The corpus-predominant gastritis index may serve as an early marker of Helicobacter pylori-infected patients at risk of gastric cancer

Authors

  • Y.-C. Tsai,

    1. Institute of Clinical Medicine, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    2. Department of Internal Medicine, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    3. Department of Pathology, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for more papers by this author
  • W.-H. Hsiao,

    1. Institute of Clinical Medicine, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for more papers by this author
  • H.-B. Yang,

    1. Department of Internal Medicine, Tainan Hospital, Department of Health, Executive Yuan, Tainan, Taiwan
    2. Department of Pathology, Ton Yen General Hospital, Hsin-Chu, Taiwan
    Search for more papers by this author
  • H.-C. Cheng,

    1. Institute of Clinical Medicine, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    2. Department of Internal Medicine, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for more papers by this author
  • W.-L. Chang,

    1. Institute of Clinical Medicine, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    2. Department of Internal Medicine, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for more papers by this author
  • C.-C. Lu,

    1. Department of Pathology, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for more papers by this author
  • B.-S. Sheu

    Corresponding author
    1. Department of Internal Medicine, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    • Institute of Clinical Medicine, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for more papers by this author

Correspondence to:

Prof. B.-S. Sheu, Department of Internal Medicine, National Cheng Kung University Hospital, 138 Sheng Li Road, 70428 Tainan, Taiwan.

E-mail: sheubs@mail.ncku.edu.tw

Summary

Background

To eradicate Helicobacter pylori before the occurrence of precancerous changes is important to prevent gastric carcinogenesis.

Aim

To validate whether the corpus-predominant gastritis index (CGI) can serve as an early marker to identify the H. pylori-infected patients at risk of gastric carcinogenesis.

Methods

This study enrolled 188 subjects, including 43 noncardiac gastric cancer patients, 63 of their first-degree relatives and 82 sex- and age-matched duodenal ulcer patients as controls. All received endoscopy to provide topographic gastric specimens to test for H. pylori infection and its related histological features, translated into the operative link on gastritis assessment (OLGA), operative link on gastric intestinal metaplasia assessment (OLGIM) stages, and the presence of CGI. Spasmolytic polypeptide-expressing metaplasia (SPEM) was assessed by immunohistochemistry staining of trefoil factor 2.

Results

Gastric cancer patients had higher prevalence of CGI and OLGIM stage II–IV, but not OLGA stage II–IV, than the controls (= 0.001, OR = 3.4[95% CI: 1.4–8.1] for CGI; OR = 5.0[95% CI: 2.0–12.8] for OLGIM). In patients with the combined presence of CGI and OLGIM stage II–IV, the risk of gastric cancer increased to 9.8 (< 0.001). The first-degree relatives of the gastric cancer patients had a higher rate of the presence of CGI, but not OLGA or OLGIM stage II–IV than the duodenal ulcer controls (= 0.001). Of the first-degree relatives, the presence of CGI increased the risk of SPEM (= 0.003, OR = 5.5[95% CI: 1.8–17.0]).

Conclusion

The corpus-predominant gastritis index, which is highly correlated to SPEM, may serve as an early marker to identify the H. pylori-infected patients at a higher risk of gastric cancer.

Introduction

Helicobacter pylori infection is a WHO type I carcinogen for gastric cancer.[1, 2] Gastric cancer may consequently occur after chronic atrophic gastritis, and intestinal metaplasia (IM) in H. pylori-infected patients.[3-5] Meining et al. address the concept to take topographic gastric samplings and reveal the correlation between the corpus inflammation and the gastric cancer risk in H. pylori infection.[6] Further studies also support the high grade corpus gastritis can serve with equivalent role as atrophy during H. pylori-related gastric carcinogenesis.[7, 8] In contrast, H. pylori infection with antrum-predominant gastritis and high gastric acid secretion into a duodenal ulcer become protective factors to gastric carcinogenesis.[9] Therefore, validation of the H. pylori-infected patients at a high risk of gastric cancer should focus on the histological predominance topographically during chronic infection.

