Introduction

Endoscopic resection, including endoscopic submucosal dissection (ESD), is currently accepted widely as an effective, minimally invasive local treatment for early gastric cancer with a negligible risk of lymph node metastasis, delivering a favorable long-term outcome [1–7]. However, the development of metachronous gastric cancer (MGC), owing to the preservation of the entire stomach after endoscopic resection, remains a possibility [8–10]. A previous multicenter prospective randomized study performed in Japan showed that eradication of Helicobacter pylori reduced the incidence of MGC after endoscopic resection for early gastric cancer [11]. However, MGC sometimes develops even after successful H. pylori eradication, with an annual incidence of 0.8–4.1 % [9–15]. In most previous studies, only incidences of MGC within a 3-year median follow-up period were reported, and a study with a longer follow-up period is needed. Additionally, only a few studies have evaluated risk factors associated with the development of MGC [12, 13].

Our group previously reported that accumulation of aberrant DNA methylation induced by H. pylori infection and chronic inflammation in noncancerous gastric mucosae causes field cancerization, and that the degree of the field is strongly associated with the risk of development of gastric cancer [16–22]. To demonstrate the usefulness of assessing the degree of epigenetic field cancerization, we recently conducted a multicenter prospective cohort study with a median follow-up of 2.97 years, and demonstrated that the methylation level of miR-124a-3 was associated with an increased risk of developing MGC [23]. The cohort established in the study has a high value because of its size, the sufficiently long period of follow-up (a median of 4.5 years), and the high patient follow-up rate (97.9 %; 809 of 826 patients).

Here, we performed a retrospective analysis to investigate the incidence of and nonepigenetic risk factors for MGC after patients had undergone endoscopic resection for early gastric cancer with successful H. pylori eradication using the above-mentioned cohort.

Patients and methods

Patients

A total of 826 patients who underwent ESD for early gastric cancer between 1999 and 2010 in one of three hospitals (National Cancer Center Hospital, Tokyo University Hospital, and Wakayama Medical University Hospital) were enrolled between 2008 and 2010 in the multicenter prospective cohort study [23]. One hundred nineteen patients who tested negative for H. pylori infection without previous eradication were excluded, as were 100 patients who had already achieved successful eradication before detection of their initial gastric cancer. Thirteen patients who were followed up for less than 1 year were also excluded. The remaining 594 patients (470, 86, and 38 patients at National Cancer Center Hospital, Tokyo University Hospital, and Wakayama Medical University Hospital, respectively) who underwent ESD and successful H. pylori eradication were analyzed retrospectively in the present study (Fig. 1). Among them, 210 were patients who gastric cancer had been newly diagnosed and treated when they were enrolled, and 384 were postresection patients who had previously undergone ESD and were in the follow-up phase. R0 resection was performed by ESD for all gastric cancer patients.

Fig. 1
figure 1

Flowchart of study patients

Full size image

Detection of Helicobacter pylori infection and its eradication

The presence of H. pylori infection was examined with the urea breath test (Otsuka, Tokushima, Japan), serum anti-H. pylori antibody test (Eiken, Tokyo, Japan), and either the culture method or the rapid urease test (Otsuka, Tokushima, Japan). A positive result in at least one of these tests was considered evidence of H. pylori infection. To eradicate H. pylori, patients were treated with lansoprazole (30 mg), amoxicillin (750 mg), and clarithromycin (200 mg), each taken twice daily for 1 week. Successful eradication was confirmed on a negative urea breath test 2 months or more after treatment. If eradication was unsuccessful, the patient received second-line eradication therapy. Implementation of third-line and further eradication therapy was at the discretion of the attending physician. To confirm continuous H. pylori negative status, the urea breath test, serum anti-H. pylori antibody test, and the culture method were performed at 3 and 5 years after the patients’ enrollment in the previous study [23]. If any of these follow-up examinations revealed positivity for H. pylori, additional eradication was performed.

Evaluation of baseline gastric mucosal atrophy, the pepsinogen test, and lifestyle habits

Baseline gastric mucosal atrophy immediately before successful H. pylori eradication was assessed endoscopically according to the Kimura-Takemoto classification [24], which correlates well with the histologic degree of atrophic gastritis. Gastric mucosal atrophy was divided into three categories: mild (closed type I and II), moderate (closed type III and open type I), and severe (open type II and III). Blood samples were collected on the day of the first endoscopy after the enrollment, and serum levels of pepsinogen (PG) I and PG II were measured by the LZ test (Eiken, Tokyo, Japan). Patients with a PG I level of 70 ng/ml or lower and a PG I/PG II ratio of 3 or less were considered positive for PG. A questionnaire-style survey on lifestyle habits (history of smoking, drinking, and green vegetable intake) was conducted for all patients when they were enrolled.

