Introduction
The prevalence of gastroesophageal reflux disease (GERD) has been increasing, especially
in Western countries [1]
[2]
[3]. Although the reported prevalence is lower in Asian populations, GERD has become
more prevalent in recent decades due to several factors, including a decrease in the
rate of Helicobacter pylori (H. pylori) infection, westernization of eating habits, and increased obesity rate [4]
[5]
[6]
[7]. Patients with GERD suffer from heartburn and regurgitation, which are caused by
reflux of stomach contents, leading to reductions in health-related quality of life
and work productivity [8]
[9]. GERD patients will usually experience relapse once the prescribed drug is discontinued,
because contributing factors, such as hiatal hernia, lower esophageal sphincter (LES)
hypotension, and delayed gastric emptying are not corrected by acid inhibition. In
addition, GERD is a risk factor for development of Barrett’s esophagus and esophageal
adenocarcinoma [10]
[11], thus longitudinal management of affected patients is considered to be important.
However, the long-term outcome of GERD patients with or without medical management
remains to be fully clarified.
GERD can be classified into non-erosive reflux disease (NERD) and erosive reflux disease,
or reflux esophagitis (RE), in which mucosal breaks are observed by endoscopy [12]. Severity in patients with the latter condition is associated with esophageal acid
exposure time [13]. Previously, we focused on the circumferential distribution of esophageal mucosal
breaks in patients with various grades of RE and found a unique asymmetrical circumferential
distribution [14]. In that study, esophageal mucosal breaks in low-grade esophagitis [Los Angeles
(LA) grade A and B] cases [15] were mainly found on the right anterior wall of the distal esophagus. In contrast,
those in cases of high-grade esophagitis (LA grade C) were mainly found on the posterior
wall of the esophagus, suggesting a difference in acid reflux pattern between low-
and high- grade RE in association with the circumferential distribution of mucosal
breaks [13]. Given RE progresses to more severe forms over time, circumferential location of
mucosal breaks may change at the time of recurrence as compared to that at the time
of initial diagnosis. However, it remains unclear whether the circumferential location
of esophageal mucosal breaks changes during a long-term course of RE.
The aim of the current study was to investigate the circumferential distribution of
mucosal breaks in patients with recurrent RE and compare to the location noted at
the initial diagnosis.
Patients and methods
Subjects
We retrospectively reviewed the medical records of patients diagnosed with recurrent
RE with LA grade A-C [15] at Shimane University Hospital between July 1996 and June 2014. All were initially
given proton pump inhibitor (PPI) treatment for at least 2 months and healing of initial
mucosal breaks (LA grade N or M) was confirmed by esophagogastroduodenoscopy (EGD)
findings. Each was followed on a long-term basis under routine care at the discretion
of the attending physician. Yearly follow-up EGD examinations were performed for screening
of gastric cancer in some cases, and repeated EGD was also performed when symptom
relapse occurred. Patients with new esophageal mucosal breaks were diagnosed as recurrent
RE, irrespective of acid suppressive treatment, and enrolled in this study.
Endoscopic images with reports were separately reviewed by 3 of the authors (N.F,
H.M, N.I). The examiners were blinded in regard to the clinical diagnosis of each
case and compared each report with related endoscopic images to confirm presence and
location of each lesion described by examining endoscopists. Endoscopic diagnosis
was established by consensus of at least 2 of the 3. The grade and location of new
mucosal breaks were evaluated to determine the circumferential distribution of the
lesions, and their location was also compared to that at the time of the initial diagnosis.
