Tel.: +81-734-47-1335, Fax: +81-734-45-3616
Development of gastric cancer in nonatrophic stomach with highly active inflammation identified by serum levels of pepsinogen and Helicobacter pylori antibody together with endoscopic rugal hyperplastic gastritis
Article first published online: 28 MAR 2012
Copyright © 2012 UICC
International Journal of Cancer
Volume 131, Issue 11, pages 2632–2642, 1 December 2012
How to Cite
Watanabe, M., Kato, J., Inoue, I., Yoshimura, N., Yoshida, T., Mukoubayashi, C., Deguchi, H., Enomoto, S., Ueda, K., Maekita, T., Iguchi, M., Tamai, H., Utsunomiya, H., Yamamichi, N., Fujishiro, M., Iwane, M., Tekeshita, T., Mohara, O., Ushijima, T. and Ichinose, M. (2012), Development of gastric cancer in nonatrophic stomach with highly active inflammation identified by serum levels of pepsinogen and Helicobacter pylori antibody together with endoscopic rugal hyperplastic gastritis. Int. J. Cancer, 131: 2632–2642. doi: 10.1002/ijc.27514
- Issue published online: 25 SEP 2012
- Article first published online: 28 MAR 2012
- Accepted manuscript online: 2 MAR 2012 02:52AM EST
- Manuscript Accepted: 14 FEB 2012
- Manuscript Received: 8 OCT 2011
- Ministry of Health, Labor and Welfare of Japan
- Helicobacter pylori;
- diffuse-type gastric cancer;
- rugal hyperplastic gastritis
This study aimed to elucidate groups at high risk of developing cancer among patients with serologically identified Helicobacter pylori infection and nonatrophic stomach. Annual endoscopy was performed for a mean of 5.4 years in 496 asymptomatic middle-aged men who were H. pylori antibody-positive and pepsinogen (PG) test-negative. Subjects were stratified according to the activity of H. pylori-associated gastritis measured by serum levels of PG and H. pylori antibody, and/or by endoscopic findings of rugal hyperplastic gastritis (RHG), and cancer development was investigated. During the study period, seven cases of cancer developed in the cohort (incidence rate, 261/100,000 person-years), with 85.7% developing in the group showing a PGI/II ratio ≤3.0, reflecting active inflammation-based high PGII levels. Cancer incidence was significantly higher in this group (750/100,000 person-years) than in groups with less active gastritis. Furthermore, cancer incidence for this group was significantly higher in the subgroup with high H. pylori antibody titers than in the low-titer subgroup. Meanwhile, endoscopic findings revealed that 11.7% of subjects showed RHG reflecting localized highly active inflammation, and cancer risk was significantly higher in patients with RHG than in patients without. Combining the two serum tests and endoscopic examination for RHG allowed identification of subjects with more active gastritis and higher cancer risk. No cancer development was observed in these high-risk subjects after H. pylori eradication. Subjects with highly active gastritis identified by the two serological tests and endoscopic RHG constitute a group at high risk of cancer development with H. pylori-infected nonatrophic stomach.
