Abstract
Results from case-control and cohort studies of the association between fruit and vegetable consumption and gastric cancer risk have been inconsistent. Cases for the current study consisted of incident distal gastric cancers identified between 1996 and 2007 among members of the Shanghai Women's Health Study (n = 206) and the Shanghai Men's Health Study (n = 132). Intakes of fruits, vegetables, and select micronutrients were assessed on the basis of validated food frequency questionnaires. Multivariate-adjusted hazards ratios and 95% confidence intervals were calculated by Cox proportional hazards regression. For women, no associations were found between gastric cancer risk and the highest intake of fruits (hazard ratio (HR) = 1.02, 95% confidence interval (CI): 0.68, 1.54; Ptrend = 0.87) or vegetables (HR = 0.89, 95% CI: 0.60, 1.31; Ptrend = 0.32). For men, increased fruit intake was associated with decreased risk of distal gastric cancer (for the highest quartile of intake, HR = 0.50, 95% CI: 0.29, 0.84; Ptrend = 0.004), but no association was seen with increased intake of vegetables (HR = 1.00, 95% CI: 0.59, 1.68; Ptrend = 0.87). The inverse association with fruit intake for men was more evident among ever smokers (Ptrend = 0.001) than never smokers (Ptrend = 0.67). No associations between dietary intakes of select antioxidant micronutrients were seen for men or women. Fruit intake is inversely associated with distal gastric cancer risk among men in Shanghai, China.
Keywords: diet, fruit, micronutrients, stomach neoplasms, vegetables
Gastric cancer is the fourth most common malignancy in the world, although the incidence rate has been on the decline for the past century (1, 2). In China today, gastric cancer is second only to lung cancer as the most frequent cancer, and it ranks third in cause of death from cancer (1). Although Helicobacter pylori, a Gram-negative microaerophilic spiral bacterium that colonizes the gastric lumen, is the strongest known risk factor for gastric cancer, the vast majority of the population infected with H. pylori never develop gastric cancer, indicating a prominent role for other risk factors (3).
The association between fruit and vegetable consumption and risk of gastric cancer has been explored in numerous studies, in both high- and low-risk countries. Case-control studies from Europe, Asia, and North America have consistently found intakes of both fruits and vegetables to be protective against gastric cancer, reducing risk by approximately 40% for fruits and 30% for vegetables for the highest versus lowest categories of intake, respectively (4–6). Case-control studies, however, face the challenges of recall and selection bias. Cohort studies on this subject, on the other hand, have been less consistent. In 2005, a meta-analysis found overall weaker associations in the reduction of gastric cancer risk, from 18% for high intake of fruits to 12% for high intake of vegetables (7). Since the publication of that meta-analysis, 5 new prospective cohort analyses of fruit and vegetable consumption and gastric cancer risk have been published; results from these studies were not entirely consistent (8–12). A recent report from an Expert Panel of the World Cancer Research Fund released in 2007 concluded that fruit and nonstarchy vegetables “probably protect against gastric cancer” (13, p. 265).
The present analysis seeks to further our understanding of the association between fruit and vegetable consumption and distal gastric cancer risk by examining this question in 2 large, population-based, prospective cohorts—the Shanghai Women's Health Study and the Shanghai Men's Health Study. These rich resources allow for a detailed investigation, including the adjustment of multiple confounders, of the relation of risk of distal gastric cancer with overall consumption of fruits and vegetables, specific groups of these food items, individual fruits and vegetables, and select antioxidant micronutrients.
MATERIALS AND METHODS
Study population
The Shanghai Women's Health Study (SWHS) and the Shanghai Men's Health Study (SMHS) are both population-based, prospective cohort studies based in urban Shanghai, China. The SWHS was begun first, having recruited 74,942 women aged 40–70 years from 1996 to 2000 (14). The SMHS recruited 61,500 men aged 40–74 years from 2002 to 2006 (15). For members of both cohorts, trained interviewers administered detailed in-person interviews at baseline, using a structured questionnaire to collect information on demographic characteristics, dietary habits, physical activity, disease and surgery history, personal habits (including smoking, alcohol consumption, tea consumption, and so on), residential history, occupational history, family cancer history, and reproductive history and hormone use (for women only). Body measurements, including height, weight, and waist and hip circumferences, were also taken at baseline. Previous publications have described information on the baseline methods in more detail for the SWHS (14) and the SMHS (15).
