0016-5107/95/4102-013053.00+ 0
GASTROINTESTINAL ENDOSCOPY
Copyright | 1995 by the American Society for Gastrointestinal Endoscopy
Short- and long-term outcome of esophageal
perforation
Richard Sawyer, MD, Clay Phillips, MD, Nimish Vakil, MD
Rochester, New York
Esophageal perforation is an important complication of endoscopy and dilation, but few data are available regarding long-term outcome in these patients.
Thirty-one consecutive patients with esophageal perforation were studied.
Long-term outcome was determined by office follow-up and questionnaires
via mail or telephone. The study group included 16 men and 15 women with
a mean age of 61 +_ 4 years. Instrumentation was the principal cause of perforation (77%), and pain was the principal symptom in the majority of cases.
However, some patients reported no pain, their only symptom being shortness
of breath or dysphagia on swallowing saliva. Contrast examinations were used
to establish the diagnosis, but results were initially negative in 3 of 27 patients
(11%). Seven patients were treated nonsurgically (nasogastric drainage, antibiotics, and intravenous alimentation). Fourteen patients underwent primary
closure of the esophageal rent. Six patients underwent surgical drainage, 3
underwent total esophagectomy, and I died during surgery before the esophagus was exposed. Thirty-day mortality was 6.4%. Mean hospital stay was
26 +_ 5 days. Postdischarge follow-up was available in 28 of the 29 patients
(97%) surviving initial hospitalization, and the mean follow-up was 47 _+ 9
months. Nineteen patients had persistent dysphagia, and 6 underwent dilation
again. One of these patients had a second perforation. Contrary to results of
some earlier studies, the mortality from esophageal perforation was low in our
study because of early intervention in suspected cases. Results of contrast
studies are sometimes negative in the early stages after perforation, and
studies may need to be repeated. Long-term dysphagia is a difficult problem
in patients with strictures, and some patients need further dilation, with the risk
for another perforation. In contrast, long-term dysphagia is rare in patients with
barogenic perforation. (Gastrointest Endosc 1995;41:130-4.)
Esophageal perforation is a serious complication of
instrumentation in the esophagus and also occurs as a
result of increased luminal pressure (barogenic perforation), usually as a result of vomiting (Boerhaave's
syndrome). 1 With advances in antibiotic therapy and
greater awareness of this complication, patients are
Received January 11, 1994. For revision February 23, 1994.
Accepted May 21, 1994.
From the Departments of Medicine and Thoracic Surgery, University of Rochester Medical Center, Rochester, New York.
Presented in part at the 58th annual meeting of the American College of Gastroenterology, New York, October 1993.
Reprint requests: Nimish Vakil, MD, University of Wisconsin
Medical School, SSMC, 945 North 12th Street, A 324 A, Milwaukee, WI 53233.
37/1/58186
130
GASTROINTESTINAL ENDOSCOPY
receiving early treatment, and many survive the initial
episode of perforation. However, few data are available
regarding long-term outcome in these patients. 2 The
aim of this study was to determine the short- and
long-term outcome of therapy for esophageal perforation.
METHODS
Strong Memorial Hospital is a regional referral center for
the treatment of esophageal perforation. From 1980 to 1993,
31 consecutive patients were treated for esophageal perforation. These patients were identified retrospectively by
searching a computerized database of patient diagnoses. The
hospital and office records of these patients were reviewed.
Long-term outcome was determined by patient interview,
review of follow-up data, and telephone or mail questionVOLUME 41, NO. 2, 1995
naires. Dysphagia was assessed and graded as follows: 0, no
dysphagia; 1, occasional dysphagia with some solids, minor
dietary limitations; 2, dysphagia with most solids, able to eat
soft solids only; 3, able to tolerate semisolids only; 4, able to
swallow liquids only; 5, severe dysphagia to solids and
liquids. Results were expressed as mean + SEM.
