Li MT, Zheng Y, Zhao JY, Wu ZQ, Lu LZ, Sun J, Li Q, Shu XC, Naji WQ, Yang J, Chen FF, Ma XM, Zhang HY, Yang FC, Du LL, Xie YD, Du PW, Kong WR, Yuan H, Ji R, Guo QH, Chen ZF, Wang YP, Zhou YN. Associations between dietary polyphenols and risks of gastric precancerous lesions and cancer. World J Gastroenterol 2026; 32(20): 117603 [DOI: 10.3748/wjg.v32.i20.117603]
Corresponding Author of This Article
Yong-Ning Zhou, MD, Professor, Department of Gastroenterology, The First Hospital of Lanzhou University, No. 1 West Donggang Road, Lanzhou 730030, Gansu Province, China. zhouyn@lzu.edu.cn
Research Domain of This Article
Gastroenterology & Hepatology
Article-Type of This Article
Observational Study
Open-Access Policy of This Article
This article is an open-access article which was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
Baishideng Publishing Group Inc, 7041 Koll Center Parkway, Suite 160, Pleasanton, CA 94566, USA
Share the Article
Li MT, Zheng Y, Zhao JY, Wu ZQ, Lu LZ, Sun J, Li Q, Shu XC, Naji WQ, Yang J, Chen FF, Ma XM, Zhang HY, Yang FC, Du LL, Xie YD, Du PW, Kong WR, Yuan H, Ji R, Guo QH, Chen ZF, Wang YP, Zhou YN. Associations between dietary polyphenols and risks of gastric precancerous lesions and cancer. World J Gastroenterol 2026; 32(20): 117603 [DOI: 10.3748/wjg.v32.i20.117603]
Ma-Tu Li, Jin-Yu Zhao, Waleed Q Naji, Jing Yang, Fei-Fei Chen, Hui-Yun Zhang, Fu-Cheng Yang, Li-Li Du, Yan-De Xie, Pan-Wang Du, The First Clinical Medical College, Lanzhou University, Lanzhou 730030, Gansu Province, China
Ma-Tu Li, Ya Zheng, Qiang Li, Xiao-Chuang Shu, Waleed Q Naji, Jing Yang, Fei-Fei Chen, Xiao-Mei Ma, Hui-Yun Zhang, Hao Yuan, Rui Ji, Qing-Hong Guo, Zhao-Feng Chen, Yu-Ping Wang, Yong-Ning Zhou, Department of Gastroenterology, The First Hospital of Lanzhou University, Lanzhou 730030, Gansu Province, China
Ma-Tu Li, Ya Zheng, Jing Yang, Fei-Fei Chen, Xiao-Mei Ma, Hui-Yun Zhang, Rui Ji, Qing-Hong Guo, Zhao-Feng Chen, Yu-Ping Wang, Yong-Ning Zhou, Gansu Province Clinical Research Center for Digestive Diseases, The First Hospital of Lanzhou University, Lanzhou 730030, Gansu Province, China
Zheng-Qi Wu, Department of Gastroenterology, Gansu Wuwei Liangzhou Hospital, Wuwei 733000, Gansu Province, China
Lin-Zhi Lu, Department of Gastroenterology, Gansu Wuwei Tumor Hospital, Wuwei 733000, Gansu Province, China
Jie Sun, Department of Geriatrics (I), The First Hospital of Lanzhou University, Lanzhou 730030, Gansu Province, China
Wan-Ru Kong, Department of Infection Management, Gansu Provincial Hospital, Lanzhou 730030, Gansu Province, China
Co-corresponding authors: Yu-Ping Wang and Yong-Ning Zhou.
Author contributions: Wang YP and Zhou YN designed the research study; Li Q, Sun J, Shu XC, Ji R, Guo QH, Chen ZF, Wu ZQ, Li MT, and Zheng Y performed the research; Li MT, Zheng Y, Lu LZ, Yuan H, Kong WR, and Zhao JY analyzed the data; Li MT and Zheng Y wrote the manuscript; Chen FF, Yang J, Ma XM, Zhang HY, Naji WQ, Yang FC, Du LL, Xie YD, Li MT, Zheng Y, Yuan H, Kong WR, and Du PW reviewed and edited the manuscript; Zhou YN and Wang YP supervised the study; Li MT and Zheng Y contributed equally to this work as co-first authors; Zhou YN and Wang YP contributed equally to this work as co-corresponding authors; all authors have read and approved the final manuscript.
Supported by the Joint Scientific Research Fund of Gansu Province, No. 23JRRA1487; Lanzhou Science and Technology Program, No. 2023-1-19; Gansu Provincial Department of Education Innovation and Entrepreneurship Education Reform Project, No. 202314-16; and National Natural Science Foundation of China, No. 82160498.
Institutional review board statement: Approval for the study was granted by the Ethics Committee of the First Hospital of Lanzhou University (approval No. LDYYLL2012001).
Informed consent statement: All participants provided written informed consent in accordance with the Declaration of Helsinki.
Conflict-of-interest statement: The authors declare that they have no conflict of interest.
STROBE statement: The authors have read the STROBE Statement—a checklist of items, and the manuscript was prepared and revised according to the STROBE Statement-a checklist of items.
Data sharing statement: Technical appendix, statistical code, and dataset available from the corresponding author at zhouyn@lzu.edu.cn.
Corresponding author: Yong-Ning Zhou, MD, Professor, Department of Gastroenterology, The First Hospital of Lanzhou University, No. 1 West Donggang Road, Lanzhou 730030, Gansu Province, China. zhouyn@lzu.edu.cn
Received: December 12, 2025 Revised: January 20, 2026 Accepted: March 16, 2026 Published online: May 28, 2026 Processing time: 159 Days and 15.6 Hours
Abstract
BACKGROUND
While numerous studies have examined the associations between dietary polyphenol subclasses (i.e., flavonoids, phenolic acids, lignans, and stilbenes) and the risk of gastric cancer (GC), evidence regarding their relationships with gastric precancerous lesions (GPL) and their subtypes [i.e., chronic atrophic gastritis, intestinal metaplasia (IM), and low-grade dysplasia (LGD)] remains extremely limited.