Surveillance endoscopy and definition of metachronous gastric cancer

Annual surveillance esophagogastroduodenoscopy was performed for all patients after endoscopic resection for initial gastric cancer. MGC was defined as a new gastric cancer detected at least 1 year after successful H. pylori eradication and located in an area other than the site of the previous endoscopic resection; this definition was in accordance with the previous study [23].

Statistical analysis

The cumulative incidence of MGC was calculated by Kaplan–Meier analysis. The follow-up period was calculated from the starting date of the latest successful H. pylori eradication to the date of detection of the initial MGC or the latest surveillance endoscopy. The person-year method was used to calculate the incidence of MGC, which included only initial MGCs; subsequent tumors were excluded if multiple MGCs developed.

We performed univariate analysis using the log-rank test to evaluate the relationship between the clinicopathological features and the probability of MGC development. We also performed multivariate using the Cox proportional hazards model for variables considered significant on univariate analysis in order to identify independent risk factors for MGC development. P values less than 0.05 were considered to be statistically significant. All statistical analyses were performed by use of the statistical analysis software (SPSS, version 20; SPSS, Tokyo, Japan) and the statistical program R, version 3.1.1 (http://cran.r-project.org).

Results

Patient characteristics and incidence of metachronous gastric cancer

Table 1 shows the clinicopathological characteristics of participating patients. The cohort was mostly male (82 %: 485 of 594 patients), and the median follow-up period after successful H. pylori eradication was 4.5 years (range 1.0–12.6 years). Severe gastric mucosal atrophy was observed endoscopically in over half of the patients (55 %: 325 of 594 patients) just before successful eradication, and 22 % of patients (130 of 594 patients) had multiple gastric cancers before successful eradication.

Table 1 Patient characteristics
Full size table

During the follow-up period, 94 MGCs were detected in 79 patients (13 %: 79 of 594 patients). Kaplan–Meier curves showed that the cumulative incidence of MGC 5 years after successful H. pylori eradication was 15.0 % (Fig. 2). The incidence of MGC calculated by the person-year method was 29.9 cases per 1000 person-years (79 cases per 2645.1 person-years). The incidence of MGC when all subsequent MGCs were included was 33.4 lesions per 1000 person-pears (94 lesions per 2817.1 person-years).

Fig. 2
figure 2

Kaplan–Meier analysis for the cumulative incidence of metachronous gastric cancer after successful Helicobacter pylori eradication

Full size image

Risk factors for the development of metachronous gastric cancer

Univariate analysis demonstrated that male sex, severe gastric mucosal atrophy, the location of the initial gastric cancer in the upper third of the stomach, multiple gastric cancers before successful H. pylori eradication, and smoking for 40 pack-years or more were significantly associated with the development of MGC (Table 2). Multivariate analysis for the risk of MGC revealed that male sex, severe gastric mucosal atrophy, and multiple gastric cancers before successful H. pylori eradication were independent risk factors (Table 3). Figure 3 shows Kaplan–Meier analyses by sex, severity of gastric mucosal atrophy, and number of gastric cancers before successful H. pylori eradication.

Table 2 Univariate analysis for the risk of metachronous gastric cancer (MGC) development
Full size table
Table 3 Multivariate analysis for the risk of metachronous gastric cancer development for five variables considered significant on univariate analysis
Full size table
Fig. 3
figure 3

Kaplan–Meier analysis for the cumulative incidence of metachronous gastric cancer after successful Helicobacter pylori eradication by sex (a), severity of gastric mucosal atrophy (b), and number of gastric cancers before H. pylori eradication (c)

Full size image

Clinicopathological characteristics of metachronous gastric cancers

Table 4 shows the clinicopathological characteristics of 94 MGCs detected in 79 patients during the follow-up period. Multiple MGCs developed in 16 % of patients (13 of 79 patients), and 11 % (10 of 94) of the MGCs were detected more than 5 years after successful H. pylori eradication. MGCs were mostly intramucosal, differentiated-type tumors that were cured by ESD performed according to the Japanese gastric cancer treatment guidelines [25], with the exception of one differentiated-type tumor that was 4 mm in size, of category T1b (submucosal invasion 80 μm), and had lymphatic involvement.

