Introduction
Helicobacter pylori, when it infects the gastric mucosa, has been recognized as a major pathogen in the
development of gastric cancer [1]. In particular, the presence of gastric intestinal metaplasia (IM) is thought to
be an important risk in the development of gastric cancer because of its strong association
with intestinal-type gastric cancer [2]
[3]. Appropriate surveillance of gastric IM could lead to earlier detection and treatment
of gastric cancer, potentially leading to more favorable patient survival outcomes
[4]. As gastric IM shows heterogeneous patchy distribution and is not distinctly visible
by conventional white-light endoscopy alone, the evaluation of gastric IM is of limited
accuracy. In the Kyoto global consensus report on H. pylori gastritis, expert panels strongly recommend modern image-enhanced endoscopy, such
as narrow-band imaging (NBI), for the accurate diagnosis of gastric IM [5].
Using magnifying NBI (M-NBI), light blue crest (LBC) has been widely accepted as an
endoscopic marker of gastric IM [6]. LBC is defined as a fine, blue-white line on the crests of the epithelial surface/gyri
observable with M-NBI and histologically indicates the brush border of IM [6]. However, in a recent study, white opaque substance (WOS) observed with M-NBI was
reported as another, novel, endoscopic marker of IM [7]. WOS on M-NBI was first reported by Yao et al. [8] as a substance present in the superficial part of gastric neoplasias that obscures
the subepithelial microvascular architecture. According to fat (oil-red O) staining
and immunohistochemical and immunoelectron microscopic studies for adipophilin, WOS
in gastric epithelial neoplasia is an optical phenomenon caused by accumulated lipid
droplets [9]
[10]. Studies of H. pylori infection-associated IM of the stomach using lipid staining and light microscopy
or electron microscopy showed that the epithelium in IM has the ability to absorb
lipid droplets [11]
[12].
Lipase plays a crucial role in the lipid digestion and absorption process and is known
to be inactivated in strong acidity [13]. Therefore, we assume that the presence of absorbed lipid droplets as WOS in IM
may represent a hypochlorhydria state caused by atrophic gastritis due to H. pylori infection and may not be a consistent phenomenon because it is influenced by acid
recovery in H. pylori-eradicated patients. In other words, WOS may not be sufficiently sensitive as an
endoscopic marker of IM in patients with recovered acid secretion after H. pylori eradication, an issue of critical importance that requires resolution. The Japanese
government approved eradication treatment for patients with H. pylori-associated chronic gastritis under the National Health Insurance System from February
2013, and the number of patients who have received H. pylori eradication therapy is increasing in Japan. Moreover, the incidence of gastric cancer
after H. pylori eradication has been increasing, especially in patients who have had observed IM
in background mucosa [14]. Using routine screening endoscopy, accurate diagnosis of IM as a precancerous lesion
in patients is required, including in those who have had H. pylori eradicated.
In this study, we aimed to clarify the clinical significance of the presence of WOS
and its limitations in the changing gastric acid conditions in various chronic gastritis
states secondary to H. pylori infection.
Methods
Study design and participants
This study was approved by the institutional review board of Oita Red Cross Hospital
(IRB no. 190) and conducted according to the Declaration of Helsinki. All patients
provided their written informed consent.
We retrospectively extracted 582 patients who underwent both upper gastrointestinal
(GI) screening endoscopy with magnifying NBI and intragastric pH assessment at Oita
Red Cross Hospital between April 2014 and July 2016. The data of fasting gastric juice
used in this study were originally collected in our previous study on investigation
of endoscopic findings that can predict gastric acid condition [15].
Subjects were excluded if they had following exclusion criteria: (1) patients who
had a history of gastric resection; (2) with severe organ failure; (3) current intake
of antiacid agents (within 5 d prior to the endoscopic examination); (4) patients
within 6 mo of prior H. pylori eradication; (5) patients with an unclear H. pylori status or suggestive of spontaneous H. pylori eradication and; (6) endoscopic images insufficient for assessment. After 81 patients
were excluded, a total of 501 patients were included in this study.
