Association of B Vitamins and Methionine Intake with the Risk of Gastric Cancer: The Japan Public Health Center–based Prospective Study

Gastric cancer is one of the most frequent cancers and a leading global cause of cancer-related mortality in both genders (1). Specific dietary patterns and high-risk genetic variants are related to the higher incidence of gastric cancer (2). Lifestyle modifications, including salt intake restriction and more fruit and vegetable consumption were associated with the reduced risk of gastric cancer (3–5). Nutrients involved in the one-carbon metabolism (OCM) pathway, such as folate, vitamin B6, vitamin B12, and methionine are involved in the DNA methylation and synthesis (6). Imbalances and deficiencies among these nutrients may lead to disturbances in DNA replication, repair, and regulation, which could result in carcinogenesis (7, 8).

Several studies have investigated the associations between B vitamins and gastric cancer risk (9–11); however, the findings remain inconsistent. For example, the dietary intakes of vitamin B6 and folate were inversely associated with the risk of gastric cancer in an American study (12); yet, folate, but not vitamin B6, was positively associated with the risk of gastric cancer in Chinese premenopausal women (13). On the other hand, a case–control study from Korea showed an inverse association between dietary folate intake and the risk of gastric cancer (14), while in contrast, two case–control studies from the Netherlands (11) and South America (15) failed to find any associations. Another case–control study from the United States reported a positive association between vitamin B12 intake and the risk of gastric cancer (12), whereas two other cohort studies from the United States (9) and Australia (10) concluded a null association. Indeed, the previous studies varied greatly in the adjustment factors. Many previous studies did not include sodium intake and the status of Helicobacter pylori (H. pylori) infection as adjustment variables (9, 11–13, 15). Moreover, a review on the role of folate in colorectal cancer pointed to potentially two different effects of folate in carcinogenesis. In normal tissue, folate may have a protective effect against carcinogenesis, while in the presence of precancerous lesions, folate may induce carcinogenesis (16).

East Asian countries, including Japan, have the highest salt and sodium intake (17, 18). Higher dietary intake of sodium could associate with the increased risk of gastric cancer due to its involvement in gastric precancerous lesions development (19–21). The pathogenesis of this association has been partially explained in an animal study that found that salty diets significantly elevated H. pylori-induced gastric carcinogenesis (22).

On the basis of these considerations, we aimed to evaluate the association between separate dietary intake of folate, vitamin B6, vitamin B12, and methionine and the risk of gastric cancer, according to sodium intake, in the Japan Public Health Center–based Prospective Study.

The Japan Public Health Center–based Prospective Study (JPHC) is an ongoing cohort study designed to investigate the association between lifestyle factors and the incidence of cancer and metabolic diseases. The study baseline between 1990 and 1994 recruited 140,420 residents ages 40–69 years in 11 Japanese public health center (PHC) areas. During 1995–1998, a 5-year follow-up survey was conducted, which obtained comprehensive information on food intake. The details of the JPHC study have been explained previously (23).

Participants registered in the Tokyo-Katsushika (n = 7,097) and Osaka-Suita (n = 16,427) PHC areas were not included in this study because of the incomplete data on the cancer incidence in the Tokyo area and a different definition of the study population in the Osaka area. Of the remaining 116,896 participants from nine PHC areas, 92,905 responded to the 5-year surveillance (response rate = 79.5%). We excluded participants of non-Japanese nationality, moved out of study areas before the 5-year follow-up, had duplicate registrations, had a history of cancer, had missing values for B vitamins and methionine dietary intakes, or had extreme energy intakes (the upper and lower 1% of the distribution). Finally, 86,820 participants were included in the analysis.

Our study focused on the dietary intakes of OCM-related nutrients such as folate, vitamin B6, vitamin B12, and methionine, assessed at the 5-year follow-up survey. Participants were asked to report the average frequency and portion size of 138 foods and beverages consumed in the past year by a self-administered food frequency questionnaire (FFQ). Energy-adjusted dietary intake of folate, vitamin B6, vitamin B12, and methionine were calculated using the residual method in reference to the Standard Tables of Food Composition in Japan (7th revised and enlarged edition) (24, 25). The FFQ-estimated nutrient intakes were validated by 28-day or 14-day dietary records in a subsample of the JPHC participants (26, 27). The Spearman correlation coefficients between FFQ-estimated intakes and the 28-day or 14-day dietary records-estimated intakes were published elsewhere (26, 27). The correlation value for folate, vitamin B6, vitamin B12, and methionine in men were 0.40, 0.45, 0.33, and 0.29, respectively, in cohort I, and 0.50, 0.36, 0.35, and 0.27, respectively, in cohort II (26, 27). The correlation coefficients for folate, vitamin B6, vitamin B12, and methionine in women were 0.35, 0.47, 0.34, and 0.27, respectively, in cohort I and 0.48, 0.40, 0.27, and 0.31, respectively, in cohort II (26, 27).