Gastric cancer is one of the leading causes of cancer-related deaths in Japan, with 50,017 deaths attributed to this cancer in 2009.1 Helicobacter pylori infection is now widely accepted as playing a major role in the development of gastric cancer in areas showing high-risk for this cancer, including Japan.2–5 H. pylori triggers chronic inflammation in the gastric mucosa, leading to a series of molecular and morphological events known as the atrophy-metaplasia-dysplasia-cancer sequence.3–5 This carcinogenic sequence is considered to represent a major route of stomach carcinogenesis, and previous studies, including our own, have identified positive correlations between extent of chronic atrophic gastritis (CAG) and risk of gastric cancer.6–13 Our longitudinal cohort study clearly indicated that progression of H. pylori-associated chronic gastritis increases cancer risk in a stepwise manner, and that patients with metaplastic gastritis, an end result of chronic H. pylori infection, show the highest annual cancer incidence rate of around 1% among asymptomatic middle-aged Japanese men.11 Other studies have indicated a different subroute, in which incomplete and unstable CAG is directly associated with cancer development.14 In fact, clinicopathological studies have indicated that 20–30% of all stomach cancers in Western countries develop from nonatrophic stomach,5, 15 and seroepidemiological studies using serum pepsinogen (PG) in Japan have revealed that the background stomach does not appear to have extensive CAG in 20–40% of cancer cases.12, 16 The bacterial load in H. pylori-associated chronic gastritis is generally reported to be larger in stomachs with mild to moderate atrophy than in stomachs with extensive atrophy, and the bacterium is spontaneously eradicated with the establishment of metaplastic gastritis.17–19 Thus, if H. pylori infection itself plays a pivotal role in the development of gastric cancer, a high-risk group should be identifiable among subjects without extensive CAG. Indeed, our previous longitudinal cohort study revealed a small group of subjects at high-risk of cancer among CAG-negative subjects identified by serum PG levels.12 This group of subjects developed cancer at an annual incidence rate of more than 0.2%, comparable to that in CAG-positive subjects identified by PG levels. Characteristics of this high-risk group with nonatrophic stomach are low serum PGI/II ratio (due to a high PGII level), reflecting a process of highly active gastric inflammation, and more importantly, frequent development of diffuse-type cancer as the histopathological-type, representing cancer with a higher malignant potential than cancer developing from the aforementioned main route of stomach carcinogenesis, i.e., intestinal-type cancer.12 The incidence rate of diffuse-type cancer in this high-risk group is 120/100,000 person-years, considerably higher than the reported rate of <5/100,000 person-years in Western countries,20 or even the rate of around 40/100,000 person-years among middle-aged Japanese men.11, 21
Highly active H. pylori-induced inflammation in the nonatrophic stomach body has also been reported to induce enlargement of the rugal folds,22–25 appearing endoscopically as rugal hyperplastic gastritis (RHG).26 This specific type of gastritis is characterized by enhancements of inflammatory cell infiltration and proinflammatory cytokine production, including interleukin (IL)-1β, which leads to the proliferation of mucosal epithelia and foveolar hyperplasia, and finally to the establishment of enlarged folds or rugal hyperplasia.23, 27 Subjects with RHG are suggested to be at a higher risk for cancer, particularly diffuse-type cancer, than subjects with CAG.28–31 However, no long-term follow-up data analyzing cancer development from RHG have been reported. Furthermore, the relationship between this serologically identified high-risk group without CAG and the group with RHG is unclear. The present longitudinal cohort study was initiated to clarify these issues and to investigate cancer development from highly active inflammation in the nonatrophic stomach as identified by serum tests and/or endoscopy.
Subjects and Methods
Between January 1999 and December 2000, a total of 36,762 middle-aged factory workers (34,610 men, 2,152 women) between 40- and 60-years old participated in an annual multiphasic health-screening program in a factory in Wakayama City, Wakayama Prefecture, Japan. This type of screening program is generally performed by various workplaces throughout Japan to detect incident diseases in the early stages, and includes an interview to ascertain general state of health, physical examination, chest radiography, electrocardiography, blood laboratory tests, urinalysis, fecal occult blood testing and gastric cancer screening by either barium X-ray or esophagogastroduodenoscopy (EGD) as selected by the individual. All the participating subjects were essentially asymptomatic and individuals presenting with symptoms requiring prompt medical care had been excluded from the screening program. The subjects could thus be considered to represent the healthy Japanese population. Of these subjects, those who had undergone barium X-ray as a part of an annual health check-up in the same workplace were investigated in the previous cohort studies for gastric cancer risk with special reference to CAG as diagnosed by serum PG levels.11–13 In the present study, 3,334 workers who selected EGD for cancer screening were candidates for inclusion, and H. pylori-infected subjects with nonatrophic stomach were selected for the study based on the two serological tests, PG and H. pylori antibody titer, as described in the following section. Figure 1 shows the schema for the selection of study subjects.
