The number of patients with gastric cancer (GC) worldwide in 2020 was 11.1 million, and GC ranked fifth and fourth regarding incidence and deaths among all cancers, respectively.¹ The incidence of GC is particularly high in East Asia and Eastern Europe.¹ Although the mortality rate of GC has decreased in Japan over the past few decades, approximately 42,000 people die annually from the disease, emphasizing the importance of GC prevention.²

Several factors contribute to the incidence of GC. Helicobacter pylori is the most established risk factor for GC and is more prevalent in endemic areas.^3-6 In previous studies, 75% to 90% of non-cardiac GC have been attributed to H. pylori.^5,7 Correa proposed that the mechanism of GC development involves the infection of the gastric mucosa with H. pylori, followed by chronic active gastritis, chronic atrophic gastritis, and finally GC.⁸ Many studies have reported that the severity of gastric mucosal atrophy and intestinal metaplasia is associated with GC development.^9,10 The Operative Link on Gastritis Assessment system for the grading and staging of the phenotypes of long-standing gastritis and the Operative Link on Gastric Intestinal Metaplasia Assessment system, which recognizes intestinal metaplasia in the gastric mucosa more easily and consistently, are used for GC risk assessment in clinical practice.^9,11 However, individual differences exist in the progression of gastric mucosal atrophy and intestinal metaplasia, and the causes remain unclear. Moreover, not all people infected with H. pylori develop GC; it develops in only 1% to 3% of infected people.¹² Therefore, identifying factors other than H. pylori that promote the development of GC is crucial.

In addition to H. pylori, older age and male sex are known risk factors for GC.¹³ However, these factors are unmodifiable. In contrast, modifiable risk factors for GC include lifestyle habits. Evidence for the role of smoking in the incidence of various cancers, including GC is unequivocal.¹⁴ Similarly, the consumption of certain foods has been reported to be associated with the incidence and prevention of several cancers. For example, studies have found that the consumption of red and processed meat promotes colorectal cancer.^15,16 Some studies have also reported an association between habitual salt intake and GC incidence.¹⁷ However, evidence for its relationship with other nutrients remains limited. Therefore, we aimed to investigate the association between daily nutrient intake and GC incidence in a Japanese population. Additionally, the relationship between nutritional intake and gastric atrophy was examined.

1. Study design, participants, and data sources

This was a post hoc analysis of a randomized controlled trial, ‘‘Gastric Cancer Screening Labeled by Serum Examination’’ in Place of Aged Gastric Cancer Organized Screening System study comparing gastrointestinal X-ray and esophagogastroduodenoscopy for GC screening.^18,19 Briefly, the trial was conducted between June 2011 and March 2013 in Akita Prefecture, Japan, among participants aged 30–74 years. Investigators excluded candidates with a history of GC, malignancies other than GC within the past 5 years, resection of the stomach or duodenum, those not expected to survive for >5 years due to serious systemic diseases, those who had difficulty providing informed consent, and those whose physicians considered it difficult for them to participate in this study. Participants were randomly assigned to the gastrointestinal X-ray and esophagogastroduodenoscopy groups and were followed up for 6 years for GC incidence. Participants with abnormal findings in each modality were finally diagnosed with GC histologically through endoscopic biopsy. This analysis further excluded participants who were diagnosed with GC at entry and those who had no follow-up after entry. All participants provided written informed consent before their inclusion in this study. The Ethics Review Board of Nihon University Hospital approved this study (approval number: 2023020201).

2. Assessment of daily nutrition

All participants completed a self-administered food frequency questionnaire (FFQ) regarding their regular dietary habits over the past 1 year at enrollment in the randomized controlled trial before GC screening began.²⁰ The FFQ comprises 47 food items, including rice, bread, noodles, margarine, butter, milk, yogurt, tofu in miso soup, cold tofu (hiyayakko), fermented soybeans (natto), fried tofu (ganmodoki), fish, small fish that can be eaten with bones, canned tuna, octopus/shrimp/crab/squid, shellfish, cod roe/salmon roe, fish paste products, eggs, chicken, beef/pork, liver, ham/sausage/salami/bacon, mayonnaise, deep-fried dishes, light-fried dishes, potatoes, pumpkin, carrot, broccoli, green-leaf vegetables, other green-yellow vegetables, cabbage, radish, dried radish (kiriboshi daikon), bamboo shot/burdock, other light-colored vegetables, mushroom, seaweed, oranges/grapefruits, other fruits, peanuts/almond, Western confectionery, Japanese confectionery, green tea, coffee, and alcohol. The FFQ also included questions regarding the average frequency of intake over the past 1 year, with the following eight options: rarely, one to three times per month, one to two times per week, three to four times per week, five to six times per week, once per day, twice per day, and ≥three times per day.

