1. Introduction
H.
pylori eradication therapy may have the potential to regress atrophic and metaplastic changes and subsequently prevent the development of gastric cancer. In fact, previous studies have suggested that
H.
pylori eradication reduces metachronous recurrence of gastric cancer after endoscopic surgery [
1,
2,
3,
4,
5]. It has been suggested that
H.
pylori eradication partly improves atrophy and metaplasia as well [
6], although there are numerous reports which failed to demonstrate such histological reversion, and indeed the improvement of metaplasia after eradication has not been confirmed by meta-analyses.
Previous studies have reported that oxyntic atrophy seems to be a pathogenic factor in gastric carcinogenesis [
7,
8,
9]. Reduced gastric acidity in the atrophic stomach leads to hypergastrinemia, which stimulates the function and proliferation of enterochromaffin-like cells located in the oxyntic mucosa, and potentially promotes gastric carcinogenesis in the oxyntic mucosa [
10]. Recently, we and others have reported that long-term use of proton pump inhibitors (PPIs) may be a risk factor for gastric cancer after
H. pylori eradication [
11,
12]. It is well known that acid suppression by PPIs and other drugs causes hypergastrinemia via a negative feedback mechanism, which may be associated with the subsequent risk of gastric cancer [
10]. In contrast, the role of metaplasia in gastric carcinogenesis has been questioned in recent years, and metaplasia may just be a marker of long-term atrophic gastritis [
8,
9,
13,
14].
In this report, we conducted a retrospective cohort study which collected clinical and histological data from H. pylori-eradicated patients. We quantitatively evaluated the reversibility of histological findings of intestinal metaplasia, atrophy, and inflammation using the updated Sydney system, and investigated the associations with several clinical and molecular parameters, focusing especially on the use of acid-suppressing drugs.
3. Discussion
We showed that
H.
pylori eradication improves histological atrophy and inflammation but does not accelerate the recovery of intestinal metaplasia in our 5.4-year observation period. In particular, PPI or H2RA use may inhibit the improvement of gastric atrophy compared with non-acid suppressant users. Therefore, gastric atrophy may be regulated by molecularly distinct mechanisms to that of intestinal metaplasia (i.e., aberrant expression of CDX proteins). In other words, improvement of gastric atrophy after
H. pylori eradication may be independent of changes in intestinal metaplasia [
15,
16,
17,
18,
19,
20], which seems to be a more sustained, stable reprogramming event of the gastric glands. In our study, the expression of CDX1 was more closely associated with changes in the intestinal metaplasia score than that of CDX2. In mice, ectopic expression of CDX1 and CDX2 in the stomach induces intestinal metaplasia, but expression of either CDX genes alone does not induce expression of the other gene [
21]. Therefore, CDX1 and CDX2 may be independently regulated. Our findings suggest that changes in CDX1 expression might precede changes in CDX2 expression.
The finding that eradication resulted in improvement of atrophy in the corpus is consistent with a previous meta-analysis of short-term observational studies [
22] and a long-term observational study in Japan [
23] but inconsistent with a randomized controlled trial in Korea [
15] and a long-term observational study [
24]. By contrast, the lack of improvement in intestinal metaplasia after eradication is inconsistent with some previous studies [
15,
22]. This may be explained by the different scoring systems used for evaluation of atrophy and metaplasia. In addition, baseline histological severities of atrophy and intestinal metaplasia were different. There are also differences in regions and countries, races and, presumably,
H. pylori strains among these studies, all of which may affect outcomes.
Previously, Annibale et al. reported that
H. pylori eradication improved gastric atrophy only in a subset of patients whose serum gastrin level became normal during the observation period, while the remaining patients showed sustained gastric atrophy with elevated serum gastrin [
25]. A reasonable interpretation of these results would be that restoration of normal acid secretion in improved patients suppresses the elevation of serum gastrin in a negative feedback manner. Nonetheless, it is possible that an elevated level of serum gastrin, perhaps due to the intake of acid-suppressing drugs, delays the recovery of gastric atrophy. In mice, hypergastrinemia induces gastric atrophy, metaplasia and, eventually, cancer [
26], and these changes are accelerated by the administration of PPIs [
27]. Long-term PPI use also promotes gastric atrophy and cancer development in
Helicobacter-infected Mongolian gerbils [
28]. In humans, long-term use of PPIs has been reported to worsen corpus atrophic gastritis in the patients infected with
H. pylori [
29,
30]. In any case, prior epidemiological retrospective cohort studies [
11,
12] and our current findings suggest that PPI use may prolong gastric atrophy after eradication, which may increase the risk for gastric cancer in post-eradicated patients [
1].
