Content uploaded by Yi-Qi Du
Author content
All content in this area was uploaded by Yi-Qi Du
Content may be subject to copyright.
ORIGINAL ARTICLE
Anti-inflammatory Effects of Rebamipide According
to Helicobacter pylori Status in Patients with Chronic Erosive
Gastritis: A Randomized Sucralfate-Controlled Multicenter Trial
in China—STARS Study
Yiqi Du ÆZhaoshen Li ÆXianbao Zhan ÆJie Chen ÆJun Gao Æ
Yanfang Gong ÆJianlin Ren ÆLiping He ÆZhijian Zhang ÆXiaozhong Guo Æ
Jianshen Wu ÆZibin Tian ÆRuihua Shi ÆBo Jiang ÆDianchun Fang Æ
Youming Li
Received: 17 September 2007 / Accepted: 20 December 2007 / Published online: 21 February 2008
ÓSpringer Science+Business Media, LLC 2008
Abstract The aim of the study was to investigate the
effects of rebamipide on symptom, histology, endogenous
prostaglandin, and mucosal oxygen free radicals in chronic
erosive gastritis (CEG) patients by using sucralfate as a
control. The trial also examined whether Helicobacter
pylori infection would affect rebamipide-induced protec-
tion. A total of 453 endoscopy-confirmed CEG patients from
11 hospitals in China were enrolled in the study. They
randomly received either rebamipide (100 mg t.i.d) or
sucralfate (1.0 t.i.d) for 8 weeks with a ratio of 3:1. Per-
protocol analysis (n=415) showed the accumulated
symptom score in the rebamipide group dropped from
5.54 ±0.97 to 0.80 ±0.47 after 8 weeks (P\0.001
versus control). The endoscopic inflammation score in
rebamipide group also decreased from 2.65 ±0.09 to
0.60 ±0.10, which showed better effects than sucralfate. It
was shown a significant improvement (P\0.01) in pros-
taglandin E2 (PGE
2
) contents in rebamipide-treated subjects
mucosa (225.4 ±18.3 pg/g versus 266.7 ±14.7 pg/g)
compared with that in sucralfate group after 8 weeks of
treatment. Malondialdehyde (MDA) contents were signifi-
cantly depressed both in the trial and control group. When
Helicobacter pylori infection was considered, no statisti-
cally difference was found in the effect of rebamipide on
either symptom or inflammation scores. In conclusion,
Rebamipide demonstrated a stronger suppressive effect on
the mucosal inflammation in chronic erosive gastritis than
sucralfate. The gastroprotection induced by rebamipide is
not influenced by H. pylori infection, which indicates its
usage in the treatment of H. pylori-associated CEG.
Z. Tian
Department of Gastroenterology, The Medical School Hospital
of Qingdao University, Qingdao 266003, China
R. Shi
Department of Gastroenterology, The First Affiliated Hospital
of Nanjing Medical University, Nanjing 210029, China
B. Jiang
Department of Gastroenterology, Nanfang Hospital,
South Medical University, Guangzhou 510515, China
D. Fang
Department of Gastroenterology, Southwest Hospital,
Third Military Medical University, Chongqing 400038, China
Y. Li
Department of Gastroenterology, Zhejiang First Hospital,
Medicine Zhejiang University, Hangzhou 310003, China
Y. Du Z. Li (&)X. Zhan J. Chen J. Gao Y. Gong
Department of Gastroenterology, Changhai Hospital,
Second Military Medical University, Shanghai 200433, China
e-mail: zhaoshenli@hotmail.com
J. Ren
Department of Gastroenterology, Zhongshan Hospital,
Xiamen University, Xiamen 361004, China
L. He
Department of Gastroenterology, Fujian Province Hospital,
Fuzhou 350001, China
Z. Zhang
Department of Gastroenterology, Fuzhou General Hospital,
Fuzhou 350025, China
X. Guo
Department of Gastroenterology, The General Hospital of
Shenyang Military Region, Shenyang 110016, China
J. Wu
Department of Gastroenterology, The First Affiliated Hospital of
Wenzhou Medical College, Wenzhou 325000, China
123
Dig Dis Sci (2008) 53:2886–2895
DOI 10.1007/s10620-007-0180-z
Keywords Chronic erosive gastritis Rebamipide
Helicobacter pylori Gastroprotection
Abbreviations
CEG Chronic erosive gastritis
PGE
2
Prostaglandin E2
MDA Malondialdehyde
NSAIDs Nonsteroidal anti-inflammatory drugs
REB Rebamipide
SUC Sucralfate
STARS Symptom Treatment and Anti-inflammatory
Effect of Rebamipide and Sucralfate for
gastritis
Chronic erosive gastritis is a very common disease puz-
zling clinicians, especially in Asian areas such as China,
Japan, and Korea. Even without risk of carcinogenesis as
atrophic gastritis, chronic erosive gastritis (CEG) always
presents with various GI symptoms and histological change
in gastric mucosa leading to decreased quality of life. The
drugs to be selected to treat the disease are less effective
than those for peptic ulcer. Traditional gastroprotective
agents such as sucralfate can inhibit inflammation while the
symptom release and duration was not satisfied. Previous
data from our center showed that endoscopy-confirmed
superficial, atrophic, and erosive gastritis cases per year
changed from 529, 881, and 2,891 in 2000 to 1,605, 686,
and 5,357 in 2006, respectively. Against this trend in
China, we tried an application of rebamipide, a gastro-
protective drug in the treatment of CEG.