AIM
To investigate the associations between dietary polyphenol subclasses and both GPL subtypes and GC in a high-risk population.
METHODS
This cross-sectional study utilized baseline data from the Wuwei Cohort. The intakes of dietary polyphenol subclasses were estimated using a food frequency questionnaire and the Phenol-Explorer database. Gastric diseases were diagnosed via endoscopic screening followed by pathological confirmation and further classified into three distinct groups: Normal control group, GPL group, and GC group. Logistic regression model and restricted cubic spline (RCS) analyses were employed to assess the associations between dietary polyphenol subclasses and the risks of GPL and GC.
RESULTS
Higher stilbenes intake was associated with lower GPL risk [odds ratio (OR) T3vsT1 = 0.78, 95% confidence interval (CI): 0.67-0.89]. For specific lesions, strong inverse associations were observed in IM and phenolic acids (ORT3vsT1 = 0.75, 95%CI: 0.62-0.91), lignans (ORT3vsT1 = 0.80, 95%CI: 0.66-0.96), and stilbenes (ORT3vsT1 = 0.58, 95%CI: 0.48-0.69). Flavonoids intake was associated with an increased risk of IM (ORperLog2 = 1.13, 95%CI: 1.02-1.25). RCS analyses revealed a reverse U-shaped relationship between flavonoids and LGD risk (P = 0.022). Regarding GC risk, an inverse association for stilbenes (ORT3vsT1 = 0.59, 95%CI: 0.42-0.84; ORperLog2 = 0.92, 95%CI: 0.86-0.97) was observed, and a positive association for flavonoids (ORperLog2 = 1.25, 95%CI: 1.02-1.55). However, these associations were attenuated and became non-significant after comprehensive covariate adjustment.
CONCLUSION
Specific dietary polyphenol subclasses, including stilbenes, lignans, and phenolic acids, are associated with reduced risks of GPL, particularly IM. Flavonoids intake may be associated with increased risks of GPL and GC.
Core Tip: This cross-sectional study based on a natural population cohort in an area with a high incidence of gastric cancer in China shows that dietary polyphenol subclasses, including stilbenes, lignans, and phenolic acids, are associated with reduced risks of gastric precancerous lesions such as chronic atrophic gastritis, intestinal metaplasia, and low-grade dysplasia. This finding highlights that dietary patterns may play an important role in the prevention of gastric cancer in Wuwei, an area with a high incidence of gastric cancer.
Citation: Li MT, Zheng Y, Zhao JY, Wu ZQ, Lu LZ, Sun J, Li Q, Shu XC, Naji WQ, Yang J, Chen FF, Ma XM, Zhang HY, Yang FC, Du LL, Xie YD, Du PW, Kong WR, Yuan H, Ji R, Guo QH, Chen ZF, Wang YP, Zhou YN. Associations between dietary polyphenols and risks of gastric precancerous lesions and cancer. World J Gastroenterol 2026; 32(20): 117603
Gastric cancer (GC) remains a major global health challenge, with an incidence ranking of fifth and a mortality ranking of fourth worldwide in 2022[1]. In China, the situation is more critical, with the third-highest mortality rate, underscoring an urgent public health challenge[2]. Since most GC cases are diagnosed at advanced stages with poor prognosis, effective prevention strategies should focus on intercepting the precancerous cascade. Gastric precancerous lesions (GPL), including chronic atrophic gastritis (CAG), intestinal metaplasia (IM), and low-grade dysplasia (LGD), represent a pivotal and intervenable stage in Correa’s cascade of gastric carcinogenesis, offering a critical window for early prevention and intervention[3].
Diet represents one of the most modifiable risk factors for GC. Plant-based diets rich in fruits, vegetables, and other plant foods have been associated with reduced risk, with their protective effects largely attributed to phytochemicals, particularly phenolic compounds[4-7]. These compounds, including flavonoids, phenolic acids, stilbenes, and lignans, demonstrate potential antioxidant and anti-inflammatory properties that may interfere with carcinogenesis. Experimental and review studies indicate that specific polyphenol subclasses, including flavonoids, phenolic acids, and stilbenes, influence key cancer-related processes such as oxidative stress, inflammation, cell proliferation, apoptosis, and metastasis, with evidence of antineoplastic activity in vitro and in vivo across multiple models of human cancers[8-10]. However, most previous studies have focused primarily on overall GC risk, and there is limited evidence regarding the effects of dietary polyphenols at the stage of GPL, particularly concerning specific polyphenol subclasses[11-13]. Furthermore, current evidence is limited regarding potential nonlinear dose-response relationships between polyphenol intake and GPL, and the modifying effects of factors such as Helicobacter pylori (H. pylori) infection remain poorly understood.
Therefore, this study utilized data from the Wuwei Cohort in Wuwei, a population-based screening cohort in a high-risk region for GC in northwestern China[14], to systematically examine the associations between main dietary polyphenol subclasses and the risks of both GC and histologically confirmed GPL, aiming to provide scientific evidence for dietary strategies targeting the early, potentially reversible stages of gastric carcinogenesis. By integrating detailed dietary assessment with histopathologically confirmed outcomes across multiple stages of GPL, this study provides a novel, stage-specific perspective on dietary prevention of GC and helps to fill current gaps in the literature.
MATERIALS AND METHODS
Study population
This cross-sectional study was based on the Wuwei Cohort, an ongoing population-based prospective study aiming to explore the relationships between dietary, lifestyle, and various other factors and the risk of GC. A total of 23346 eligible participants aged 35-70 years comprised the Wuwei Cohort. Detailed information about the Wuwei Cohort has been published elsewhere[14]. Among the cohort, 9509 participants with a pathological biopsy were included. Exclusions were made for 36 individuals with incomplete dietary data and 966 participants with reported total energy intake falling outside the plausible range (men: < 800 kcal/day or > 4000 kcal/day; women: < 500 kcal/day or > 3500 kcal/day)[15]. Subsequently, a total of 8507 participants were incorporated into the final analysis. The detailed filtering procedure is illustrated in Figure 1. Approval for the study was granted by the Ethics Committee of the First Hospital of Lanzhou University (approval No. LDYYLL2012001), and all participants provided written informed consent in accordance with the Declaration of Helsinki[16].