Although
H. pylori infection appears to be inversely associated with gastroesophageal reflux disease due to low acid secretion from the atrophic stomach, certain populations of
H. pylori-infected patients continue to suffer from acid reflux symptoms. In our cohort, the gastrointestinal symptom rating scale (GSRS) scores of upper abdominal pain, heartburn, abdominal distention, and bloating were temporarily decreased (improved) after
H. pylori eradication (
Tables S4 and S5), while these scores eventually returned to their original levels after the mean 5-year observation period. Thus, PPIs or H2RAs are often prescribed for patients with chronic acid reflux symptoms after eradication; however, our data suggest that the use of H2RAs or PPIs may be more cautiously considered given the potential effects on gastric atrophy and subsequent cancer risk.
The strengths of this study are the long-term, comprehensive assessment of clinical, endoscopic, histological, and molecular findings. Our analysis is quantitative and has extensively investigated the long-term effects of PPI and H2RA use following H. pylori eradication. However, this study also had limitations. Firstly, although we conducted long-term follow-up, the analyses were not prospective. The number of biopsy specimens and the observation period were limited, and some patients were lost to follow-up; secondly, our immunohistochemical stain data were limited to certain patients, associated with potential selection bias; thirdly, our sample size might be too small to reliably evaluate further associations between duration of drug use and histological score changes; fourthly, NSAIDs were more commonly taken in PPI and H2RA users than in non-drug users, and it might act as a confounder and affect the outcome; finally, our cohort lacked consideration of confounding factors such as smoking status, alcohol intake, and diet.
In conclusion, gastric atrophy but not intestinal metaplasia may be improved in patients who underwent successful H. pylori eradication, especially in patients with non-acid-suppressing drug use. Long-term acid suppression therapy might be associated with sustained gastric atrophy following eradication.
4. Methods
4.1. Study Design and Setting
We collected consecutive data on patients who underwent
H.
pylori eradication therapy and gastric cancer surveillance at Tokyo University Hospital from 1996 to 2015 (
Figure 1A, study flow chart). We performed retrospective cohort analyses of these data. This study was approved by the Institutional Review Board of the University of Tokyo (no. 2058-2, 17 November 2017) and was performed in accordance with an assurance filed with and approved by the Japanese Ministry of Health, Labour, and Welfare.
4.2. Participants
Patients who were positive for H. pylori infection and atrophic gastritis by endoscopy, and had undergone successful eradication therapy, were eligible for inclusion in this study. H. pylori infection was diagnosed by urea breath test, culture, rapid urease test, or serum antibody test. An exclusion criterion was a non-eradication assessment by urea breath test.
4.3. Eradication Treatment and Assessment
As the first-line treatment, amoxicillin (750 mg), clarithromycin (500 mg), and the PPI lansoprazole (30 mg) were administered twice daily for 7 days. As the second-line treatment, metronidazole (250 mg), amoxicillin (750 mg), and the PPI lansoprazole (30 mg) were given twice daily for 7 days. The H. pylori eradication status was determined based on the urea breath test.
After eradication, all patients underwent upper gastrointestinal endoscopy evaluation every 1 year up to 2017. Data were censored at the time of the last endoscopic examination or the last hospital visit for patients who were lost to follow-up or who withdrew from the study (
Figure 1B, study time course).
4.4. Upper Gastrointestinal Endoscopy and Histological Examination
All upper gastrointestinal endoscopies were performed using an electronic video endoscope (Olympus Medical System, Tokyo, Japan).