Endoscopic procedure and pH measurement of fasting gastric juice
A high-resolution magnifying upper GI endoscope (GIF-Q240Z; Olympus, Tokyo, Japan)
or a high-definition magnifying upper GI endoscope (GIF-H260Z, Olympus) and an electronic
endoscopy system (EVIS LUCERA Spectrum, Olympus) were used. A soft black hood (MAJ-1989
for the GIF-Q240Z, MAJ-1990 for the GIF-H260Z; Olympus) was mounted at the tip of
the endoscope to enable the endoscopist to fix a consistent focal distance between
the tip of the endoscope and the gastric mucosa. Measurement of the pH of fasting
gastric juice and M-NBI examinations were carried out by endoscopists K. T., T. U.,
K. W., H. H., S. I., M. F., and Y. Y.
Endoscopic study was performed following an overnight fast of 12 h or more. To avoid
contamination affecting the accuracy of pH measurements, premedication (such as mucolytic
or deforming agents) was not administered, except for local pharyngeal anesthesia
with 8 % lidocaine pump spray. Following endoscope insertion into the stomach, gastric
juice pooled in the stomach was aspirated and collected endoscopically according to
a previously described technique [16]. The pH of the collected sample was immediately measured using a 14-step pH test
strip (pH 1 – 14; AS ONE Co., Osaka, Japan). Following this, the observer evaluated
the presence or absence of WOS at 2 regions of interest using M-NBI. The observer
then recorded the results on a digital filing system (NEXUS; Fujifilm Medical Co.,
Tokyo, Japan), and the images were subsequently double-checked by an experienced M-NBI
endoscopist (T. U.).
Definitions of WOS
The 2 regions of interest identified for assessment of presence of WOS were a 2 × 2 cm
area at the lesser curvature of the antrum, 2 cm oral to the pylorus ring, and a 2 × 2 cm
area at the lesser curvature of the gastric corpus, 4 cm oral to the angulus according
to a previous report [17].
We defined the region of interest as WOS-positive when M-NBI findings showed a white
substance obscuring the subepithelial capillaries of the intervening parts between
crypt openings regardless of the degree or distribution and regardless of LBC coexistence
([Fig. 1a]). A normal mucosa or LBC-positive but WOS-negative mucosa was defined as WOS-negative
([Fig. 1b] and [Fig. 1c]). We divided the cases into WOS-negative (WOS-0), for which WOS was negative at
both the pyloric gland area and the fundic gland area, WOS-pyloric (WOS-P), for which
WOS was positive at the pyloric gland area only, WOS-fundic (WOS-F), for which WOS
was positive at the fundic gland area only, and WOS-pyloric plus fundic (WOS-P + F),
for which WOS was positive at both the pyloric gland area and the fundic gland area.
Fig. 1 Definition of WOS-positive or WOS-negative. a WOS was defined as positive
when a white substance was observed by M-NBI that obscured the subepithelial capillaries
of the
intervening part between crypt openings in the region of interest, regardless of LBC
coexistence. b Normal mucosa. c LBC positive but WOS-negative mucosa was defined as WOS-negative.
Definitions of H. pylori status
To detect H. pylori infection, the following tests were used: serum Ig G antibody against H. pylori, 13C-urea breath test (UBT), urine antibody and stool antigen against H. pylori, and histological examination. A positive result for any test was diagnosed as “H. pylori-infected.” A negative result for least 1 test, without endoscopic atrophy, was diagnosed
as “H. pylori-uninfected.” When a UBT yielded negative results with a history of eradication therapy,
a diagnosis of “H. pylori-eradicated” was given. Subjects with a negative result for at least 2 of the tests,
with endoscopic atrophy and no previous history of H. pylori eradication suggestive of spontaneous eradication, were excluded from the study.
Proton pump inhibitor administration in 29 H. pylori-eradicated cases with gastric epithelial neoplasia
We assumed that WOS would not be highly sensitive as an endoscopic marker for IM in
patients after H. pylori eradication, especially in cases with acid recovery. To clarify this, we investigated
the relationship between the presence of WOS and the pH level of fasting gastric juice
before and after proton pump inhibitor (PPI) administration in 29 H. pylori-eradicated cases with gastric epithelial neoplasia. Of these cases, 9 adenomas and
20 differentiated adenocarcinomas were detected at least 6 mo after successful H. pylori eradication, and these patients received endoscopic resection of neoplasia. For the
treatment of gastric ulcer after endoscopic resection, we administered PPIs. In all
29 cases, we measured fasting gastric juice pH and observed for WOS before and after
PPI administration (mean 21 d) according to the aforementioned procedure.