Figure 1 Flowchart of participant selection.
FFQ: Food frequency questionnaire.
Data collection
Ascertainment of covariates: Covariates were collected by trained staff via a face-to-face interview using self-reported general questionnaires, including age (continuous variable), sex (male or female), educational attainment (illiterate, primary school, secondary high school, or tertiary education and above), occupation (farmer or other), annual household income (low: < 20000 Chinese yuan, moderate: 20000-30000 Chinese yuan, or high: > 30000 Chinese yuan), smoking status (never, current, or former), alcohol consumption (yes or no), and family history of digestive cancer (yes or no). Body mass index (BMI) was calculated according to the formula weight (kg)/height squared (m2) and categorized according to the Chinese cut-points: BMI < 18.5 kg/m2 for underweight, 18.5 kg/m2 ≤ BMI < 24 kg/m2 for normal-weight, 24 kg/m2 ≤ BMI < 28.0 kg/m2 for overweight, and BMI ≥ 28.0 kg/m2 for obesity[17]. Daily energy intake was estimated in kcal using a food frequency questionnaire (FFQ). The H. pylori infection was detected by a 14C-urea breath test.
Dietary assessment: Dietary data were evaluated using an FFQ that included questions on the consumption frequency of food items categorized into ten distinct groups. Detailed descriptions of the Wuwei Cohort study design and characteristics of the participants have been provided elsewhere[14]. After enrollment, participants completed the FFQ under the face-to-face guidance of a qualified dietitian. In detail, participants were queried about their frequency of “never or once per month” to “two or more times per day”. Participants then provided information on the quantity of food consumed in grams per unit of time. Additionally, the frequency of food intake was categorized as modicum, moderate, or sufficient based on the average level of the local population. The energy intake of each food and beverage was calculated according to the Chinese food composition[18,19] and dietary data.
Estimation of dietary polyphenol intake: An advanced search was carried out in the Phenol-Explorer Database (http://www.phenol-explorer.eu/contents) to retrieve mean content values for all types of polyphenols contained in foods, and individual polyphenol intake from each food was calculated by multiplying the content of polyphenol by the daily consumption of each food. Ultimately, the daily intakes of main polyphenol subclasses were estimated, including flavonoids (chalcones, flavanols, isoflavonoids, anthocyanins, flavanols, flavones, and so on), phenolic acids (hydroxybenzoic acids, hydroxycinnamic acids, and hydroxy phenylacetic acids), lignans (lariciresinol, matairesinol, pinoresinol, secoisolariciresinol, medioresinol, and syringaresinol), and stilbenes (resveratrol, piceatannol, and d-viniferin). The classification of polyphenol compounds contained in polyphenol subclasses was referred to the Phenol-Explorer Database. All subclasses of polyphenols were evaluated as energy-adjusted densities, per 1000 kcal/day (4184 kJ/day)[20].
Case ascertainment: Gastroscopic examination was performed by trained physicians using conventional white light GIFH 260 or 290 gastroscopes (Olympus, Japan), and biopsy specimens were obtained from the gastric body, angulus, and antrum according to the study protocol adapted from the updated Sydney system[21]. If an obvious lesion was observed in other areas, additional biopsies were taken. Histological biopsy specimens were examined by a panel of experienced pathologists based on the World Health Organization classification[22]. The final diagnosis of cancer was made independently by three pathologists. In case of disagreement, a consensus diagnosis was sought by consultation. To be specific, the endoscopic pathological diagnosis encompassed a range of gastric mucosal conditions, including normal mucosa, superficial gastritis, CAG, IM, LGD, high-grade dysplasia, early GC, and advanced GC. These were further classified into three distinct groups: The normal control (NC) group comprising normal gastric mucosa and superficial gastritis, the GPL group consisting of CAG, IM, and LGD, and the GC group encompassing gastric adenocarcinoma and high-grade dysplasia. Due to the limited number of high-grade dysplasia, early-stage GC, and advanced-stage GC cases, these three pathological categories were combined into a single GC group for statistical analysis. High-grade dysplasia, which may share the same risk factors as cancer[23], was included in the GC group in this study.
Statistical analysis
Participants were stratified according to endoscopic diagnosis. Continuous variables with a normal distribution were described as mean ± SD[24], and those with a non-normal distribution were described as medians and interquartile ranges. Categorical variables were described as n (%). Between group differences of polyphenol subclasses were analyzed by the Kruskal-Wallis test followed by Dunn’s multiple comparison test. Correlations between polyphenol subclasses were evaluated using Pearson correlation analysis. Multivariable multinomial logistic regression was used to evaluate the associations between dietary polyphenols and the risks of GPL and GC, with the NC group as the reference. In this analysis, polyphenol intake was analyzed both as log2-transformed continuous variables and categorized into tertiles. Covariates were adjusted in a stepwise manner across three models: The crude model included energy-standardized polyphenol intake (per 1000 kcal/day); Model 1 further adjusted for sex, age, education level, occupation, annual household income, smoking status, alcohol consumption, BMI, H. pylori status, and family history of digestive cancers; Model 2 additionally included all other polyphenol subclasses to account for potential mutual confounding among different polyphenol types. Moreover, potential nonlinear relationships between dietary polyphenol intake and the risks of GPL and GC, including specific GPL subtypes such as CAG, IM, and LGD, were examined using restricted cubic splines (RCS). Three knots were placed at the 5th, 50th, and 95th percentiles of the log2-transformed polyphenol intake. The reference value [odds ratio (OR) = 1.00] was set at the median intake level (50th percentile).