Other variables such as smoking status, alcohol intake habits, physical activity, family history of cancer, height, and weight were obtained from the 5-year follow-up questionnaire survey.

An enzyme immunoassay kit (E plate “Eiken” H. pylori Antibody II; Eiken Kagaku), with 91.2% sensitivity and 97.4% specificity, was used to detect H. pylori infection status among the 15,839 participants with available data of H. pylori infection (28). Seropositivity for anti-H. pylori antibody was defined as an IgG titer of ≥10 U/mL. Serum levels of pepsinogen I and pepsinogen II were measured by latex agglutination (LZ test “Eiken” Pepsinogen I, II; Eiken Kagaku; ref. 29). Atrophic gastritis was defined among participants with a PGI/PGII ratio of three or more (30). The validation of the cut-off value of IgG titer of H. pylori and PGI/PGII in the JPHC study's population has been done in the ROC analysis, elsewhere (31).

Participants were followed from the date of responding to the 5-year follow-up survey until censored on the date of gastric cancer diagnosis, date of death, date of migration out of the study area, or the end of follow-up (December 31, 2013)—whichever occurred first.

Data on the incidence of gastric cancer cases were identified from the medical registers of major local hospitals in the study areas and by data linkage with population-based cancer registries, with permission from the local authorities responsible for the registries. Death certificate information was used as a Supplementary Data source. All cases of gastric cancer were defined according to the Third Edition of the International Classification for Oncology as code C16.0–C16.9 (32).

Participants were divided into five groups according to B vitamins and methionine intake, with the lowest consumption group as the reference. Cox proportional hazards regression analyses were used to estimate the HRs and 95% confidence intervals (CI) of gastric cancer according to B vitamins and methionine consumption. The dietary intake of folate, vitamin B6, vitamin B12, and methionine were analyzed separately, and the results were also generated for each of those nutrients separately.

In model 1, the adjusted factors were sex, age, and PHC area. Model 2 was further adjusted for quintiles of the body mass index (BMI) in kg/m², physical activity in metabolic equivalent of task (MET) unit (quintiles), alcohol consumption (not current drinker, current drinker of ≤150 g/week, ≤151–300 g/week, <301–450 g/week, and ≥450 g/week, or missing), smoking (never, former, current smoker of ≤19 cigarettes/day, 20–29 cigarettes/day, and ≥30 cigarettes/day, or missing), family history of cancer (yes or no), and energy-adjusted sodium intake (quintiles). The P-values for the linear trend analysis were assessed by entering each group's median value into the model.

We performed stratified analyses based on sodium intake (4,500 mg/day <, and ≥ 4,500 mg/day), considering the median value of energy-adjusted sodium intake (4,835 mg/day) in our study. We investigated the interaction effects (P_interaction) for cross-product terms of sodium and B vitamins and methionine intakes.

A total of 15,839 participants had serum data on H. pylori infection and level of pepsinogen. Among them, we conducted a subgroup analysis by adding data on H. pylori infection status and atrophic gastritis as adjustment factors. A stratified analysis based on sodium intake was also performed in this subgroup. All P values presented were two sided, and values <0.05 were considered statistically significant. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc.).

The study was conducted in compliance with the provisions of the Declaration of Helsinki.

JPHC study was launched in 1990 before the enactment of ethical guidelines in Japan (2002), and thus obtaining written informed consent was not mandatory. The participants were informed of the objectives of the study, and that completion of the survey questionnaire was regarded as providing consent to participate. The study protocol including informed consent was approved by the Institutional Review Board of the National Cancer Center, Tokyo, Japan (approval number: 2001-013, 14-038). The current study protocol was approved by the Ethical Review Board of Osaka University, Osaka, Japan (approval number: 14020-9).