Follow-up by endoscopy
As described above, the 3,334 subjects underwent annual check-up by panendoscopy (Type XQ200; Olympus, Tokyo, Japan). The day of gastric cancer identification was defined as the day of the health check-up when cancer was detected. The observation period was calculated for each subject from the time of the baseline survey to the diagnosis of gastric cancer. The location of cancer detected in the stomach was classified as cardia or noncardia based on clinical or histopathological records. Resected specimens of gastric cancer obtained by surgery were histopathologically assessed and classified as intestinal or diffuse-type, according to the classification described by Lauren.32 RHG was diagnosed based on endoscopic findings of severely enlarged tortuous folds in the gastric body that did not disappear despite adequate insufflation, according to the classification of the Sydney system.26
Aliquots of separated sera from fasting blood samples collected as routine laboratory tests for the aforementioned general health check-up were stored at -20°C and used for the measurement of serum levels of PG and H. pylori antibody. Serum PGI and PGII levels were measured using a modification (RIAbeads Kit; Dainabott, Tokyo, Japan) of our previously reported radioimmunoassay.33 Subjects with extensive CAG were diagnosed based on PG test-positive criteria of PGI ≤50 ng ml−1 and PGI/II ratio ≤3.0.16 These criteria offer 69% sensitivity and 80% specificity for the diagnosis of extensive CAG.16 Serum H. pylori antibody levels were measured using enzyme-linked immunosorbent assay (ELISA) (MBL, Nagoya, Japan).34 Subjects with antibody titers >50 U ml−1 were classified as H. pylori-infected. Sensitivity and specificity of the ELISA used in the present study were 93.5 and 92.5%, respectively.34 Based on the difference in cancer risk according to the antibody titer level as described in our previous study,13 H. pylori-infected subjects were further divided into two groups depending on the antibody titer: a high-titer group comprising subjects with titers >500 U ml−1; and a low-titer group.
Subjects with a previous history of gastric cancer or adenoma, surgical resection of the stomach, H. pylori eradication or renal failure, and those who had been prescribed medications that might affect gastrointestinal function, such as proton pump inhibitors, adrenocortical steroids or nonsteroidal antiinflammatory drugs, were excluded from the study. As a result, among the 3,334 subjects who selected EGD for cancer screening, a total of 2,236 subjects (2,212 men, 24 women) opted to pay the extra charge for measurement of serum levels of PG and H. pylori antibody, in addition to the ordinary screening program (Fig. 1). Female subjects were subsequently excluded from analysis in this study because of their small number.
The remaining 2,212 male subjects were classified into one of the following four groups based on the results of the two serum tests for PG and H. pylori antibody: Group A, H. pylori-negative and CAG-negative; Group B, H. pylori-positive and CAG-negative; Group C, H. pylori-positive and CAG-positive; and Group D, H. pylori-negative and CAG-positive. This classification reflects each stage of serial changes in the stomach mucosa induced by the progression of H. pylori-associated chronic gastritis, with Group A including subjects with H. pylori infection-free healthy stomach, Group B including H. pylori-infected subjects without CAG, Group C including subjects with H. pylori-induced CAG, and Group D including subjects with extensive CAG together with widespread intestinal metaplasia.11
In the present cohort, 937 subjects were classified into Group B, and were advised to attend the clinic at their workplace. At the clinic, they were told about the possible roles of H. pylori-infection in peptic ulcer disease or gastric cancer, and also about the possible effects of H. pylori eradication, including costs and potential adverse effects. Thereafter, during the period of enrollment 399 subjects underwent H. pylori eradication with dual therapy consisting of the proton pump inhibitor omeprazole at 20 mg twice a day and amoxicillin at 750 mg or 500 mg twice a day for 2 weeks or with triple therapy consisting of omeprazole at 20 mg twice a day, amoxicillin at 750 mg twice a day and clarithromycin at 200 mg twice a day for 1 week. These patients were followed separately for analysis of the effects of eradication. In study subjects, H. pylori status including the completeness of eradication was assessed by serum H. pylori antibody level at all annual health check-ups during the study period.