Techniques used for dilation, postdilation management,
and treatment of suspected and proven perforation are described below.
Dilation technique and immediate postdilation
management
Balloon dilation for achalasia. This procedure was
performed on an outpatient basis, usually early in the morning to permit observation for several hours. All patients underwent a contrast esophagogram immediately after the
procedure. If no evidence of perforation was found, patients
were given water to drink after recovery from anesthesia.
Pain or dysphagia after water ingestion was considered suggestive of perforation, and these patients then underwent
another contrast examination. Both the cases in this study
were detected by contrast studies immediately after dilation.
M a l o n e y a n d S a v a r y dilation. Savary dilation was
always performed under fluoroscopic guidance after endoscopic or combined endoscopic and fluoroscopic placement
of a guidewire. Maloney dilation was usually not performed
under fluoroscopic guidance. After dilation, patients were
observed for at least 30 to 45 minutes if unsedated or until
recovery from sedation. If the patient reported chest pain or
dysphagia on swallowing saliva, a contrast esophagogram
was immediately performed. Patients with minor throat
discomfort or no symptoms were given a glass of water to
swallow before discharge. Pain, dysphagia, or shortness of
breath on ingestion of water were indications for a contrast
examination.
Criteria for treatment of perforation
The choice of therapy for perforation was based on clinical criteria. Patients were treated nonsurgically if they met
the following criteria adapted from Cameron et al.3:
1. Recent perforation (within 24 hours);
2. No food intake after the perforation;
3. Perforation not proximal to a high-grade stenosis;
4. Minor symptoms (pain or dysphagia) without clinical
signs of sepsis or hemodynamic compromise;
5. Perforation contained within the mediastinum without
contamination of adjacent body cavities (e.g., pleural
space);
6. Contrast studies showing a small perforation with good
drainage of contrast material back into the esophagus.
In surgically treated patients, primary closure was usually
selected for patients with recent perforations and a normal
esophagus or those with recent perforations and minor
strictures with a reasonable prospect of good esophageal
function. Esophagectomy was performed if the perforation
was longstanding within a suppurative environment and if
the underlying disease made reasonable function unlikely
even with closure of the perforation (e.g., mega-esophagus,
neoplasm, and chronic refractory stricture). Decisions were
individualized and also based on the general condition of the
patient.
VOLUME 41, NO. 2, 1995
RESULTS
Patient characteristics
S t u d y subjects included 16 m e n and 15 women.
T h e i r m e a n age was 61 _+ 4 years, with a range of 21 to
89 years.
Etiology and presentation
I n s t r u m e n t a t i o n was the most c o m m o n cause of
esophageal perforation, accounting for 23 patients
(77%). Seventeen patients u n d e r w e n t dilation for
esophageal strictures proved to be benign by previous
endoscopic biopsy and cytology studies. Sixteen of
these were chronic strictures caused by reflux esophagitis, and 1 was a radiation stricture. In these 17 patients, perforations developed in 5 with Savary dilators; all of these procedures were p e r f o r m e d u n d e r
fluoroscopic guidance. P e r f o r a t i o n developed in 6
patients after dilation with Maloney dilators; these
procedures were p e r f o r m e d without fluoroscopy. Perforation developed in 1 p a t i e n t after through-the-endoscope balloon dilation of a benign esophageal stricture. P e r f o r a t i o n occurred in 2 patients after pneumatic dilation for achalasia and in I as a complication
of sclerotherapy. A t t e m p t s at i n t u b a t i o n caused perforation in 3 patients: 1 during intubation with a rigid
esophagoscope, 1 during endotracheal intubation, and
1 during E R C P with a flexible endoscope, after which
a high cervical perforation of a Zenker's diverticulum
developed. T h e latter was the only case of perforation
after diagnostic flexible endoscopy during this period.