Based on the updated Sydney system to assess gastric specimens topographically, the operative link on gastritis assessment (OLGA) staging system has been validated in several studies.[10-13] However, due to high levels of inter-observer disagreement in assessing atrophy, the accuracy of the OLGA staging system is controversial.[10-13] The operative link on gastric intestinal metaplasia assessment (OLGIM) staging system, which uses intestinal metaplasia instead of atrophy, has been shown to improve accuracy.[14, 15] As H. pylori eradication should be started before IM to achieve a significant prevention of gastric cancer,[16] an early marker would thus be beneficial in identifying the high-risk H. pylori-infected patients. As ‘corpus-predominant gastritis’, based on acute inflammatory scores and chronic inflammatory scores in updated Sydney system, can be accessed in early stages of chronic H. pylori infection before the occurrence of precancerous changes, we thus devised a simple new corpus-predominant gastritis index (CGI) from the topographical gastric features of acute and chronic inflammatory scores assessed by the commonly used updated Sydney system. Based on novel design to include H. pylori-infected subjects with different risks of gastric cancer, we validated whether CGI could differentiate H. pylori-related histology among these subjects.

Because of familial clustering in gastric cancer[17, 18] and the higher occurrence of precancerous lesions in the relatives of H. pylori-related gastric cancer patients,[19, 20] it is reasonable to assume that the relatives of gastric cancer patients are at an intermediate to high risk of gastric carcinogenesis. Therefore, this study enrolled not only high-risk cancer patients and low-risk duodenal ulcer patients but also the first-degree relatives of the gastric cancer patients. We hypothesised that the rate of CGI would rise in parallel with the increase in cancer risk of the different H. pylori-infected groups.

Spasmolytic polypeptide-expressing metaplasia (SPEM) has been identified with trefoil factor 2 (TFF2)-expressing cells over the deep antral-like oxyntic glands of the gastric corpus.[21] It can be promoted by H. pylori-related chronic inflammation to induce loss of parietal cells and chief cell transdifferentiation with TFF2 expressions in the whole oxyntic gland.[22, 23] SPEM is now recognised as a precancerous lesion appearing even earlier than IM in gastric cancer.[24-28] As SPEM can be promoted by H. pylori-related chronic inflammation, we also determined whether the CGI correlated to the presence or advance of SPEM, especially in the first-degree relatives of the gastric cancer patients.

The study is highly original in illustrating that the CGI alone can serve as an early marker before IM to identify the patients at high-risk of gastric carcinogenesis, and especially in the relatives of gastric cancer patients. Moreover, the presence of CGI after H. pylori infection could identify the advancement of SPEM in the relatives of gastric cancer patients. The application of the CGI as a marker to identify patients at high-risk of H. pylori-related cancer, thereby indicating immediate H. pylori eradication, may be promising for the control of gastric cancer.

Materials and methods

Patients and study design

This study consecutively enrolled three groups of patients with H. pylori infection, including those with noncardiac adenocarcinoma, first-degree relatives of the gastric cancer patients, and those with benign duodenal ulcers as the controls. All of the patients gave written informed consent to receive H. pylori screening by 13C-urea breath test and gastric endoscopy to provide topographic gastric biopsy specimens for histological assessment. The informed consent forms and study design were reviewed by the Research Ethics Committee of the institute (certification code: HR-98-023). In this study, the presence of H. pylori infection was confirmed by positive histology and 13C-urea breath test. All of the participants were free from recent exposure to bismuth, antibiotics, probiotics and antisecretory agents within 2 weeks before the 13C-urea breath test.

We enrolled first-degree relatives of noncardiac gastric adenocarcinoma patients from the gastric cancer data bank of National Chen Kung University into a screening project for the H. pylori-infected relatives of the gastric cancer patients (NHRI-EX101-9908 BI) by 13C-urea breath test. For those with positive 13C-urea breath test, we invited then to receive gastric endoscopy to provide topographic gastric biopsy specimens. The duodenal ulcer patients were enrolled to serve as age- and gender-matched controls.