An endoscopic biopsy was performed to evaluate intestinal metaplasia, atrophy, neutrophil and mononuclear cell infiltration, and
H.
pylori scores according to the updated Sydney system [
31]. These parameters were scored as follows: normal, 0 points; mild, 1 point; moderate, 2 points; and marked, 3 points. Biopsy specimens were obtained from the antrum and the middle corpus of the greater curvature based on our previous works [
1,
12,
32]. Incomplete metaplasia was also evaluated by hematoxylin and eosin staining. Histological findings were assessed and confirmed by two experienced pathologists without any disagreement.
4.5. Outcomes and Variables
The primary outcome was change in the intestinal metaplasia, atrophy, neutrophil and mononuclear cell infiltration, and H. pylori scores in patients with PPI, H2 receptor agonist (H2RA), and non-acid suppressant use. The secondary outcome was change in these scores in patients according to the duration of acid suppressing therapy. We also evaluated associations in changes of histology and CDX1 and CDX2 expression between the first and final biopsies in the cases whose intestinal metaplasia or atrophy score changed during study periods. For intestinal metaplasia and atrophy, improvement and exacerbation were defined as a decrease and increase, respectively, in the updated Sydney system score between the first and final biopsies.
We evaluated age, sex, medication use (PPIs, H2RAs, and non-steroidal anti-inflammatory drugs (NSAIDs)), and acid-reflux symptoms. Age was categorized into five groups. Use of PPIs, H2RAs, and NSAIDs was defined as regular oral administration for at least 30 days during the follow-up period. Based on an appropriate receiver operating characteristic curve model for predicting atrophy execration, duration of PPIs and H2RAs was categorized as follows: PPI short-term use, <90 days; PPI long-term use, ≥90 days; H2RA short-term use, <485 days; H2RA long-term use, ≥485 days. Use of NSAIDs included low-dose aspirin use. Acid-reflux symptoms were evaluated using the upper gastrointestinal related gastrointestinal symptom rating scale (GSRS): upper abdominal pain, heartburn, acid reflux, abdominal distention, hunger pain, rumbling, bloating, and burning, administered as a seven-point Likert scale self-reported questionnaire [
33].
4.6. Immunohistochemical Stain Analyses of CDX1 and CDX2 Expression
Sections (3 mm thick) were deparaffinized, rehydrated in phosphate-buffered saline (PBS), placed in 10 mM citrate buffer (pH 6.0), and heated to 120 °C for 5 min to recover antigenicity. Sections were preincubated with blocking buffer (3% hydrogen peroxide) for 5 min at room temperature. The primary anti-CDX1 polyclonal antibody (1:100; BioGenex, Fremont, CA, USA) and anti-CDX2 monoclonal antibody (1:100; Abcam, Cambridge, UK) were diluted in PBS and incubated with the sections overnight at 4 °C. Sections were then washed in PBS and incubated with Histofine Simple Stain Max-Po (Multi) (Nichirei Biosciences Inc., Tokyo, Japan). After development with 3,3,9-diaminobenzidine tetrahydrochloride (DAB Substrate Kit, Nichirei Biosciences), sections were counterstained with hematoxylin and viewed under a light microscope. CDX1 and CDX2 expression were assessed as proportions of CDX1- and CDX2-expressing gastric glands by one researcher (RN) (
Figure 2).
4.7. Statistical Analyses
We calculated the coefficients and 95% confidence intervals (CIs) of annual changes in the atrophy, intestinal metaplasia, mononuclear cell and neutrophil infiltration, and H. pylori scores in biopsies of the antrum and middle corpus of the greater curvature and compared these scores among PPI, H2Rs, and non-acid suppressant users based on a generalized linear mixed model. We also compared the coefficients and 95% CIs of the annual changes of atrophy, intestinal metaplasia, mononuclear cell and neutrophil infiltration, and H. pylori scores between short- and long-term users of PPIs and H2RAs using a generalized linear mixed model. In addition, we evaluated changes between the initial and final CDX1 and CDX2 expression levels using paired t-tests in each metaplasia and atrophy change subgroup. Furthermore, we evaluated changes between the initial and final incomplete metaplasia rates using chi-squared tests in each metaplasia subgroup. We evaluated changes in the GSRS before and 3 months after eradication and changes in the GSRS during all follow-up periods using paired t-tests. Statistical analyses were performed using SAS software v. 9.4 (SAS Institute, Cary, NC, USA). A p-value < 0.05 was considered to indicate statistical significance.