Acute necrotizing esophagitis: role of nonsteroidal anti-inflammatory drugs.

Abstract

Acute necrotizing esophagitis (ANE) is a rarely described entity that is thought to be a cause of upper gastrointestinal (UGI) bleeding, The present study examined the incidence of ANE among patients with UGI bleeding, as well as the clinical features of ANE, and the coexisting illnesses and medication histories of ANE patients.
A retrospective analysis of clinical and endoscopic findings and the clinical course in 16 patients with ANE was carried out over a 3-year period.
We observed 16 patients (6%) of ANE in 239 patients with UGI bleeding during the 3-year period. The average age of the patients was 62.5 years. The lesions predominantly affected the lower third of the esophagus, and hiatal hernia was the most common (63%) coexisting endoscopic finding. All patients had coexisting disease. Fifty percent of patients with ANE (eight patients) had taken nonsteroidal anti-inflammatory drugs (NSAIDs). ANE also occurred in four patients with diabetic ketoacidosis. Supportive therapy, including parenteral nutrition and administration of a proton pump inhibitor, was effective.
ANE is more common than has been previously reported, and it should be included in the differential diagnosis of UGI bleeding. ANE could be characterized as an "acute esophageal mucosal lesion," particularly in aged patients with hiatal hernia and among those who consume NSAIDs.

Full text

Page 1

J Gastroenterol 2006; 41:193–197
DOI 10.1007/s00535-005-1741-6
Acute necrotizing esophagitis: role of nonsteroidal
anti-inflammatory drugs
Hiroshi Yasuda, Masaya Yamada, Yutaka Endo, Kazuaki Inoue, and Makoto Yoshiba
Division of Gastroenterology, Showa University Fujigaoka Hospital, 1-30 Fujigaoka, Yokohama 227-8501, Japan
and variceal bleeding2 have been well characterized as
causes of UGI bleeding from the esophagus, bleeding
associated with esophagitis or esophageal ulcer is less
understood. Acute necrotizing esophagitis (ANE) has
been reported as a rare cause of esophageal bleeding.
ANE becomes evident with UGI bleeding, and at
endoscopy, the esophagus appears dark black in color.
Goldenberg et al.3 reported the first endoscopic descrip-
tion of ANE. Since then, several patients of ANE have
been reported.4,5 Recently, a relatively large study of 29
cases of ANE was reported.6 In that study, ANE was
found in 0.28% of all UGI endoscopies, and most pa-
tients had comorbid conditions such as diabetes melli-
tus, hypertension, and congestive heart failure.
Nonsteroidal anti-inflammatory drugs (NSAIDs) are
one of the most important factors in the pathogenesis
of gastric and duodenal ulcers.7 The use of these drugs
has also been associated with ulceration in the small
and large bowel. Several studies have suggested that
NSAIDs play a role in the etiology of benign esoph-
ageal stricture and esophageal ulceration.8–12 The role of
gastroesophageal reflux and NSAIDs in causing bleed-
ing esophagitis is poorly understood.
We observed 16 cases of ANE among 239 patients
with UGI hemorrhage during a 3-year period. The
present study examined clinical features, coexisting ill-
nesses, and patient use of medications, NSAIDs in par-
ticular, among these cases.
Patients and methods
Between January 2002 and December 2004, 239 admis-
sions for UGI hemorrhage were identified at Showa
University Fujigaoka Hospital, Yokohama, Japan. All
patients with UGI bleeding manifested by coffee-
ground emesis, hematemesis, or melena and who under-
went endoscopy were included in this study. Informed
consent was obtained from all patients before enroll-
Background. Acute necrotizing esophagitis (ANE) is a
rarely described entity that is thought to be a cause
of upper gastrointestinal (UGI) bleeding, The present
study examined the incidence of ANE among patients
with UGI bleeding, as well as the clinical features of
ANE, and the coexisting illnesses and medication histo-
ries of ANE patients. Methods. A retrospective analysis
of clinical and endoscopic findings and the clinical
course in 16 patients with ANE was carried out over a 3-
year period. Results. We observed 16 patients (6%) of
ANE in 239 patients with UGI bleeding during the 3-
year period. The average age of the patients was 62.5
years. The lesions predominantly affected the lower
third of the esophagus, and hiatal hernia was the most
common (63%) coexisting endoscopic finding. All pa-
tients had coexisting disease. Fifty percent of patients
with ANE (eight patients) had taken nonsteroidal anti-
inflammatory drugs (NSAIDs). ANE also occurred
in four patients with diabetic ketoacidosis. Supportive
therapy, including parenteral nutrition and administra-
tion of a proton pump inhibitor, was effective. Conclu-
sions. ANE is more common than has been previously
reported, and it should be included in the differential
diagnosis of UGI bleeding. ANE could be characterized
as an “acute esophageal mucosal lesion,” particularly in
aged patients with hiatal hernia and among those who
consume NSAIDs.
Key words: acute necrotizing esophagitis, nonsteroidal
anti-inflammatory drugs, diabetic ketoacidosis
Introduction
Upper gastrointestinal (UGI) bleeding can result from
a wide variety of causes. While Mallory-Weiss tear1
Received: May 2, 2005 / Accepted: November 9, 2005
Reprint requests to: H. Yasuda

