Original article

Evaluation of a protocol for the non-operative management of

perforated peptic ulcer
C. Marshall, P. Ramaswamy, F. G. Bergin, I. L. Rosenberg and D. J. Leaper
Department of Surgery, North Tees General Hospital, Stockton-on-Tees TS19 8PE, UK
Correspondence to: Professor D. J. Leaper

Background: The non-operative management of perforated peptic ulcer has previously been shown to
be both safe and effective although it remains controversial. A protocol for non-operative management
was set up in this hospital in 1989. Adherence to the guidelines in the protocol has been audited over a
6-year period with a review of outcome.
Methods: The case-notes of patients with a diagnosis of perforated peptic ulcer were reviewed. Twelve
guidelines from the protocol were selected for evaluation of compliance to the protocol.
Results: Forty-nine patients underwent non-operative treatment initially. Eight patients failed to
respond and underwent operation. Complications included abscess formation (seven patients), renal
failure (one), gastric ileus (one), chest infection (two), and cardiac failure and stroke (one). Four deaths
occurred in this group. Adherence to certain protocol guidelines was poor, notably those concerning
prevention of thromboembolism, use of antibiotics, use of contrast examination to conrm the diagnosis
and referral for follow-up endoscopy. Two gastric cancers were detected on subsequent endoscopy.
Conclusion: This experience demonstrates that non-operative treatment can be used successfully in a
general hospital. Adherence to protocol guidelines was found to be variable and the protocol has
therefore been simplied. This study highlights the need for an accurate diagnosis and the importance
of follow-up endoscopy.

Paper accepted 19 August 1998

Introduction

Current treatment of perforated peptic ulcer still remains

largely surgical despite the fact that the non-operative
treatment of this condition has been shown in a randomized controlled trial to be both safe and effective in
selected patients1. It is known that perforated ulcers frequently seal spontaneously by the adherence of omentum
or adjacent organs to the ulcer and the rationale of nonoperative management is that in these patients operation
can be avoided2. The current guidelines for non-operative
management advocate vigorous uid resuscitation with
regular, close observation of the patient's haemodynamic
status and general condition by a senior surgeon, placement of a nasogastric tube, use of broad-spectrum antibiotics and H2-receptor antagonists and thromboembolic
prophylaxis. Follow-up oesophagogastroduodenoscopy at
46 weeks is recommended to monitor ulcer healing3.
Using this regimen the majority of patients presenting
with a clinical diagnosis of perforated peptic ulcer can
avoid laparotomy, and the morbidity and mortality rates

1999 Blackwell Science Ltd

British Journal of Surgery 1999, 86, 131134

of patients treated in this way are similar to those of

patients treated surgically1.
A protocol for the non-operative management of perforated peptic ulceration was set up at this hospital shortly
following the publication by Crofts and colleagues1, and it
became the policy to treat all patients with a clinical
diagnosis of perforated peptic ulcer by a trial of nonoperative treatment. Since its introduction the protocol
has been subject to repeated internal review as part of the
ongoing process of clinical audit; the authors' preliminary
experience has been reported previously4.
Patients and methods

The aims of this study were to evaluate the protocol in

terms of adherence to protocol guidelines and to review
the cases treated using this regimen. Guidelines for active
resuscitation included the use of antibiotics for at least
5 days, administration of crystalloid or colloid with central venous pressure monitoring if the patient was aged
over 60 years and administration of intravenous raniti-

British Journal of Surgery 1999, 86, 131134

131

132

Non-operative management of perforated peptic ulcer C. Marshall, P. Ramaswamy, F. G. Bergin et al.

dine. Exclusion criteria were a diagnosis of perforation of

longer than 24 h duration, patients on antiulcer medication aged over 70 years and failure to place a nasogastric
tube satisfactorily. During non-operative management all
patients were reviewed regularly by the surgical team and
evidence of clinical deterioration indicated by increasing
pain, rising pulse or temperature, fall in blood pressure or
generalized peritonitis was an indication for immediate
laparotomy or laparoscopy with peritoneal lavage. The
management protocol was issued to every trainee doctor
working on the surgical unit.
Patients with a diagnosis of perforated peptic ulcer
treated during the period August 1991 to August 1997
were identied using the patient database. The case-notes
of these patients were reviewed using a standard pro forma
and the data collected were entered into a Microsoft
Access database (version 7.0) for subsequent analysis.
Information gathered included patients' demographic
details, details of co-morbidity, previous history of peptic
ulceration, risk factors, outcome and complications. To
assess adherence to the protocol 12 guidelines were
selected to compare with an ideal audit standard of 100
per cent. These were that patients should have a history of
less than 24 h duration; have a nasogastric tube inserted;
undergo erect chest radiography; receive antibiotics for
5 days; receive intravenous ranitidine; receive both antiembolism stockings and subcutaneous calcium heparin;
undergo central venous pressure monitoring if aged over
60 years; undergo a water-soluble contrast study; have the
position of the nasogastric tube conrmed; receive antiulcer medication on discharge; and undergo follow-up
endoscopy at 6 weeks.
Results