Rebamipide is one of the most commonly used gastro-
protective agents in East Asia [1]. It has been widely
proven in animal model that the drug exhibits preventive
effects in gastric mucosa by increasing endogenous pros-
taglandin or by suppressing oxygen free radicals, as well as
increasing blood flow [2–5]. Though much evidence has
demonstrated that rebamipide improves histological gas-
tritis in vivo, more clinical evidence is needed to confirm
its effects on chronic gastritis. Unlike in Western countries,
gastric atrophy is more prominent in Asia, which indicates
that mucosal protection is more important than mono anti-
acid therapy. Furthermore, Helicobacter pylori and non-
steroidal anti-inflammatory drug (NSAID) are two major
causes related to gastric injury. Preclinical data has dem-
onstrated various mechanisms involved in rebamipide
effects on H. pylori-associated gastritis, including distur-
bance of the adhesion of H. pylori to gastric epithelial cells
and inhibitory effects on H. pylori-induced neutrophil
activation or interleukin-8 secretion [6–8]. A few reports
have even suggested that rebamipide may have a potential
anti-H. pylori role [9] while this was not confirmed in large
clinical trials [10]. However, we still lack clinical data on
rebamipide for the treatment of H. pylori-related gastritis.
Although up to 50% of nonulcer dyspepsia (NUD)
patients have H. pylori infection and underlying chronic
gastritis, it remains controversial whether the bacteria
influences the pattern of gastric symptoms [11]. However,
in one recent published meta-analysis from China, the
summary odds ratio for improvement in dyspeptic symp-
toms in patients with functional dyspepsia in whom H.
pylori was eradicated was 3.61 (95% CI: 2.62–4.98,
P\0.00001) and the difference in the follow-up period
did not influence the final outcomes [12]. Thus H. pylori
status needs to be determined first in the current study
before the effect of rebamipide could be evaluated.
Moreover, chronic gastritis is characterized by the accu-
mulation of oxidative DNA damage [13] and antral
prostaglandin E2 basal levels appear to be important for the
development of aspirin-induced gastric damage in subjects
without H. pylori infection [14]. In this study, two indi-
cators of oxygen free radicals and gastroprotective agents
were investigated after rebamipide administration. Sucral-
fate, a traditional gastroprotective agent, has shown
affirmative effects on chronic gastritis by means of anti-
acid. Without the molecular mechanism of PG and oxida-
tion adjustment, it becomes an ideal candidate for the
control of rebamipide.
Therefore, the major aim of the trial was to evaluate the
effect of rebamipide on CEG in Chinese patients. Due to
the less common use of NSAID in China than in Western
countries, NSAID-induced gastric injury was excluded
from the study. There is a great deal of evidence showing
rebamipide’s protection on NSAID-induced gastric injury,
so the current study was designed to illustrate the role of
the drug on NSAID-unrelated gastric pathology.
Materials and Methods
Patients
The study was carried out as an open, randomized, positive
drug parallel-controlled and subgrouped clinical trial per-
formed in 11 centers between October 2004 and December
2005. Each center was expected to complete 60 cases of
CEG therapy with either rebamipide or sucralfate (with a
ratio of 3:1). The sample size and ratio were determined by
a power study by statistics specialists.
Criteria for inclusion were age 18–65 years, diagnosed
chronic erosive gastritis by endoscopy within 1 week, and
having at least two symptoms of abdominal pain, disten-
sion, acid reflux, and belching. Exclusion criteria were:
Dig Dis Sci (2008) 53:2886–2895 2887
123
(i) patients with malignancy diseases, (ii) peptic ulcer, (iii)
patients who has been administrated with drugs that may
affect evaluation during two weeks before enrolled
(NSAID, proton pump inhibitor, H
2
-antagonist, anti-acid
regents, and antibiotics, etc.), (iv) severe heart or pul-
monary disease, (v) pregnancy, (vi) allergic habitus, and
(vii) other situations that the investigators considered
unsuitable for the study. Previous H. pylori infection his-
tory and any NSAID usage during 1 month were also
recorded. All patients signed written informed consent
prior to the study and the whole protocol was approved by
the ethics committee in each participating institute.
Erosive gastritis was determined by endoscopy accord-
ing to the Sydney system and modified endoscopic chronic
gastritis classification consensus (2003, China). Three
biopsy samples were taken from antrum during endoscopy,
one from the erosion area and another from a normal area
with an extra sample for rapid urease testing. H. pylori
positivity was defined as positive for both the rapid urease
test and
13
Cor
14
C urea-breath test. In two centers, an extra
two samples were taken in each subject for quantification
of mucosal PGE
2
and the oxygen free radical product
malondialdehyde (MDA).
Study Design
Finally 453 patients were enrolled, 15 of which did not
have a definite H. pylori status. To evaluate the effect of
rebamipide on H. pylori-associated gastritis, 438 patients
were classified to three kinds of conditions before admin-
istration of gastroprotective agents: (i) H. pylori positive
but not eradicated; (ii) H. pylori positive but eradicated,
patients received a base therapy consisting of 1 week
omeprazole 20 mg b.i.d. plus amoxicillin 1.0 b.i.d. and
clarithromycin 500 mg b.i.d. prior to randomization; (iii)
H. pylori negative. This subgrouping will lead to a clear
interpretation of the effect of the gastroprotective agents
beyond the influence of H. pylori status. The homogeneity
between each subgroup was first analyzed as the 1 week
eradication therapy in H. pylori-positive subjects may
affect the dyspeptic symptom during the baseline periods.
If there was any significant difference between the base-
lines of the subgroups, multivariant analysis was
conducted. All 453 patients were randomized into two
groups, the REB group receiving rebamipide 300 mg/d
(100 mg po t.i.d.) (Mucosta
Ò
, Otsuka, Japan) and the SUC
group receiving sucralfate 3.0/d (1.0 po t.i.d.) (Shu Ke
Fei
Ò
, Hefeng Pham. Ltd, Shanghai) for 8 weeks (Fig. 1).
GI symptom changes in the first week were recorded
daily on diary paper by patients and the following 2, 4, 6,
and 8 weeks of information was written by investigators.
Cure of H. pylori infection was determined by repeated
13
C
or
14
C UBT at the end of study. Moreover, aiming to obtain
evidence for the rebamipide-induced molecular mecha-
nism, mucosal PGE
2
and MDA concentration were
detected before and after 8 weeks therapy in 2 of the 11
centers (Changhai & Zhongshan Hospitals; the other cen-
ters did not have the facilities to test this).