End points
The end points of the study were (1) the prevalence of WOS and the relationship between
the presence of WOS and fasting gastric juice pH level in 3 groups (H. pylori-uninfected, H. pylori-infected, and H. pylori-eradicated); (2) the relationship between the presence of WOS and fasting gastric
juice pH level before and after PPI administration in 29 H. pylori-eradicated cases.
Statistical analysis
Continuous variables were expressed as mean ± standard deviation (SD) or 95 % confidence
interval (CI). For parametric variables, Student’s t-test was used to compare the
means between 2 groups; otherwise, a Wilcoxon’s rank-sum test was used to compare
the means of the groups. The chi-squared test or Fisher’s exact test was used for
comparisons of prevalence among the groups. McNemar’s test was used to compare the
prevalence of WOS before and after PPI administration in the 29 H. pylori-eradicated cases. A 2-sided p-value of less than 0.05 was considered statistically
significant. Statistical analyses were performed using JMP-9 software (SAS Institute,
Cary, NC).
Results
Clinical features of patients divided into 3 groups by H. pylori status
[Table 1] shows the characteristics of patients by dividing into 3 groups by H-pylori status (H. pylori-uninfected, H. pylori-infected, and H. pylori-eradicated). Among 501 patients included in this study, the mean ± SD age of the
patients was 57.2 ± 13.4 y. The male to female ratio was 303:198. Of the 501 patients,
206 (41.1 %) were H. pylori-uninfected, 109 (21.8 %) were H. pylori-infected, and 189 (37.1 %) were H. pylori-eradicated. The mean period (mo) after H. pylori eradication in subjects was 36 mo (range: 7 – 180 mo).
Table 1
Clinical characteristics of patients divided into 3 H. pylori status.
|
H. pylori-uninfected (n = 206)
|
H. pylori-infected (n = 109)
|
H. pylori-eradicated (n = 186)
|
Total (n = 501)
|
|
Mean age ± SD (y)
|
49.0 ± 10.5
|
62.0 ± 13.1
|
64.4 ± 11.7
|
57.2 ± 13.4
|
|
Male/female, n
|
109/97
|
74/35
|
120/66
|
303/198
|
|
Gastric mucosal atrophy none/positive atrophy (closed or open type)
|
205/1
|
4/105
|
15/171
|
224/277
|
|
|
1 (0.5)
|
30 (27.5)
|
96 (51.6)
|
127 (25.3)
|
|
|
0 (0)
|
75 (68.8)
|
75 (40.3)
|
150 (30.0)
|
|
Current or previous history of gastric tumor, n (%)
|
0 (0)
|
17 (15.6)
|
52 (27.9)
|
69 (13.7)
|
According to the Kimura-Takemoto classification system for degree of gastritis [18], 224 (44.7 %) had no atrophy, 127 (25.3 %) had closed-type atrophic gastritis, and
150 (30.0 %) had open-type atrophic gastritis. Between the H. pylori-infected group and the H. pylori-eradicated group, there was no significant difference in the presence of atrophy
(P = 0.2179). The presence of atrophy was 96.3 % (closed type: 27.5 % + open type: 68.8 %)
in the H. pylori-infected group and was 91.9 % (closed type: 51.6 % + open type: 40.3 %) in the H. pylori-eradicated group.
Sixty-nine (13.7 %) patients had a history of treatment of gastric epithelial neoplasia.
Between the H. pylori-infected group and the H. pylori-eradicated group, history of treatment of gastric epithelial neoplasia was significantly
higher in the H. pylori-eradicated group (27.9 %) compared to the H. pylori-infected group (15.6 %) (P = 0.0158).
The prevalence of WOS in H. pylori-uninfected, H. pylori-infected, and H. pylori-eradicated groups
Prevalence of WOS was 0 % (0/206) in the H. pylori-uninfected group, 28.4 % (31/109) in the H. pylori-infected group, and 3.2 % (6/186) in the H. pylori-eradicated group ([Fig. 2]). A statistically significant difference in the prevalence of WOS was observed between
the H. pylori-uninfected and H. pylori-infected groups (P < 0.0001) between the H. pylori-infected and H. pylori-eradicated groups (P < 0.0001) and between the H. pylori-uninfected and H. pylori-eradicated groups (P = 0.0109). The prevalence of WOS-P and WOS-P + F was 22.9 % and 5.5 %, respectively,
in the H. pylori-infected group and 2.1 % and 1.1 %, respectively, in the H. pylori-eradicated group. There were no WOS-F cases.