To explore whether associations differ across populations, we evaluated the potential modification effects of age, gender, and H. pylori using a likelihood ratio test in subgroup analyses. In addition, we set dietary polyphenol intake to quartiles and performed multivariable logistic regression to confirm the stability of the results. All statistical analyses were performed using R version 4.2.2. A two-tailed P value < 0.05 was considered statistically significant for the hypothesis test.
RESULTS
Characteristics of study population
Table 1 shows the baseline characteristics of the study population. A total of 8507 individuals were included, with a mean age of 51.93 years and 54.2% being male. Among these, 1570 (18.45%) participants were diagnosed with normal or superficial gastritis (NC group), 6713 (78.91%) with GPL, and 224 (2.63%) with GC. Participants with GC or GPL tended to be older, male, farmers, current smokers, had lower education and household income levels, and were infected with H. pylori. Participants with GPL had the highest rates of farming (92.5%) and H. pylori infection (56.1%). Median daily energy intake was highest in the GC group (1687.5 kcal/day).
Table 1 Baseline participant characteristics and polyphenols intake by pathological diagnosis, mean ± SD/n (%).
Regarding dietary polyphenols, the medians of dietary flavonoids, phenolic acids, lignans, and stilbenes were 102.30 mg/1000 kcal, 76.76 mg/1000 kcal, 30.26 mg/1000 kcal, and 5.21 μg/1000 kcal per day, respectively (Table 1). Figure 2 describes the dietary polyphenol intake profiles across different gastric disease populations. Compared with the NC group, flavonoid intake was lower in the GPL group (Figure 2A), particularly in the LGD subgroup (Figure 2B). No significant between group differences were observed for phenolic acids or lignans intake (Figure 2C-F). Stilbene intake was generally reduced in both the GPL and GC groups (Figure 2G), including in the CAG, IM, and LGD subgroups (Figure 2H). These findings are largely consistent with the crude models (without covariate adjustment) in the subsequent logistic regression analysis. Furthermore, correlation analysis among polyphenol subclasses revealed weak positive correlations between different polyphenol subclasses (0 < r < 0.4, P < 0.001), except between phenolic acids and stilbenes (P > 0.05) (Figure 2I). This provides a rationale for adjusting polyphenol intake in the subsequent adjusted logistic regression model (model 2).
Figure 2 Distribution of dietary polyphenol intake across different gastric disease groups.
A-H: Violin plots show the estimated daily intake of four polyphenol subclasses: Flavonoids (A and B), phenolic acids (C and D), lignans (E and F), and stilbenes (G and H). The midline indicates the median; the upper and lower edges correspond to the first and third quartiles, respectively. Between group differences were assessed using the Kruskal-Wallis test followed by Dunn’s multiple comparison test; I: Heatmap of Pearson correlation coefficients among the four polyphenol subclasses. Correlations were evaluated using Pearson correlation analysis. aP < 0.05. bP < 0.01. cP < 0.001. P vs normal control group. Values in the heatmap represent the correlation coefficient r. NC: Normal control; GPL: Gastric precancerous lesions; GC: Gastric cancer; CAG: Chronic atrophic gastritis; IM: Intestinal metaplasia; LGD: Low-grade dysplasia.
Multivariate logistics regression analyses
Dietary polyphenols and risk of GPL:Table 2 shows the associations between dietary polyphenol subclasses (flavonoids, phenolic acids, lignans, and stilbenes) and the risks of GPL and GC. Intake was modeled both as tertiles [with T1 (lowest) as the reference] and as log2-transformed continuous variables. For the overall GPL outcome, inverse associations were observed for lignans and stilbenes, but not for flavonoids or phenolic acids. Specifically, higher stilbene intake was consistently associated with a lower risk of GPL across all models. In the fully adjusted model (model 2), the OR for the highest vs lowest tertile was 0.78 [95% confidence interval (CI): 0.67-0.89; P < 0.001], and each doubling of intake was associated with a 5% reduction in ORs (P < 0.001). Lignans also showed a significant inverse association with GPL risk in the categorical analysis (ORT3vsT1 = 0.83, 95%CI: 0.72-0.96; P = 0.010). However, when analyzed as a continuous variable, the association was attenuated and only marginally significant (ORperLog2 = 0.94, 95%CI: 0.89-1.00; P = 0.053).
Table 2 Odds ratios and 95% confidence intervals for the associations of dietary polyphenol intake with gastric precancerous lesions and gastric cancer.
We further analyzed the associations between dietary polyphenols and pathological subtypes of GPL, including CAG, IM, and LGD, as presented in Table 3. For CAG and LGD, stilbene intake showed inverse associations in the log2-transformed (CAG: OR = 0.97, 95%CI: 0.94-0.99, P = 0.040; LGD: OR = 0.95, 95%CI: 0.92-0.98, P = 0.002), although the highest tertile (T3) did not reach statistical significance for CAG after full adjustment. However, for LGD, the highest tertile (T3) in model 2 showed a significant inverse association (OR = 0.78, 95%CI: 0.66-0.94, P = 0.007), indicating a consistent trend in both continuous and categorical analyses. In contrast, for IM, stronger and more consistent inverse associations were observed. Participants in the highest tertile (T3) of phenolic acids, lignans, and stilbenes intake had substantially lower ORs of IM compared with those in the lowest tertile (T1). In the fully adjusted model (model 2), the ORs for T3 vs T1 were 0.75 (95%CI: 0.62-0.91, P = 0.004) for phenolic acids, 0.80 (95%CI: 0.66-0.96, P = 0.017) for lignans, and 0.58 (95%CI: 0.48-0.69, P < 0.001) for stilbenes, respectively. Consistent results were observed in the log2-transformed analyses, with each doubling of phenolic acids (OR = 0.78, 95%CI: 0.67-0.90), lignan (OR = 0.91, 95%CI: 0.85-0.98), and stilbenes (OR = 0.91, 95%CI: 0.87-0.94) intake being associated with lower ORs of IM. Notably, flavonoid intake was positively associated with IM risk in the log2-transformed model (OR = 1.13, 95%CI: 1.02-1.25, P = 0.017), while the association was not statistically significant in the tertile analysis (OR = 1.07, 95%CI: 0.89-1.28, P = 0.474).