Daily intake of the following nutrients was calculated from the above responses: energy, protein, lipid, carbohydrate, sodium, potassium, calcium, iron, carotene, retinol equivalent, vitamin D (VD), vitamin C, vitamin E, vitamin B1, vitamin B2, folate, saturated fatty acid, monounsaturated fatty acid, polyunsaturated fatty acid, n-3 highly unsaturated fatty acid, cholesterol, soluble dietary fiber (SDF), and insoluble dietary fiber .

3. Assessment of other GC-related factors, gastric atrophy, and pathology

All participants were assessed for H. pylori immunoglobulin G (IgG) levels and underwent serological evaluation of gastric mucosal atrophy at enrollment. H. pylori status was evaluated through specific H. pylori IgG antibodies using an enzyme immunoassay kit (E-plate; Eiken Kagaku, Tokyo, Japan). Serological evaluation of gastric mucosal atrophy was based on pepsinogen (PG) I and II levels, which were measured using a radioimmunoassay (pepsinogen kit; BML Inc., Akita, Japan). H. pylori IgG antibodies were defined as positive and negative when their concentrations were ≥10 U/mL and <10 U/mL, respectively. Based on previous reports, PG I and PG I/PG II ratios were used to evaluate gastric atrophy, with positive gastric atrophy defined as PG I ≤70 ng/mL with a PG I/PG II ratio ≤3.²¹ Expert pathologists performed pathological diagnosis, which was classified according to the Japanese classification of gastric carcinoma.²² Differentiated types included papillary adenocarcinoma in addition to well or moderately tubular adenocarcinoma, while undifferentiated types encompassed poorly adenocarcinoma, signet ring cell carcinoma, and mucinous adenocarcinoma. Mixed types were determined based on the predominant tissue present.

4. Outcome and statistical analysis

This study primarily aimed to investigate the factors associated with the incidence of GC, including nutrients. Continuous variables are presented as medians (interquartile range). Categorical variables are presented as numbers (percentages). Participant characteristics were examined using the Mann-Whitney U and Fisher exact tests for continuous and categorical variables, respectively.

For the continuous variables of nutrient intake, we performed receiver operating characteristic analysis and calculated the area under the curve. The cutoff values were determined using the Youden index.²³ We dichotomized each nutrient into high and low categories using the cutoff values calculated above, which were subsequently investigated for their association with GC incidence using Cox proportional hazards regression analysis. All nutrient intake variables and participant characteristics (age, sex, smoking and alcohol habit, H. pylori IgG antibody, PG I, and PG II) with a p-value <0.05 in the crude analysis were forcibly entered in multivariate Cox proportional hazards analysis. Next, we examined the association between the above-mentioned nutrients and gastric atrophy at baseline using a bivariate logistic regression model. All nutrient intake variables and participant characteristics with a p-value <0.05 in the bivariate analysis were used for multivariate logistic regression analysis to identify the factors associated with gastric atrophy. Before each multivariate analysis, the entered variables were tested for multicollinearity by calculating variance inflation factors (VIFs) using linear regression models, where factors with a VIF >10 were excluded from the corresponding multivariate analysis.

Statistical analyses were performed using IBM SPSS Statistics 24.0 (IBM Corp., Armonk, NY, USA) and EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), graphical user interfaces for R (R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was set at p<0.05.

1. Participant demographics

Overall, 1,206 participants were enrolled in the original study. Fifty-nine participants were excluded due to lack of follow-up, leaving 1,147 enrolled in this analysis. The median age was 62 years (interquartile range, 58 to 67 years), and 581 participants (50.7%) were men. GC was observed in 25 participants during the follow-up period (median, 2,184 days; interquartile range, 2,153 to 2,237 days), resulting in an incidence rate of 25 GC per 6,721 person-years.

The baseline characteristics of the participants with and without GC are presented in Table 1. The median age of the participants with GC was higher than that of those without GC (66 [61 to 70] years vs 62 [58 to 67] years, p=0.006). There were significant differences in smoking, positivity for H. pylori IgG antibody, and PG I levels between the two groups (smoking, p<0.001; positivity for H. pylori IgG antibody, p=0.004; PG I level, p=0.032). Regarding GC pathology, there were 21 (86.0%) and four (14.0%) cases of the differentiated and undifferentiated types, respectively.