Randomization
Randomization was designed to determine the subjects to
receive rebamipide or sucralfate therapy. Randomization
was performed after the H. pylori status of subjects had
been determined. The subjects were then recruited
according to a randomization schedule produced by sta-
tistics software. A randomization number associated with
either rebamipide or sucralfate was assigned to each patient
in the study. An allocation ratio of 3:1 for the two treatment
groups was set according to the power study, and each
center used its own randomization number. Randomization
numbers were generated by using the SAS program.
Clinical Effect Evaluation
The symptoms during the first week and at the end of 2, 4,
6, and 8 weeks were monitored by scoring symptoms
including pain, distension, reflux, and belching. Each
H.pylori positive
and eradicated
H.pylori positive
and non-eradicated H.pylori negative
History inquiry, endoscopy, pathology, mucosal PGE2 and MDA
(one week before enrolled)
One-week triple therapy
Randomization (3:1)
Symptoms recording
(at 1, 2, 4, 6, 8 weeks)
Endoscopy, pathology, mucosal PGE2 and MDA
(at the end of study)
Rapid urease test, 13C or 14C-UBT
Rapid urease test, 13C or 14C-UBT
(only in H.pylori positive and eradicated subjects)
Trial group:
Rebamipide 0.1 t.i.d.×8w
Control group:
Sucralfate 1.0 t.i.d.×8w
Fig. 1 Study regimen
2888 Dig Dis Sci (2008) 53:2886–2895
123
symptom was graded as 0 (none), 1 (mild), 2 (obvious,
partially disturbing daily life), and 3 (severe, disturbing
daily life and needing drugs). The total scores of symptoms
were then calculated to evaluate the effect on symptom
change by treatment. Each patient was taught how to
evaluate and make a record on a diary card.
Endoscopic images before and after therapy were
considered another indicator of effective treatment.
Alterations in gastric mucosa images could be quantified
according to the modified Lanza standard as following: 0
(no erosion), 1 (one or two erosive lesions limited in one
area as antrum, corpus or fundus), 2 (three to five lesions
but in the same area), 3 (lesions involving two areas but
fewer than ten), and 4 (extensive lesions or more than
ten). Endoscopists in each center received training before
the trial started and two copies of pictures from every
patient were sent to the leading center for repeat
reviewing.
The specimens taken from antrum were used to study
the histological effects of rebamipide on inflammation in
the gastric mucosa. The strategy of taking both erosive
and nonerosive lesions at the same time avoided mis-
counting of the inflammatory score. Histological findings
on the activity and chronic inflammation were graded as
0 (normal), 1 (mild), 2 (moderate), or 3 (marked)
according to the updated Sydney system [15]. Specimens
were treated with hematoxylin and eosin, and Giemsa
stains. One pathologist, having no information on the
subjects, performed uniform histological grading. Safety
monitoring was also conducted according to the recorded
adverse events.
Mucosal PGE
2
and MDA Measuring
Two mucosal specimens taken from erosive and nonerosive
lesion (within 2 cm of the erosive area) in each patient
were collected before and after treatment. Mucosal PGE
2
and MDA concentration were measured by the RIA and
thiobarbituric acid (TBA) methods as previously published,
respectively.
End Points
The primary endpoint was symptom change and histolog-
ical remission at 8 weeks. The secondary endpoint was the
change of PGE
2
and MDA concentration in the gastric
mucosa after the therapy. To exclude the effect of H. pylori
infection, subgroup analysis was carried out after the
analysis of the whole group.
Statistical Analysis
All the case report forms (CRFs) were sent to the Depart-
ment of Statistics, Second Military Medical University by
the end of study. Data were then computerized and analyzed
with SAS 8.2 software (SAS Institute Inc., North Carolina,
US). The full analysis set (FAS) consists of the randomized
453 patients. The per-protocol (PP) analysis was performed
by using data only from subjects characterized by the cri-
teria: (i) completion of the whole treatment; (ii) availability
of results useful for the primary aim; and (iii) no major
protocol violations. The demographic characteristics of the
two groups before treatment were compared using the
Student’s t-test, chi-squared test, and Mann–Whitney Utest
according the character of index. Within each group,
symptom scoring, endoscopic, and histological grading were
compared before and after therapy by using the Wilcoxon
rank sum test for their categorical character. For the
PGE
2
and MDA results, the paired t-test was used to compare
the alteration. When comparing the effects between two
groups or three subgroups, we used covariance analysis or
Cochran-Mantel-Haenszel (CMH) method. All statistical
tests were two-sided, with a 5% level of significance.
Results
Demographic Characteristics of Subjects
A total of 453 CEG patients were enrolled in the study and
randomized into two groups at the beginning as follows:
rebamipide 342 cases and sucralfate 110 cases. One
patient’s grouping information was not recorded on the
CRF and could not be put in FAS. H. pylori status was not
determined for another 14 patients in the rebamipide group,
which hence could not be included in the PP analysis
(Fig. 2). The number of cases in the subgroups according
to H. pylori status was 150 (eradicated), 129 (not eradi-
cated), and 159 (negative). Then 438 patients were
randomized as 331 cases in rebamipide group and 107 in
sucralfate group. Twenty-three patients (REB: 13; SUC:
10) were excluded from the PP analysis for the following
reasons: incomplete histology results (n=16); failure of
eradication assessment (n=2); lost to follow-up (n=2);
adverse events (n=3). Finally the PP set consisted of 415
patients (REB: 318; SUC: 97). The adverse events (AE)
occurring during study includes one case of eczema, one
case of abnormal hepatic function, and another hospital-
ization for acute appendicitis.
The comparison of demographic characteristics
showed that there were no significant differences between
the REB and SUC groups in terms of age, sex, stature,
Dig Dis Sci (2008) 53:2886–2895 2889
123
body weight, disease course, smoking or alcohol habit
and current H. pylori status, as shown in Table 1.