Endoscopy and topographic gastric histology grading

During endoscopy, each participant provided five gastric biopsy specimens, topographically from the antrum (×2), corpus (×2) and high corpus near to cardia (×1) to confirm the presence of H. pylori infection and the related gastric histological features.[29-33] The exact locations of topographic samplings were shown in Figure S1. Each specimens of the topographic site was reviewed in a blinded manner by the pathologist to assess the H. pylori-related histological features using the modified updated Sydney system.[34] For each topographic site, the gastric histological features included the acute inflammation score (range: 0–3), chronic inflammation score (range: 1–3), H. pylori density (range: 0–5), atrophic change (AT, absence: 0; presence: score 1–3) and intestinal metaplasia (IM, absence: 0; presence: score 1–3). Namely, for the antrum and lower corpus, either the acute or chronic inflammation score was an overall assessment to both bits of biopsy. In this study, the presence of IM was strictly defined as at least one gastric biopsy sample showing metaplastic cells (goblet cells) involving at least 5% of the upper third of the gastric mucosa.[31-33]

OLGA, OLGIM and CGI

Based on the topographical features of the histology reviewed by the updated Sydney system over the antrum and corpus, we translated each the histological findings for each patient into OLGA and OLGIM stages by methods described in previous studies.[11-13] This study then devised a new CGI from the combination of acute inflammation score and chronic inflammation score as assessed by the updated Sidney system. The presence of CGI was defined as the occurrence of either: 1) the combination of acute and chronic inflammation scores in the high corpus being equal to or larger than the antrum, but not equal to 1 in score; or 2) the combination of scores in the body being larger than the antrum. All the other conditions were defined as an absence of the CGI, including the combination of acute inflammation score and chronic inflammation score of the antrum and high corpus being equal to 1 in score.

TFF2 immunohistochemistry for SPEM grading

As SPEM is the TFF2-expressing metaplastic cell lineage transdifferentiated from chief cells after chronic inflammation,[21-24] we performed TFF2 immunohistochemistry to assess SPEM in the topographical tissues from the antrum, corpus and high corpus of the 63 first-degree relatives, and randomly selected 52 duodenal ulcer controls. The study was performed using mouse antihuman spasmolytic polypeptide (TFF2) antibody (GE16C; Leica Biosystem Newcastle Ltd, New Castle, UK) and a commercial kit (NovoLink Polymer Detection System; Leica Biosystem Newcastle Ltd). The protocol was modified from the instructions of the manufacturer to use deparaffinised sections. Antigen retrieval was done with EDTA at a pH of 9 (Novocastra Epitope Retrieval Solution; Leica Biosystem Newcastle Ltd). After antigen retrieval, peroxidase block was done with 3% hydrogen peroxide for 5 min, and protein block was done for 10 min with 0.4% casein in phosphate-buffered saline. The slides were then incubated with mouse monoclonal IgM antihuman TFF2 (1:50) for 1 h. After post primary block with 10% animal serum for 10 min, the slides were incubated with antimouse IgG (NovoLink Polymer) for 15 min, and then incubated in DAB chromogen for 10 min. These sections were then stained with 0.02% hematoxylin for the review by the same pathologist blinded to the patient group.

TFF2 staining was characterised in the neck cells of normal oxyntic glands by a ‘scattered’ pattern between parietal cells (Figure 1a). In the presence of chronic inflammation and mucosal atrophy, there was a loss of parietal cells, with TFF2-expressing cells distributed over both the middle and lower parts of the glands (Figure 1b), as shown by torturous deep glands in the corpus mucosa in an ‘antrum-like’ pattern (Figure 1c). Based on these TFF2-expressing features, this study defined the presence of SPEM as glands formed by TFF2-expressing cells detected in the specimens of either the corpus or high corpus. Advanced SPEM was defined as either glands in the corpus or high corpus mucosa with an ‘antrum-like’ pattern, or a diffuse TFF2-expression into the lower one-third of the mucosa with the loss of parietal cells.