Table 1 . Patient Characteristics.

Characteristic	Overall (n=1,147)	GC (+) (n=25)	GC (–) (n=1,122)	p-value
Age, yr	62 (58–67)	66 (61–70)	62 (58–67)	0.006
Male sex	581 (50.7)	22 (88.0)	559 (49.8)	<0.001
Body mass index, kg/m2	23.4 (21.2–25.3)	24.5 (22.4–25.9)	23.3 (21.2–25.3)	0.202
Alcohol	652 (56.8)	15 (60.0)	637 (56.8)	0.840
Smoking	460 (40.1)	19 (76.0)	441 (39.3)	<0.001
H. pylori status
Pepsinogen I, ng/mL	44.2 (29.8–62.3)	35.6 (17.8–65.8)	44.3 (30.3–62.2)	0.032
Pepsinogen II, ng/mL	15.6 (9.6–26.7)	17.6 (10.9–24.8)	15.4 (9.5–26.7)	0.452
H. pylori IgG antibody	652 (56.8)	21 (84.0)	631 (56.2)	0.004
History of H. pylori eradication	314 (27.4)	7 (28.0)	307 (27.4)	0.549
Pathology	-		-	-
Differentiated type		21 (84.0)
Undifferentiated type		4 (16.0)

Data are presented as median (interquartile range) or number (%)..

GC, gastric cancer; H. pylori, Helicobacter pylori; IgG: immunoglobulin G..

2. Factors of GC incidence in nutrient intake

Nutrient intake was calculated using the FFQ (Supplementary Table 1). Incidence rates of GC for each nutrient in both high and low intake groups are presented in Supplementary Table 2. As shown in Fig. 1, the bivariate analysis revealed that the intake of sodium was the most strongly associated item with GC incidence (hazard ratio [HR], 4.530; 95% confidence interval [CI], 1.953 to 10.508; p<0.001), followed by energy (HR, 3.181; 95% CI, 1.270 to 7.966; p=0.014), carbohydrate (HR, 3.767; 95% CI, 1.504 to 9.435; p=0.005), and VD (HR, 2.661; 95% CI, 1.175 to 6.028; p=0.019). In contrast, higher intake of vitamin C (HR, 0.307; 95% CI, 0.106 to 0.896; p=0.031), monounsaturated fatty acid (HR, 0.269; 95% CI, 0.092 to 0.785; p=0.016), insoluble dietary fiber (HR, 0.323; 95% CI, 0.129 to 0.809; p=0.016), and SDF (HR, 0.133; 95% CI, 0.018 to 0.982; p=0.048) had an inverse association with GC incidence. Using linear regression models, we calculated the VIF for these eight nutrients and patient characteristics with significant associations as follows: age, male sex, PG I, positivity for H. pylori IgG antibody, and smoking. The highest VIF value observed was 6.21 for carbohydrates, and all other variables had VIF values <10.

Figure 1. Cox proportional hazards regression analysis of the relationship between gastric cancer incidence and nutrient intake variables. Hazard ratios (HRs) were calculated for the high-intake group when the low-intake group was set as the reference for each dichotomized nutrition intake variable. The X-axis is log-transformed. BMI, body mass index; H. pylori, Helicobacter pylori; IgG, immunoglobulin G; SFA, saturated fatty acid; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid; n-3 HUFA, n-3 highly unsaturated fatty acid.

Variables that were significantly associated with GC incidence in the crude analysis and with a VIF <10 were entered into the multivariate Cox proportional hazards regression model. As shown in Fig. 2, sodium intake (adjusted HR [aHR], 3.905; 95% CI, 1.520 to 10.035; p=0.005) and VD (aHR, 2.747; 95% CI, 1.111 to 6.788; p=0.029) were positively associated and SDF (aHR, 0.104; 95% CI, 0.012 to 0.905; p=0.040) was inversely associated with GC incidence. Regarding factors unrelated to nutrition, higher age (aHR, 2.478; 95% CI, 1.040 to 5.905; p=0.041), positivity for H. pylori IgG antibody (aHR, 4.159; 95% CI, 1.393 to 12.419; p=0.011), and lower PG I level (aHR, 2.688; 95% CI, 1.156 to 6.250; p=0.022) were all positively associated with GC incidence.