The remaining 538 subjects did not undergo eradication therapy, and the prevalence of RHG on endoscopy at the time of enrollment was analyzed. The diagnosis of RHG was made by each endoscopist based on endoscopic findings according to the Sydney system.26 In addition, the recorded endoscopic images of each subject were examined, and the diagnosis of RHG was confirmed retrospectively by two other endoscopists, who were blinded to information about study subjects, including cancer development. After excluding those with poor-quality recorded endoscopic images or with an inappropriately distended stomach on examination, interpretation of the endoscopic images was possible in the remaining 496 Group B subjects. As described previously, CAG-free subjects (Groups A and B) serologically identified by PG test-negative criteria consist of three distinct groups from the perspectives of gastritis activity and cancer risk: Groups α, β and γ.12 The 496 Group B subjects were further classified into the following three groups: Group Bα (n = 111), with PGI ≤50 ng ml−1 and PGI/II ratio >3.0; Group Bβ (n = 235), with PGI >50 ng ml−1 and PGI/II ratio >3.0; and Group Bγ (n = 150), with PGI >50 ng ml−1 and PGI/II ratio ≤3.0. These subjects were followed annually to detect endoscopically incident cancer, as described above. The ethics committee of Wakayama Medical University approved the study protocols, and informed consent was obtained from all subjects prior to participation.
Data were analyzed using SPSS 11.0 software (SPSS, Chicago, IL) and STATA software (STATA, College Station, TX). Differences in continuous values were tested for significance using the t test for comparisons between two groups. The χ2 test was used to compare categorical variables. Long-term effects of endoscopic findings of RHG and/or serum levels of PG and H. pylori antibody on gastric cancer development were analyzed using the Kaplan–Meier method and statistical differences between curves were tested using the log-rank test. Hazard ratios (HRs) were calculated using Cox proportional hazards modeling. For all comparisons, probability values <5% (p < 0.05) were considered statistically significant.
A total of 496 H. pylori-infected subjects without CAG (i.e., Group B subjects) were followed for a mean [standard deviation (SD)] period of 5.4 (4.0) years (Table 1). During the study period, eight cases of gastric cancer developed in study subjects. Among these was one case of cancer that developed within the first year of follow-up. That case was excluded from the study, since the cancer might have been present from the start of the study. The overall cancer incidence rate was thus 261/100,000 person-years. All seven cancers detected in the present study were localized in the middle- or upper-third of the stomach, and none involved the gastric cardia. Histopathologically, all were early-stage cancers confined to the mucosa or submucosa, and six cases (85.7%) were diffuse-type (three mucosal cancers, three submucosal cancers), while the remaining case (14.3%) was intestinal-type with submucosal invasion. All cancer cases were therefore treated by surgical resection and the patients all remain alive under follow-up at our clinic.