Eighty patients with a diagnosis of perforated peptic ulcer

were identied; 73 sets of case-notes (91 per cent) were
available for review. A trial of non-operative management
was started in 49 patients and 21 patients who fullled the
exclusion criteria underwent primary surgical treatment.
Three patients who had an American Society of Anesthesiologists grade of 5E on admission were excluded
from the analysis. Details of the patients in the group who
had non-operative management are shown in Table 1. The
median stay in hospital was 7 (range 077) days. Of the 47
patients who had an erect chest radiograph free gas was
demonstrated in 37. Twenty patients had a water-soluble
contrast study and this demonstrated a leak of contrast in
ve cases.
Complications arising in the group of patients treated
non-operatively included intra-abdominal abscess formation in seven patients. Three patients required percutaBritish Journal of Surgery 1999, 86, 131134

Table 1 Details of patients treated by non-operative

management
No. of patients
Mean (range) age (years)
Sex ratio (M:F)
Risk factors
Smoking
History of non-steroidal
anti-inammatory drug intake
Previous history of peptic
ulceration or dyspepsia
Previous treatment with
H2 antagonist
ASA grade 3
ASA grade 4
Median (range) duration
of perforation

49
52 (1680)
34:15
29
18
15
15
7
2
6 h (range 1 h to 5
weeks)

ASA, American Society of Anesthesiologists

neous drainage and one patient required a laparotomy for

drainage. One subphrenic abscess was noted coincidentally during laparotomy for a bleeding ulcer. In addition,
one patient developed a serous pelvic collection requiring
percutaneous drainage. Other complications included
acute renal failure (one patient), prolonged ileus (one),
chest infection (one), atrial brillation (one), and cardiac
failure and stroke (one).
Three patients treated in the non-operative group died:
one patient died on the day following admission with a
myocardial infarction; one patient developed intra-abdominal abscesses and a chest infection, and died 4 days
following admission; and one death occurred in a patient
who presented with a 3-week history of peritonitis but
who was treated conservatively.
Eight patients failed a trial of non-operative management and proceeded to operation. In two of these the
reasons behind the decision to perform laparotomy were
not documented in the case-notes. Four patients required
surgical intervention owing to failure to settle with nonoperative treatment. In addition, one patient underwent
laparotomy at 8 days for oversewing of a perforated duodenal ulcer which bled acutely and one patient required a
late laparotomy for drainage of intra-abdominal abscesses.
Complications in the group of patients who failed initial
non-operative management included wound infection
(one patient), gastric ileus (one), and symptoms of alcohol
withdrawal and chest infection (one). One patient in this
group, who underwent laparotomy 36 h after admission
for failure to settle, subsequently developed systemic
inammatory response syndrome and multiple organ
dysfunction syndrome and died 8 days after operation.
Only 22 patients in the non-operative group had a
follow-up oesophagogastroduodenoscopy. Eleven of these
1999 Blackwell Science Ltd

C. Marshall, P. Ramaswamy, F. G. Bergin et al. Non-operative management of perforated peptic ulcer 133

procedures showed a positive nding: gastric carcinoma

(two patients), evidence of previous duodenal ulceration
(four), duodenitis (two), gastritis and gastric ulceration
(two), positive urease test for Helicobacter pylori (one) and
Candida oesophagitis (one).
Of the 21 patients who underwent primary operative
treatment, complications included abscess formation (one
patient), atrial brillation (one), duodenal stula, sepsis
and empyema (one), chest infection (one) and symptoms
of alcohol withdrawal (one). Three patients died in this
group; one as a result of abscess formation and subsequent
septicaemia; one patient died 5 days after operation (cause
of death not recorded in notes); and one patient died
following discharge but within 30 days of surgery (cause
of death not known).
Of the seven sets of case-notes which were unobtainable, limited information regarding patient management
and outcome was available from the Micromed (Medical
Systems, Great Missenden, UK) database. Three of these
patients had successful non-operative management although one died within 30 days of discharge from a dissecting aortic aneurysm. Two patients underwent surgery
for perforated ulcer and two died shortly after admission
with renal failure and sepsis. It is not clear whether these
four patients received a trial of non-operative management.
Assessment of adherence to protocol guidelines using
the 12 selected audit standards is shown in Table 2.