Fig. 2 Prevalence of WOS in H. pylori-uninfected, H. pylori-infected, and H. pylori-eradicated groups.
Relationship between the presence of WOS and pH level of fasting gastric juice in
H. pylori-uninfected, H. pylori-infected, and H. pylori-eradicated groups
In the H. pylori-uninfected group, all cases were WOS-negative and the fasting gastric juice pH levels
showed strong acidity (mean pH: 1.1, 95 % CI: 1.0 – 1.1). In the H. pylori-infected group, mean pH level (95 % CI) of fasting gastric juice in the WOS-positive
cases was significantly higher than that of the in WOS-negative cases (mean pH: 7.4
[7.1 – 7.6] vs. 4.6 [3.9 – 5.2], P < 0.0001). Similar to the H. pylori-infected group, the mean pH level of fasting gastric juice in the H. pylori-eradicated group was also significantly higher in WOS-positive cases than WOS-negative
cases (mean pH: 7.2 [6.1 – 8.2] vs. 1.6 [1.4 – 1.8], P < 0.0001). We observed a few WOS-positive cases in the H. pylori-eradicated group, comprising 3.2 % (n = 6) of the total, and all of these fasting
gastric juices demonstrated a neutral pH level (mean pH: 7.2, range: 6 – 8) ([Fig. 3]).
Fig. 3 Relationship between the presence of WOS and fasting gastric juice pH level in H. pylori-uninfected, H. pylori-infected, and H. pylori-eradicated groups.
Relationship between the presence of WOS and pH level of fasting gastric juice before
and after PPI administration in 29 H. pylori-eradicated cases with gastric epithelial neoplasia
We investigated the relationship between the presence of WOS and fasting gastric juice
pH before and after PPI administration in 29 H. pylori-eradicated cases with gastric tumor to clarify that presence of WOS may not be a
highly sensitive endoscopic marker of IM in patients after H. pylori eradication, especially in cases with acid recovery. The mean fasting gastric juice
pH level in these 29 cases increased from 1.1 before PPI administration to 6.9 after.
WOS positivity increased from 0 % (0/29) before PPI administration to 45 % (13/29)
after ([Fig. 4]). Of the 13 cases in which WOS appeared after PPI administration, the prevalence
of WOS-P and WOS-P + F was 11/13 and 2/13, respectively ([Fig. 4] and [Fig. 5]). These findings suggested that the presence of WOS is not highly sensitive as an
endoscopic marker for IM in patients after H. pylori eradication, especially in patients with acid recovery.
Fig. 4 Relationship between the presence of WOS and fasting gastric juice pH before and
after PPI administration in 29 H. pylori-eradicated cases with gastric epithelial neoplasia.
Fig. 5 A representative H. pylori-eradicated case with WOS appearance: endoscopic findings before and after PPI administration.
WOS is not observed before PPI administration (a – d). However, WOS appears dramatically after 14 d PPI administration (a’ – d’). Fasting gastric juice pH changed from pH1 to pH7 following PPI administration.
Discussion
In this study, we clarified the clinical significance of WOS as an optical marker
for IM and its limitations in the changing gastric acid conditions in various H. pylori states. The presence of WOS in IM was thought to be closely associated with a hypochlorhydria
state in atrophic gastric mucosa caused by H. pylori infection. However, WOS may not be highly sensitive as an endoscopic marker for IM
in patients with acid recovery after H. pylori eradication.
The study results demonstrated that the presence of WOS in IM was closely associated
with a hypochlorhydria state caused by H. pylori infection. WOS, as observed by M-NBI, was most frequently found in the H. pylori-infected group (28.4 %), infrequently observed in H. pylori-eradicated cases (3.2 %), and not found at all in H. pylori-uninfected cases. In addition, there was a close relationship between the presence
of WOS and fasting gastric juice pH level. Among the 295 H. pylori-infected and H. pylori-eradicated cases, fasting gastric juice pH level was significantly higher in WOS-positive
cases (mean pH 7.4 and 7.2, respectively) than in WOS-negative cases (mean pH 4.6
and 1.6, respectively). Interestingly, All of WOS-positive cases in the H. pylori-eradicated group (3.2 % of the total) showed neutral fasting gastric juice pH levels
(mean pH 7.2, range 6 – 8), even after successful H. pylori eradication. From these findings, we concluded that a WOS-positive state was closely
associated with neutralization of intragastric pH in H. pylori-infected and H. pylori-eradicated cases.