Table 1 shows the details of cancer development among the 496 subjects in the serologically defined Groups Bα, Bβ and Bγ. Mean age and percentages of smokers and alcohol drinkers did not differ significantly among these three groups. Mean serum levels of PG II and H. pylori antibody showed significant stepwise increases from Bα to Bγ, whereas mean PGI level increased significantly from Bα to Bβ, then remained fairly constant in Bγ. As a result, PGI/II ratio showed a marked and significant reduction from Bβ to Bγ. Among the 7 cases of cancer examined in the present study, 6 (85.7%) were from Group Bγ, with 83.3% (5/6) showing diffuse-type histopathology. These developed steadily throughout the study period after 1.5 years of follow-up (Fig. 2a). The remaining one case was from Group Bβ, with no cancer development observed in Group Bα. Cancer incidence rates were thus 0, 78/100,000 person-years and 750/100,000 person-years in Groups Bα, Bβ and Bγ, respectively, with a significant difference apparent between Groups Bβ and Bγ [HR, 10.00; 95% confidence interval (CI), 1.19–84.05]. High H. pylori antibody titer (>500 U ml−1) was seen in 32.1% (159/496) of study subjects, and the percentage of the high-titer group increased in a stepwise manner from Group Bα (13.5%) to Group Bβ (33.6%) to Group Bγ (43.3%). No significant differences in mean age or alcohol-drinking status were seen between the high- and low-titer groups, but presence of a smoking habit was significantly less frequent in the high-titer group, reflecting mainly smoking status in Group Bβ rather than that in Group Bγ. Cancer incidence was significantly higher in the high-titer group (577/100,000 person-years) than in the low-titer group (110/100,000 person-years) (HR, 6.51; 95%CI, 1.24–34.17), and 71.4% (5/7) of cancers developed in the high-titer group. All cancers that developed in the high-titer group belonged to Group Bγ, and 80% (4/5) showed diffuse-type histopathology. A marked increase in cancer development was apparent in Group Bγ with an increase in antibody titer (p < 0.05, Fig. 2b); cancer incidence rates in Group Bγ with high and low titers were 1,524/100,000 person-years and 212/100,000 person-years, respectively (Table 1). Meanwhile, endoscopic findings revealed RHG in 11.7% (58/496) of study subjects, and the percentage of RHG increased in a stepwise manner with the transition of serologically classified groups from Bα to Bγ, with 0% (0/111) in Group Bα, 7.7% (18/235) in Group Bβ and 26.7% (40/150) in Group Bγ (Table 1). In both Groups Bβ and Bγ, RHG was more prevalent in that subgroup of individuals with a high titer of H. pylori antibody. The percentage of RHG was thus highest in Group Bγ with high titer (35.4%).
As shown in Table 2, mean age, alcohol drinking and smoking habits did not differ significantly between subjects with and without RHG. Serum levels of PGI, PGII and H. pylori antibody were significantly higher, and the PGI/II ratio was significantly lower in subjects with RHG. During the study period, continuous development of cancer was observed in subjects with RHG (Fig. 3), and a total of 85.7% (6/7) of the detected cancers derived from these subjects, 83.3% (5/6) of which were diffuse-type cancers. Cancer incidence was significantly higher in subjects with RHG (1,749/100,000 person-years) than in subjects without (43/100,000 person-years, HR, 43.32; 95% CI, 5.16–363.41, Table 2).
As described above and listed in Tables 1 and 2, inclusion in Group Bγ, having a high titer of H. pylori antibody or presence of RHG on endoscopy were revealed as independent risk factors for gastric cancer by Cox proportional hazards modeling after adjusting for age, smoking habit and alcohol drinking habit. As a result, 150 subjects from Group Bγ were stratified based on serum H. pylori antibody titers and endoscopic finding of RHG, and cancer development was investigated (Table 3). Among the four stratified groups, mean age and percentages of smokers and alcohol drinkers tended to be highest in subjects with high titer and RHG, but no significant differences were evident among the four groups. In both high- and low-titer groups, serum levels of PGI, PGII and H. pylori antibody were higher, and PGI/II ratio was lower in subjects with RHG than in those without, and the difference in PGII level was significant between high-titer groups with and without RHG. Cancer development was higher in the high-titer group with RHG (2,857/100,000 person-years) than in the high-titer group without RHG (532/100,000 person-years) or in the low-titer group with RHG (1,064/100,000 person-years) (Table 3). In the present cohort, 57.1% (4/7) of cancers developed from the H. pylori antibody high-titer group with RHG. Meanwhile, no cancer development was seen in the low-titer group without RHG.