Page 2

194H. Yasuda et al.: ANE and NSAIDs
ment. The causes of UGI bleeding in the 239 admissions
are presented in Table 1. The mean patient age was 65
years (range, 18–92 years).
The following diagnostic criteria for ANE, as de-
scribed by Augusto et al.,6 were used: acute presenta-
tion of hematemesis or melena; endoscopic finding of a
diffuse or spotty black-appearing esophageal mucosa,
which is more accentuated in the lower third and which
terminates abruptly at the esophagocardiac junction;
and complete necrosis of the mucosa, sometimes
extending to the muscular wall in histological examina-
tion. Other causes of acute esophagitis, such as caustic,
infectious, traumatic, or radiotherapy factors, were ex-
cluded. Hiatal hernia was classified endoscopially
according to Makuuchi:13 Grade 0, normal (mucosal
junction is seen below hiatus); grade C, minor hernia
(gastric mucosa is seen partially above hiatus); grade B,
minor hernia (gastric mucosa is seen circularly less than
3cm above hiatus); grade A, definite hiatal hernia (sac-
shaped portion covered with gastric mucosa is seen
more than 3cm above hiatus). Helicobacter pylori infec-
tion was assessed by enzyme-linked immunosorbent as-
say (ELISA) for ant-H. pylori IgG antibody.
Results
Among the 239 patients of UGI bleeding, ANE was
diagnosed in 16 patients (Table 2). These patients in-
cluded 12 men and 4 women with a mean age of 62.5
years (range, 32–85 years). Coffee-ground emesis was
the most common manifestation of hemorrhage (10
of 16 patients, 63%). Hematemesis was seen in four
patients (25%), and melena was seen in two patients
(12.5%). Endoscopic examination revealed diffuse or
spotty circumferential necrosis of the mucosa with
mucosal bleeding, especially in the lower third of the
esophagus. There was an abrupt interruption at the Z
line. One patient developed stricture, and balloon dila-
tation was performed. Several patients had coexisting
abnormal endoscopic findings. Ten patients (62%) had
grade B to A (minor to definite) hiatal hernia; however,
Table 1. Causes of UGI bleeding
CauseCases
Gastric ulcer
Duodenal ulcer
Variceal bleeding
Mallory-Weiss
ANE
Gastric cancer
Other
Total
131
38
31
12
16
9
2
239
UGI, upper gastrointestinal; ANE, acute necrotizing esophagitis
Table 2. Summary of the case records
Case
Age/sex
NSAIDs
Coexisting diseases
Helicobacter pylori
Treatment
Hiatal herniaa
1
76/M
Aspirin, 100mg/day
Diabetic ketoacidosis
+
PPI
A
2
67/M
Aspirin, 81mg/day
Osteoarthritis
+
PPI, balloon dilatation
B
3
65/M
Aspirin, 81mg/day
Ischemic heart disease
ND
PPI
C
4
82/F
Aspirin, 81mg/day
Hypertension
ND
PPI
B
5
75/M
Tiaramide hydrochloride, 300mg/day
Osteoarthritis
−
PPI
A
6
78/M
Loxopronin, 180mg/day aspirin, 81mg/day
Rheumatoid arthritis
+
HPU, PPI
A
7
84/M
—
Atrial fibrillation
−
PPI
A
8
68/M
Loxopronin, 180mg/day
Pancratic carcinoma
−
PPI
C
9
60/M
—
Diabetes mellitus
−
PPI
C
10
68/M
—
Lymphoma
ND
PPI
B
11
26/M
—
Diabetic ketoacidosis
−
PPI
A
12
54/F
—
Diabetic ketoacidosis
ND
Hemoclipping, PPI
C
13
32/M
—
Diabetic ketoacidosis
−
PPI
C
14
44/M
—
Cholelithiasis
−
PPI
C
15
36/F
Tiaramide hydrochloride, 300mg/day
Ovarian carcinoma
−
PPI
B
16
85/M
—
Stroke
−
PPI
B
NSAIDs, nonsteroidal anti-inflammatory drugs; ND, no data; PPI, proton pump inhibitor; HPU, coagulation therapy with heater probe unit
aClassified endoscopically according to Makuuchi13 (see Patients and Methods)