Information regarding clinical parameters, including demographics (age, sex), PPI
or H2 receptor antagonist (H2RA) administration at the time of recurrence, presence of hiatal hernia, and gastric
mucosal atrophy, was reviewed. A hiatal hernia was diagnosed when its length was greater
than 2 cm. The degree of gastric mucosal atrophy was endoscopically graded according
to the Kimura-Takemoto classification. A grade of C1 – 3 for closed type indicated
milder gastric mucosal atrophy, while a grade of O1 – 3 for open type indicated more
severe gastric mucosal atrophy. Patients with a history of upper gastrointestinal
surgery were excluded. Patients who had taken any PPI or H2RA within 4 weeks prior to the initial diagnosis were also excluded. The protocol
was approved by the ethics committee of Shimane University School of Medicine and
the study was carried out in accordance with the principles of the Helsinki Declaration.
Assessment of circumferential distribution of mucosal breaks
The circumferential locations of mucosal breaks on the esophageal wall were evaluated
as on a clock face, as follows. With the anterior wall of the esophagus always positioned
at 0 o’clock, the 3 o’clock position was defined as the right lateral wall of the
esophagus aligned with the lesser curvature of the stomach. Examining endoscopists
described a clock face orientation of the esophageal lumen when they found the mucosal
breaks in the endoscopic report. Patients with insufficient endoscopic reports and
those with mucosal breaks whose location could not be clearly shown by endoscopic
imaging were excluded. Subjects with recurrent RE were divided into 2 groups; recurrence
at the same location and that at a different location. If mucosal breaks were observed
at more than 1 location at the time of recurrence, those with at least 1 at the same
location were defined as having lesions at the same location.
Representative endoscopic findings in a patient with mucosal breaks at the same locations
are shown in [Fig. 1], while those in a patient with mucosal breaks in a different location are shown
in [Fig. 2]. The image in [Fig. 1a] was obtained at the time of the initial diagnosis and shows a mucosal break (arrowhead)
at the 2 o’clock location. Following PPI treatment, healing of the mucosal break seen
at the initial diagnosis was confirmed ([Fig. 1b]). At the time of recurrence, a new mucosal break was observed in the same location
(2 o’clock) as the initial lesion ([Fig. 1c]). [Fig. 2a] presents an image of mucosal breaks (arrowhead) located at 0, and 3 o’clock that
were observed at the initial diagnosis. Following PPI treatment, healing of the mucosal
breaks seen at the initial diagnosis was confirmed ([Fig. 2b]). At recurrence, a new mucosal break was observed at position corresponding to 6
o’clock with severe hiatal hernia ([Fig. 2c]), indicating recurrent mucosal break at different location.
Fig. 1 Circumferential distribution of esophageal mucosal breaks in a patient with recurrent
RE. Shown are representative endoscopic findings of esophageal mucosal breaks in the
same location at recurrence. a Esophageal mucosal break (arrowhead) observed at 2 o’clock at initial diagnosis.
b Following PPI treatment, healing was confirmed. c At recurrence, a new mucosal break (arrowhead) was again observed at 2 o’clock, which
was defined as recurrence at the same location as noted at the time of the initial
diagnosis.
Fig. 2 Circumferential distribution of esophageal mucosal breaks in a patient with recurrent
RE. Shown are representative endoscopic findings of esophageal mucosal breaks at a
different location at recurrence. a Esophageal mucosal break (arrowheads) was observed at 0, and 3 o’clock at the initial
diagnosis. B Following PPI treatment, healing was confirmed. c At recurrence, a new mucosal break (arrowhead) was observed at 6 o’clock with severe
hiatal hernia, which was defined as recurrence at a different location as compared
to that noted at the time of initial diagnosis.
Background factors, including sex, age at initial diagnosis, age at recurrence, time
to recurrence, LA grade at the initial diagnosis and at recurrence, administration
of PPI or H2RA at recurrence, and the presence of hiatal hernia and atrophic gastritis were compared
between cases with recurrent lesions at the same location and those with such recurrence
at different locations.
Statistical analysis
Statistical analyses were performed using chi-squared and Mann – Whitney U-tests.