Effects of Rebamipide on GI Symptoms
The records on daily and weekly symptoms revealed that
treatment with rebamipide or sucralfate could dramatically
alleviate the four common CEG symptoms of pain, reflux,
distension, and belching, as shown in Fig. 3. Improvement
in symptoms showed a tendency to be acceptable (score
lower than one) after 1 week therapy and be continuous for
8 weeks time. Subsequently, when compared with the
baseline, the accumulated scores were significantly reduced
both in the REB (2.49 ±0.54 versus 5.54 ±0.97,
P\0.001) and SUC (3.11 ±0.47 versus 5.95 ±0.83,
P\0.001) groups at the end of week 1. However, the
score difference between the baseline and week 8 in the
REB group was significantly larger than that in the SUC
group (P\0.001, Fig. 4), while there was no significant
difference between the two baselines (P=0.189). Though
the median together with 25% and 75% intervals instead of
the mean should be used to represent the symptom score
for its categorical native, the mean value was preferred to
show the trend clearly.
Effects of Rebamipide on Gastric Mucosal
Inflammation
The visible improvement of mucosal inflammation under
endoscopy is one of the most important indexes in the
evaluation of the therapeutic effect on CEG. Both REB and
SUC could significantly decrease the endoscopic score with
8 weeks therapy as shown in Fig. 5a. Before therapy the
median endoscopic score in both the REB and SUC groups
was 3. At 8 weeks, the median scores became 0 in the REB
and 1 in the SUC group (0.60 ±0.10 versus 2.65 ±0.09,
P\0.001 in the REB group and 1.05 ±0.19 versus
2.53 ±0.14, P\0.001 in the SUC group, mean ±SEM,
Wilcoxon signed rank test). The Wilcoxon rank sum
analysis showed that the difference between the two groups
was statistically significant (P\0.001) after 8 weeks
therapy while no significant difference could be observed
at baseline (P=0.219). Furthermore, Fig. 5b, c shows the
mean inflammation and activity scores by the updated
Sydney system in the histological examination of the gas-
tric mucosa at baseline and at the end of week 8. Both the
inflammation and activity scores of the two groups at
baseline were comparable (P=0.078 and P=0.851). In
both groups, the scores of chronic inflammation signifi-
cantly decreased after treatment (median dropped from 3 to
1.5 in the REB group and from 2 to 2 in the SUC group,
P\0.001 by Wilcoxon signed rank test, respectively) and
no significant differences could be observed between the
Missing H.pylori information: n=14
H.pylori positive
and eradicated
n= 150
H.pylori positive
and non-eradicated
n= 129
H.pylori negative
n= 159
Patients enrolled and radomized
n= 453
REB: n= 117
SUC: n= 33
One case missed grouping information
REB: n= 92
SUC: n= 37
REB: n= 122
SUC: n= 37
PP analysed: n= 140
REB: n= 110
SUC: n= 30
Excluded from analysis:
n= 10
(No complete histology 8,
no final H.pylori status 2)
PP analysed: n= 124
REB: n= 91
SUC: n= 33
Excluded from analysis:
n= 5
(No complete histology 3,
AE 2)
PP analysed: n= 151
REB: n= 117
SUC: n= 34
Excluded from analysis:
n= 8
(No complete histology 5,
lost to follow-up 2, AE 1)
PP analysed: n= 415
(Rebamipide 318, Sucralfate 97)
FAS analysed: n= 452
(Rebamipide 345, Sucralfate 107)
Fig. 2 Flow diagram showing enrolled, information missed, and
subjects for PP analysis (FAS: full analysis set, PP: per protocol,
REB: rebamipide, SUC: sucralfate, AE: adverse effect)
Table 1 Comparison of patient characteristics between the reb-
amipide (REB) and sucralfate (SUC) groups (FAS, n=452)
REB
(n=345)
SUC
(n=107)
P-value
Age (mean ±SD,
years)
45.4 ±12.5 44.0 ±14.2 0.338
a
Sex (male, %) 201 (58.3%) 65 (60.7%) 0.736
b
Stature (mean ±SD, cm) 166.8 ±7.2 167.2 ±8.0 0.693
a
Body weight
(mean ±SD, kg)
62.7 ±11.0 62.0 ±11.0 0.533
a
Disease course
(mean ±SD, months)
23.8 ±52.0 23.0 ±55.3 0.892
c
Smoking habit (+ve rate) 74 (21.4%) 24 (22.4%) 0.893
b
Alcohol habit (+ve rate) 59 (17.1%) 23 (21.5%) 0.316
b
Current H. pylori
infection
(+ve rate)
209 (63.1%)
d
70 (65.4%) 0.729
b
a
Student t-test,
b
Chi-squared test,
c
Mann-Whitney Utest,
d
four-
teen patients in REB did not have H. pylori status information
(n=331)
2890 Dig Dis Sci (2008) 53:2886–2895
123
Abdominal pain
0
0 7 14 28 42 56 0 7 14 28 42 56
0 7 14 28 42 560 7 14 28 42 56
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
REB ( n=318 )
SUC ( n=97 )
REB ( n=318 )
SUC ( n=97 )
REB ( n=318 )
SUC ( n=97 )
REB ( n=318 )
SUC ( n=97 )
Distension
Acid reflux
2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2Belching
Fig. 3 The trend of CEG-
associated GI symptoms in the
rebamipide (REB) and
sucralfate (SUC) groups. Time
courses of the mean value of
symptom scores are shown in
the graphs. The X-axis
represents days from baseline
and the Y-axis represents
symptom score. (The lines
represent the trend rather than a
continuous variable)
0
1
2
3
4
5
6
7
8
Baseline Week1 Week2 Week4 Week6 Week8
REB ( n=318 )
SUC ( n=97 )
*P< 0.001 < 0.001 < 0.001 < 0.001 < 0.001
**P= 0.183 = 0.004 < 0.001 < 0.001 < 0.001 < 0.001
Fig. 4 The reduced total scores of CEG-associated GI symptoms
with rebamipide (REB) and sucralfate (SUC) treatment by PP
analysis. The X-axis indicates the time courses after therapy while the
Y-axis represents accumulated symptom score. The error bars
represent the standard deviation (SD) and the height of the bars
represents the mean score. * P-value produced by comparing with the
baseline within each group (Wilcoxon signed rank test), ** P-value
between the two groups by Wilcoxon rank sum analysis
0
1
2
3
4
Pre-
therapy
Post-
therapy
Pre-
therapy
Post-
therapy
Pre-
therap
y
Post
-
therap
y
REB ( n=318 ) SUC ( n=97 )
A Endoscopic B Inflammation C Activity
*P< 0.001 < 0.001 < 0.001
**P< 0.001 < 0.001 0.166
***P= 0.219 < 0.001 0.078 0.545 0.851 0.877
Fig. 5 Inhibition of gastric mucosal inflammation with rebamipide
(REB) and sucralfate (SUC) treatment by PP analysis. The three
regions in the graph represent endoscopic inflammation score (a)),
histological chronic inflammation score (b), and histological activity
score (c). The X-axis indicates the time courses before and after
8 weeks therapy while the Y-axis represents the inflammation score.