Dietary assessment
At baseline, a comprehensive food frequency questionnaire was administered in person for members of both cohorts. The 81-item SMHS food frequency questionnaire was adapted from the 77-item SWHS food frequency questionnaire with the following modifications: 1) The SWHS food frequency questionnaire includes sweet potatoes, and the SMHS food frequency questionnaire does not; 2) the SWHS food frequency questionnaire combines fresh soybeans, fresh peas, and fresh broad beans into one category, and the SMHS food frequency questionnaire collects information on these items separately; and 3) the SMHS includes the additional food items, not included in the SWHS, of garland chrysanthemum, shepherd's purse, clover, and amaranth (all cruciferous vegetables), pea shoots, and pig's ham hock. Therefore, the food frequency questionnaires used in these 2 studies are virtually identical. Both the SWHS food frequency questionnaire and the SMHS food frequency questionnaire have been validated (16, 17). For each individual food item or group of foods, participants were first asked how frequently they consumed the food or food group (daily, weekly, monthly, annually, or never) during the past year. Then they were asked the amount of that food item they typically consumed per unit of time, in liang (1 liang = 50 g). For seasonal food consumption in the SWHS, a subset of women (approximately 1,000 per season) were asked about the number of months out of the year that food item was eaten, and then those results were used to calibrate intake of seasonal foods for all SWHS members. In the SMHS, all men were asked how many months of the year each food item was eaten. To calculate individual nutrient intakes, the amount of each food item was multiplied by the nutrient content per gram of the food as obtained from the Chinese Food Composition Tables (18).
Case identification
Incident gastric cancer cases were identified through a combination of annual record linkage with the Shanghai Cancer Registry and the Shanghai Vital Statistics database and biennial in-home interviews. For the SWHS, the in-person follow-up rates for the first (2000–2002), second (2002–2004), and third (2004–2007) surveys were 99.8%, 98.7%, and 96.7%, respectively. For the SMHS, the first follow-up (2004–2008) has just been completed, with a response rate of 97.6%. After identification through registry matching, all possible incident cancer cases were verified by conducting home visits and reviewing medical charts from the diagnostic hospital. Of the 373 gastric cancer cases identified, 338 (or 90.6%) were distal gastric cancer, defined as having an International Classification of Diseases for Oncology (ICD-O) code 161–166, 168, or 169 vs. gastric cardia cancer (ICD-O code 160). To be pure, a subanalysis was performed excluding the 21% of cases (70/338) with ICD-O code 168 or 169, which corresponds to “overlapping lesion of the stomach” and “stomach not otherwise specified.” In summary, the present study included all incident distal (i.e., noncardia) gastric cancer cases diagnosed between the date of baseline enrollment and December 2007 (SWHS, n = 206; SMHS, n = 132).
Statistical analysis
For the present study, we excluded individuals with a prior history of cancer (SWHS, n = 1,576; SMHS, n = 0, as this was part of the exclusion criteria to participation in the SMHS); a history of a gastrectomy (SWHS, n = 239; SMHS, n = 1,323); or a report of an implausible total energy intake, defined as an average daily energy intake of <500 or >4,000 kcal (SWHS, n = 50; SMHS, n = 90). Additionally, after determining that educational level was a confounder in the association between intakes of fruits and vegetables and gastric cancer risk, we excluded cohort members who were missing information on the highest level of education attained (SWHS, n = 13; SMHS, n = 840). The resulting cohort population used in these analyses consists of 73,064 women and 59,247 men. All analyses are stratified by sex as the intake levels of fruits and vegetables vary substantially between the men and women in the study. Additionally, for the main association of fruit and gastric cancer risk, an interaction by sex was suggested (Pinteraction = 0.06).