Figure 2. Multivariate Cox proportional hazards regression analysis of the factors associated with gastric cancer incidence. Adjusted hazard ratios (aHRs) were calculated for the higher group when the lower group was set as the reference for each dichotomized variable. The X-axis is log-transformed. H. pylori, Helicobacter pylori; IgG, immunoglobulin G; MUFA, monounsaturated fatty acid.

3. Relationship between nutrition intake and gastric atrophy

Of the 1,147 patients, 493 had gastric atrophy. In bivariate analysis, a higher intake of sodium (odds ratio [OR], 1.473; 95% CI, 1.145 to 1.896; p=0.003), saturated fatty acid (OR, 1.397; 95% CI, 1.033 to 1.889; p=0.030), iron (OR, 1.057; 95% CI, 1.007 to 1.109; p=0.024), protein (OR, 1.009; 95% CI, 1.002 to 1.016; p=0.008), higher age (OR, 1.918; 95% CI, 1.485 to 2.477; p<0.001), positivity for H. pylori IgG antibody (OR, 9.160; 95% CI, 6.864 to 12.224; p<0.001), and history of H. pylori eradication (OR, 2.076; 95% CI, 1.596 to 2.701; p<0.001) were positively associated with gastric atrophy. Using linear regression models, we calculated the VIF values for these seven factors. The highest observed VIF value was 2.28 for protein, and all other variables had VIF values <10. Multivariate analysis of these factors revealed that higher intake of sodium (OR, 1.436; 95% CI, 1.011 to 2.038; p=0.043), higher age (OR, 2.021; 95% CI, 1.488 to 2.745; p<0.001) and positivity for H. pylori IgG antibody (OR, 9.027; 95% CI, 6.645 to 12.262; p<0.001) were positively associated with gastric atrophy (Table 2).

Table 2 . Nutrition Intake and Gastric Atrophy.

Variable	Atrophy (+) (n=493)	Atrophy (–) (n=654)	Bivariate			Multivariate
			OR* (95% CI)	p-value		OR* (95% CI)	p-value
Age, yr	64 (60–68)	61 (56-65)	1.918 (1.485–2.477)	<0.001		2.021 (1.488–2.745)	<0.001
Male sex	261 (52.9)	320 (48.9)	1.174 (0.929–1.484)	0.179
Alcohol	293 (59.4)	359 (54.9)	1.204 (0.950–1.525)	0.125
Smoking	197 (40.0)	263 (40.2)	0.989 (0.779–1.256)	0.931
H. pylori IgG antibody	414 (84.0)	238 (36.4))	9.160 (6.864–12.224)	<0.001		9.020 (6.645–12.262)	<0.001
History of H. pylori eradication	176 (35.7)	138 (21.1)	2.076 (1.596–2.701)	<0.001		1.267 (0.932–1.724)	0.131
Energy, kcal	1,791 (1,601–2,039)	1,750 (1,555–1,981)	1.243 (0.984–1.571)	0.069
Protein, g	63.2 (54.2–73.1)	58.8 (51.6–69.4)	1.009 (1.002–1.016)	0.008		0.994 (0.983–1.004)	0.228
Lipid, g	44.9 (37.6–53.8)	43.7 (36.8–52.0)	1.203 (0.951–1.522)	0.124
Carbohydrate, g	263.5 (230.0–312.0)	256.3 (220.6–301.6)	1.244 (0.983–1.573)	0.069
Sodium, mg	2,627 (1,981–3,178)	2,462 (1,858–2,970)	1.473 (1.145–1.896)	0.003		1.436 (1.011–2.038)	0.043
Potassium, mg	2,441 (2,126–2,774)	2,373 (2,043–2,766)	1.251 (0.989–1.582)	0.061
Calcium, mg	550 (454–662)	541 (450–641)	1.278 (0.972–1.680)	0.079
Iron, mg	8.6 (7.2–10.1)	8.2 (6.9–10.1)	1.057 (1.007–1.109)	0.024		1.019 (0.948–1.095)	0.612
Carotene, μg	2,917 (2,140–4,037)	2,917 (2,114–4,037)	0.965 (0.763–1.220)	0.776
Retinol equivalent, μgRE	932 (659–1,240)	896 (648–1,181)	1.177 (0.931–1.487)	0.174
Vitamin D, μg	10 (7–12)	8 (7–12)	1.018 (0.999–1.037)	0.070
Vitamin E, mg α-TE	8.2 (7.0–9.9)	8.2 (7.0–9.7)	0.957 (0.722–1.269)	0.760
Vitamin B1, mg	0.68 (0.62–0.76)	0.67 (0.62–0.74)	1.060 (0.826–1.360)	0.646
Vitamin B2, mg	1.09 (0.92–1.32)	1.07 (0.90–1.27)	1.266 (0.860–1.866)	0.232
Vitamin C, mg	80 (62–105)	81 (63–108)	0.954 (0.749–1.215)	0.702
Folate, μg	307 (243–378)	302 (246–381)	1.060 (0.836–1.342)	0.632
SFA, g	11.8 (9.7–13.7)	11.5 (9.4–13.1)	1.397 (1.033–1.889)	0.030		1.465 (0.985-2.179)	0.060
MUFA, g	16.7 (14.3–20.4)	16.6 (14.1–19.6)	0.963 (0.563–1.647)	0.891
PUFA, g	15.4 (13.0–18.4)	15.2 (12.8–17.4)	1.130 (0.892–1.430)	0.312
n-3 HUFA, mg	994 (755–1312)	855 (738–1,299)	1.254 (0.991–1.585)	0.059
Cholesterol, mg	249 (208–319)	249 (200–312)	1.039 (0.709–1.521)	0.845
Insoluble dietary fiber, g	2.3 (1.8–2.7)	2.2 (1.8–2.6)	0.993 (0.753–1.311)	0.963
Soluble dietary fiber, g	7.9 (6.5–9.6)	7.9 (6.5–9.7)	1.070 (0.847–1.352)	0.571