The data generated and/or analyzed in this study are not publicly available due to participant privacy, according to ethical guidelines in Japan. In addition, the informed consent we obtained does not include a provision for publicly sharing data. The data are available upon reasonable request from the JPHC Study Group at https://epi.ncc.go.jp/en/jphc/805/8155.html or the Office of The JPHC Study Group at [email protected].

The median age at baseline was 57 years. Participants with higher dietary intake of methionine and B vitamins were older than those with lower intake, except for the dietary intake of folate. There was an increased intake of sodium across the increasing categories of B vitamins and methionine intake. While alcohol intake decreased with higher dietary intakes of folate (Table 1).

A total of 2,269 cases of gastric cancer were identified during 1,337,035 person-years of follow-up (median follow-up period, 15.4 years) for 86,820 participants (40,563 men and 46,257 women). There were no significant associations between any of dietary intakes of B vitamins and methionine and the risk of gastric cancer. The multivariable HRs (95% CIs) in the highest versus lowest quintiles of intake were 1.11 (0.96–1.28; P_trend = 0.08) for folate, 1.03 (0.89–1.19; P_trend = 0.50) for vitamin B6, 0.94 (0.82–1.08; P_trend = 0.75) for vitamin B12, and 0.91 (0.79–1.05; P_trend = 0.19) for methionine (Table 2).

The stratified analysis of the association between folate intake and gastric cancer risk by sodium intake is shown in Fig. 1. We observed a positive association between higher intakes of folate and the risk of gastric cancer among participants in the high sodium intake group (≥4,500 mg/day): HR (95% CI) = 1.28 (1.06–1.56), P_trend = 0.001, but not among those in the low sodium intake group: HR (95% CI) = 0.94 (0.73–1.21), P_trend = 0.49; P_interaction = 0.02. There was no significant association between vitamin B12, vitamin B6, or methionine with the risk of gastric cancer in stratified analysis by sodium intake (Supplementary Table S1).

Supplementary Table S2 shows the characteristics of the 15,836 participants from four PHC areas for whom H. pylori infection and atrophic gastritis data were available. In this subgroup, 372 gastric cancer cases were identified, and the adjustment for H. pylori infection and atrophic gastritis did not change the observed null associations (Table 3). In this subgroup, the stratified analysis based on sodium intake showed an increased risk of gastric cancer in participants with higher folate intake similar to that observed for the whole cohort (Q5 vs. Q1) who consumed high amounts of sodium; HR = 2.60 (95% CI, 1.18–5.72; P_trend = 0.08; P_interaction = 0.28; Table 4). Moreover, we observed the high gastric cancer risk among participants with high folate intake who had a positive H. pylori infection and consumed high amounts of sodium. Meanwhile, there was no association between higher folate intake and the risk of gastric cancer among participants who consumed high amounts of sodium with no H. pylori infection (Table 5; Supplementary Table S3). Furthermore, there was no association between folate intake and the risk of gastric cancer in the stratified analysis by H. pylori infection (Supplementary Table S4). In the stratified analysis by alcohol intake, we found no associations of any dietary intakes of folate, vitamin B6, vitamin B12, and methionine with the risk of gastric cancer (Supplementary Table S5).

None of the dietary intakes of folate, vitamin B6, vitamin B12, or methionine were associated with the risk of gastric cancer in Japanese men and women. However, we observed an increased risk of gastric cancer for the highest dietary intake of folate among participants with high sodium intake.

Our study findings were consistent with several previous studies conducted in the United States, Australia, and China that did not find any association between higher dietary intake of vitamin B12, vitamin B6, and methionine with gastric cancer risk, including the NIH-AARP Diet and Health Study and the Melbourne Collaborative Cohort Study (MCCS) (9–11, 13). There was only one case–control study, by Mayne and colleagues, that reported a positive association between vitamin B12 and the risk of gastric cancer (12). However, the number of participants in that study were limited (n = 1,782) compared with our large sample size. A meta-analysis of three prospective studies and 13 retrospective studies found no association between folate intake and gastric cancer risk (33). In contrast, a case–control study by Kim and colleagues found an inverse association between dietary folate intake and the risk of gastric cancer [OR = 0.40 (95% CI, 0.20–0.83); P = 0.01] (14); while on the other hand, the Shanghai Women's Health Study reported an increased risk of gastric cancer with high folate intake in premenopausal women [HR and 95% CI = 2.62 (1.04–6.59)] (13).