The error bars represent the standard error of the mean (SEM) and the
height of bars represents the mean score. * P-value produced by
comparing with the baseline in the REB group (Wilcoxon signed rank
test), ** P-value produced by comparing with the baseline in the SUC
group (Wilcoxon signed rank test), *** P-value between the two
groups by Wilcoxon rank sum analysis
Dig Dis Sci (2008) 53:2886–2895 2891
123
REB and SUC groups (P=0.545). However, the REB
group showed a significant inhibition on inflammatory
activity (median of 1 versus 0, P\0.001 by Wilcoxon
signed rank test) while the SUC group did not, even though
the difference between the two groups remained nonsig-
nificant (P=0.877 by Wilcoxon rank sum analysis).
Rebamipide Increased PGE
2
and Depressed MDA
Content in Gastric Mucosa
The PGE
2
content in gastric mucosa has been wildly accepted
as an indicator of mucosal protecting agent level, which
indeed reflects the effect of REB or SUC. Considering that the
distribution of inflammation in antrum was not homogeneous,
we compared the effect of REB and SUC on the PGE
2
level in
the erosive and nonerosive regions, respectively. In contrast to
expectations, the nonerosive area had a slightly but nonsig-
nificantly higher PGE
2
level than the erosive area before
therapy. After 8 weeks of treatment, REB induced a signifi-
cantly increased PGE
2
levelinbothareas,from
236.2 ±17.9 pg/g to 261.2 ±21.6 pg/g in the nonerosive
area and from 225.4 ±18.3 pg/g to 266.7 ±14.7 pg/g in the
erosive area (P\0.001 by paired t-test, Table 2). SUC also
increased PGE
2
level in both areas. However the difference
between REBand SUC on PGE
2
inductionwas only evident in
the erosive area (P=0.002 by covariance analysis).
Furthermore, MDA, a metabolite of oxygen free radicals,
showed a lower concentration in the nonerosive area than in
the erosive area at baseline (203.8 mmol/g versus
316.5 mmol/g, P\0.01). Depressed MDA level could be
observed in both the erosive and nonerosive areas in the REB
group (216.5 ±61.5 mmol/g and 177.6 ±32.5 mmol/g,
Table 2). The MDA content in the SUC group was decreased
in both areas but only significantly in the erosive area. When
compared with SUC, REB showed a statistically better effect
on MDA inhibition in the erosive area (P=0.046 by
covariance analysis).
Further Analysis in Subgroups According to H. pylori
Status
In the PP set (Fig. 2), at the start of the trial 201 of the 318
subjects in the REB group and 63 of the 97 subjects in the
SUC group were H. pylori-infected patients, with an
infection rate of 63.2% and 64.9%, respectively
(P=0.810 by chi-squared test). For study purposes,
54.7% (n=110) of the H. pylori-positive patients in the
RUB group and 47.6% (n=30) of the H. pylori-positive
patients in the SUC group received triple eradication
therapy. There was no significant difference between the
REB and SUC groups in terms of the ratio of antibiotics-
treated patients before the formal treatment started
(P=0.386). While the H. pylori-positive rate became
24.5% (n=78) in the REB group and 36.1% (n=35) in
the SUC group when the H. pylori status was redetected
after 8 weeks of therapy. The positive rate in the REB
group was significantly lower than that in the SUC group,
with a Pvalue of 0.027 by the chi-squared test.
A subgroup analysis by H. pylori status associated with
REB or SUC effect on symptom change is shown in Fig. 6.
The tendency cause by REB was not affected by whether H.
pylori eradication was performed or not (P[0.05). Inter-
estingly, the significant difference between the REB and
SUC groups could be observed 1 week earlier in H. pylori-
negative patients than in H. pylori-positive patients, in
whom the differences could only be observed after 1 month.
Moreover, endoscopic scores revealed that, even in the H.