Table 4 Characteristics of 94 metachronous gastric cancers (MGCs) in 79 patients
Full size table

Discussion

Our investigation clearly demonstrated the incidence of and risk factors for MGC in patients who underwent endoscopic resection for early gastric cancer with successful H. pylori eradication using data derived from a large-scale, multicenter prospective cohort study [23]. Several studies have reported annual incidences of MGC after H. pylori eradication, ranging from 0.1 to 0.5 % in healthy individuals or peptic ulcer patients [26–33], and from 0.8 to 4.1 % in patients after endoscopic resection for early gastric cancer [9, 11–15]. Individuals after endoscopic resection had a higher risk of MGC compared with those who were healthy or had peptic ulcers. Most of these previous studies, however, reported the incidence of MGC during a median follow-up of 3 years. A multicenter prospective randomized study [11] from Japan showed that the incidence of MGC in 255 patients who underwent H. pylori eradication was 14.1 cases per 1000 person-years during a 3-year follow-up period. The incidence of MGC in our present study (29.9 cases per 1000 person-years) was higher than that of the previous study. This variation may be due to differences in patient characteristics, as our study included a higher percentage of male patients as well as patients with severe gastric mucosal atrophy than the previous study; these factors are independent risk factors for MGC. Hence, H. pylori eradication aimed at reducing the risk of MGC development after endoscopic resection may have had a more limited effect in such patients. Similarly, a recent meta-analysis indicated that H. pylori eradication was less effective in preventing gastric cancer in patients with intestinal metaplasia [34]. Moreover, Kaplan–Meier analysis showed a linear increase in the cumulative incidence of MGC (Fig. 2); 11 % of MGCs (10 of 94) were detected more than 5 years after successful H. pylori eradication (Table 4). Several cases of MGC detected more than 10 years after successful eradication of H. pylori were also reported [35, 36]. It is therefore important to perform surveillance endoscopy continuously.

The present study clarified independent risk factors of MGC (including male sex, severe gastric mucosal atrophy, and multiple gastric cancers before successful H. pylori eradication) by multivariate analysis using the Cox proportional hazards model (Table 3). The sex difference in gastric cancer patients was clearly demonstrated in previous epidemiological studies [37]. Several studies have also reported that severe gastric mucosal atrophy was a risk factor for MGC development [12, 13, 29]. Our group has previously demonstrated that inflammation triggered by H. pylori induced aberrant DNA methylation in gastric mucosae; the accumulation of such methylation was strongly associated with a risk of gastric cancer development (epigenetic field cancerization) [17–23]. It is estimated that patients with severe gastric mucosal atrophy have high methylation levels because of long-term H. pylori infection and chronic inflammation. Additionally, we reported that patients with multiple gastric cancers had significantly higher methylation levels than those with a single gastric cancer [38]. These data indicate that severe gastric mucosal atrophy and multiple gastric cancers before successful H. pylori eradication are risk factors for MGC.

To the best of our knowledge, this study includes the largest number of subjects in published studies demonstrating MGC after endoscopic resection with H. pylori eradication. Moreover, strictly observed H. pylori infection by repeatedly using three examination methods in combination during the follow-up period.

The limitations of the current study include a heterogeneous patient population: both patients with newly diagnosed gastric cancer and postresection follow-up patients were enrolled. Therefore, individual patients had differing lengths of the interval between detection of the initial gastric cancer and successful H. pylori eradication. However, univariate analysis showed that any relationship between these differing lengths of the intervals and the probability of MGC development was not statistically significant (Table 2). A portion of our study patients who had already achieved successful H. pylori eradication before enrollment had the potential to show a decreased PG value because it was measured at the time of enrollment. It was impossible to evaluate the efficacy of the PG method precisely, although several studies have reported that PG is a predictive factor for gastric cancer development [12, 39–41].

In conclusion, annual and continuous surveillance endoscopy for MGC in patients who have undergone endoscopic resection for early gastric cancer should be performed even after successful H. pylori eradication. This is particularly true for male patients, those with extensive gastric mucosal atrophy, and those who developed multiple gastric cancers before H. pylori eradication.