Table 2 Compliance with protocol guidelines for selected audit

standards in 49 patients who underwent non-operative treatment
Audit standard

No. achieving
audit standard
(n = 49)

History of less than 24 h duration

Nasogastric tube inserted
Erect chest radiograph performed
Antibiotics for 5 days
Intravenous ranitidine
Anti-embolism stockings
Subcutaneous calcium heparin
CVP monitoring if aged over 60 years
(n = 22)
Contrast study performed
Nasogastric tube position conrmed
Discharged on H2 antagonist or PPI
(n = 45)
Follow-up endoscopy performed
(n = 45)

44
48
47
25
40
13
20
1

(90)
(98)
(96)
(51)
(82)
(27)
(41)
(5)

20 (41)
21 (43)
35 (78)
22 (49)

Values in parentheses are percentages. CVP, Central venous pressure;

PPI, proton pump inhibitor

1999 Blackwell Science Ltd

Discussion

The feasibility of non-operative management of perforated

peptic ulcer has been known since 19515, but evidence that
this is a safe method of treating selected patients was not
provided until the rst randomized controlled trial of this
treatment. However, non-operative management has not
been adopted widely and still remains controversial. The
reasons for this may include the repeated patient review
required by a senior surgeon, the potential problems with
misdiagnosis, the lack of opportunity for performing denitive ulcer surgery and an individual surgeon's preference for surgical treatment. The present experience
demonstrates that a dened protocol for non-operative
management can be applied effectively in the setting of a
general hospital and can achieve results comparable to
those of previously published work. The study by Crofts
et al.1 showed a failure of non-operative management in 28
per cent of cases (11 of 40) with a morbidity rate of 50 per
cent and a mortality rate of 5 per cent. In the group randomized to surgery the morbidity rate was 40 per cent and
the mortality rate 5 per cent. In the present patients selected for non-operative treatment there was a conversion
rate to operation of 16 per cent (eight of 49), a morbidity
rate of 24 per cent and a mortality rate of 8 per cent. In
patients undergoing primary surgery the morbidity rate
was 24 per cent (ve of 21) and the mortality rate 14 per
cent. This high mortality rate in the patients treated by
primary laparotomy probably reects the fact that patients
who full the exclusion criteria for non-operative treatment are at higher risk.
Despite the widespread availability and high awareness
of the presence of this protocol among the surgeons on
the unit treating this condition, adherence to the guidelines of the protocol is disappointing. The reasons for the
large numbers of protocol violations are unclear but
probably represent a lack of familiarity with the guidelines
and variations in clinical practice between surgeons. These
results highlight the need for regular detailed audit using
achievable standards in order to identify problems with an
existing protocol. To improve compliance with the
guidelines for treatment the protocol has recently been
reviewed and simplied as a result of this work. In future
every patient presenting with a clinical diagnosis of
perforated peptic ulcer will have a guideline check list
inserted in the case-notes so that each intervention can be
ticked off as it is carried out, making omissions in treatment more clearly identiable. Of particular importance is
the need for all patients to be referred for follow-up
endoscopy to monitor ulcer healing, test for H. pylori and
to provide an accurate diagnosis which is not always

British Journal of Surgery 1999, 86, 131134

134

Non-operative management of perforated peptic ulcer C. Marshall, P. Ramaswamy, F. G. Bergin et al.

possible before discharge from hospital. This is illustrated

by the fact that although only 22 of 49 patients in the
present study underwent follow-up endoscopy two gastric
cancers were picked up during this procedure.
Simplications to the revised protocol now include the
use of a water-soluble contrast study only when the
diagnosis is in doubt. The use of central venous pressure
monitoring is now recommended based on clinical indications rather than in every patient aged over 60 years.
Follow-up endoscopy at 6 weeks is mandatory and
H2-receptor antagonists or a proton pump inhibitor
should be stopped 2 weeks before the procedure in order
to increase the sensitivity of Helicobacter detection.
In conclusion, this study has demonstrated that the
authors' current policy of non-operative management of
selected patients with perforated peptic ulcer is both safe
and efcacious although adherence to the guidelines of the

British Journal of Surgery 1999, 86, 131134

protocol could be improved substantially to optimize

patient care.
References
1 Crofts TJ, Park KGM, Steele RJC, Chung SSC, Li AKC. A
randomized trial of nonoperative treatment for perforated
peptic ulcer. N Engl J Med 1989; 320: 9703.
2 Donovan AJ, Vinson TL, Maulsby GO, Gewin JR. Selective
treatment of duodenal ulcer with perforation. Ann Surg 1979;
189: 62736.
3 Raimes SA, Devlin HB. Perforated duodenal ulcer. Br J Surg
1987; 74: 812 (Editorial).
4 Rigg KM, Stuart RC, Rosenberg IL. Conservative management of perforated peptic ulcer. Lancet 1990; 335: 673
(Letter).
5 Taylor H. Aspiration treatment of perforated ulcers. Lancet
1951; i: 712.

1999 Blackwell Science Ltd