During the study period, a separate group of 399 subjects underwent H. pylori eradication, with successful treatment in 327 subjects. Of these, 149 subjects belonged to Group Bγ, and baseline characteristics of the group did not differ significantly from the above-described 150 Group Bγ subjects without eradication, except for a significantly higher mean age in the latter group (Table 3). Serum levels of H. pylori antibody and PGI/II ratio were significantly higher in the eradication group. H. pylori eradication led to marked regression of endoscopic findings of severely enlarged folds in all subjects with RHG. In addition, reductions in the observed high serum levels of PG and H. pylori antibody, which were characteristic of Group Bγ subjects, were observed in all subjects (not shown). As a result, no cancer development was observed in these subjects by the end of the study period, regardless of H. pylori antibody level or endoscopic findings of RHG at baseline. The number needed to treat (NNT) for 1 year to prevent development of a single case of cancer in Group Bγ was thus estimated to be around 133. Furthermore, if the target of eradication is set for Group Bγ subjects with both high-titer H. pylori antibody and RHG, the 1-year NNT would be around 35.
In addition to initiating a cascade of events leading mainly to intestinal-type cancer—that is, the atrophy-metaplasia-dysplasia-cancer sequence—the inflammatory process induced by H. pylori infection is postulated to directly induce cancer, particularly diffuse-type cancer, from the nonatrophic stomach without passing through this sequence.4, 5 Although H. pylori-induced chronic inflammation is recognized as a universal precursor condition, other precancerous or high-risk conditions for cancers developing from nonatrophic stomach have not been described in detail. The present long-term follow-up study clearly demonstrated that a group of subjects in the cohort with highly active gastritis identified by the two serological tests of PG and H. pylori antibody are at high-risk for development of gastric cancer, particularly diffuse-type cancer with higher malignant potential. Moreover, among these serologically defined subjects at high risk of cancer, those with endoscopic RHG are at even higher risk.
In subjects with nonatrophic stomach, the establishment of H. pylori infection increases serum levels of PG, particularly PGII, together with H. pylori antibody, and the extent of the increase in these serum markers reportedly shows positive correlations with the activity of H. pylori-associated inflammation.35, 36 The basic mechanisms underlying the elevation of serum PG levels are not fully known. While H. pylori antibody level reflects the complex interaction between bacterial infection and immunological host response, PG levels are considered to reflect local mucosal damage. These two serum tests thus appear to indicate two different aspects of H. pylori infection. As described above, H. pylori-infected subjects with nonatrophic stomach were classified into the three groups of Bα, Bβ and Bγ, and the activity of H. pylori-associated gastritis appeared to increase in the order of Group Bα, Bβ and Bγ, as revealed by serum levels of PGII and H. pylori-antibody. Cancer development occurred in the same order, and Group Bγ showed the highest cancer risk among these three groups. Furthermore, stratification of Group Bγ based on the H. pylori-antibody level led to the identification of subjects with more active inflammation and higher cancer risk, especially for diffuse-type cancer, showing good accordance with the reported chronic active inflammation-based carcinogenesis for this type of cancer.5, 15
Meanwhile, highly active inflammation in the corpus of the nonatrophic stomach reportedly leads to the formation of enlarged gastric folds.22, 23 Although enlarged folds can be caused by various pathological processes, including inflammatory, or tumorous infiltration, or by hyperplasia of the foveolae or glands,37, 38 H. pylori infection is considered to represent the leading cause of gastric enlarged folds, that is, gastritis known variously as RHG or enlarged-fold gastritis based on endoscopic or barium X-ray findings, respectively.22–31 Previous studies analyzing the relationship between enlarged gastric folds on barium X-ray and gastric cancer in a case-and-control setting have suggested the possibility that enlarged fold gastritis is associated with high-risk of cancer.28–30 The present study is the first to report results after a long-term follow-up of subjects with RHG, clearly demonstrating that H. pylori-infected subjects with RHG are indeed at higher risk of cancer.