After identification of significant predictors by univariate analysis, logistic regression
analysis was performed to calculate the odds ratio with confidence interval (CI) for
independent predictors of a different location of mucosal breaks at recurrence. A
value of P < 0.05 was considered to indicate a statistically significant difference. All statistical
analyses were performed using the SPSS statistical analysis software package (version
22.0 for the PC, Chicago, IL, USA).
Results
Patient profiles
The clinical characteristics of patients with recurrent reflux esophagitis are shown
in [Table 1]. We enrolled 114 patients (76 males, 38 females) whose mean (± SD) age was 65.5 ± 12.5
years (range 34 – 86 years) at the time of the initial diagnosis. The number of cases
with LA grade A was 71, while 34 were LA grade B and 9 were LA grade C, while those
numbers at recurrence were 68, 38, and 8, respectively. The mean time to recurrence
after the initial diagnosis was 39.4 ± 31.4 months (range 3 – 142 months). PPIs were
administrated to 44 patients (37 with regular or double dose, 7 with half dose), while
H2 receptor antagonists were administered to 10 patients at the time of recurrence.
Table 1
Clinical characteristics of patients with recurrent reflux esophagitis.
|
Number of patients
|
114
|
|
Male, no. (%)
|
76 (67)
|
|
Age at initial diagnosis, years, mean ± SD (range)
|
65.5±12.5 (34 – 86)
|
|
Age at recurrence, years, mean ± SD (range)
|
68.7 ± 12.2 (42 – 89)
|
|
Time to recurrence, months, mean ± SD (range)
|
39.4 ± 31.4 (3 – 142)
|
|
LA grade at initial diagnosis, no. (A/B/C)
|
71/34/9
|
|
LA grade at recurrence, no. (A/B/C)
|
68/38/8
|
|
PPI administration at recurrence, no. (%)
|
44 (39)
|
|
H2RA administration at recurrence, no. (%)
|
10 (9)
|
|
Hiatal hernia, no. (%)
|
88 (77)
|
|
Atrophic gastritis (open type), no. (%)
|
23 (20)
|
PPI, proton pump inhibitor; H2RA, histamine H2 receptor antagonist; LA, Los Angeles classification of gastroesophageal reflux disease
Change in severity of LA grade at recurrence
First, we evaluated the change in severity of LA grade at recurrence ([Table 2]). Fifty-six (78.9 %) of our RE patients with LA grade A had recurrence with the
same grade, while 15 (21.1 %) progressed to a more severe form (14 grade B, 1 grade
C). Likewise, 22 (64.7 %) with LA grade B had the same grade at recurrence, while
only 2 (5.9 %) progressed to grade C. In contrast, 10 (29.4 %) had a milder form of
RE (grade A) at recurrence. Taken together, only 17 (14.9 %) progressed to a more
severe grade of RE, while the majority (72.8 %) had recurrence at the same LA grade
as at the time of the initial diagnosis. When the patients were divided into those
with PPI administration at recurrence (n = 44) or those without administration at
recurrence (n = 70), the majority had the same LA grade at recurrence in both groups
(63.6 % of subjects with PPI administration, 78.6 % of subjects without PPI), while
the ratio of patients with improvement to a milder grade was higher in those with
PPI administration than in those without PPI administration (20.5 % of subjects with
PPI administration, 7.1 % of subjects without PPI, p < 0.05).
Table 2
Change in severity of LA grade at recurrence.