Table 3 Odds ratios and 95% confidence intervals for the associations of dietary polyphenol intake with subtypes of gastric precancerous lesions.
Overall, higher intakes of stilbenes, phenolic acids, and lignans were inversely associated with the risks of GPL, particularly IM, whereas flavonoid intake was positively associated with IM risk.
Dietary polyphenols and risk of GC: Regarding the risk of GC (Table 2), an inverse association was observed between dietary stilbene intake and GC risk in the crude model. The highest tertile of stilbene intake was associated with lower ORs of GC (OR = 0.59, 95%CI: 0.42-0.84, P = 0.003), and each doubling of intake was associated with an 8% risk reduction (ORperLog2 = 0.92, 95%CI: 0.86-0.97, P = 0.006). However, these associations were attenuated and became non-significant after full adjustment for covariates. For flavonoids, a transient positive association was observed in model 1 (ORperLog2 = 1.25, 95%CI: 1.02-1.55, P = 0.034), but this association did not persist in the fully adjusted model. No significant associations were found for phenolic acids or lignans in either the crude or adjusted models. In summary, inverse associations between dietary stilbenes and GC risk were observed only in crude analyses, whereas no statistically significant associations remained after comprehensive covariate adjustment.
Dose-response relationships via RCS analysis
Figure 3A-H illustrates the dose-response relationships between continuous polyphenol intake and both GPL and GC after adjusting for covariates in model 2. For GPL, a statistically significant linear inverse association was observed for stilbene intake (P for overall = 0.001), which is consistent with the results presented in Table 2. No significant associations were found for flavonoids, phenolic acids, and lignans with either GPL or GC (all P for overall > 0.05). Further analysis using RCS revealed distinct dose-response patterns between dietary polyphenol subclasses and specific pathological subtypes of GPL (Figure 3I-T). For IM, significant linear inverse associations similar to the results in Table 3 were observed with phenolic acids (P for overall < 0.001), lignans (P for overall = 0.040), and stilbenes (P for overall < 0.001), which is consistent with the results in Table 2. Notably, the non-linear relationship for phenolic acids (P for nonlinear = 0.011) was characterized by a threshold effect around log2 > 6. Beyond this point, lower ORs of IM were observed. For LGD, significant associations were observed for both flavonoids and stilbenes. Flavonoids exhibited a non-linear relationship (P for nonlinear = 0.022) characterized by a reverse U-shaped curve, while stilbenes showed a significant linear inverse association (P for overall = 0.006). No significant associations were observed between any polyphenol subclass and CAG risk. Overall, these findings highlight the linear and non-linear protective effects of dietary phenolic acids, lignans, and stilbenes in subtypes of GPL, particularly for IM and LGD.
Figure 3 Restricted cubic spline curves.
A-H: For the relationships between dietary polyphenol intake and the risks of gastric precancerous lesions (GPL) and gastric cancer (GC). A: Flavonoids and GPL risk; B: Phenolic acids and GPL risk; C: Lignans and GPL risk; D: Stilbenes and GPL risk; E: Flavonoids and GC risk; F: Phenolic acids and GC risk; G: Lignans and GC risk; H: Stilbenes and GC risk; I-T: For the relationships between dietary polyphenol intake and the risks of subtypes of gastric precancerous lesions. I: Flavonoids and chronic atrophic gastritis (CAG) risk; J: Phenolic acids and CAG risk; K: Lignans and CAG risk; L: Stilbenes and CAG risk; M: Flavonoids and intestinal metaplasia (IM) risk; N: Phenolic acids and IM risk; O: Lignans and IM risk; P: Stilbenes and IM risk; Q: Flavonoids and low-grade dysplasia (LGD) risk; R: Phenolic acids and LGD risk; S: Lignans and LGD risk; T: Stilbenes and LGD risk. The adjusted odds ratios (solid lines) and 95% confidence intervals (shaded areas) were calculated based on restricted cubic spline models for log2-transformed dietary polyphenol intake. The median intake (50th percentile) was set as the reference value (odds ratio = 1.00), with knots placed at the 5th, 50th, and 95th percentiles. All models were adjusted for sex, age, education, occupation, income, smoking, alcohol consumption, body mass index, Helicobacter pylori status, family history of digestive cancer, and mutually for all other polyphenol subclasses. OR: Odds ratio; CI: Confidence interval.
Subgroup analysis
Subgroup analyses revealed significant effect modifications by age and H. pylori status (Figure 4A). Overall, flavonoid intake showed positive associations with the risks of GC (OR = 1.68, 95%CI: 1.18-2.41) and GPL (OR = 1.18, 95%CI: 1.05-1.32) only in H. pylori-positive individuals (P for interaction = 0.045 and 0.003, respectively), while lignans showed an inverse association with GPL risk only among H. pylori-negative individuals (OR = 0.91, 95%CI: 0.84-0.99; P for interaction = 0.022). In participants aged ≥ 50 years, flavonoid intake was positively associated with GPL (OR = 1.18, 95%CI: 1.06-1.32), whereas phenolic acid intake was inversely associated with GPL (OR = 0.77, 95%CI: 0.66-0.90) with a statistically significant interaction by age (P for interaction < 0.05).
Figure 4 Subgroup analysis of dietary polyphenols and the risks.
A: Of gastric cancer and gastric precancerous lesions; B: Of subtypes of gastric precancerous lesions. Odds ratios and 95% confidence intervals were calculated using the logistic regression models adjusting for sex, age, education level, occupation, annual household income, smoking status, alcohol consumption, body mass index, Helicobacter pylori status, and digestive cancer family history, and mutual adjustment by flavonoids, phenolic acids, lignans, and stilbenes. P values for interaction were calculated using a likelihood ratio test. OR: Odds ratio; CI: Confidence interval.