Serum levels of PG and H. pylori antibody were significantly more elevated in H. pylori-infected subjects with RHG than in subjects without such gastritis, consistent with the highly active H. pylori-induced inflammatory process that is reported to be involved in the genesis of RHG.22, 23 The percentage of RHG was highest in Group Bγ with a high-titer, representing the subgroup considered to show the most active gastritis among the groups defined by the two serological tests. Moreover, stratification of each serologically defined group based on endoscopic findings of RHG revealed that the activity of gastritis in Group Bγ with high-titer antibody together with RHG was probably the highest of all the groups in the present cohort according to the two serological tests. In the overall cohort, gastric cancer development, particularly that of diffuse-type cancer, appeared to correlate closely with the activity of gastritis, and was highest in this serologically and endoscopically defined subgroup, with an incidence rate of 2,857/100,000 person-years. All cancers that developed in this subgroup were located in the corpus of the stomach, and 75% of these cancers were diffuse-type, showing quite a high incidence rate of 2,143/100,000 person-years; compared with the aforementioned incidence for diffuse-type cancer in Japan21 or in Western countries,20 the rate in the high-risk group was about 50- or 400-fold higher, respectively. These results strongly suggest that highly active inflammation in the corpus mucosa with RHG represents a major driving force for the carcinogenic pathway in these cancers, consistent with the previous finding that individuals with corpus-predominant gastritis are at high risk for stomach cancer.10 That is, extensive increases in the mediators of inflammatory responses such as cytokines, reactive oxygen species, nitric oxide, prostaglandins and growth factors can induce genetic and epigenetic changes in the corpus mucosa, which will lead to neoplastic transformation and subsequent progression, and finally to the establishment of cancer with high malignant potential. Indeed, mounting evidence indicates that H. pylori infection frequently induces alterations in DNA methylation of various gene promoter regions in the stomach mucosa, including enlarged folds,39–41 and that inflammation-induced DNA methylation is deeply involved in the development of diffuse-type gastric cancer.42–44 In the present study, only a small proportion (<5%) of H. pylori-infected subjects developed this type of active gastritis. Various cofactors thus appear to be involved in the establishment of this highly active inflammation of stomach. Because almost all H. pylori isolates from East Asia, including Japan, possess the Cag A gene,45 the differences in cancer risk among the subgroups in this study are highly unlikely to be caused by differences in Cag A expression in the infecting bacteria. Further analyses of other bacterial factors together with host genetic and environmental factors modulating the response to H. pylori infection are required.
In line with the above-described active inflammation-based carcinogenesis, the improvement of active gastritis on eradication of H. pylori reduced the cancer incidence to zero among 149 Group Bγ subjects during the study period. These results strongly indicate that H. pylori eradication has the potential to prevent gastric cancer development in high-risk subjects with highly active inflammation in nonatrophic stomach, consistent with the current available data that H. pylori eradication could reduce the cancer risk, especially in subjects without extensive CAG.46, 47 Because other subjects at high-risk of cancer with extensive CAG, which constitutes a main route of stomach carcinogenesis, remain at high risk even after eradication,46, 47 different mechanisms are highly likely to be involved in carcinogenesis in the nonatrophic stomach. Recent evidence indicates that H. pylori eradication not only leads to marked reductions in inflammatory cell infiltration, inflammatory cytokine production and oxidative stress in the stomach mucosa, but also induces the regression of aberrant methylation in various gene regions including the CDH1 (E-cadherin) promoter in the H. pylori-infected stomach mucosa of both humans and experimental animals, if implemented before the establishment of permanent epigenetic changes, which marks the point of no return.39–41, 48, 49 All of these findings together indicate the possibility that most of these high-risk subjects are in the stage before the point of no return in terms of the process of H. pylori-induced stomach carcinogenesis.