These inconsistent findings of different studies warrant further discussion. First, there was a diverse range of dietary folate intake among the studies. The similarity between our study and those studies which found no association between folate intake and gastric cancer risk was that even in the lowest folate intake group, the dietary intake levels exceeded the recommended dietary intakes (9, 10, 34), even though the FFQ was not intended to assess absolute nutrient intakes. On the other hand, there was a relatively low dietary folate intake in the highest tertile of cases (327 mcg/day), and controls (354 mcg/day) of the Korean study that reported an inverse association between dietary folate intake and risk of gastric cancer (14) compared with the median intake in the highest quintile in our study (595.8 mcg/day). This could imply that the JPHC study population has a relatively sufficient folate intake, which might have affected the inability to detect a significant association between insufficient dietary intake of folate and the risk of gastric cancer.

Second, the potential effect modification by salt intake and H. pylori infection has not been studied in previous studies. For example, a study from China by Kweon and colleagues that found that there was an increased risk of distal gastric cancer in premenopausal women, included salt-preserved food intake as one of the adjustment factors, but not H. pylori infection status (13). However, they did not examine the potential effect modification of salt intake or sodium intake by performing stratified analyses (13). The NIH-AARP Diet and Health Study found a null association between dietary folate intake and the risk of gastric cancer, and they did not include salt intake or H. pylori as adjustment factors (9). Meanwhile, the MCCS also showed no association between folate intake and gastric cancer risk after adjustment for H. pylori infection status, but not sodium or salt intake (10). We performed a stratified analysis by sodium intake, H. pylori infection, and the combination of both these factors. In the stratified analysis according to dietary sodium intake, both in total participants and in the subgroup that provided H. pylori infection data, we found a higher gastric cancer risk by an increase in folate intake only among participants with high sodium intake. Furthermore, in the stratified analysis by sodium intake and H. pylori infection, we found that the adverse effect of higher folate intake on the risk of gastric cancer was only observed in the participants who had a positive H. pylori infection and at the same time consumed higher amounts of sodium; however, the P_interaction did not reach statistical significance.

There is abundant evidence on how salty food and high sodium intake can promote the risk of gastric cancer by disrupting the mucous barriers of the stomach that lead to decreased protection and enable H. pylori colonization (19, 21–23, 35–37). A Korean endoscopy-based study implied that high-salt intake may be associated with an increased risk of atrophic gastritis with intestinal metaplasia (38). Similarly, salt intake may increase the risk of atrophic gastritis or gastric dysplasia among participants with H. pylori infection after treatment or who had persistent H. pylori infection (39). Moreover, an animal study showed that H. pylori infection was more potent than high-salt administration in enhancing gastric cancer development; yet, the group with the highest risk of gastric cancer was Mongolian gerbils which had a positive H. pylori infection and high-salt diet (37).

Our study may be in line with the two potentially different effects of folate regarding carcinogenesis (16, 40); an adequate folate intake could inhibit carcinogenesis in the intact tissue with no precancerous lesions by maintaining normal DNA methylation (41, 42). To the contrary, after neoplasia foci have been formed, folate intake might induce carcinogenesis by increasing cell growth and metabolic activity in the abnormal tissue (43–45). For example, several animal studies suggested that folate intake might accelerate carcinogenesis of breast cancer (43, 44) and colorectal cancer (45, 46) by increasing cell growth and metabolic activity in tissues with precancerous lesions (43, 44). In our study, among the subsample with available data of atrophic gastritis, the prevalence of atrophic gastritis was 4,631(57%) in the persons with high sodium intake group and 3,494 (43%) in persons with low sodium intake. Therefore, in our study, there was a larger prevalence of atrophic gastritis, which is one of the main premalignant lesions of gastric cancer (47), in the higher sodium intake group. Accordingly, the reason why high folate intake was associated with an increased risk of gastric cancer among those with a high-salt intake was that they might have had precancerous lesions or precursor of cancer on the gastric mucosa, which attributed to the high-salt intake (46, 48).

On the other hand, our study result was discrepant with the findings from animal and clinical studies (8, 49). An animal study of mice with gastric dysplasia showed that folic acid supplementation can slow or prevent the development of gastric cancer by reversing DNA hypomethylation (8). Consistent with that study, a small prospective randomized trial of 38 patients with premalignant gastric lesions found that folic acid supplementation had chemopreventive effects on the development of gastric cancer by inducing epithelial apoptosis and increasing the tumor suppressor p53 expression in gastric mucosa (49). The effect of folate intake and folic acid supplementation in the development of precancerous gastric lesions needs further investigation.

