Perforated Gastric Ulcers - Surgery - University of Cincinnati

Perforated Gastric Ulcers
A Plea for Management by Simple Closures
William W. Turner, Jr, MD; William M. Thompson, Jr, MD;
Erwin R. Thal, MD
\s=b\One hundred seven patients with perforated gastric ulcers
were treated by either simple closures (omental patches, 81
or ulcer
patients; primary suture without patches, 13 patients;
excisions with closures, two patients) or primary gastric
resections (11 patients). The latter were performed when
ulcers were too large to be treated by simple closures. The
mortality rate after omental patches or ulcer excisions with
closures was 12%, while that following primary gastric resections
was 45%. Patients who underwent closures with suturing
only had a mortality rate of 62%, which was significantly
higher than the mortality rate following patch closures. Gastric
outlet obstructions developed following 15% of simple closures
of prepyloric ulcers. Closures of perforated gastric
ulcers with omental patches or ulcer excisions can be undertaken
with low mortality and morbidity rates. Primary gastric
resections are reserved for patients with ulcers that are large
or located in the prepyloric area.
(Arch Surg 1988;123:960-964)
The operative treatment of acutely perforated peptic
ulcers with either simple closures or "definitive" pri¬
mary gastric resections has been debated for decades. Few
reports have focused on perforated gastric ulcers. In one
of the earliest series from which data on patients with
perforated gastric ulcers can be examined, a mortality rate
of 20% was reported following simple closures,1 in contrast
to that following primary gastric resections, where no
deaths occurred. Furthermore, 50% of the patients who
survived simple closures required subsequent gastric re¬
sections for persistent gastroduodenal symptoms. Unfor¬
tunately, follow-up data on the patients who underwent
resections were not presented. The recommendation that
primary gastric resections be considered more frequently
in the treatment of perforated gastric ulcers was echoed in
later studies.2-4
In the only study that attempted to randomize patients
with perforated gastric ulcers to treatment via simple
Accepted for publication Jan 5, 1988.
From the Department of Surgery, University of Texas Southwestern
Medical Center, Dallas.
Read before the 95th Annual Meeting of the Western Surgical Association,
Dallas, Nov 18, 1987.
Reprints not available.
closures or primary gastric resections, gastroduodenal
symptoms recurred in 53% of patients after simple closures
and in 57% of patients following resections.5 Unfortunately,
the data concerning operative mortality were not presented
in such a way as to allow comparison ofthe two procedures.
In recent reports from McGee and Sawyers,6·7 postoper¬
ative mortality and complication rates were higher follow¬
ing simple closures (29% and 17%, respectively) than after
primary gastric resections (15% and 5%, respectively).
Only the differences in postoperative complication rates
were significant. These results differ from those reported
in a group of elderly patients with perforated gastric
ulcers.8 In this study, 47 patients underwent simple closures
with a 32% mortality rate, and seven patients had gastric
resections with a 43% mortality
Downloaded from
www.archsurg.com at University of Cincinnati, on March 20, 2011
rate (differences not
statistically significant). Postoperative complications oc¬
curred with equal frequencies in the two groups.
The problems in drawing conclusions from the above
studies are the small numbers of patients in some studies,
the frequent management of patients by simple closures
when they were judged "too ill to undergo resections," and
too little information presented to assess comparability of
operative groups. The purpose of this review was to answer
two questions: (1) what are the postoperative complication
rates of selective management of perforated gastric ulcers
(ie, simple closures where technically feasible or primary
gastric resections in patients with large ulcers), and (2)
can patients be identified who are likely to develop recur¬
rent gastroduodenal symptoms?
PATIENTS AND METHODS
One hundred seven patients with perforated gastric ulcers
underwent operations in our institution from 1955 to 1987. There
were 76 men and 31 women, ranging in age from 20 to 83 years
(mean ± SD, 49 ± 15 years). Perforations were treated by simple
closures in 96 patients and by primary gastric resections in 11
patients. Simple closures involved omental patches (81 patients),
suture without omental patches (13 patients), and local ulcer
excisions and closures (two patients). One of the latter patients
had the perforation closed as a pyloroplasty, accompanied by a
truncal vagotomy. All gastric resections were of the Billroth II
variety and were accompanied by truncal vagotomies in five
patients.