As the degree of mucosal atrophy progresses, acidity in the stomach is neutralized
to a pH level of approximately 7.0 [19]. In contrast, H. pylori eradication therapy contributes to the normalization of gastric juice pH owing to
a recovery of acid secretion, not only in the intact stomach but also in the remaining
stomach after gastrectomy [20]
[21]. However, some patients with profound hypochlorhydria may not recover because of
progression to an irreversible stage [22]. It is well known that chronic gastric acid hypochlorhydria and achlorhydria are
important risk factors for the development of gastric cancer [23]
[24]. Therefore, endoscopic investigations that allow speculation of gastric acid conditions
are highly desirable. The mechanism of WOS expression suggests that presence of WOS
could be a marker representing neutralization of intragastric pH due to H. pylori infection. Accordingly, WOS might be an important marker for predicting high risk
for the development of gastric cancer.
Furthermore, the study results demonstrate the limitation of WOS as a marker for IM.
That is, WOS might not be highly sensitive as an endoscopic marker for IM in patients
after H. pylori eradication with acid recovery. Between the H. pylori-infected group and the H. pylori-eradicated group, there was no significant difference in the presence of atrophy
by Kimura-Takemoto classification system (P = 0.2179). The presence of atrophy was 96.3 % in the H. pylori-infected group and was 91.9 % in the H. pylori-eradicated group. On the other hand, treatment history of gastric neoplasia was significantly
higher in the H. pylori-eradicated group (27.9 %) compared to the H. pylori-infected group (15.6 %) (P = 0.0158). However, WOS was significantly lower (3.3 %) in the H. pylori-eradicated group compared to the H. pylori-infected group (28.4 %) in spite of there being no significant difference on the
presence of atrophy representative of a surrogate marker of IM. In addition, we investigated
the relationship between the presence of WOS and fasting gastric juice pH level before
and after PPI administration in 29 H. pylori-eradicated cases with gastric epithelial neoplasias. Although we could not observe
any WOS in these H. pylori-eradicated cases with acid recovery, WOS dramatically appeared in 45 % of these cases
following neutralization of intragastric pH with PPI administration. From these findings,
we concluded that the presence of WOS is not a consistent phenomenon. Therefore, in
our routine screening endoscopy, we should pay attention for the evaluation of IM
by WOS findings if a patient has a history of H. pylori eradication. If WOS is not detected, the existence of histological IM cannot be completely
excluded. If WOS is detected in patients even after H. pylori eradication, not only is the existence of histologic IM suggested, but also that
acid secretion has not yet recovered. In other words, such patients may have a high
risk for developing gastric cancer despite the eradication of H. pylori, since chronic gastric acid hypochlorhydria and achlorhydria are important risk factors
for the development of gastric cancer [23]
[24].
Kanemitsu et al., [7] recently demonstrated that the presence of WOS, as observed by M-NBI as a novel
endoscopic parameter for the histological diagnosis of IM, had a sensitivity and specificity
of 50 % and 100 %, respectively. However, in patients with H. pylori eradication, diagnosing IM using WOS findings alone may not be sufficiently sensitive.
Our results suggested that WOS was not highly sensitive despite histological IM because
WOS is affected by some of the intragastric environment. Ohtsu et al. [25] revealed that the oral ingestion of foods containing fats increased the prevalence
of WOS in gastric epithelial neoplasia. We believe that the intragastric pH environment
is also an important factor, with the activation of lipase, including pancreatic,
lingual and gastric types, increasing the prevalence of WOS. It may be necessary to
clarify how the combination of antacid and emulsified fat foods contribute to the
emergence of WOS.
The study had several limitations. First, it was retrospective and conducted at a
single center. Second, gastric acid conditions were only evaluated by fasting gastric
juice pH level. Although pH measurement of fasting gastric juice is one of reliable
information for gastric acid condition, it is not an ideal method for evaluation of
acid secretion itself. Third, we did not study the relationship of other optical markers
such as LBC with histological IM, regardless of H. pylori eradication status. Finally, the patients included in this study received endoscopy
for mainly screening purposes, so we could not examine IM histologically. A further
well-designed prospective study with histologic examination is expected to clarify
these issues.
In conclusion, the results of this study suggest that the presence of WOS in H. pylori-infected gastritis is closely associated with a neutralization of intragastric pH
due to H. pylori infection but is not a highly sensitive endoscopic marker of IM in patients after
H. pylori eradication.