Figure 1.

The schematic charts and immunohistochemical histograms of the gastric oxyntic glands in the Helicobacter pylori-infected corpus mucosa, with and without spasmolytic polypeptide-expressing metaplasia (SPEM). (a) Left: the schematic gastric oxyntic glands. Right: the gastric trefoil factor 2 (TFF2)-expressing in the neck cells of normal oxyntic glands had a ‘scattered’ pattern (yellow arrows) among the parietal cells (red arrows) in the middle third of mucosa layer (10× & 20×). (b) Left: the gastric oxyntic glands with some loss of parietal cells (red arrow) and chief cell transdifferentiation into TFF2-positive stained gland cell (yellow arrow). Right: the presence of SPEM defined by TFF2-expressing cells over the whole gland in the atrophic mucosa (10× & 20×). (c) Left: torturous gastric oxyntic glands with loss of parietal cells and TFF2-positive staining in whole gland cells indicating advanced SPEM. Right: immunohistochemistry revealed the corpus mucosa had become ‘antrum-like’ with glands composed of TFF2-expressing cells (yellow arrow) distributed in both the middle third to lower third layers of the corpus mucosa (10× & 20×).

Statistics

All analyses were performed using spss version 12.0 for Windows (SPSS Inc, Chicago, IL, USA). The Student's t-test was applied when appropriate to determine parametric differences such as age. The Mann–Whitney U-test was applied to compare the differences in pathological severity by the updated Sydney system. Pearson's χ2 test and Fisher's exact test were used to verify nonparametric proportions such as prevalence of different stages of OLGA, OLGIM, CGI and SPEM in the cases and controls, and to assess odds ratios and 95% confidence intervals (95% CI) between the two groups. All tests were two-tailed with s significance set at P < 0.05.

Results

Patient enrolled in the study

Forty-three patients with H. pylori-related noncardiac gastric adenocarcinoma were enrolled. In addition, 236 first-degree relatives of 151 gastric cancer index patients from the gastric cancer data bank of National Cheng Kung University were enrolled to receive H. pylori screening by UBT. Sixty-three of these 236 first-degree relatives had a positive [13]C-urea breath test, and then received gastric endoscopy. To serve as the age- and gender-matched controls, 130 duodenal ulcer patients (48 for gastric cancer patients, and 82 for the first-degree relatives) were enrolled.

Histological features by the updated Sydney system among the study groups

There were no significant differences in the demographic data between the gastric cancer index patients, first-degree relatives of the gastric cancer patients and their age- and gender-matched duodenal ulcer controls (P > 0.05) (Table 1). The duodenal ulcer patients had higher acute inflammation score and H. pylori density than those of the gastric cancer patients in the antrum (< 0.05), but not in the corpus or high corpus (P > 0.05), supporting the characteristic antrum-predominant gastritis in the duodenal ulcer patients (Table 1). There was no significant difference in the severity of gastric atrophy between the gastric cancer patients and controls (P > 0.05). In contrast, the gastric cancer patients had more severe IM (= 0.001) than the duodenal ulcer patients.

Table 1. Comparisons of the pathologic presentations in gastric cancer index patients, the first-degree gastric cancer relatives and their age and sex matched duodenal ulcer controls
 GCA vs. DUFirst-degree relatives vs. DU
Parameters Mean (25–75th quartile)GCA patients (= 43)DU patients (= 48)P valueRelatives (= 63)DU patients (= 82)P value
  1. GCA, gastric cancer index patient; DU, duodenal ulcer; AIS, acute inflammatory score; CIS, chronic inflammatory score; IM, intestinal metaplasia; s.d., standard deviation.