A history of at least one risk factor for ulcer perforation was
present in 71 of the patients:
Risk Factor No. of Patients
52
Smoking
Ethanol
Ulcerogenic drugs
39
25
Twenty-nine patients had histories of either gastric or duodenal
ulcers, including one patient who had both. Twenty-seven of these
patients underwent simple closures, and two patients underwent
gastric resections.
Presenting symptoms in 107 patients with perforated gastric
ulcers were as follows:
Symptom
Abdominal pain
Pneumoperitoneum
Nausea/vomiting
Hypotension
Hematemesis
No. of Patients
100
73
58
15
14
Melena 13
Fever 10
Diarrhea 6
Hematochezia 1
Ages, sexes, histories, presenting symptoms, and ethnic back¬
grounds were similar among patients from the various operative
groups. Operative procedures were performed with equal frequen¬
cies during the four decades included in this study. Postoperative
complications, including death, also occurred with equal frequen¬
cies throughout the four decades and among the various ethnic
groups.
Most perforations occurred in the prepyloric area (defined as
that area within 3 cm of the pylorus). Ulcer sizes were recorded
in 87 patients. Biopsy was performed in 59% of patients, and in
only one instance was a malignant neoplasm identified. Concurrent
pyloric or duodenal ulcers were noted in 19% of patients. Ulcer
characteristics in patients with perforated gastric ulcers were as
follows:
Characteristic No. of Patients
Ulcer location
Prepyloric 49
Antral 40
Fundic 13
Unknown 5
Anterior 84
Posterior 17
Histological findings
Benign 62
1
Malignant
Not documented 44
Early postoperative complications were defined as those occur¬
ring during hospitalizations or within 30 days following operations.
Follow-up information was available for 71 of the 84 surviving
patients, ranging over periods from one week to 36 years (median,
7.25 months; 75th percentile, 5.5 years). Thirty-seven percent of
the patients were followed up for at least ten months.
Statistical analyses were performed using Fisher's exact and
Student's t tests.
RESULTS
Early postoperative complications, including deaths, oc¬
curred in 40 patients (37%), 34 (35%) of whom underwent
simple closures and six (55%) of whom had primary gastric
resections (not statistically significant). The incidences of
early postoperative complications among the various op¬
erative subgroups are listed in Table 1. Suture closures of
gastric perforations without omental patches led to signif¬
icantly more postoperative complications than did omental
patch closures.
The overall operative mortality rate was 21%, including
18 deaths (19%) in the patients who had simple closures
and five deaths (45%) in the patients who had gastric
resections (P = .056). The postoperative mortality rate was
significantly higher among the patients with suture clo¬
sures of perforations without omental patches than among
the patients with patch closures (Table 2). There were
significantly more postoperative deaths (45%) among the
patients who underwent primary gastric resections than
in those who had patch closures (12%). Postoperative
deaths occurred more frequently in women (32%) than in
men (17%) (P = . 073).
Among patients in whom ulcer sizes were recorded,
ulcers ranged from 0.1 to 10 cm (median, 1.0 cm; 75th
percentile, 2.0 cm). Ulcer sizes were significantly larger in
the 11 patients who underwent primary gastric resections
(median, 5.0 cm; 75th percentile, 5.0 cm) than in the 76
patients who underwent simple closures (median, 1.0 cm;
75th percentile, 1.5 cm) (P = .001). Ulcers were signifi¬
cantly larger in patients who developed early postoperative
complications (P = .004) and among patients who died
(P

Table 1.—Early Postoperative Complications by
Operative Procedures
No. of Patients With
Type of Closure No. of Patients Complications*
Simple closure
Omental patch 81 24
Suture 13 10
Excision and closure
Primary gastric resection
Total 107 40
mental patch vs suture, = .002; suture vs primary gastric resection,
097.
Table 2.—Postoperative Deaths by Operative Procedures
Type of Closure No. of Patients (Deaths)*
Simple closure
Omental patch 81 (10)
Suture 13 (8)
Excision and closure 2 (0)
Primary gastric resection 11 (5)
Total 107 (23)
*Omental patch vs suture, P10mo

ventions were significantly different in men and women.