Data are presented as median (interquartile range) or number (%)..

OR, odds ratio; CI, confidence interval; H. pylori, Helicobacter pylori; IgG, immunoglobulin G; α-TE, α-tocopherol equivalent; SFA, saturated fatty acid; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid; n-3 HUFA, n-3 highly unsaturated fatty acid..

^*ORs were calculated for the higher intake group when the lower intake group was set as the reference for each dichotomized nutritional intake variable..

This study demonstrated that high daily intake of sodium and VD and low SDF intake were associated with GC incidence after adjusting for other significant factors including age, H. pylori status, gastric mucosal atrophy level, and smoking. Additionally, a higher sodium intake was positively correlated with gastric atrophy.

The positive correlation of high salt or sodium intake with GC and gastric atrophy is consistent with the findings of previous reports.^17,24 In an endoscopy-based study, Song et al.²⁵ reported that high salt intake increased the risk of gastric atrophy with intestinal metaplasia.

Regarding the mechanism underlying the relationship between salt intake and GC incidence, in vitro experiments with rats revealed that salt intake causes gastritis and enhances the action of carcinogens such as N-methyl-N-nitro-N-nitrosoguanidine.²⁶ Moreover, some studies have reported that a high salt intake increases colonization by H. pylori and induces mucosal damage in persistent H. pylori infection.²⁷ Therefore, salt intake might promote Correa’s cascade steps, where chronic active gastritis leads to chronic atrophic gastritis and, ultimately, GC. Furthermore, reports exist that high salt intake increases the GC mortality rate.²⁸ Japanese people consume more salt than people in other countries because traditional Japanese seasonings, such as soy sauce, miso, and dried fish, contain a significant amount of salt. Particularly, people in the northeast region, where Akita Prefecture is located and where this study was conducted, have a higher salt intake and GC mortality rate than those in other regions of Japan.^2,29 The World Health Organization recommends a salt intake of <5.0 g/day for adults.³⁰ In Japan, salt intake targets are set at <7.5 g/day and 6.5 g/day for healthy men and women, respectively, because of the above eating habits.³¹ These numerical targets are solely for preventing cardiovascular or cerebrovascular diseases and do not consider GC. In this study, when the amount of sodium calculated from FFQ was converted to salt content, a statistically significant difference was found between GC+ and GC– (7.8 g/day vs 6.4 g/day, p=0.002). Moreover, it was 6.3 g/day and 6.2 g/day for men and women without GC, respectively, and 7.6 g/day and 7.1 g/day for men and women with GC, respectively, which were higher than the aforementioned target values in Japan.³¹ Our results suggest that the target salt intake recommended in Japan is appropriate for GC prevention.