P e r f o r a t i o n occurred after vomiting (barogenic perforation) in 5 patients, and no obvious cause (nonbarogenic spontaneous perforation) could be found in
1. P e r f o r a t i o n occurred in 1 patient after foreign body
ingestion and in a n o t h e r after blunt t r a u m a to the
chest.
T h e perforation rate with individual procedures
could not be d e t e r m i n e d with accuracy for the entire
s t u d y group because 12 of the 31 patients in this s t u d y
were t r a n s f e r r e d to our hospital from other institutions and details of annual procedure rates were not
available. Accurate d a t a regarding individual procedures from our institution ( J a n u a r y 1992 to J u n e 1993)
provide some estimate of the frequency of perforation.
During this period, 1844 UGI endoscopic examinations were p e r f o r m e d with no perforations, 6 balloon
dilations for achalasia with 1 perforation, 57 pneumatic dilations of stenoses with no perforations, 235
Maloney dilations with no perforations, and 99 Savary
dilations with 2 perforations.
Symptoms
Pain, the most f r e q u e n t presenting symptom, was
n o t e d in 28 of the 31 cases (90%). Twelve patients
GASTROINTESTINAL ENDOSCOPY
131
Table 2.
Long-term dysphagia and type of surgery
(n 28*)
Table 1.
Cause of perforation and type of surgery
(n = 30*)
Cause
Esophageal
dilation
(n = 19)*
Intubation
=
Nonsurgical Primary Surgical Esophagectomy
therapy
closure drainage
4
9
3
2
0
3
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
2
1
1
1
0
0
0
(n = 3)
Sclerotherapy
(n = 1)
Blunt trauma
(n = 1)
Foreign body
(n = 1)
Barogenic
(n = 5)
Spontaneous
(n = 1)
*One patient died during surgery before the esophagus was
exposed; 2 patients undergoing balloon dilation for achalasia had
myotomy in addition to closure of the perforation.
(36%) had dysphagia and 6 patients (18%) complained of shortness of breath.
Diagnosis
Chest radiographs were performed in 24 patients
and revealed mediastinal air in 2 of them. Contrast
studies were initially diagnostic in 24 of 27 subjects in
whom they were performed. Results of barium swallow
studies were initially normal in 2 cases but when
repeated several hours later demonstrated perforation. The third patient with a normal barium swallow
study had mediastinal air on the chest radiograph and
did not undergo another contrast study. Three patients were diagnosed by gastroscopy, and 1 perforation was recognized at the time of rigid endoscopy. Of
the 31 perforations, 6 occurred in the upper esophagus,
4 in the midesophagus, and 21 in the lower esophagus.
Treatment
Twenty-nine of the 31 patients presented within 24
hours, and all suspected perforations were initially
treated with nasogastric aspiration and intravenous
antibiotics. Seven patients were treated nonsurgically
(nasogastric drainage, antibiotics, and intravenous alimentation) using the criteria outlined above.
Fourteen patients underwent primary closure of the
esophageal rent. Six patients underwent surgical drainage, and 3 patients underwent total esophagectomy
(Table 1). Gross contamination of the pleural space
and mediastinitis were noted in i of these cases; in another purulent mediastinitis was present and the exact location of the perforation could not be identified.
The third patient had a distal esophageal stricture and
a rent that could not be adequately treated with a
132
GASTROINTESTINAL ENDOSCOPY
Dysphagia
grade
Nonsurgical
therapy
Primary
closure
Surgical
drainage
0
1
2
3
4
5
4
0
0
2
1
0
1
3
1
3
5
0
2
2
0
0
1
1
Es~176
2
0
0
0
0
0
*Two patients died within 30 days of treatment and 1 patient
was lost to follow-up.
fundal patch. Hyperalimentation or feeding through a
jejunostomy tube inserted at surgery was used in all
surgically treated patients. Six patients with small
perforations who were treated medically could be
managed with peripheral alimentation alone.
Early outcome
Two deaths occurred in the surgically treated patients and none in the medically treated patients. One
patient died during surgery of hemorrhage from large
collateral vessels near the gastroesophageal junction.