Further analysis of GPL subtypes revealed more detailed and subtype-specific association patterns (Figure 4B). Phenolic acid intake was inversely associated with both CAG (OR = 0.80, 95%CI: 0.68-0.96) and LGD (OR = 0.77, 95%CI: 0.63-0.93) among participants aged ≥ 50 years, with statistically significant interactions by age (P for interaction = 0.017 and 0.012, respectively). In contrast, flavonoid intake was positively associated with CAG (OR = 1.16, 95%CI: 1.03-1.31), IM (OR = 1.22, 95%CI: 1.06-1.42), and LGD (OR = 1.18, 95%CI: 1.02-1.36) exclusively among H. pylori-positive individuals, with significant interaction effects by H. pylori status (P for interaction = 0.006, 0.013, and 0.044, respectively). Conversely, lignan intake was inversely associated with IM only among H. pylori-negative individuals (OR = 0.86, 95%CI: 0.78-0.96, P for interaction = 0.01).
Sensitivity analysis
We set dietary polyphenol intake to quartiles to confirm the reliability of tertile logistic regression results (Tables 2 and 3) and present similar results (Table 4). After adjustment for covariates identical to those included in model 2, dietary phenolic acid intake was inversely associated with IM (ORQ4vsQ1 = 0.74, 95%CI: 0.59-0.93, P for trend = 0.003). Similarly, lignan intake was inversely associated with IM (ORQ4vsQ1 = 0.76, 95%CI: 0.61-0.94, P for trend = 0.012) and LGD (ORQ4vsQ1 = 0.80, 95%CI: 0.65-0.99, P for trend = 0.082). Stilbene intake showed inverse associations across all GPL subtypes, including CAG (ORQ4vsQ1 = 0.88, 95%CI: 0.74-1.06, P for trend = 0.031), IM (ORQ4vsQ1 = 0.58, 95%CI: 0.47-0.72, P for trend < 0.001), and LGD (ORQ4vsQ1 = 0.82, 95%CI: 0.67-1.00, P for trend = 0.001). In contrast, no significant associations were observed between dietary flavonoids and the risks of GPL or GC subtypes (all P for Q4 vs Q1 > 0.05, all P for trend > 0.05).
Table 4 Sensitivity analysis of the associations between dietary polyphenol and the risks of gastric cancer and gastric precancerous lesions based on quartiles of polyphenol intake.
Our study reveals distinct associations between dietary polyphenol subclasses and gastric pathologies, with more robust protective effects observed for GPL than GC. Specifically, stilbenes demonstrated a consistent inverse association with the overall and specific subtype GPL risks, supported by a linear dose-response relationship, while phenolic acids exhibited a non-linear, threshold-based protection against IM. Conversely, flavonoids may increase the risks of both GPL and GC.
A notable finding of this study was the consistent inverse association between stilbenes intake and the risks of GPL, particularly IM and LGD, which persisted across multiple adjustment models and subgroup analyses. The protective effects of stilbenes observed in our study are supported by existing mechanistic evidence. Specifically, resveratrol, a prominent stilbene, has been demonstrated to inhibit bile acid-induced gastric IM through the phosphatidylinositol 3-kinase/protein kinase B/p-FoxO4 signaling pathway[25], suggesting a potential reversal of gastric IM. Moreover, recent reviews have summarized that resveratrol may modulate key pathways implicated in gastric carcinogenesis, including inflammation- and oxidative stress-related signaling, supporting its biological plausibility as a chemo-preventive agent in the stomach[26,27]. While direct human evidence on gastric lesions remains limited, population-based studies have linked dietary polyphenol subclasses, including stilbenes, to GC risk, providing epidemiological context for our findings[11,28]. Taken together, these findings are consistent with the concept that natural products such as resveratrol may act along the Correa’s cascade, potentially intervening at early and potentially reversible stages of gastric lesion development[29]. Experimental evidence from other upper gastrointestinal models further supports this notion, as resveratrol supplementation has been reported to suppress the initiation of metaplasia and its progression to esophageal adenocarcinoma[30]. Accordingly, the linear dose-response pattern observed in our analysis further supports a potential protective role of stilbenes in mitigating early gastric lesions.
The non-linear, threshold effect observed for phenolic acids, particularly with respect to IM, is another critical finding. It suggests that a certain intake level (log2 > 6) must be reached to activate their protective mechanisms, which may be related to factors such as bioavailability and the metabolic capacity required for conversion into active compounds[31]. This interpretation is supported by a pharmacokinetic study in adults aged ≥ 50 years, which demonstrated that phenolic compounds derived from cranberry juice reached peak plasma concentrations only after 8-10 hours and exhibited substantial interindividual variability, thereby providing a plausible explanation for the intake threshold observed in our study[32]. Additionally, phenolic acids may confer protection by modulating gastric mucosal barrier function. Experimental evidence suggests that phenolic acid compounds, such as caffeic acid, can significantly influence the expression of mucins associated with IM, including MUC2, MUC13, and MUC17[33]. Moreover, hydroxycinnamic acids have been reported to help maintain intestinal epithelial barrier integrity by reducing oxidative stress and inflammation and modulating tight-junction-related pathways, while ferulic acid has been shown to suppress endoplasmic reticulum stress-related inflammatory responses in intestinal epithelial cells[34-37]. This coordinated regulation of mucosal and epithelial barrier function, together with the inhibitory effects of phenolic acids on the Wnt/β-catenin signaling pathway[38], may act synergistically to delay the progression from GPLs to GC[29].
Lignans intake was inversely associated with the risk of IM. In our analysis, tertile-based comparisons indicated approximately a 20% risk reduction in the highest intake group, while continuous analyses showed a 9% reduction in IM risk per two-fold increase in intake. A dose-response analysis further confirmed a linear association. Although considerable heterogeneity exists across existing literature on this topic, evidence from biomarkers of internal exposure supports the anti-inflammatory potential of lignans. Specifically, data from the National Health and Nutrition Examination Survey documented significant inverse correlations between circulating enterolignan levels and markers of chronic inflammation, particularly C-reactive protein[39]. While urinary or circulating biomarker measurements reflect metabolized lignans and differ from dietary intake estimates, they similarly suggest the biological activity of lignans in humans. Notably, the protective association between lignans intake and IM observed in our study appeared to be population-specific, as it was significant only among H. pylori-negative individuals. This finding is consistent with previous reports describing population-specific protective associations of lignans[40]. One possible explanation for this specificity lies in the critical role of gut microbiota in converting dietary lignans into their bioactive enterolignans, which is essential for lignan bioactivity in humans[41]. As H. pylori infection has been shown to alter gastrointestinal microbial ecology[42,43], it may interfere with this bioactivation process, thereby attenuating the protective effects of lignans in infected individuals.