Page 3

H. Yasuda et al.: ANE and NSAIDs195
none of the patients had previously taken medication
for reflux symptoms. One patient had multiple gastric
erosions and ulcers, as well as duodenitis. While ANE
presented with UGI bleeding in all cases, the effect on
hemoglobin levels was not significant except in two pa-
tients. In these patients, endoscopic examination re-
vealed arterial spurting in one patient and a visible
vessel in the other. They were treated by endoscopic
hemoclipping or coagulation with a heater probe, and
were given red-cell transfusions. To treat ANE, empiri-
cal supportive therapy was provided, which included
parenteral nutrition for 4–5 days and administration of a
proton pump inhibitor. ANE resolved without any seri-
ous complications in all cases. There were no episodes
of rebleeding, and there were no bleeding-associated
deaths.
All patients had coexisting disease (Table 2). Eight of
16 patients (50%) took NSAIDs because of underlying
disease such as ischemic heart disease, rheumatoid ar-
thritis, or cancer. When we compared the medication
history of ANE patients with that of patients with
bleeding gastric/duodenal ulcers (Table 3), we found
that the incidence of NSAID use was similar between
patients with esophagitis (50%) and those with UGI
ulcer bleeding (34%). None of the patients had taken
ticlopidine or warfarin for ANE. H. pylori was tested
for in 12 patients, among which three were positive
(25%).
ANE also occurred in four patients (25%) with dia-
betic ketoacidosis (DKA). We had 29 admissions for
DKA during this period. Endoscopy was performed
in seven patients because of UGI symptoms, such
as coffee-ground emesis, frequent vomiting, or
epigastralgia; four of these patients had ANE, and in
three patients, the esophagus was normal.
Discussion
We observed 16 cases of ANE (6%) among 239 patients
with UGI hemorrhage during a 3-year period. In agree-
ment with the findings of Augusto et al.,6 our results
showed that ANE was more common than has been
generally recognized. In clinical practice, more atten-
tion should be paid to ANE as a cause of UGI bleeding.
ANE was found in 50% of the 16 patients who took
NSAIDs. NSAID-induced gastric and duodenal ulcers
have received the most attention, but esophageal
complications due to NSAIDs have been also recog-
nized.10–12,14 The best-known target of NSAIDs is
cyclooxygenase.7 In the human esophagus, however,
the main pathway for prostaglandin (PG) synthesis
is reported to involve lipoxygenase rather than
cyclooxygenase.15 With respect to lower esophagus
sphincter (LES) pressure, PGs and NSAIDs have little
effect on LES pressure and motility of the esophageal
body. NSAIDs do not provoke gastroesophageal reflux
Table 3. Drug and medication usage in patients with acute necrotizing esophagitis
(ANE) and bleeding gastric/duodenal ulcers
Patients with
ANE (n = 16)
n (%)
Patients with gastric/
duodenal ulcers (n = 169)
n (%)
Low-dose aspirin
NSAIDs
Ticlopidine/aspirin
Prednisolone
Ticlopidine
Warfarin
Total
4 (25)
4 (25)
0 (0)
1 (6)
0 (0)
0 (0)
9 (56)
11 (7)
30 (19)
12 (7)
5 (3)
9 (5)
2 (1)
69 (41)
Fig. 1. Black esophagitis in a patient with diabetic
ketoacidosis