To create categories of consumption of fruits (all fruit, all fruit excluding watermelon, citrus fruit, and noncitrus fruit), vegetables (all vegetables, cruciferous vegetables, green leafy vegetables, allium vegetables, legumes, and other vegetables), and select micronutrients (vitamin A, vitamin C, vitamin E, carotene, retinol, selenium, and folic acid), we used baseline dietary information to create quartiles of intake based on the distribution in each cohort as a whole. All fruit excluding watermelon was created as a separate group because watermelon, although eaten seasonally only, accounts for 44%–52% of total fruit intake, and the quantification of this fruit eaten in large quantities can be difficult to assess, potentially leading to significant exposure measurement error. (Intake of watermelon alone was not associated with gastric cancer risk.) Confounding variables, chosen because they were a priori believed to be confounders and because they were associated with both the exposure and the outcome in the data, included age, education, smoking, and total energy intake. For the models used with the SWHS, the age at cohort entry was adjusted for continuously, the highest level of education attained was adjusted for categorically (elementary school or less, junior high, high school, and professional education or above), smoking was adjusted for as a binary variable (ever/never), and total energy intake was adjusted for continuously (in average daily kilocalories). For the SMHS, all confounders were adjusted in the same manner except for smoking, which was categorized as not a current smoker, current smoker of <20 cigarettes a day, and current smoker of ≥20 cigarettes a day. Adjusting for smoking in this more detailed way in the SMHS yielded a meaningful difference in the odds ratios compared with the binary categorization used in the SWHS, in which only 3% of subjects ever smoked cigarettes regularly. More detailed adjustment for smoking did not affect the odds ratios for the main risk factors.
Cox proportional hazards regression was used to compute hazard ratios and 95% confidence intervals, with age as the time metric. Age at the date of completion of the baseline questionnaire was the definition of entry age, and exit age was defined as the age at the first of the following outcomes: 1) cancer diagnosis; 2) death; or 3) last date of follow-up. Indicator variables were entered into the models to represent quartiles of intake, using the lowest quartile as the reference category. To test for a linear trend across levels of consumption of each food group or nutrient, a variable was created and assigned the median value of consumption for each quartile of food group or nutrient. Schoenfeld residuals were assessed to test the proportional hazards assumption. To evaluate the potential of effect modification, we also analyzed the association of fruits, vegetables, and nutrients with distal gastric cancer risk in separate models stratified by smoking status and educational level. All analyses were conducted using SAS, version 9.2, software (SAS Institute, Inc., Cary, North Carolina).
RESULTS
Compared with all members in their respective cohorts, the women and men diagnosed with distal gastric cancer in this study were older, more likely to have ever smoked, more likely to regularly be physically active, and less likely to have obtained at least a high school education (Table 1). In the SWHS, the cases were more likely than the cohort as a whole to have a first-degree family history of gastric cancer, but in the SMHS, cases were less likely than the average cohort member to have a family history of gastric cancer.
Table 1.
Baseline Demographic Characteristics of the Shanghai Women's (1996–2000) and Men's (2002–2006) Health Studies, Comparing the Entire Cohorts With Distal Gastric Cancer Cases
| Women |
Men |
|||||||
| Cohort (n = 73,064) |
Cases (n = 206) |
Cohort (n = 59,247) |
Cases (n = 132) |
|||||
| Mean (SD) | % | Mean (SD) | % | Mean (SD) | % | Mean (SD) | % | |
| Age, years | 52.5 (9.1) | 58.2 (9.0) | 55.2 (9.7) | 62.1 (9.8) | ||||
| Body mass index, kg/m2 | 24.0 (3.4) | 24.5 (3.4) | 23.8 (3.1) | 23.7 (2.9) | ||||
| Ever smoked | 2.7 | 5.8 | 69.5 | 73.5 | ||||
| Regular alcohol use | 2.2 | 2.4 | 33.7 | 34.9 | ||||
| Regular physical activity | 35.0 | 45.2 | 35.4 | 40.9 | ||||
| Energy intake, kcal/day | 1,676 (399) | 1,634 (412) | 1,907 (476) | 1,925 (459) | ||||
| Highest educational level achieved | ||||||||
| Elementary school or less | 21.3 | 43.7 | 6.6 | 14.4 | ||||
| Junior high school | 37.2 | 26.7 | 33.5 | 42.4 | ||||
| High school | 27.9 | 19.4 | 36.1 | 25.0 | ||||
| Professional education or above | 13.5 | 10.2 | 23.8 | 18.2 | ||||
| Family history of gastric cancer | 5.8 | 8.3 | 6.2 | 2.3 | ||||
Abbreviation: SD, standard deviation.