Some limitations in the present study methods must be considered. First, all study subjects were asymptomatic, cancer-susceptible aged factory workers who were self-referred for EGD. As such, selection bias was unavoidable, as subjects may have been more health-conscious or had undisclosed reasons for suspecting the presence of upper gastrointestinal disease. These subjects may have had overall prevalences and risk profiles for gastric cancer differing from those of the general working population. The reported prevalence of enlarged gastric folds among asymptomatic middle-aged Japanese men who underwent gastric cancer screening by barium X-ray in the early 1990s was about 11%, representing a level similar to that in our subjects.24 Considering the recent trends toward rapid decreases in the prevalence of H. pylori infection, our study subjects appear to have had a higher risk of gastric cancer than the general population. Second, RHG was diagnosed based on endoscopic findings of enlarged folds or hyperrugosity, not on histopathology of the stomach mucosa, according to the Sydney system.26 Normal folds in a partially contracted stomach can sometimes be interpreted as being thickened. Although endoscopists endeavor to avoid such misinterpretations, the diagnosis of RHG must still be considered somewhat subjective. The risk of developing cancer from RHG might therefore have been underestimated. Third, although the present study prospectively investigated the possibility of gastric cancer control in Group Bγ subjects using H. pylori eradication, eradication therapy was not allocated in a randomized manner. However, randomization was neither feasible nor ethical, as a considerable number of eligible subjects were unwilling to remain untreated for long periods, particularly given the knowledge that H. pylori represents a major risk factor for the development of gastric cancer.
The present results strongly indicate that mucosal inflammation plays a pivotal role in carcinogenesis for the H. pylori-infected nonatrophic stomach, and that higher activity is associated with a higher risk of cancer, particularly diffuse-type cancer with higher malignant potential. In addition, the combination of two serological tests, for PG and H. pylori antibody, together with an endoscopic finding of RHG can identify a group at high risk of cancer (that is, a group of subjects with highly active H. pylori-associated gastritis in the nonatrophic stomach) that has not been fully described until now. In the present study, ∼57% of total cancers and 50% of diffuse-type cancers developed from this serologically and endoscopically defined high-risk group, which represented only 4.6% (23/496) of the study cohort. Careful endoscopic follow-up of these subjects thus appears warranted from the perspective of early detection of cancer. Furthermore, H. pylori eradication markedly reduced the risk of cancer development in these subjects. Our previous long-term follow-up study revealed that eradication was effective in controlling cancer development among PG test-negative subjects with nonatrophic stomach (Group B subjects), while most PG test-positive subjects with CAG have passed beyond the point of no return in the process of stomach carcinogenesis.47 However, the cost-effectiveness of eradication targeting middle-aged Group B subjects, who constitute a major proportion of the 60 million people infected with H. pylori in Japan,11, 50 is low, as the NNT for 1 year to prevent a single cancer development is around 2,000.47 Based on the present results, strategies for cancer control by H. pylori eradication appear more cost-effective when targeting Group Bγ subjects (NNT = 133), particularly those with high H. pylori antibody titer together with RHG (NNT = 35). We thus believe that the cancer high-risk group identified by the two serological tests together with endoscopic RHG represents a good target for H. pylori eradication.
The authors express their deepest thanks to Ms. Kazu Konishi for her excellent secretarial assistance. None of the authors have any conflicts of interest to declare.
- 1Vital Statistics of Japan. Statistics and Information Dept., Minister's Secretariat, Ministry of Health, Labour and Welfare 2009; 1: 300–1.
- 2Hamilton S, Aaltonen L, eds. Pathology and genetics. Tumors of the digestive system. WHO classification of tumours. Lyon: IARC Press, 2000.
- 28Helicobacter pylori-induced enlarged-fold gastritis is associated with increased mutagenicity of gastric juice, increased oxidative DNA damage, and an increased risk of gastric carcinoma. J Gastroenterol Hepatol 2003; 18: 1384–91., , , , , , , , , , .
- 29The association between tumour necrosis factor-alpha gene polymorphism and the susceptibility to rugal hyperplastic gastritis and gastric carcinoma. Eur J Gastroenterol Hepatol 2004; 43: 653–6., , , , , , , , , , .
- 38Nonneoplastic lesions of the stomach. In: Alexander R, Margulis H, Burhenne J, eds. Alimentary tract radiology. St Louis: Mosby, 1989. 551–93., .