Age (year)65.1 (15.3)62.6 (6.8)0.33140.8 (10.8)42.6 (10.2)0.293
Gender (F: M)18: 2516: 320.51535: 2835: 470.135
Smoking (n)1080.784490.728
H. pylori (n)28340.82163821
AIS
Antrum0.9 (0–2)1.3 (0–2)0.0441.6 (1–2)1.6 (1–2)0.931
Corpus0.7 (0–2)0.5 (0–1.8)0.3230.8 (0–2)0.8 (0–2)0.880
High corpus0.7 (0–2)0.6 (0–1.8)0.5661.1 (0–2)0.6 (0–2)0.002
CIS
Antrum2.2 (1–3)2.4 (1.3–3)0.2662.8 (3–3)2.8 (3–3)0.926
Corpus2.1 (1–3)2.0 (1–3)0.3942.1 (1–3)2.2 (1–3)0.581
High corpus2.1 (1–3)1.9 (1–3)0.2702.2 (1–3)1.9 (1–3)0.028
H. pylori density
Antrum1.6 (0–3)2.6 (0–4)0.0223.1 (2–4)3.2 (2–4)0.772
Corpus1.7 (0–3)2.3 (0–4)0.1523.0 (2–4)3.0 (2–4)0.842
High corpus1.5 (0–3)1.8 (0–3)0.4522.6 (2–4)2.4 (2–3)0.322
Atrophy
Antrum1.2 (0–2)1.1 (0–2)0.5250.9 (0–1)0.8 (0–1)0.447
Corpus0.4 (0–1)0.2 (0–0)0.1140.1 (0–0)0.1 (0–0)0.967
High corpus0.2 (0–0)0.2 (0–0)0.5780.2 (0–0)0.1 (0–0)0.050
IM
IM score2.3 (0–3)0.8 (0–1)0.0010.4 (0–0)0.6 (0–0.3)0.298
Antrum1.3 (0–3)0.7 (0–1)0.0260.3 (0–0)0.5 (0–0)0.269
Corpus0.5 (0–0)0.2 (0–0)0.0680.1 (0–0)0.1 (0–0)0.583
High corpus0.5 (0–0)0 (0–0)0.0040 (0–0)0 (0–0)1.000

The first-degree relatives of the gastric cancer patients had higher acute inflammation score and chronic inflammation score in the high corpus than those of the matched duodenal ulcer controls (AIS, = 0.002; CIS, P = 0.028) (Table 1). However, there were no differences in the severity of gastric atrophy and intestinal metaplasia at the different topographical sites between the relatives of the gastric cancer patients and the controls.

Distributions of OLGA stage, OLGIM stage and CGI in the different groups

In Table 2, the gastric cancer patients had a higher proportion of OLGIM stage II–IV than the duodenal ulcer controls (= 0.005). However, there were similar proportions of OLGA stages between the gastric cancer patients and their duodenal ulcer controls (P > 0.05). We thus simplified the OLGIM stage into early modified OLGIM stage (including the initial OLGIM stages as 0 or I), and late modified OLGIM stage (initial OLGIM stages from II to V). The gastric cancer patients had a higher rate of late modified OLGIM stage than the controls (53.5% vs. 18.8%; = 0.001; OR 5.0; 95% CI: 2.0–12.8) (Table 2). The gastric cancer patients also had a higher rate of the presence of CGI than the controls (58.1% vs. 29.2%; = 0.006; OR 3.4; 95% CI: 1.4–8.1) (Table 2). Combining late modified OLGIM stage and CGI showed an increased risk of gastric cancer to 9.8 (< 0.001).

Table 2. OLGA and OLGIM stages and the corpus-predominant gastritis index (CGI) in the study groups
Parameters % (n)GCA vs. DUFirst-degree relatives vs. DU
GCA patients (= 43)DU patients (= 48)P value OR (95% CI)aRelatives (= 63)DU patients (= 82)P value OR (95% CI)a
  1. A, operative link on gastritis assessment; OLGIM, operative link on gastric intestinal metaplasia assessment; CGI, corpus predominant gastritis index; GCA, gastric cancer index patient; DU, duodenal ulcer patients.

  2. COC denotes combined simplified OLGIM stage and CGI. Presence of COC indicates the presence of CGI plus a late simplified OLGIM stage. Absence of COC indicates all conditions other than the presence of COC.