The strengths of our study include its national-based prospective study design with a long follow-up duration and a large number of participants. We adjusted for a wide range of hypothesized confounding factors, including the energy-adjusted sodium intake. Furthermore, we provide a subgroup analysis that adjusted for the status of H. pylori infection in the model, which is another important risk factor for gastric cancer, and atrophic gastritis, which is a precancerous lesion of gastric cancer (47, 50). In addition, the large cohort size granted us the possibility to conduct stratified analyses by two strong potential effect modifiers; sodium intake and H. pylori infection. On the other hand, data from subgroup analysis that provided H. pylori infection and atrophic gastritis might contain bias because not all of the PHC areas provided data of blood samples with H. pylori serology (there were only four PHC areas that provided H. pylori infection data: Ibaraki, Nigata, Kochi, and Nagasaki). However, we found similar associations of B vitamins and methionine with gastric cancer risk in analyses of total participants and on a group of participants with available data on H. pylori infection. Our study has some limitations that warrant further investigation. First, there was a lack of data on the consumption of vitamin B supplements; less than 5% of the participants answered the question on vitamin consumption in the questionnaire. Second, we did not have any data on OCM-related gene polymorphism. Folate and other OCM-related nutrients are influenced by genes encoding OCM enzymes such as methylenetetrahydrofolate reductase, methionine synthase, methionine synthase reductase, and thymidylate synthetase (51). Thus, gene polymorphisms may influence the impact of OCM-related nutrient intake and the risk of cancer.

In conclusion, we found no association between any dietary intakes of folate, vitamin B6, vitamin B12, and methionine with gastric cancer risk. Our findings suggest that the increased intake of B vitamins and methionine in populations with adequate dietary intake of these nutrients may not provide further protection in terms of gastric cancer risk. Future studies should investigate the effect of folate supplementation on the risk of gastric cancer in patients with precancerous lesions such as atrophic gastritis or gastric dysplasia.

N. Sawada reports grants from National Cancer Center Research and Development Fund during the conduct of the study. T. Yamaji reports grants from Ministry of Health, Labour and Welfare of Japan during the conduct of the study. No disclosures were reported by the other authors.

P. Khairan: Conceptualization, formal analysis, investigation, visualization, methodology, writing–original draft, writing–review and editing. T. Sobue: Conceptualization, supervision, validation, investigation, visualization, methodology, writing–review and editing. E.S. Eshak: Conceptualization, supervision, investigation, visualization, methodology, writing–review and editing. L. Zha: Writing–review and editing. T. Kitamura: Supervision, project administration, writing–review and editing. N. Sawada: Resources, data curation, funding acquisition, validation, methodology, project administration, writing–review and editing. M. Iwasaki: Resources, validation, methodology, project administration, writing–review and editing. M. Inoue: Resources, data curation, validation, methodology, project administration, writing–review and editing. T. Yamaji: Resources, validation, methodology, project administration, writing–review and editing. H. Iso: Conceptualization, resources, supervision, validation, investigation, methodology, writing–review and editing. S. Tsugane: Resources, data curation, funding acquisition, validation, project administration, writing–review and editing.

JPHC members are listed on the following website (as of September 2019): https://epi.ncc.go.jp/en/jphc/781/8390.html. We are indebted to the Aomori, Akita, Iwate, Ibaraki, Niigata, Osaka, Kochi, Nagasaki, and Okinawa Cancer Registries for providing their incidence data. N. Sawada and S. Tsugane reports grants from National Cancer Center Research and Development Fund. P. Khairan is the recipient of the scholarship of the Ministry of Education, Culture, Sports, Science, and Technology-Japan for PhD program at Osaka University, Japan.

This study was supported by the National Cancer Center Research and Development Fund (since 2011) [23-A-31(toku), 26-A-2 and 29-A-4], Grant-in-Aid for Cancer Research from the Ministry of Health, Labour and Welfare of Japan (from 1989 to 2010; 19shi-2) and the Ministry of Agriculture, Fishery and Forestry, Japan (JPJ005336). All grants listed were received by N. Sawada and S. Tsugane. The funders played no role in the design, interpretation, or publication of study findings.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.