No attempt was made in our series to tailor the operative
procedure based on a history of ulcer disease or on the
duration of symptoms preceding hospitalization. A history
of ulcer disease can be unreliable, and there is a poor
correlation with mortality in patients with perforated
peptic ulcers.1·10 A history of ulcer disease was identified
in only 28% of our patients, and there was no relationship
to mortality or early postoperative morbidity. Delay in
presentation to the hospital after the acute onset of
symptoms and delay in operative therapy once patients
were admitted to the hospital were associated with signif¬
icant increases in early postoperative complications, par¬
ticularly death. This relationship has been noted previously
in patients with perforated peptic ulcers.10
1. Donaldson G, Jarrett F: Perforated gastroduodenal ulcer disease at
the Massachusetts General Hospital from 1952 to 1970. Am J Surg
1970;120:306-311.
2. Rees J, Thorbjarnarson B: Perforated gastric ulcer. Am J Surg
1973;126:93-97.
3. Hahnloser P, Bruttin J, Cavin R, et al: Perforation ofcancer or stomach
ulcer: A one- or two-stage operation (in French). Helv Chir Acta 1980;47:565\x=req-\
574.
4. Wilson-Macdonald J, Mortensen N, Williamson R: Perforated gastric
ulcer. Postgrad Med J 1985;61:217-220.
5. Skarstein A, Hoisaeter P: Perforated peptic ulcer: A comparison of
long term results following partial gastric resection or simple closure. Br J
J. Bradley Aust, MD, San Antonio, Tex: I think the title of
this report is misleading. The authors claim that it is a plea for
simple closure, and yet the report really suggests that judgment
at the time of surgery should play a major role in the way in which
perforated gastric ulcers are treated. I certainly concur with that
approach. The omental patch was devised by Graham because
perforated duodenal ulcers frequently leak because the duodenal
wall is thin and inflamed and holds sutures poorly. It is, therefore,
of some surprise to me to find out that in this series gastric suture
has similar limitations since the gastric wall is much thicker and
certainly holds stitches much better than the duodenal wall.
My own philosophy in the treatment of gastric ulcers is to "make
the punishment fit the crime," if I can.
Prepyloric ulcers are treated in a manner similar to duodenal
ulcers. If there is little abdominal contamination and the patient
is not late in arriving in the operating room, I believe that
definitive surgery is reasonable, and if the patient is late and
there is much contamination, suture and patching is in order.
For primary gastric ulcers, we all agree that medical treatment
is less effective than for duodenal ulcers. Therefore, resectional
therapy is in order if deemed safe by the lack of delay or
contamination.
Giant ulcers mandate gastric resection in spite of their greater
risk. Even so, in our own environment, delays and degree of
contamination lead to far more local excisions of the ulcer for
histologie confirmation, suture, and patching than definitive sur¬
gery.
I became interested in what we might have done in the past two
years and looked up our experience with gastric ulcer. We had 17
perforated gastric ulcers. I was amazed to find that seven of these
were in infants, leaving only ten adult perforations in two years.
George L. Jordan, Jr, MD, Houston: We are indebted to this
group for a very detailed presentation ofthe problem of perforating
gastric ulcers. However, their conclusions are not justified, either
on the basis of their own data or review of the literature and
certainly not from our experience.
They do suggest a selective approach, and that concept is one
with which I fully agree and we have advocated for years, but
their selection is not very helpful.
If you look at their mortality rates, you will find that their
highest mortality rate is in a group treated by a special technique
of primary closure that they do not recommend any more. To
References
Discussion
Perforated gastric ulcers are uncommon, and the fund of
knowledge from which to make judgments about operative
management is still growing. A limited number of retro¬
spective studies report conflicting results after various
operative procedures. Experience gained from the treat¬
ment of more frequent duodenal ulcer perforations may
not be applicable to the management of perforated gastric
ulcers.
The results of our study indicate that simple closures of
perforated gastric ulcers with omental patches or local
excisions with closures can be performed with low mortal¬
ity and low early morbidity rates. Primary gastric resec¬
tions can be reserved for those patients who have large
ulcers. Our data suggest that patients with prepyloric
perforations may benefit from primary gastric resections.
Surg 1976;63:700-703.
6. McGee G, Sawyers J: Primary gastric resection for perforated gastric
ulcer. Surg Rounds 1987;10:19-26.
7. McGee GS, Sawyers JL: Perforated gastric ulcers: A plea for management
by primary gastric resection. Arch Surg 1987;122:555-561.
8. St James Collier D, Pain J: Perforated gastric ulcer: A reappraisal of
the role of biopsy and oversewing. J R Coll Surg Edinb 1985;30:26-29.
9. Brown R, Langman M, Lambert P: Hospital admissions for peptic
ulcer 1958-1972. Br Med J 1976;1:35-37.