Page 4

196H. Yasuda et al.: ANE and NSAIDs
Fig. 2. A Diffuse esophageal ne-
crosis in a chronic nonsteroidal
anti-inflammatory drug user. B
There was an abrupt interrup-
tion at the Z-line with normal-
appearing cardiac mucosa. C
Healed esophageal mucosa 4 days
after proton pump inhibitor treat-
ment. D Biopsy specimen of acute
phase showed esophageal necro-
sis with abundant neutrophils and
mononuclear cells
Fig. 3. Exposed vessel at the esophagocardiac junction.
Successful coagulation therapy with a heater probe was
performed
A
B
C
D
in the pathogenesis of esophageal ulceration. Rather,
direct cellular toxicity has been suggested as an impor-
tant mechanism in the pathogenesis of esophagitis.
Advanced age, recumbence, and motility disturbance
associated with GERD may cause prolonged contact
of NSAIDs with the esophageal mucosa, which can
worsen a preexisting nonsymptomatic reflux. The
acidic pH of the lower esophagus favors the formation
of lipid-soluble forms of NSAIDs, thereby enhancing
their entry into the mucosal cells and causing cellular
toxicity.18
ANE has been reported in association with several
diseases. Interestingly, we observed four cases of ANE
with DKA. Faigel and Metz19 observed 26 cases of UGI
hemorrhage in 330 admissions for DKA. Augusto et al.6
observed 12 cases of hyperglycemia and 1 case of DKA
among 29 cases of ANE. Gastric stasis, which can occur
in the absence of autonomic neuropathy during DKA,20
may induce massive gastroesophageal reflux and cause
ANE. To our knowledge, no clinical trials for the treat-
ment of ANE have been published previously. To treat
ANE, discontinuation of the offending medication and
administration of proton pump inhibitor was effective
in most cases, which suggests that acute exacerbation of
gastroesophageal reflux is the predominant mechanism
of esophageal injury in ANE.
Acute gastric mucosal lesion, which is an impor-
tant cause of UGI bleeding,21 is characterized by
disease (GERD) in healthy subjects.16 The effect of
synthetic PGE1 was no different from that of a placebo
when administered for the treatment of reflux esophagi-
tis.17 PGs, therefore, do not seem to play a major role