  3. a

    Odds ratio (95% confidence interval) assessed by two-tailed Fisher's exact test.

OLGA
023.3 (10)35.4 (17)0.46838.1 (24)42.7 (35)0.680
I39.5 (17)29.2 (14) 36.5 (23)40.2 (33) 
II20.9 (9)14.6 (7) 14.3 (9)9.8 (8) 
III11.6 (5)18.8 (9) 11.1 (7)7.3 (6) 
IV4.7 (2)2.1 (1) 0 (0)0 (0) 
OLGIM
044.2 (19)72.9 (35)0.00582.5 (52)75.6 (62)0.515
I2.3 (1)8.3 (4) 3.2 (2)7.3 (6) 
II14.0 (6)0 (0) 7.9 (5)6.1 (5) 
III30.2 (13)12.5 (6) 6.3 (4)8.5 (7) 
IV9.3 (4)6.3 (3) 0 (0)2.4 (2) 
Simplified OLGIM
Late53.5 (23)18.8 (9)0.00114.3 (9)16.5 (14)0.819
Early46.5 (20)81.3 (39)5.0 (2.0–12.8)85.7 (54)83.5 (68)0.8 (0.3–2.1)
CGI
Presence58.1 (25)29.2 (14)0.00655.6 (35)29.3 (24)0.002
Absence41.9 (18)70.8 (34)3.4(1.4–8.1)44.4 (28)70.7 (58)3.0 (1.5–6.0)
COC
Presence39.5 (17)6.3 (3)<0.00111.1 (7)3.7 (3)0.103
Absence60.5 (26)93.8 (459.8(2.6–36.7)88.9 (56)96.3 (79)3.3 (0.8–13.3)

The differences in the distribution of OLGA stage, OLGIM stage, and the presence of CGI between the first-degree relatives of the gastric cancer patients and their duodenal ulcer controls are also shown in Table 2. The presence of CGI was more prevalent in the first-degree relatives than in the controls (55.6% vs. 29.3%; = 0.002; OR 3.0; 95% CI: 1.5–6.0). In contrast, there were no differences in the distribution of OLGA and OLGIM stage between the first-degree relatives and their matched controls (> 0.05).

Presence of CGI correlated with SPEM in the first-degree relatives of gastric cancer patients

As the first-degree relatives of the gastric cancer patients had a higher rate of the presence of CGI than the controls, we further validated whether any specific demographic or histological feature was related to the presence of CGI in these subjects. There were no differences in the distribution of OLGA and OLGIM stage between the first-degree relatives with and without CGI (> 0.05) (Table 3). In contrast, the first-degree relatives with CGI had a higher prevalence of SPEM (80% vs. 42.3%; = 0.003; OR 5.5; 95% CI: 1.8–17.0) and advanced SPEM (31.4% vs. 7.4%; = 0.002; OR 5.7; 95% CI: 1.1–28.6) than those without CGI (Table 3). Moreover, the H. pylori density was higher in the first-degree relatives with CGI than in those without, especially over the corpus (P = 0.008) and high corpus (< 0.001) (Figure 2). The duodenal ulcer controls had a significantly lower rate of SPEM (19%, 10/52) than the first-degree relatives (< 0.001). However, only 1.9% (1/52) of the H. pylori-infected duodenal ulcer patients had advanced SPEM.

Table 3. Differences in the demographics, OLGA and OLGIM stages, and SPEM between the first-degree relatives with and without CGI
Parameters % (n)With CGI (n = 35)Without CGI (n = 28)P value OR (95% CI)a
  1. OLGA, operative link on gastritis assessment; OLGIM, operative link on gastric intestinal metaplasia assessment; s.d., standard deviation; CGI, corpus-predominant gastritis index; SPEM, spasmolytic polypeptide-expressing metaplasia.

  2. a

    Indicates odds ratio (95% confidence interval) assessed by two-tailed Fisher's exact test.

  3. † Two subjects and ‡ One study subject without CGI who only had available samples from corpus, but not the high corpus, and thus were attributed as missing data to assess the rate of SPEM and advanced SPEM, respectively.