10. Sirinek K, Levine B, Swesinger W, et al: Simple closure of perforated
peptic ulcer: Still an effective procedure for patients with delay in treatment.
Arch Surg 1981;116:591-596.
condemn gastric resection as having a high mortality rate when
certain types of simple closure also have a high mortality rate does
not make a strong point for the differential approach.
The high mortality rate in both of those groups, that is, the
resection group and the closure group, almost surely relates to
selection of patients more than it does to the procedure employed,
and it would appear that the omental patch closures were most
likely performed in the patients with smaller ulcers, which as a
rule are most easily treated.
The authors decry the absence of data for long-term studies in
resection, but present none. We have a tremendous number of
long-term data concerning surgery for gastric ulcer with other
complications and there is really no reason to presume that
treatment of this complication will produce a different long-term
result.
They present only a short follow-up, and yet, as I understand it,
19 of their 63 patients treated by simple closure have already had
trouble, with an average follow-up time of only seven months.
With follow-up times of four or five years, I predict that you are
going to find up to 75% or 80% recurrence rates.
The literature does not support the concept that gastric resec¬
tion for perforated ulcer in selected patients is associated with a
high mortality rate. This one is high, I am sure, because of
selection of patients, not because of the procedure.
Our reported experience with perforated ulcer includes 1183
patients, 244 of whom had gastric ulcers. Our most recent (1987)
report was by David Feliciano, MD. He reported an 11-year
experience from the Ben Taub Hospital that I think would correlate
with your patient population; 60 of those patients had gastric
ulcers and there were no carcinomas in the group.
We resected 57% of these ulcers and closed 43%. The mortality
rate in the resected group was 8.8%, and our overall mortality
rate for all 60 patients for all types of perforations of gastric ulcer
was only 13.3%. This mortality rate is not significantly different
for our total experience from that of the best group reported here
with simple closure.
We have looked specifically at the patients over 60 years of age
and reported our experience with those patients in 1984. In that
particular report we had 26 gastric ulcers in patients ranging from
60 to 97 years of age. We performed resections in 13 and closed
the ulcers in 13 of the patients. In these patients over 60 years of
age, our mortality rate was 19.2%.
Downloaded from
www.archsurg.com at University of Cincinnati, on March 20, 2011

I have several questions for the authors. First, what was the
actual cause of death? Our studies have not indicated that the
procedure is really the cause of death. The cause of death relates
to other problems in these patients.
Second, I am concerned about drawing conclusions on the basis
of a short follow-up period. I am really concerned about this is¬
sue because, according to the report, histologie nature was deter¬
mined in only 63% of the ulcers, so I wish you would describe your
follow-up procedure in patients in whom you performed simple
closure.
Third, I am surprised that since you did include one patient
with carcinoma that the incidence was that low. Our incidence of
carcinoma, which is not included in our data (we break those up
into a separate group), is approximately 5%. There are not many
perforated gastric ulcers per year; therefore, treatment of those
patients is done by a multitude of surgeons over a long period of
time, hence the difficulty in recording a large series. Nevertheless,
the data on gastric ulcer are pretty clear relative to long-term
results of definitive surgery, which are good, and I think they are
pretty clear relative to the long-term results of simple closure,
which are not good.
Robert E. McCurdy, MD, Denver: It is a delight for people
like me who grew up in the era of gastric surgery in the late 1940s,
1950s, and 1960s to find out that there is still some pathology
occurring in the stomach.
It has been one of the great mysteries to me as to why we have
had this tremendous decline in peptic ulcer disease, and this
happened even before the advent of cimetidine.
I tell our residents that I do not know of any condition in the
abdomen thatjustifies more good surgicaljudgment than selecting
the operation for a patient with perforated ulcer.
I would like to present
an anecdotal case from our former
president, Tom Throckmorton, MD. He was in the midst of a
family vacation in Winter Park, Fla, two years ago when he called
me and said, "My daughter just perforated a gastric ulcer. Would
you send a helicopter up to Winter Park to pick her up?"
Later that day, at surgery, she was found to have a high
perforated gastric ulcer. We performed biopsy, vagotomy, and
pyloroplasty.
I said, "Tom you told me over the phone that she had had a
perforated gastric ulcer. What made you think that it wasn't just
routine duodenal ulcer?"
He looked at me and said just one word: "Experience."