Fruit intake, total or categorized as citrus-only or noncitrus, was not associated with distal gastric cancer risk for women in this study (Table 2). For men, however, increasing intake of all fruits except watermelon led to decreasing risks of distal gastric cancer (Ptrend = 0.004), with a 50% (95% confidence interval (CI): 0.29, 0.84) reduction observed for men in the fourth quartile, compared with men in the lowest, first quartile of fruit intake. Comparing citrus with noncitrus fruits did not explain the apparent protection offered by total fruit intake, as neither was significantly associated with gastric cancer risk in men. Individual types of fruit—in the categories as they appeared on the questionnaire: apples; pears; tangerines, oranges, grapefruits; bananas; grapes; watermelon; peaches; other fruits (e.g., strawberries, cantaloupe)—were each associated with a reduced distal gastric cancer risk in men. Only apples, however, showed a significant association with risk at P ≤ 0.05 (data not shown). However, the protection offered by apples did not fully explain the all-fruit association. Excluding male cases diagnosed with distal gastric cancer within 1 year of enrollment (n = 45) did not change the significant inverse association found with fruits excluding watermelon (for men in the fourth quartile, the hazard ratio (HR) = 0.48, 95% CI: 0.25, 0.93; Ptrend = 0.01). These detailed and subgroup analyses, however, did not reveal any apparent association of fruit intake with gastric cancer risk in women.
Table 2.
Hazard Ratios of Distal Gastric Cancer by Quartiles of Fruit Group Intake in the Shanghai Women's and Men's Health Studies, 1996–2008
| Q1 (Referent) |
Q2 |
Q3 |
Q4 |
Ptrend | ||||||||||||
| No. of Cases | Range, g/day | Hazard Ratio | No. of Cases | Range, g/day | Hazard Ratio | 95% CI | No. of Cases | Range, g/day | Hazard Ratio | 95% CI | No. of Cases | Range, g/day | Hazard Ratio | 95% CI | ||
| Women | ||||||||||||||||
| All fruits | 67 | ≤134.2 | 46 | >134.2–238.3 | 51 | >238.3–357.8 | 42 | >357.8 | ||||||||
| Age-adjusted results | 1.00 | 0.82 | 0.56, 1.20 | 0.97 | 0.67, 1.41 | 0.88 | 0.59, 1.30 | 0.68 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.88 | 0.60, 1.29 | 1.07 | 0.73, 1.57 | 0.98 | 0.65, 1.49 | 0.89 | ||||||||
| All fruits (except watermelon) | 66 | ≤61.5 | 48 | >61.5–126.1 | 48 | >126.1–208.0 | 44 | >208.0 | ||||||||
| Age-adjusted results | 1.00 | 0.86 | 0.59, 1.25 | 0.91 | 0.62, 1.32 | 0.90 | 0.61, 1.34 | 0.64 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.92 | 0.63, 1.34 | 1.00 | 0.68, 1.47 | 1.02 | 0.68, 1.54 | 0.87 | ||||||||
| Citrus fruitb | 64 | ≤6.1 | 49 | >6.1–17.7 | 51 | >17.7–31.9 | 42 | >31.9 | ||||||||
| Age-adjusted results | 1.00 | 0.94 | 0.65, 1.37 | 0.95 | 0.65, 1.37 | 0.84 | 0.57, 1.25 | 0.44 | ||||||||
| Fully adjusted resultsa | 1.00 | 1.00 | 0.68, 1.46 | 1.05 | 0.71, 1.53 | 0.94 | 0.62, 1.42 | 0.86 | ||||||||
| Noncitrus fruitc | 67 | ≤121.6 | 50 | >121.6–217.6 | 44 | >217.6–328.2 | 45 | >328.2 | ||||||||
| Age-adjusted results | 1.00 | 0.89 | 0.62, 1.29 | 0.84 | 0.57, 1.23 | 0.94 | 0.64, 1.39 | 0.64 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.95 | 0.66, 1.38 | 0.92 | 0.62, 1.37 | 1.05 | 0.70, 1.58 | 0.92 | ||||||||
| Men | ||||||||||||||||
| All fruits | 41 | ≤56.5 | 35 | >56.5–128.3 | 25 | >128.3–215.7 | 31 | >215.7 | ||||||||