The relationship between dietary flavonoids and gastric pathologies appeared to be complex, with the most consistent finding being effect modification by H. pylori infection status across all endpoints. Subgroup analyses revealed that the observed positive associations between flavonoids and gastric lesions were predominantly confined to H. pylori-positive individuals, which may reflect altered flavonoid bioactivity within the infected gastric microenvironment[44,45]. H. pylori infection is known to induce persistent oxidative stress and chronic inflammation in the gastric mucosa[46], resulting in a microenvironment enriched in reactive oxygen species and transition metal ions. This inflammatory and oxidative milieu may substantially alter the redox balance and chemical stability of dietary polyphenols, thereby affecting their biological activity within the gastric mucosa[47]. Accordingly, dietary flavonoids may undergo autoxidation or enzymatic oxidation, leading to a shift from antioxidant to prooxidant activity[48-51]. Such prooxidant effects of flavonoids have been reported to exacerbate oxidative damage under conditions of chronic inflammation, particularly in the presence of transition metal ions, which may contribute to mucosal injury[52,53]. In addition, plant-based polyphenols, including flavonoids, have been shown to influence H. pylori infection and modulate gut microbial ecology, highlighting their context-dependent effects on bacterial colonization and gastric inflammation[54]. Collectively, this well-recognized dual behavior of polyphenols provides a biologically plausible explanation for our findings, whereby compounds that may exert protective effects in a relatively normal gastric environment could have adverse effects in the context of H. pylori associated pathological alterations.
Although the association between stilbenes and GC became non-significant after full adjustment in our study, the substantial effect size we observed in the crude model (OR = 0.59, 95%CI: 0.42-0.84) aligns remarkably with findings from the MCC-Spain study, which reported a statistically significant and stronger inverse association (ORQ4vsQ1 = 0.470)[28], supporting the biological plausibility of a protective role of dietary stilbenes in gastric carcinogenesis. Several meta-analyses have suggested an overall inverse association between dietary polyphenol intake and GC risk, while pooled evidence from the Stomach Cancer Pooling Project highlights heterogeneity across populations and exposure patterns[11,55]. Against this background, the lack of statistical significance observed in our fully adjusted models may reflect limited statistical power, given the relatively small number of GC cases, as well as attenuation due to extensive covariate adjustment, dietary measurement error, and residual confounding commonly encountered in nutritional epidemiology[56-58]. Notably, the null finding for polyphenols and GC contrasts with the significant and consistent inverse associations observed for GPL within the same population. This stage-specific pattern suggests that gastric mucosal alterations during early phases of carcinogenesis may be more biologically susceptible to dietary modulation[59], whereas the influence of dietary factors may diminish once malignant transformation has occurred[60]. The attenuated association with advanced cancer may further reflect both the reduced impact of dietary exposures in established malignancies and the accumulation of complex genetic and epigenetic alterations that characterize late-stage disease[61]. Although reverse causality cannot be entirely ruled out due to the cross-sectional design, dietary modifications related to subclinical symptoms or occurring around the time of disease detection may contribute to exposure misclassification and attenuation of associations for overt GC[62].
We acknowledge some clear strengths of this study. First, the analysis utilized data from the Wuwei Cohort, a large, systematically sampled population from a region with one of the highest GC incidence rates in China, providing a valuable context for investigating dietary risk factors. Second, the endoscopic and histopathological confirmation of all gastric lesions by multiple pathologists ensures robust case identification and classification. Most importantly, this is among the first studies to comprehensively evaluate the associations of multiple specific polyphenol subclasses with the spectrum of GPL, providing novel insights into the early stages of gastric carcinogenesis. Finally, the application of multiple analytical models and adjustment for key covariates, including socioeconomic factors, enhances the reliability of our findings. Several limitations of this study should be acknowledged. First, the cross-sectional design based on baseline data limits the ability to establish temporal relationships and precludes causal inference. Second, dietary polyphenol intake was estimated using FFQ data rather than objective biomarkers, which may introduce recall bias and does not fully capture interindividual variation in bioavailability related to food processing, cooking methods, and metabolism[24]. Third, although socioeconomic status was adjusted for using household income, residual confounding related to unmeasured aspects of socioeconomic status and healthcare access cannot be entirely excluded. Fourth, the high correlations among phytochemicals present in plant-based foods complicate the disentanglement of the independent effects of specific polyphenol subclasses. Finally, the observed associations should be interpreted as hypothesis-generating rather than conclusive evidence of causal relationships. Future analyses using longitudinal follow-up data from the ongoing Wuwei Cohort are warranted to clarify temporal relationships between dietary polyphenol intake and gastric carcinogenesis. In addition, biomarker-based exposure assessments and mechanistic studies incorporating gut microbiota and H. pylori status may help elucidate the underlying biological pathways. Multi-center studies in other regions and populations will be needed to evaluate the generalizability of these findings.
CONCLUSION
This study highlights the protective role of dietary stilbenes against GPL, supporting a potential role of stilbene intake in mitigating early gastric mucosal pathological progression. In contrast, associations between other polyphenol subclasses and gastric lesions appeared to be more heterogeneous and varied across different gastric pathological stages, suggesting that distinct polyphenols may play different roles during gastric carcinogenesis. Notably, the more consistent and robust associations observed for precancerous lesions, compared with GC, support the concept that dietary polyphenols may exert their primary effects during the early and potentially reversible stages of gastric carcinogenesis. Although the cross-sectional nature of the present analysis precludes definitive causal inference, our findings provide population-based epidemiological evidence from a GC high-risk setting linking specific polyphenol subclasses to gastric mucosal pathological changes. Overall, these results contribute to a more nuanced understanding of the role of dietary polyphenols in gastric carcinogenesis and suggest that dietary strategies targeting the precancerous stage may hold promise for GC prevention. Further confirmation using longitudinal follow-up data and mechanistic studies is warranted to validate these findings and clarify the underlying biological pathways.