Robert J. Baker, MD, Chicago: I don't think most of us would
try to do a Tanner or some other fancy sort of gastrectomy in a
patient with a very high lying perforation, and in that circum¬
stance, Dr Turner and his colleagues have certainly shown that a
patch plus some definitive ulcer operation, usually a parietal cell
vagotomy in our center, can remedy a difficult situation.
I would like to add a word of caution to what has been said by
the previous speaker concerning gastric cancer. We have had four
telling experiences in the last three years with what were thought
to be perforations of benign gastric ulcers. Two of these were
resected, and the pathological specimen revealed infiltrating
adenocarcinoma. Unfortunately, the other two patients were so ill
and the perforations so small that the ulcers were patched; when
the patients underwent endoscopy in the postoperative period,
one at about two months' and one at about four months' follow-up,
obvious carcinomas were seen. These patients all underwent
operation by experienced surgeons with a senior surgical resident;
this proves again that it is not always a simple matter to ascertain
at the operating table whether a gastric ulcer is benign or
malignant. I would strongly urge that emphasis be placed on
determination of malignancy, as about 12% of our perforated
gastric ulcers prove to be malignant. We must always think of
carcinoma and add that factor to the other data that have already
been presented in support of resection. A perforated gastric ulcer
is a condition that usually requires a definitive operation.
I would like to ask a question of the authors. When in the
postoperative period did these patients who were treated by patch
undergo endoscopy? The prepyloric ulcers are not a problem, as
those are really duodenal ulcers in behavior and treatment. They
end up feeling like and undergoing a postoperative course like a
perforated duodenal ulcer does. But if one is going to look for
persistent disease or more importantly for carcinoma, an optimal
follow-up is required. If the follow-up is not secured, there is a
potential, or about one chance in eight, that the patient harbors a
carcinoma.
Dr Turner: Dr Aust has pointed out a real problem we face in
our hospital concerning delays both in getting the patients to the
hospital and in getting the complication diagnosed, and subse¬
quently getting the patient into the operating room once in the
hospital. That plagues our series, as pointed out.
He also reiterated the rarity of gastric ulcer perforations; that
has been our experience as well. We had only 107 patients over 33
years.
We have all sat at the font of knowledge that Dr Jordan
represents, and I was privileged to have him discuss my study.
We do not condemn the high mortality we found for primary
gastric resection in our study. Quite the contrary; these were
highly selected patients with larger ulcers. Larger ulcers tended
to occur in older patients, and it is not surprising at all to us that
the mortality rate in our group with primary gastric resections
was quite high.
The patients who died did so primarily of ventilatory failure and
pneumonia and not of complications directly related to the ulcer
disease in terms of leaks or the like.
The only possible exception was in the patients who had simple
closures, and I would just reiterate that I think the decision to
use simple closures in these patients was bad judgment in the
context of our advocacy of judgment's importance here. In three
of those patients intra-abdominal abscesses developed, which may,
in fact, have been related to leaks at ulcer repair sites.
Much has been said about follow-up of these patients, and all I
can say about our patients is that in our county hospital setting,
mea culpa. We advocate close follow-up, particularly endoscopy
within four weeks of closures by patch plications. However, many
of these patients do not come back for follow-up. It is a justifiable
criticism of conservative management.
Dr McCurdy mentioned the importance of judgment, and I
would reiterate that we believe this as well. We believe it was bad
judgment to perform simple closures in some patients.
Why the stomach, Dr Aust, does not tolerate simple closure, I
cannot answer other than from my own personal experience. When
you look at these ulcer margins, it does not surprise me that they
do not hold sutures very well.
Dr Baker, your question was very interesting and your comments
about dealing with high gastric perforations are intriguing. It is
precisely that problem that led me to question Dr Sawyers' report
last year at the Western Surgical Association meeting. When I
asked how they treated a high gastroesophageal junction perfo¬
ration, his answer was total gastrectomy. That led us to look at
our experience with patch closures and gastrectomies.
When do we perform endoscopy? As I mentioned, at four weeks.
Incorrect Reference Information. \p=m-\Inthe article entitled "Safe Management of the Impossible
Duodenum: Risk Avoidance in Surgery of Peptic Ulcer," published in the May 1988
Archives (1988;123:558-562), an incorrect publication date appeared in reference 8. The year of
publication should be 1972, not 1932 as printed.
Downloaded from
www.archsurg.com at University of Cincinnati, on March 20, 2011