Another patient died of sepsis in the postoperative
period (2 weeks after esophagectomy). The 30-day
mortality was therefore 6.4 %. All the other patients
recovered and were discharged from the hospital.
Mean duration in the hospital after perforation was
26 + 5 days, with a range of 5 to 115 days.
Long-term results
Of the 29 patients surviving initial hospitalization,
follow-up information was available for 28 (97%).
Three patients died during the follow-up period, 2
from unrelated malignancies and 1 from multiple organ failure, also unrelated to the perforation. The
mean follow-up was 47 + 9 months, with a range of 3
to 150 months. Nine patients reported no dysphagia
after surgery; 1 of these patients had a primary diagnosis of achalasia, and the others had benign strictures
of the esophagus. The other patients reported varying
grades of dysphagia (Tables 2 and 3). Treatment of the
dysphagia was individualized according to underlying
disease, overall medical condition of the patient, and
patient choice. Understandably, patients were reluctant to undergo any procedures despite symptoms.
Dysphagia was sufficiently severe to warrant esophageal dilation in 6 patients. Five of these patients had
a benign esophageal stricture secondary to reflux, and
1 had a stricture after sclerotherapy for esophageal
varices. All patients underwent esophagoscopy and
barium swallow studies, and surgical therapy was considered in all. Three patients were determined to be
unfit for surgery, 1 because of advanced liver disease
and portal hypertension and 2 because of their poor
VOLUME 41, NO. 2, 1995
Table 3.
Long-term dysphagia and cause of perforation
(n = 28*)
Grade Grade Grade Grade Grade Grade
Cause
0
1
2
3
4
5
Esophageal dilation 4
2
1
4
5
1
(n = 17)
Intubation
2
1
(n = 3)
Sclerotherapy
1
(n = 1)
Blunt trauma
(n = 1)
Foreign body
(n = 1)
Spontaneous
1
1
(n = 5)
*Two patients died within 30 days of treatment and 1 patient
was lost to follow-up.
medical condition. Three patients refused further
surgical intervention. Five patients had no complications from subsequent dilation, but another perforation developed in 1 patient as a result of subsequent
dilation. A second barogenic perforation, requiring
surgery, also occurred in a patient with barogenic perforation.
DISCUSSION
As in other studies, esophageal dilation was the most
frequent cause of esophageal perforation, 4-6 and the
most frequent site of perforation was the thoracic
esophagus. The only case of perforation caused by fiberoptic endoscopy, a high cervical perforation of a
Zenker's diverticulum, occurred during blind intubation for ERCP. Diverticula have been described to
predispose to perforation in previous studies; the incidence has ranged from 2 % to 15 % .7, s The absence of
patients with malignancy in our study deserves some
comment. In part it represents the low incidence of
squamous cell carcinoma in our patient population
and concurrent trials of radiotherapy and chemotherapy for the palliation of dysphagia. Some studies have
suggested that bougienage does not necessarily entail
a higher risk for perforation in malignancy.9
Barogenic perforation refers to esophageal disruption caused by an abrupt rise in esophageal pressure.
In our study, forceful vomiting was the most common
cause of barogenic perforation in otherwise healthy
patients; 1 patient had nonbarogenic spontaneous
perforation of undetermined cause. Pain, dysphagia,
and shortness of breath were the most frequent symptoms, but it is noteworthy that even in the cases of
acute perforation after dilation, some patients did not
report pain but reported dysphagia when swallowing
saliva. Results of initial contrast studies were negative
in 3 of 27 patients (11% ), and this figure is similar to
data reported in other studies. 1~ All 3 had received
V O L U M E 41, NO. 2, 1995
Table 4.
