surgery for
morbid obesity in patients older than 50 years. Obesity Surg
5:399-402, 1995
13. Sarr MG: How I do it: vertical disconnected Roux-en-Y gastric
bypass. Dig Surg 13:45-49, 1996
14. Smith CD, Herkes SB, Behrns KE, et al: Gastric acid secretion and
vitamin B12 absorption after vertical Roux-en-Y gastric bypass for
morbid obesity. Ann Surg 218:91-96, 1993