Page 5

H. Yasuda et al.: ANE and NSAIDs 197
severe epigastralgia, with sudden onset, vomiting,
hematemesis, and melena following probable causes
such as NSAIDs, and it is diagnosed on the basis of
endoscopic findings of multiple gastric erosions, hemor-
rhagic gastritis, or gastric ulcer. ANE is characterized by
acute presentation of hematemesis or melena following
NSAID intake or from underlying disease such as
DKA. ANE could also be referred to as an acute esoph-
ageal mucosal lesion, which seems to better describe
both the hemorrhagic and necrotic endoscopic findings
of ANE.
References
1. Sugawa C, Benishek D. Mallory-Weiss syndrome. Am J Surg
1983;145:30–3.
2. Jensen DM. Endoscopic screening for varices in cirrhosis:
findings, implications, and outcomes. Gastroenterology 2002;122:
1620–30.
3. Goldenberg SP, Wain SL, Marignani P. Acute necrotizing esoph-
agitis. Gastroenterology 1990;98:493–6.
4. Moreto M, Ojembarrena E, Zaballa M, Tanago JG, Ibanez S.
Idiopathic acute esophageal necrosis: not necessarily a terminal
event. Endoscopy 1993;25:534–8.
5. Cattan P, Cuillerier E, Cellier C, Carnot F, Landi B, Dusoleil A,
et al. Black esophagitis associated with herpes esophagitis.
Gastrointest Endosc 1999;49:105–7.
6. Augusto F, Fernandes V, Cremers MI, Oliveira AP, Lobato C,
Alves AL, et al. Acute necrotizing esophagitis: a large retrospec-
tive case series. Endoscopy 2004;36:411–5.
7. Wolfe MM, Lichtenstein DR, Singh GS. Gastrointestinal toxicity
of nonsteroidal anti-inflammatory drugs. N Eng J Med 1999;
340:1888–99.
8. Heller SR, Fellows MG, Ogilvie AL, et al. Non-steroidal anti-
inflammatory drugs and benign esophageal stricture. Br Med J
1982;285:167–8.
9. Wilkins WE, Ridley MG, Pozniak AL. Benign stricture of the
oesophagus: role of non-steroidal anti-inflammatory drugs. Gut
1984;25:478–80.
10. El-Serag HB, Sonnenberg A. Association of esophagitis and
esophageal strictures with diseases treated with nonsteroidal anti-
inflammatory drugs. Am J Gastroenterol 1997;92:52–6.
11. Semble EL, Wu WC, Castell DO. Nonsteroidal anti-inflammatory
drugs and esophageal injury. Semin Arthritis Rheum 1989;19:
158–62.
12. Avidan B, Sonnenberg A, Schnell TG, Sontag SJ. Risk factors
for erosive reflux esophagitis: a case–control study. Am J
Gastroenterol 2001;96:41–6.
13. Makuuchi H. Clinical study of esophageal hiatal hernia. Jpn J
Gastroenterol 1982;79:1557–67.
14. Mason JC. NSAIDs and the oesophagus. Eur J Gastroenterol
Hepatol 1999;11:369–73.
15. Alber D, Moussard C, Toubin M, Henry JC, Ottingnon Y,
Deschamps JP. Gas chromatographic/mass spectrometric quanti-
tative analysis of eicosanoids in human esophageal mucosa.
Biomed Environ Mass Spectrom 1988;16:299–304.
16. Scheiman JM, Patel PM, Henson EK, Nostrant TT. Effect of
naproxen on gastroesophageal reflux and esophageal function:
a randomized, double-blind, placebo-controlled study. Am J
Gastroenterol 1995;90:754–7.
17. Purushotham AD, Gray GR, Hansell DT. Prostaglandins in the
treatment of reflux esophagitis: double-blind placebo controlled
clinical trial. J R Coll Surg Edinb 1991;36:216–8.
18. Shirazi SS, Zike WL, Brubacher M, DenBesten L. Effect of aspi-
rin, alcohol and pepsin on mucosal permeability of esophageal
mucosa. Surg Forum 1974;25:335–7.
19. Faigel DO, Metz DC. Prevalence, etiology, and prognostic
significance of upper gastrointestinal hemorrhage in diabetic ke-
toacidosis. Dig Dis Sci 1996;41:1–8.
20. Diabetic ketoacidosis. In: Pickup UC, Williams G, editors. Text-
book of diabetes, 2nd ed. Oxford: Blackwell Science; 1997.
21. Katz D, Siegel HI. Erosive gastritis and acute gastric mucosal
lesion. In: Glass GB, editor. Progress in gastroenterology, vol. 1.
New York: Grune and Stratton; 1968. p. 67–96.