Age (year), mean (s.d.)39.9 (9.9)41.8 (12.0)0.485
Gender (F: M)21: 1414: 140.456
OLGA stage (n)
0: I: II: III: IV12: 14: 6: 3: 012: 9: 3: 4: 00.689
OLGIM stage (n)
0: I: II: III: IV28: 0: 4: 3: 024: 2: 1: 1: 00.223
SPEM
Presence80 (28)42.3 (11)†0.003
Absence20 (7)57.7 (15)†5.5 (1.8–17.0)
Advanced SPEM
Presence31.4 (11)7.4 (2)‡0.028
Absence68.6 (24)92.6 (25)‡5.7 (1.1–28.6)
Figure 2.

The Helicobacter pylori density in the topographic gastric samples of the first-degree relatives of gastric cancer patients with and without corpus-predominant gastritis index (CGI). The H. pylori density was significantly higher in the first-degree relatives with CGI than in those without, especially over the corpus (3.5, 25%–75% quartile 2–5 vs. 2.3, 25%–75% quartile 1–4, P = 0.008) and high corpus (3.1, 25%–75% quartile 2–4 vs. 1.9, 25%–75% quartile 1–2, < 0.001).

Discussion

Among the many people infected with H. pylori, identifying those at high-risk of cancer so that early H. pylori eradication can be initiated requires reliable markers, especially before the formation of precancerous changes. This study demonstrates that the CGI histological index may be potentially useful in the identification of H. pylori-infected patients at high risk of gastric carcinogenesis. The CGI index was designed on the basis that corpus predominant gastritis carries a higher risk of gastric carcinogenesis during long-term H. pylori infection. Moreover, the CGI can be widely applied, because it is composed of the topographical gastritis features assessed by the commonly applied updated Sydney system. The CGI can be either combined with OLGIM to differentiate the high-risk patients, or used alone as an early marker before precancerous changes to identify the first-degree relatives of H. pylori-infected gastric cancer patients at high risk.

In our study, there were similar acute inflammatory score, chronic inflammatory score, and OLGA stages between the gastric cancer patients and their age- and sex-matched duodenal ulcer controls. This indicates that the duodenal ulcer patients, generally recognised as having the least risk of gastric carcinogenesis after H. pylori infection, still develop gastric inflammation and gastric atrophy in old age. Accordingly, it limits the accuracy of OLGA stage to differentiate the risk of gastric cancer in H. pylori-infected subjects, which is consistent with previous studies.[11-13]

The gastric cancer patients had significantly higher rates of the late modified OLGIM stage (= 0.001; OR 5.0; 95% CI: 2.0–12.8) and the presence of CGI (= 0.006; OR 3.4; 95% CI: 1.4–8.1) compared with the controls. This suggests that OLGIM stage can be an effective marker to differentiate the risk of gastric cancer in H. pylori-infected subjects, consistent with previous studies.[14, 15] Moreover, the results of this study support that the corpus distribution of gastritis has an impact on H. pylori-related carcinogenesis. Combining the presence of CGI with the OLGIM as stage II–IV, the odds ratio of gastric cancer risk increased up to 9.8 (< 0.001). Accordingly, there should be either a synergistic or sequential interaction between the corpus distribution of gastritis or intestinal metaplasia to promote gastric carcinogenesis.

The presence of CGI was more prevalent in the first-degree relatives than in the duodenal ulcer controls (= 0.002). In contrast, there were no differences in the distribution of OLGA stage and OLGIM stage between the first-degree relatives and their duodenal ulcer controls (> 0.05). This supports that CGI may serve as an early marker before the presence of gastric precancerous lesions to differentiate the H. pylori-infected subjects at high risk, and especially in the first-degree relatives. The CGI may thus offer the potential to screen out the subjects at high risk enabling them to receive H. pylori eradication as early as possible to prevent gastric carcinogenesis.