|
LA classification at initial diagnosis
|
Classification at recurrence
|
|
LA-A
(n = 68)
|
LA-B
(n = 38)
|
LA-C
(n = 8)
|
|
LA-A (n = 71)
|
56
(78.9 %)
|
14
(19.7 %)
|
1
(1.4 %)
|
|
LA-B (n = 34)
|
10
(29.4 %)
|
22
(64.7 %)
|
2
(5.9 %)
|
|
LA-C (n = 9)
|
2
(22.2 %)
|
2
(22.2 %)
|
5
(55.6 %)
|
LA, Los Angeles classification of gastroesophageal reflux disease
Circumferential distribution of mucosal breaks
Next, we assessed the circumferential distribution of mucosal breaks in the esophageal
wall in our patients at the time of initial diagnosis and at recurrence of RE. That
distribution at the initial diagnosis is shown in [Fig. 3]. In patients with LA grade A, mucosal breaks were frequently located at the 2 o’clock
position, while those in patients with grade B were located at positions from 2 to
3 o’clock. In contrast, in patients with grade C, the mucosal breaks were mainly located
on the posterior wall of the esophagus. These findings were consistent with those
obtained in our previous study [14]. As shown in [Fig. 4], the circumferential distribution of mucosal breaks at recurrence was similar to
the pattern seen at the initial diagnosis, suggesting that it was consistent and closely
associated with RE severity. When that distribution were compared between those with
PPI (n = 44) and without PPI at recurrence (n = 70), the pattern of that distribution
at recurrence was not different between the groups (data not shown), suggesting that
PPI administration was not strongly associated with the circumferential distribution
of mucosal breaks.
Fig. 3 Circumferential distribution of esophageal mucosal breaks in patients with RE and
LA grade of A, B, or C (image A, B, C, respectively) at the time of the initial diagnosis.
Locations are shown in terms of clock face orientation.
Fig. 4 Circumferential distribution of esophageal mucosal breaks in patients with RE and
an LA grade of A, B, or C (image A, B, C, respectively) at the time of recurrence.
Locations are shown in terms of clock face orientation. The location was similar to
that noted at the time of initial diagnosis in each patient.
Comparison of circumferential location of mucosal breaks at initial diagnosis and
recurrence
New esophageal mucosal breaks in the same location as the initial lesions ([Fig. 1]) were found in 96 patients (84.2 %), while those in a different location ([Fig. 2]) were found in 18 (15.8 %), indicating that the majority had recurrence at the same
location. To evaluate the relationship between mucosal break location and change in
severity of LA grade at recurrence, we divided the patients into 3 groups based on
findings at recurrence; same LA grade as at initial diagnosis (n = 83), progression
to more severe grade (n = 17), and improvement to milder grade (n = 14) ([Fig. 5]). Interestingly, in most of the patients (89.2 %) with the same severity as seen
at the initial diagnosis, mucosal breaks were observed in the same location as at
the initial diagnosis, while those with either progression or improvement in severity
had lesions in the same location as at the initial diagnosis less frequently (29.4,
and 28.6 %, respectively P < 0.05). At recurrence, PPIs were more frequently administered to patients with improvement
to a milder grade as compared to the others (64.3 % of patients with milder grade,
33.7 % of patients with same grade, 41.2 % of patients with progression to more severe
grade; P < 0.05).
Fig. 5 Relationship between location of mucosal break and change in severity of LA grade
at recurrence. Patients were divided into three groups; same LA grade as at initial
diagnosis (n = 83), progression to more severe grade (n = 17), and improvement to
milder grade (n = 14). Most of the patients (89.2 %) with the same grade at the time
of recurrence showed mucosal breaks in the same location as noted at the initial diagnosis,
while those with progression or improvement of mucosal breaks at recurrence less frequently
had lesions in the same location as noted at the initial diagnosis (29.4, and 28.6,
respectively).
We also compared clinical parameters between patients with recurrent lesions in the
same location as the initial lesions (n = 96) and patients with those in different
locations (n = 18) ([Table 3]). There were no significant differences in regard to age, time period to recurrence,
administration of PPIs at recurrence, or presence of hiatal hernia and atrophic gastritis
between the groups. However, recurrence with a different LA grade severity was significantly
associated with a different location of mucosal breaks at recurrence (P < 0.05). Logistic regression analysis findings confirmed that a change in severity
of lesions was a positive independent factor for different location of recurrent mucosal
breaks (OR 6.07, 95 %CI 1.73 – 21.28, P < 0.01).