The association between low dietary fiber intake and GC incidence is controversial. Although some studies on SDF have indicated a relationship between low dietary fiber intake and GC, others have reported no relationship.^32-35 The association between dietary fiber intake and colorectal cancer incidence has been well-established in vitro and in vivo. Butyric acid is an important short-chain fatty acid that is generated from fermentable fibers and consists mostly of SDF produced by the colonic microbiota. It has been demonstrated that butyric acid has anti-tumor properties.³⁶ Regarding GC, although a few studies exist on its association with the microbiota, Matthews et al.³⁷ reported that butyric acid induces apoptosis and cell cycle alterations in GC cells in vivo. Moreover, we considered another mechanism involving low SDF intake as an inverse factor for GC from a metabolic perspective. Dietary fiber, particularly SDF from vegetables and fruits, can slow the absorption of starch, resulting in a reduced glycemic load and hyperinsulinemia.³⁸ Some studies have reported that the glycemic load and hyperinsulinemia contribute to GC development.^39-41 Thus, low SDF intake can be a risk factor for GC. Considering the type of dietary fiber, this result may elucidate the association between dietary fiber and GC.

Many studies have examined the relationship between VD and the incidence of GC. VD is converted into 25-hydroxyvitamin D (calcidiol) in the liver, which is converted into 1,25-dihydroxyvitamin D (1,25-(OH)2D, calcitriol) in the kidney.⁴² Preclinical studies in cells and animal models have indicated that 1,25-(OH)2D binds to the VD receptor and regulates the transcription of hundreds of genes with VD response elements, thereby mediating the cell cycle, proliferation, differentiation, and apoptosis, and exhibiting anti-tumor effects.^43-46 Regarding colorectal cancer, numerous studies have reported that VD inhibits or prevents carcinogenesis.^47,48 However, the results of epidemiological studies have been inconsistent, and no randomized controlled trials in humans have been conducted to definitively support the beneficial role of VD in GC. Several meta-analyses have reported no significant relationship between GC incidence and VD levels.⁴⁹ Other studies have found a positive association between VD intake and GC risk, as in this study.⁵⁰ In contrast, some studies have suggested that VD deficiency reduces the incidence and mortality rate of GC.⁵¹ The conflicting findings obtained in previous preclinical and epidemiological studies may be attributed to various factors involved in the in vivo VD pathway, such as sunlight exposure and various hormones. However, only VD values calculated from the FFQ were obtained in this study, and no values for other hormones involved in the VD pathway were available. Conversely, speculating further on the association between VD and GC is challenging; therefore, additional research is necessary for a more comprehensive understanding.

Regarding GC prevention, the World Cancer Research Fund/American Institute of Cancer Research report provides guidance to individuals on personal goals for health promotion and educates healthcare providers on evidence-based interventions for their patients.⁵² Although it concluded that no factors qualify as “convincing evidence factors” for the association between GC and dietary intake, the panel stated that several nutrition factors are “probable evidence factors” that play an important role in GC prevention and development. Among the three nutrition factors that were associated with GC incidence in this study, sodium was classified as a “probable evidence factor,” and VD and SDF were classified as “limited evidence factors.” Additionally, recommendations that salt and salt-preserved foods should not be eaten were included in the third report by World Cancer Research Fund/American Institute of Cancer Research.⁵³ Based on the results of our study, further intervention studies are desirable to clarify whether the limitation of salt and VD intake and active intake of SDF contribute to GC prevention.

The strength of our study is that the sample was recruited from the community, and the results are likely to be relevant in the general population. Additionally, we analyzed the data from a longitudinal follow-up study of 1,147 continuous cases that prospectively observed GC incidence. Moreover, to the best of our knowledge, this is the first exploratory study to examine the relationship between GC incidence and nutritional intake calculated using FFQ. However, this study had some limitations. First, it was conducted in a specific region of Japan, which may have introduced bias regarding the amount and type of nutritional intake. Second, the nutrient intake data used in this analysis were calculated using the FFQ, which has questions about what the participants had eaten during the entire year. Therefore, there may be some recall bias in the data. Furthermore, this dietary recall over the past 1 year may not accurately represent the continuous dietary pattern over 6-year period of follow-up.

In conclusion, we investigated the association between dietary nutrient intake and the incidence of GC. We found that higher sodium and VD intake and lower SDF intake were associated with the incidence of GC. We believe that our study provides definitive evidence for conducting future GC prevention studies.

Supplementary materials can be accessed at https://doi.org/10.5009/gnl230354.

No potential conflict of interest relevant to this article was reported.

Study concept and design: A.T., T.G., S.S. Data acquisition: C.K., S.S. Data analysis and interpretation: A.T., S.S., C.G., H.I. Drafting of the manuscript: A.T. Critical revision of the manuscript for important intellectual content: T.G., S.S. Statistical analysis: A.T., S.S. study supervision: T.G., H.K. Approval of final manuscript: all authors.