The first-degree relatives of the patients with gastric cancer in this study had a mean age of around 40 years, and a prevalence rate of CGI of around 55%. Whether age was a factor in determining the presence of CGI was investigated. Among the first-degree relatives, there was no difference in the mean age between those with and without the presence of CGI (> 0.05). In addition, among duodenal ulcer controls, the prevalence of CGI was similar at around 20–30% between those aged older and younger than 40 years (> 0.05). It would be interesting to investigate specific host factors that predispose to the dynamic shift to corpus predominant gastritis in the first-degree relatives after chronic H. pylori infection in future study.

The novel findings of this study support that the CGI may serve as an earlier marker than the presence of gastric precancerous lesions to differentiate the H. pylori-infected subjects at high risk, and in the first-degree relatives of patients with gastric cancer. As SPEM is a corpus specific precancerous finding, found even earlier than intestinal metaplasia,[25-29] the study included TFF2 immunohistochemistry to define SPEM in the topographic gastric specimens of the first-degree relatives with and without CGI. The results showed that the first-degree relatives with CGI had a higher prevalence of SPEM (= 0.003; OR 5.5) and advanced SPEM (= 0.002; OR 5.7) than those without CGI. Furthermore, the presence of CGI was correlated with higher H. pylori density in the corpus and high corpus (< 0.05). These data link a higher H. pylori load in the corpus with more severe corpus predominant gastritis, and therefore a higher prevalence of SPEM. As SPEM can dynamically progress under conditions of chronic inflammation, it is reasonable to assume that the more severe corpus inflammation led to the first-degree relatives with CGI having a nearly 5-fold increased risk to have an advanced pattern of SPEM.

Mening et al. has addressed a gastritis index including both corpus gastritis and the intestinal metaplasia to define the different degrees and distribution of H. pylori-infected gastric histology.[6] However, once intestinal metaplasia has happened, it shall be somewhat too late for H. pylori eradiation to counteract the gastric cancer risk totally. So the current index as CGI is designed to serve as an earlier marker without including the intestinal metaplasia, different to the initial index by Meining et al.[6]

The limitations to this study are the cross-sectional approach of the design. Due to a lack of longitudinal follow-up, this study cannot answer whether the presence of CGI or SPEM can be corrected after H. pylori eradication in specific groups. It is rather important to apply this potential useful earlier marker as CGI to validate the large scale dyspepsia patients with just gastritis to define whether CGI can indicate the ‘risk gastritis’ after H. pylori infection. Accordingly, a larger scale study applying CGI to validate the risk of gastric carcinogenesis in longitudinal surveillance for the dyspeptic patients would be of clinical interest.

In summary, the presence of CGI combined with OLGIM can be more specific to differentiate the patients at high risk of gastric cancer. The presence of CGI, correlated to the presence or advancement of SPEM, may serve as an earlier marker before precancerous changes to identify H. pylori-infected subjects at high risk of gastric cancer. Clinically, applying CGI to identify high-risk H. pylori-infected dyspepsia patients for early H. pylori eradication may be a promising approach to control the incidence of gastric cancer.

Authorship

Guarantor of the article: Dr Sheu BS.

Author contributions: Dr Sheu BS initiated the study with the grant application and coordinated the conduct of the whole study. Dr Tsai YC composed the original CGI and prepared the draft of the manuscript. Ms. Hsiao WS screened the H. pylori infection in the first-degree relatives and stained the TFF2 to define SPEM. Drs Chang WL and Dr Cheng HC conducted endoscopic biopsies. Dr Yang HB reviewed the TFF2 stains to define SPEM, and Dr Lu CC reviewed the pathology for composition of the topographical histology into OLGA and OLGIM stages. All authors approved the final version of the manuscript.

Acknowledgements

Declaration of personal interests: None.

Declaration of funding interests: This study was supported by grants (98-2628-B-006-013-MY3 & 101-2314-B-006-016-MY3) from the National Scientific Council, NHRI-EX99-9908BI from the National Health Research Institute, and DOH100-TD-C-111-003 from the Department of Health, Taiwan.

Ancillary