Long-term dysphagia and site of perforation
Dysphagia
Cervical
Distal
grade
esophagus Midesophagus esophagus
0
2
1
6
1
4
0
1
2
0
0
1
3
0
1
4
4
0
1
6
5
0
1
0
barium as the contrast agent. It is important to recognize that this may occur when studies are performed
soon after perforation, perhaps because of edema at
the site of a small perforation, and that patients should
be presumptively treated if a perforation is suspected.
The contrast study should then be repeated after several hours. Foley et al. u described 6 patients in whom
water-soluble contrast failed to demonstrate a perforation, which could then be demonstrated by administration of barium. Our study demonstrates that the
same may also be true with barium. Some experts recommend that the initial contrast examination be performed with a water-soluble contrast agent, which is
thought to cause less irritation of the mediastinum; if
results are negative, this study should be followed by
administration of barium, n Others recommend the
use of thinned barium as the agent of choice. 12 In either case, it is important to repeat the study after several hours if the clinical picture is suggestive of perforation but results of the initial examination are negative. Esophagoscopy can be used to identify the site
and size of the perforation, t but visualization is quite
difficult if a contrast study precedes this examination.
Additionally, insuffiation of air may increase contamination of the mediastinum if a perforation is present
and convert a small perforation that may be treated
medically into an overt perforation requiring surgery.
The symptoms of perforation can be subtle, and the
low mortality in this study is in part related to early
investigation in patients with dysphagia, chest pain, or
shortness of breath after dilation. A thoracic surgeon
was involved in the care of each patient early in the
course, and decisions to treat nonoperatively were
made in consultation with the surgeon. This guards
against the natural tendency of the endoscopist causing the perforation to favor conservative therapy while
precious time is lost. Delay in treatment of more than
24 hours is associated with a greater risk of complications and a higher mortality. In our study, patients
who did not meet criteria for nonsurgical management
underwent surgery promptly (within 24 hours in the
majority of cases), and no patient initially treated with
medical therapy required subsequent surgery.
The mortality from esophageal perforation is quite
variable and has ranged from 7% to 46% .5,13 In a colGASTROINTESTINAL ENDOSCOPY
133
lective review of 450 patients, Jones and Ginsberg 12
reported that iatrogenic and instrumental perforations were associated with a mortality of 19 %, whereas
spontaneous perforations had a mortality of 39 %. Two
other studies 14,15 have, however, shown no difference
between spontaneous and instrumental perforation. In
our study, 1 patient with spontaneous and 1 with instrumental perforation died after surgery. One of these
patients had previously unrecognized portal hypertension and died of hemorrhage during surgery, and
the other died of sepsis in the postoperative period.
The higher mortality in the surgical group is to be
expected because these patients tend to have large
perforations with gross contamination of the mediastinum, a delayed diagnosis, and life-threatening
complications that should never be present in the patients treated medically.
Few data on the long-term results of surgical therapy for esophageal perforation are found in the literature. Saabye et al. 2 reported 35 patients who survived
simple closure and drainage of the esophagus. They
found that long-term dysphagia was rare in the patients who had a normal esophagus before perforation.