Table 3
Characteristics of patients with recurrent RE in same and different locations.
|
Same
(n = 96)
|
Different
(n = 18)
|
P value
|
|
Male, no. (%)
|
67 (70)
|
9 (50)
|
0.10
|
|
Age at initial diagnosis, years, mean ± SD
|
64.8 ± 13.1
|
69.2 ± 8.8
|
0.25
|
|
Age at recurrence, years, mean ± SD
|
68.0 ± 12.8
|
72.5 ± 8.0
|
0.23
|
|
Time to recurrence, months, mean ± SD
|
39.4 ± 30.4
|
39.5 ± 37.2
|
0.68
|
|
LA grade at initial diagnosis, no. (A/B/C)
|
59/28/9
|
12/6/0
|
0.20
|
|
Recurrence with different LA grade, no. (%)
|
22 (23)
|
9 (50)
|
0.02
|
|
PPI administration at recurrence, no. (%)
|
37 (39)
|
7 (39)
|
0.98
|
|
H2RA administration at recurrence, no. (%)
|
7 (7)
|
3 (17)
|
0.95
|
|
Hiatal hernia, no. (%)
|
75 (78)
|
13 (72)
|
0.58
|
|
Atrophic gastritis (open type), no. (%)
|
17 (18)
|
6 (33)
|
0.13
|
PPI, proton pump inhibitor; LA, Los Angeles classification of gastroesophageal reflux
disease
Discussion
Patients with RE, easily recognized based on EGD findings, often develop complications,
such as bleeding, deep ulceration, strictures, Barrett’s esophagus, and esophageal
adenocarcinoma [10]
[11]. However, there is scant information available regarding their long-term clinical
course or type of recurrence. In the present study, RE patients were found to usually
experience recurrence at the same LA grade, as most did not progress to a more severe
grade during the follow-up observations.
Prevalence of GERD has been increasing in Asian populations in recent decades, mainly
due to decreases in rates of H. pylori infection [4] and increases in obesity [6]
[7]
[16], though the rate of obesity (body mass index ≥ 30 kg/m2) is much lower in Asian as compared to Western countries [17]
[18]. Gastric acid secretion is another important factor for development of GERD. Studies
conducted in the 1990 s of Japanese subjects showed that gastric acid secretion levels
increased from the 1970 s to 1990 s, though they remained lower as compared to Europeans
and North Americans [19]
[20]
[21]. In addition, those levels have not increased over the past 2 decades in Japanese
[22]
[23]. These factors may be related to the lower prevalence of GERD in Asian individuals
as compared to Western populations. Consistent with the present results, approximately
90 % of RE cases in Japan have been found to be low grade, such as LA grade A or B
[3].
As for the natural history of GERD, Manabe et al. followed 105 patients with LA grade
A or B RE who did not receive medical treatment for that condition for a mean 5.5
years (range 2.0 – 8.8 years) [24]. In that study, only 10.5 % of the patients progressed to a more severe grade (mostly
grade C) and none progressed to Barrett’s esophagus. They also evaluated factors associated
with development of severe esophagitis. Risk factors for progressive disease were
found to be increased age, female sex, presence of symptoms at initial diagnosis,
presence of hiatal hernia, no atrophic gastritis, and no H. pylori infection. Our results were consistent with theirs and suggest little progression
in severity over time regardless of grade. Furthermore, findings from a large multicenter
study conducted in Europe of 2721 patients who completed the 5-year follow-up protocol
suggest that most GERD patients remain stable or show improvement when receiving current
routine clinical care [25]. Taken together, in most patients, low-grade RE does not progress over time, though
factors predictive of progression to more severe esophagitis must be clarified in
order to identify those at risk.