pylori noneradicated subgroup, REB yielded improvement
in inflammation at week 8 with a significantly lower score
(0.57 ±0.85, P\0.001) compared with baseline
(2.71 ±0.84). REB also performed effectively in terms of
reducing endoscopic scores both in the H. pylori natively or
acquired negative subgroups (0.57 ±0.77 and 0.64 ±0.94,
P\0.001 compared with baseline, respectively). No sig-
nificant difference could be found among the three
subgroups by Wilcoxon rank sum analysis. The results from
Table 2 Effect of rebamipide (REB) and sucralfate (SUC) treatment on the PGE
2
and MDA concentration in gastric antral mucosa (n=75)
Mucosa Group nBaseline Week 8 P
1
P
2
PGE
2
Nonerosive REB 52 236.2 ±17.9 261.2 ±21.6 \0.001 0.643
SUC 23 242.9 ±10.5 263.8 ±9.8 \0.001
Erosive REB 52 225.4 ±18.3 266.7 ±14.7 \0.001 0.002
SUC 23 235.1 ±17.0 258.4 ±12.1 \0.001
MDA Nonerosive REB 52 212.4 ±46.8 177.6 ±32.5 \0.001 0.130
SUC 23 195.1 ±27.7 184.1 ±32.4 0.057
Erosive REB 52 325.9 ±65.6 216.5 ±61.5 \0.001 0.046
SUC 23 307.0 ±68.8 237.3 ±65.1 \0.001
The units for PGE
2
are pg per gram tissue and the units for MDA are mmol per gram tissue. The P
1
-value compares week 8 and baseline results
(paired t-test) within each group; the P
2
-value represents significance between the trial and control group using covariance analysis. No
significant differences were found between the REB and SUC groups at baseline
2892 Dig Dis Sci (2008) 53:2886–2895
123
histological inflammation or activity paralleled the result
from endoscopy. In 52 subjects tested for MDA and PGE
2
in
the REB group, the subgroups status was 18, 16, and 18 for
H. pylori noneradicated, eradicated, and negative status,
respectively. In the erosive lesions, the PGE
2
content
increased from 234.7 ±27.0 pg/g to 254.8 ±32.1 pg/g
(P\0.05) in H. pylori-eradicated subgroup, with no sig-
nificant difference when compared with the noneradicated
and negative subgroups (220.6 ±25.3 pg/g to 263.8 ±
19.8 pg/g and 240.9 ±24.5 pg/g to 260.6 ±25.8 pg/g,
respectively, P[0.05). For MDA analysis, REB signifi-
cantly decreased mucosal MDA in all three subgroups and
no H. pylori-related effects were found.
Safety Assessment
Only three cases of adverse effects (AEs) occurred, in
0.62% of the REB group (n=2) and 1.03% (n=1) of the
SUC group (n=1) and were excluded from PP analysis,
including one case of rash, one case of abnormal hepatic
function, and another hospitalization for acute appendicitis.
All of the AEs developed in the period from the start to
week 4. No gastrointestinal symptoms, including diarrhea,
vomiting, and stomatitis, were found. All the events were
judged nonserious and not relevant to the study. Two
patients (one each in the REB and SUC groups) discon-
tinued the study drug and were lost to the follow-up. These
results demonstrated that treatment with rebamipide did not
affect patient safety when compared with the previously
proved safe drug sucralfate (P=0.551).
Discussion
One of the major roles of rebamipide [2-(4-chloro-
benzoylamino)-3[2(1H)-quinolonin-4-yl] propionic acid,
Mucosta
TM
] is to stimulate the generation of endogenous
prostaglandins in the gastric mucosa and it has been
reported to facilitate and accelerate ulcer healing.
Depending on the extensive published studies in vitro or
in vivo, this clinical trial represents a reliable result of
rebamipide on chronic gastritis. Rebamipide shows satis-
fied CEG-associated symptom attenuation with 8 weeks
therapy, with a rapid effect during 1 week. Most impor-
tantly, this effect has been proven by using sucralfate, a
traditional gastroprotective regent as a positive control. In a
previous placebo-controlled multicenter study [16], reb-
amipide showed no significant improvement in individual
symptom scores, except for a significantly reduced belch-
ing score in rebamipide 100 mg and 200 mg groups at
week 2 in Helicobacter pylori-positive patients. Though
symptom scores are sometimes subjective and vary
between studies, they may provide direct evidence in
clinical trials. Rebamipide showed a better effect than
sucralfate on symptom attenuation, including stomach ache
and distension, as well as acid reflux and belching after the
baselines were calibrated.
The mechanism by which rebamipide induces gastric
mucosa protection was also proved in this study. Reb-
amipide leads to upregulated PGE
2
and downregulated
oxygen free radicals level in the whole area of antrum.
Furthermore, the anti-inflammation effects were proved by
both macro and micro inspection. The inflammation scores
were significantly reduced. Rebamipide might also perform
gastroprotection via more pathways, such as downregula-
tion of intercellular adhesion molecule-1 (ICAM-1)
expression, inhibition of mitochondrial damage, lipid per-
oxidation, and apoptosis in indomethacin-induced injuries
[17–19]. Compared with sucralfate, rebamipide shows a
more advanced role in the inhibition of activity, which is
characterized by neutrophil infiltration. It has been reported
that rebamipide decreased the susceptibility of gastric
mucosa to acid-induced injury by inhibiting neutrophil
activation in rats [20] while similar evidence in humans is
rare. Rebamipide plays an extra role of stimulation of
endogenous PG excretion and anti-oxide action compared
with sucralfate, which led to the more obvious improvement
of gastric inflammation in this study. Interestingly, reb-
amipide performed better than sucralfate on the level of
PGE
2
and MDA only in the erosive region in the stomach.
The potential causes for this lesion-dependant effect might
be the higher oxygen free radical level in the erosive area
and that rebamipide has stronger effect on COX-2 synthesis
and radical scavenging [21–23]. Moreover, other roles
involved in rebamipide-induced gastric ulcer healing, such
as stimulating angiogenesis, may also lead to the disap-
pearance of erosive lesions in CEG [24]. Other than
inflammation indexes, MDA and PGE
2
, epithelial barrier
function or paracellular permeability have been considered
0
2
4
6
8
014284256
REB ( n=91 )
SUC ( n=33 )
REB ( n=110 )
SUC ( n=30 )
REB ( n=117 )
SUC ( n=34 )
**
014284256
**
0 1428 4256
** ** ** ** ** ** **
**
H. pylori non-eradicated H. pylori eradicated H. pylori negative
Fig. 6 The decrease of total symptom scores with rebamipide (REB)
or sucralfate (SUC) therapies in subgroups according to H. pylori
status. The time courses of the mean symptom score value are shown
in the graphs. The X-axis represents days from baseline and the Y-axis
represents mean symptom score. * P\0.05 and ** P\0.01
between the REB and SUC groups by Wilcoxon rank sum analysis.