Severe reflux esophagitis and an esophageal stricture
that required surgery developed in 1 of the 7 patients
with barogenic perforation in that study; long-term
dysphagia did not occur in any of the 8 patients with
instrumental perforation and a normal esophagus. Of
11 patients who had esophageal strictures, 6 had no
postoperative dysphagia, 4 required long-term dilation, and 1 required later esophagectomy. These data
are unusual because simple closure of esophageal perforations, particularly in patients with a distal esophageal stricture, has been associated with a poor outcome, and many surgeons recommend more definitive
surgery for the stricture in one or two stages. 4 The decision to operate once again on a patient with dysphagia must be determined on a case-by-case basis. If significant dysfunction is present after the initial repair,
antireflux surgery or esophagectomy may be indicated
depending on the prospect of improved function. An
antireflux procedure would be appropriate in a patient
with persistent reflux and moderate dysfunction, but
dilation might be more effective in the setting of
reduced acid reflux. Esophagectomy would be indicated in patients with severe dysfunction and poor
prospects of recovery, provided the patient had an acceptable risk for surgery. Both trans-hiatal esophagectomy and Ivor Lewis esophagectomy carry the risk of
serious morbidity that must be considered in making
this decision. In the follow-up of our patients, we found
it noteworthy that some patients refused further therapy with dilation or surgery despite persistent dysphagia and the continued need for supplemental nu-
134
GASTROINTESTINAL ENDOSCOPY
trition by jejunostomy feedings. Others with severe
dysphagia refused surgery b u t submitted to dilation,
and a small number of patients were treated with dilation because they were considered poor risks for surgery. As in Saabye's study, we found that long-term
dysphagia was very rare in patients with an intrinsically normal esophagus. In 1 patient with perforation
after sclerotherapy for esophageal varices, a postsclerotherapy stricture and significant dysphagia developed, and a reflux stricture developed in a patient with
barogenic perforation. Serious dysphagia (grade 3 or
more) was confined to midesophageal or distal esophageal lesions.
Establishing fixed criteria for the assessment of patients in the postdilation recovery period and recognizing that the symptoms of perforation may be quite
subtle may help in the early diagnosis and management of this devastating complication.
ACKNOWLEDGMENT
We thank Dr. Richard Feins of the Department of
Thoracic Surgery for assistance with data acquisition.
REFERENCES
1. Curci JJ, Horman MJ. Boarhaave's syndrome: the importance
of early diagnosis and treatment. Ann Surg 1976;183:401-8.
2. Saabye J, Nielsen HO, Andersen K. Long-term observation following perforation and rupture of the esophagus. Scand J Thorac Cardiovasc Surg 1988;22:79-80.
3. Cameron JL, Kieffer RF, Hendrix TR, Mehigan DG, Baker RR.
Selective nonoperative management of contained intrathoracic
esophageal disruptions. Ann Thorac Surg 1978;27:404-8.
4. Moghissi K, Pender D. Instrumental perforations of the esophagus and their management. Thorax 1988;43:642-6.
5. Kim-Deobald J, Kozarek RA. Esophageal perforation: an 8-year
review of a multispecialty clinic's experience. Am J Gastroenterol 1992;87:1112-9.
6. Goldstein LA, Thompson WR. Esophageal perforations: a 15
year experience. Am J Surg 1982;143:495-503.
7. Larsen K, Jensen BS, Axelson F. Perforation and rupture of the
esophagus. Scand J Thorac Cardiovasc Surg 1983;17:311-8.
8. Schulze S, Pederson VM, Hoier-Madsen K. Iatrogenic perforation of the esophagus. Acta Chir Scand 1982;148:679-82.
9. Heit HA, Johnson LF, Siegel SR, Boyce HW. Palliative dilation
for dysphagia in esophageal carcinoma. Ann Intern Med 1978;
89:623-31.
10. Sarr HG, Pemberton JH, Payne WS. Management of instrumental perforations of the esophagus. J Thorae Cardiovase Surg
1982;84:211-8.
11. Foley MJ, Ghahremani G, Rogers L. Reappraisal of contrast
media used to detect upper gastrointestinal perforations. Radiology 1982;144:231-7.
12. Jones WG, Ginsberg RJ. Esophageal perforation: a continuing
challenge. Ann Thorac Surg 1992;53:534-43.
13. Flynn AE, Verrier ED, Way LW, Thomas AN, Pellegrini CA.
Esophageal perforation. Arch Surg 1989;124:1211-4.
14. Graeber GM, Niezgoda JA, Burton NA, Collins GJ, Zajtcbuk R.
A comparison of patients with endoscopic esophageal perforations and patients with the Boerhaave's syndrome. Chest 1987;
92:995-8.
15. Triggiani E, Belsey R. Oesophageal trauma: incidence, diagnosis and management. Thorax 1977;32:241-9.
VOLUME 41, NO. 2, 1995