ACKNOWLEDGEMENTS
All authors express gratitude to all participants in the Wuwei Cohort for their invaluable contribution to the study.
Ji R, Zhang Z, Zhang J, Wu Z, Li M, Ye Y, Li X, An F, Xu C, Lu L, Fan P, Chen Z, Guo Q, Wang Y, Wang X, Li Y, Zhang J, Guan Q, Hu X, Yang T, Liu J, Qiao L, Zheng Y, Zhou Y. Cohort Profile: A population-based cohort for the study of gastric cancer in northwest area of China (Wuwei Cohort).Int J Epidemiol. 2021;50:1433-1442.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 6][Reference Citation Analysis (0)]
Chen Z, Zheng Y, Fan P, Li M, Liu W, Yuan H, Liu X, Zhang Z, Wu Z, Wang Y, Ji R, Guo Q, Ye Y, Zhang J, Li X, An F, Lu L, Li Y, Wang X, Zhang J, Guan Q, Li Q, Liu M, Ren Q, Hu X, Lu H, Zhang H, Zhao Y, Gou X, Shu X, Wang J, Hu Z, Xue S, Liu J, Zhou Y. Risk factors in the development of gastric adenocarcinoma in the general population: A cross-sectional study of the Wuwei Cohort.Front Microbiol. 2022;13:1024155.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 14][Reference Citation Analysis (0)]
Zhou BF; Cooperative Meta-Analysis Group of the Working Group on Obesity in China. Predictive values of body mass index and waist circumference for risk factors of certain related diseases in Chinese adults--study on optimal cut-off points of body mass index and waist circumference in Chinese adults.Biomed Environ Sci. 2002;15:83-96.
[PubMed] [DOI]
Zamora-Ros R, Agudo A, Luján-Barroso L, Romieu I, Ferrari P, Knaze V, Bueno-de-Mesquita HB, Leenders M, Travis RC, Navarro C, Sánchez-Cantalejo E, Slimani N, Scalbert A, Fedirko V, Hjartåker A, Engeset D, Skeie G, Boeing H, Förster J, Li K, Teucher B, Agnoli C, Tumino R, Mattiello A, Saieva C, Johansson I, Stenling R, Redondo ML, Wallström P, Ericson U, Khaw KT, Mulligan AA, Trichopoulou A, Dilis V, Katsoulis M, Peeters PH, Igali L, Tjønneland A, Halkjær J, Touillaud M, Perquier F, Fagherazzi G, Amiano P, Ardanaz E, Bredsdorff L, Overvad K, Ricceri F, Riboli E, González CA. Dietary flavonoid and lignan intake and gastric adenocarcinoma risk in the European Prospective Investigation into Cancer and Nutrition (EPIC) study.Am J Clin Nutr. 2012;96:1398-1408.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 84][Cited by in RCA: 69][Article Influence: 4.9][Reference Citation Analysis (0)]
Rubín-García M, Vitelli-Storelli F, Molina AJ, Zamora-Ros R, Aragonés N, Adarnaz E, Castaño-Vinyals G, Obón-Santacana M, Gómez-Acebo I, Molina-Barceló A, Fernández-Tardón G, Jiménez-Moleón JJ, Alguacil J, Chirlaque MD, Toledo E, Pérez-Gómez B, Pollán M, Kogevinas M, Martín V. Association between Polyphenol Intake and Gastric Cancer Risk by Anatomic and Histologic Subtypes: MCC-Spain.Nutrients. 2020;12:3281.
[RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)][Cited by in Crossref: 3][Cited by in RCA: 12][Article Influence: 2.0][Reference Citation Analysis (0)]
Marchica A, Havlik J, Doskocil I, Hroncova Z, Volstatova T, Bernardi R.
Effects of selected dietary polyphenols on mucin expression in epithelial cells. Proceedings of the LX SIGA Annual Congress; 2016 Sep 13-16; Catania, Italy.
[PubMed] [DOI]
Ma C, Liang Z, Wang Y, Luo H, Yang X, Yao B, Tu T. p-Hydroxycinnamic Acids: Advancements in Synthetic Biology, Emerging Regulatory Targets in Gut Microbiota Interactions, and Implications for Animal Health.J Nutr. 2025;155:1041-1056.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 1][Reference Citation Analysis (0)]
Vitelli-Storelli F, Rossi M, Pelucchi C, Rota M, Palli D, Ferraroni M, Lunet N, Morais S, López-Carrillo L, Zaridze DG, Maximovich D, Rubín García M, Castaño-Vinyals G, Aragonés N, Garcia de la Hera M, Hernández-Ramírez RU, Negri E, Bonzi R, Ward MH, Lagiou A, Lagiou P, López-Cervantes M, Boffetta P, Camargo MC, Curado MP, Zhang ZF, Vioque J, La Vecchia C, Martín Sánchez V. Polyphenol Intake and Gastric Cancer Risk: Findings from the Stomach Cancer Pooling Project (StoP).Cancers (Basel). 2020;12:3064.
[RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)][Cited by in Crossref: 5][Cited by in RCA: 19][Article Influence: 3.2][Reference Citation Analysis (0)]
Correa P. Human gastric carcinogenesis: a multistep and multifactorial process--First American Cancer Society Award Lecture on Cancer Epidemiology and Prevention.Cancer Res. 1992;52:6735-6740.
[PubMed] [DOI]
P-Reviewer: Pădureanu V, MD, Postdoctoral Fellow, Romania; She XK, PhD, Postdoctoral Fellow, United States S-Editor: Fan M L-Editor: A P-Editor: Lei YY