(The lines represent the trend rather than a continuous variable)
Dig Dis Sci (2008) 53:2886–2895 2893
123
new targets for rebamipide effect in animal models [25–27],
although these indexes need to be designed practically
before being used in future clinical trials.
Because predominant clinical CEGs are H. pylori asso-
ciated, the bacterial factor should not be ignored in this study.
As estimated, the H. pylori infection rate in the population
was over 60% before randomization. H. pylori-associated
gastritis or ulcer is always harder to cure than unassociated
cases, as ulcers should be treated after anti-H. pylori triple
therapy. However, rebamipide did not show a worse result
with H. pylori infection compared with eradicated or nega-
tive subjects despite this concern. Actually rebamipide acted
in a H. pylori-independent manner. However the results
could not be explained by the hypothesis that rebamipide has
the function of eliminating H. pylori, even though we have
found the H. pylori infection rate in the rebamipide group to
be lower than in the sucralfate group. What we can conclude
from our study is that rebamipide may improve outcomes of
H. pylori-associated gastritis, as previously published
[28–31]. And more long-term, well-designed controlled tri-
als need to be carried out in the future to confirm the drug’s
relationship with H. pylori status [32].
Importantly, no serious side-effects were experienced by
any subjects in the study. Unlike misoprostol, rebamipide
resulted in less adverse diarrhea side-effects. The data in
the current study indicates the safety of rebamipide.
In conclusion, rebamipide leads to rapid, obvious, and
long-term improvement of upper GI symptoms in CEG
patients. Rebamipide can ameliorate inflammation in gas-
tric mucosa by induction of gastric mucosal PGE
2
synthesis
and inhibition of free-radical activity. Most importantly,
this effect was not influenced by H. pylori status and thus
rebamipide can be used in the treatment of H. pylori-
associated gastritis.
Acknowledgements The authors would like to thank all the insti-
tutes participating this study, as well as Prof. Jia He, Yalin Sun M.D.,
and Na Liu M.D. from the Statistics Department of SMMU for data
collection and analysis. Credits should also go to Otsuka Pharma-
ceutical Co., Ltd. for providing the drugs for the study.
References
1. Arakawa T, Kobayashi K, Yoshikawa T, Tarnawski A (1998)
Rebamipide, overview of its mechanisms of action and efficacy in
mucosal protection and ulcer healing. Dig Dis Sci 43:5S–13S
2. Arakawa T, Higuchi K, Fujiwara Y, Watanabe T, Tominaga K,
Sasaki E, Oshitani N, Yoshikawa T, Tarnawski AS (2005) 15th
anniversary of rebamipide: looking ahead to the new mechanisms
and new applications. Dig Dis Sci 50:3S–11S
3. Genta RM (2003) Review article: the role of rebamipide in the
management of inflammatory disease of the gastrointestinal tract.
Aliment Pharmacol Ther 18:8S–13S
4. Murata H, Yabe Y, Tsuji S, Tsujii M, Fu HY, Asahi K, Eguchi H,
Kawano S, Hayashi N (2005) Gastro-protective agent rebamipide
induces cyclooxygenease-2 (COX-2) in gastric epithelial cells.
Dig Dis Sci 50:70S–75S
5. Sakurai K, Osaka T, Yamasaki K (2005) Rebamipide reduces
recurrence of experimental gastric ulcers: role of free radicals and
neutrophils. Dig Dis Sci 50:90S–96S
6. Azuma T, Yamazaki S, Yamakawa A, Ito Y, Ohtani M, Dojo M,
Yamazaki Y, Higashi H, Hatakeyama M (2003) The effects of
cure of Helicobacter pylori infection on the signal transduction of
gastric epithelial cells. Aliment Pharmacol Ther 18:39S–44S
7. Yoshida N, Ishikawa T, Ichiishi E, Yoshida Y, Hanashiro K,
Kuchide M, Uchiyama K, Kokura S, Ichikawa H, Naito Y,
Yamamura Y, Okanoue T, Yoshikawa T (2003) The effect of
rebamipide on Helicobacter pylori extract-mediated changes of
gene expression in gastric epithelial cells. Aliment Pharmacol
Ther 18:63S–75S
8. Kim JS, Kim JM, Jung HC, Song IS (2003) The effect of
rebamipide on the expression of proinflammatory mediators
and apoptosis in human neutrophils by Helicobacter pylori
water-soluble surface proteins. Aliment Pharmacol Ther
18:45S–54S
9. Hahm KB, Kim DH, Lee KM, Lee JS, Surh YJ, Kim YB, Yoo
BM, Kim JH, Joo HJ, Cho YK, Nam KT, Cho SW (2003) Effect
of long-term administration of rebamipide on Helicobacter pylori
infection in mice. Aliment Pharmacol Ther 18:24S–38S
10. Fujioka T, Arakawa T, Shimoyama T, Yoshikawa T, Itoh M,
Asaka M, Ishii H, Kuwayama H, Sato R, Kawai S, Takemoto T,
Kobayashi K (2003) Effects of rebamipide, a gastro-protective
drug on the Helicobacter pylori status and inflammation in the
gastric mucosa of patients with gastric ulcer: a randomized
double-blind placebo-controlled multicentre trial. Aliment Phar-
macol Ther 18:146S–152S
11. Talley NJ, Quan C (2002) Review article: Helicobacter pylori
and nonulcer dyspepsia. Aliment Pharmacol Ther 16:58S–65S
12. Jin X, Li YM (2007) Systematic review and meta-analysis from