Consistent with our previous findings [14], the current study showed that the circumferential distribution of esophageal mucosal
breaks varies based on LA grade showing the severity of esophagitis, regardless of
the timing of the endoscopic observation (initial diagnosis or recurrence) and presence
or absence of PPI administration. Esophageal mucosal breaks were mostly found on the
right anterior wall (2 o’clock) of the esophagus in RE patients with LA grade of A
or B, while those were often seen on the posterior wall of the esophagus in patients
with high-grade RE (grade C). Potential pathophysiologic explanations underlying the
circumferential distribution of mucosal breaks in RE include the functional structure
of the lower esophageal sphincter and nonuniform asymmetric distribution of esophageal
acid exposure [26]
[27]
[28]. Interestingly, circumferential intraesophageal pressure at the LES has been reported
to be asymmetrical [29], while a recent study that used 3-dimensional high-resolution manometric imaging
revealed decreased competency within the right-side portion of the LES pressure zone
as compared to the left side [30]. These findings suggest that asymmetrical LES pressure caused by anatomical proportion
may be a major cause of asymmetrical gastroesophageal acid reflux and is correlated
with the predominant circumferential location of mucosal breaks in the esophagus.
Patients with grade C or D esophagitis show decreased LES pressure with an impaired
reflux barrier caused by the factors including hiatal hernia, thus they often have
severe nocturnal gastroesophageal reflux when lying in bed [13]. Therefore, mucosal breaks are prone to be found on the posterior wall of the esophagus.
The current study is the first to show that most patients with recurrent RE develop
mucosal breaks at the same circumferential location as noted at the time of initial
diagnosis. When clinical parameters in recurrent cases were compared between those
with the same and different locations of new lesions, recurrence with a changed LA
severity grade, irrespective of improvement or progression, was the only significant
parameter. Our results suggest that acid reflux mechanisms in patients with RE differ
among the different grades and the pathogenetic mechanism does not usually change
over the long term. When a new lesion is found in a different circumferential location,
a different acid reflux mechanism, such as having hiatal hernia, should be suspected
as predominant in that patient. Thus, the circumferential location of lesions observed
during follow-up endoscopic examinations may be an important parameter to diagnose
regarding disease progression.
Another importance of this study is that persistent higher level of gastric acid and/or
bile acid reflux at the same circumferential location may result in mucosal breaks
associated with Barrett’s esophagus (BE), and BE-associated adenocarcinoma. We previously
investigated the location of short segment BE (SSBE) and early neoplastic lesion in
patients with SSBE, and reported their predominant distribution on the right anterior
wall of the esophagus [31]. Our finding was confirmed by subsequent several promising studies [32]
[33]. In addition, recent report by Omae et al. revealed that the location of superficial
BE-associated adenocarcinoma mostly corresponds to the direction of maximal total
duration of acid or non-acid reflux [34]. Because the prognosis of patients with advanced-stage BE-associated adenocarcinoma
remains poor, early detection of neoplastic lesion in those with BE has led to recent
interest in effective treatment. Our data suggest that precise inspection of this
area is important both in patients with suspected BE and in patients undergoing their
surveillance of BE.
There are several limitations associated with this study, including its retrospective
nature and all patients being treated at the same tertiary referral center. Although
the circumferential distribution of esophageal mucosal breaks was determined retrospectively
based on recorded endoscopic images with reports, we focused on lesions in the area
of the esophago-gastric junction (EGJ). We previously reported the clinical significance
of asymmetrical distribution of lesions in the EGJ, including mucosal breaks in patients
with RE, SSBE, esophageal adenocarcinoma, Mallory-Weiss tears, and raptured esophageal
varices [27]
[35]. Therefore, we consider that our institution has adequate experience with and uses
a standardized protocol for use of endoscopic photo series in our endoscopy database,
and the results obtained in this study are thought to be reliable. Nevertheless, a
prospective study is needed to clarify the time-course changes in circumferential
distribution of esophageal mucosal breaks in patients with RE.
Conclusions
In summary, we found that most of our patients with recurrent RE developed lesions
in the same circumferential location as noted at the initial diagnosis. Those in different
locations were associated with change of LA grade with recurrence.