Chinese literature: the association between Helicobacter pylori
eradication and improvement of functional dyspepsia. Helico-
bacter 12:541–546
13. Farinati F, Cardin R, Degan P, Rugge M, Mario FD, Bonvicini P,
Naccarato R (1998) Oxidative DNA damage accumulation in
gastric carcinogenesis. Gut 42:351–356
14. Rokkas T, Simsek I, Ladas S (2007) Helicobacter pylori and non-
malignant diseases. Helicobacter 12:20S–22S
15. Rugge M, Genta RM (2005) Staging and grading of chronic
gastritis. Hum Pathol 36:228–233
16. Talley NJ, Riff DS, Schwartz H, Marcuard SP (2001) Double-
blind placebo-controlled multicentre studies of rebamipide, a
gastroprotective drug, in the treatment of functional dyspepsia
with or without Helicobacter pylori infection. Aliment Pharmacol
Ther 15:1603–1611
17. Hiratsuka T, Futagami S, Shindo T, Hamamoto T, Ueki N,
Suzuki K, Shinji Y, Kusunoki M, Shinoki K, Wada K, Miyake K,
Gudis K, Tsukui T, Sakamoto C (2005) Rebamipide reduces
indomethacin-induced gastric injury in mice via down-regulation
of ICAM-1 expression. Dig Dis Sci 50:84S–89S
18. Nagano Y, Matsui H, Muramatsu M, Shimokawa O, Shibahara T,
Yanaka A, Nakahara A, Matsuzaki Y, Tanaka N, Nakamura Y
(2005) Rebamipide significantly inhibits indomethacin-induced
mitochondrial damage, lipid peroxidation, and apoptosis in gas-
tric epithelial RGM-1 cells. Dig Dis Sci 50:76S–83S
19. Naito Y, Kajikawa H, Mizushima K, Shimozawa M, Kuroda M,
Katada K, Takagi T, Handa O, Kokura S, Ichikawa H, Yoshida N,
Matsui H, Yoshikawa T (2005) Rebamipide, a gastro-protective
drug, inhibits indomethacin-induced apoptosis in cultured rat
gastric mucosal cells: association with the inhibition of growth
arrest and DNA damage-induced 45 alpha expression. Dig Dis
Sci 50:104S–112S
2894 Dig Dis Sci (2008) 53:2886–2895
123
20. Harada N, Okajima K, Liu W (2005) Rebamipide decreases
the susceptibility of gastric mucosa to acid-induced injury in
rats by inhibiting neutrophil activation. Dig Dis Sci 50:
56S–62S
21. Hiratsuka T, Futagami S, Tatsuguchi A, Suzuki K, Shinji Y,
Kusunoki M, Shinoki K, Nishigaki H, Fujimori S, Wada K,
Miyake K, Gudis K, Tsukui T, Sakamoto C (2005) COX-1 and
COX-2 conversely promote and suppress ischemia-reperfusion
gastric injury in mice. Scand J Gastroenterol 40:903–913
22. Bamba H, Ota S, Kato A, Miyatani H, Kawamoto C, Yoshida Y,
Fujiwara K (2003) Effect of rebamipide on prostaglandin
receptors-mediated increase of inflammatory cytokine production
by macrophages. Aliment Pharmacol Ther 18:113S–118S
23. Sakurai K, Sasabe H, Koga T, Konishi T (2004) Mechanism of
hydroxyl radical scavenging by rebamipide: identification of
mono-hydroxylated rebamipide as a major reaction product. Free
Radic Res 38:487–494
24. Tarnawski AS, Chai J, Pai R, Chiou SK (2004) Rebamipide
activates genes encoding angiogenic growth factors and Cox2 and
stimulates angiogenesis: a key to its ulcer healing action? Dig Dis
Sci 49:202–209
25. Haruma K, Ito M (2003) Review article: clinical significance of
mucosal-protective agents: acid, inflammation, carcinogenesis
and rebamipide. Aliment Pharmacol Ther 18:153S–159S
26. Matysiak-Budnik T, Heyman M, Megraud F (2003) Review
article: rebamipide and the digestive epithelial barrier. Aliment
Pharmacol Ther 18:55S–62S
27. Joh T, Takezono Y, Oshima T, Sasaki M, Seno K, Yokoyama Y,
Ohara H, Nomura T, Alexander JS, Itoh M (2003) The protective
effect of rebamipide on paracellular permeability of rat gastric
epithelial cells. Aliment Pharmacol Ther 18:133S–138S
28. Terano A, Arakawa T, Sugiyama T, Yoshikawa T, Haruma K,
Asaka M, Shimosegawa T, Sakaki N, Ishii H, Sakamoto C,
Takahashi S, Kinoshita Y, Fujioka T, Kobayashi K (2006) A pilot
study to evaluate a new combination therapy for gastric ulcer:
Helicobacter pylori eradication therapy followed by gastropro-
tective treatment with rebamipide. J Gastroenterol Hepatol
21:103–109
29. Higuchi K, Tanigawa T, Hamaguchi M, Takashima T, Sasaki E,
Shiba M, Tominaga K, Fujiwara Y, Oshitani N, Matsumoto T,
Watanabe T, Arakawa T (2003) Comparison of the effects of
rebamipide with those of cimetidine on chronic gastritis associ-
ated with Helicobacter pylori in Mongolian gerbils. Aliment
Pharmacol Ther 18:1S–7S
30. Choi KW, Lee YC, Chung IS, Lee JJ, Chung MH, Kim NY, Kim
SW, Kim JG, Roe IH, Lee SW, Jung HY, Choi MG, Hahm KB,
Hong WS, Kim JH (2002) Effect of rebamipide in treatment of
Helicobacter pylori-associated duodenal ulcer: attenuation of
chemokine expression and nitrosative damage. Dig Dis Sci
47:283–291
31. Adachi K, Fujishiro H, Mihara T, Komazawa Y, Kinoshita Y
(2003) Influence of lansoprazole, famotidine, roxatidine and
rebamipide administration on the urea breath test for the diag-
nosis of Helicobacter pylori infection. J Gastroenterol Hepatol
18:168–171
32. Haruma K, Ito M, Kido S (2002) Long-term rebamipide therapy
improves Helicobacter pylori-associated chronic gastritis. Dig
Dis Sci 47:862–867
Dig Dis Sci (2008) 53:2886–2895 2895
123