| Age-adjusted results | 1.00 | 0.79 | 0.51, 1.25 | 0.58 | 0.35, 0.95 | 0.69 | 0.43, 1.11 | 0.07 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.88 | 0.56, 1.39 | 0.66 | 0.40, 1.11 | 0.79 | 0.48, 1.30 | 0.23 | ||||||||
| All fruits (except watermelon) | 46 | ≤20.1 | 35 | >20.1–55.1 | 26 | >55.1–104.2 | 25 | >104.2 | ||||||||
| Age-adjusted results | 1.00 | 0.70 | 0.45, 1.09 | 0.50 | 0.31, 0.80 | 0.45 | 0.28, 0.74 | 0.0004 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.75 | 0.48, 1.18 | 0.55 | 0.33, 0.90 | 0.50 | 0.29, 0.84 | 0.004 | ||||||||
| Citrus fruitb | 39 | ≤1.6 | 30 | >1.6–≤6.3 | 39 | >6.3–≤18.0 | 24 | >18.0 | ||||||||
| Age-adjusted results | 1.00 | 0.77 | 0.48, 1.24 | 0.90 | 0.58, 1.40 | 0.60 | 0.36, 1.01 | 0.11 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.84 | 0.52, 1.36 | 1.02 | 0.65, 1.62 | 0.70 | 0.41, 1.18 | 0.34 | ||||||||
| Noncitrus fruitc | 43 | ≤50.3 | 35 | >50.3–116.7 | 24 | >116.7–198.6 | 30 | >198.6 | ||||||||
| Age-adjusted results | 1.00 | 0.76 | 0.49, 1.19 | 0.53 | 0.32, 0.88 | 0.65 | 0.40, 1.04 | 0.03 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.84 | 0.53, 1.31 | 0.61 | 0.36, 1.01 | 0.72 | 0.44, 1.19 | 0.11 | ||||||||
Abbreviations: CI, confidence interval; Q, quartile.
Adjusted for age, education, smoking, and total energy intake.
Includes tangerines, oranges, and grapefruit.
Includes apples, pears, bananas, grapes, watermelon, peaches, strawberries, and cantaloupe.
Vegetable intake was not associated with distal gastric cancer risk for women or men (Table 3). When examined in groups defined by their botanical similarity and nutrient content, including cruciferous, green leafy, allium, and legumes, as well as a catch-all category of “other” vegetables, similarly no associations were found. For women in the second quartile of consumption of green leafy vegetables (i.e., bok choy and spinach), a reduction in risk of 39% was observed (95% CI: 0.41, 0.91), but no further reduction in risk was seen for women with greater intake (for women in the third and fourth quartiles, respectively, HR = 0.84, 95% CI: 0.58, 1.21 and HR = 0.76, 95% CI: 0.52, 1.10), and there was no evidence of a trend with increasing intake (Ptrend = 0.30).
Table 3.
Hazard Ratios of Distal Gastric Cancer by Quartile of Vegetable Group Intake in the Shanghai Women's and Men's Health Studies, 1996–2008
| Q1 (Referent) |
Q2 |
Q3 |
Q4 |
Ptrend | ||||||||||||
| No. of Cases | Range, g/day | Hazard Ratio | No. of Cases | Range, g/day | Hazard Ratio | 95% CI | No. of Cases | Range, g/day | Hazard Ratio | 95% CI | No. of Cases | Range, g/day | Hazard Ratio | 95% CI | ||
| Women | ||||||||||||||||
| All vegetables | 74 | ≤179.5 | 44 | >179.5–261.3 | 36 | >261.3–373.7 | 52 | >373.7 | ||||||||
| Age-adjusted results | 1.00 | 0.67 | 0.46, 0.97 | 0.56 | 0.38, 0.84 | 0.83 | 0.58, 1.19 | 0.17 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.69 | 0.47, 1.01 | 0.59 | 0.39, 0.90 | 0.89 | 0.60, 1.31 | 0.32 | ||||||||
| Cruciferous vegetablesb | 62 | ≤51.2 | 50 | >51.2–82.7 | 37 | >82.7–129.6 | 57 | >129.6 | ||||||||
| Age-adjusted results | 1.00 | 0.82 | 0.56, 1.18 | 0.65 | 0.43, 0.97 | 0.92 | 0.64, 1.32 | 0.44 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.83 | 0.57, 1.21 | 0.66 | 0.44, 1.00 | 0.95 | 0.65, 1.37 | 0.55 | ||||||||
| Green leafy vegetablesc | 67 | ≤37.2 | 39 | >37.2–67.4 | 51 | >67.4–108.2 | 49 | >108.2 | ||||||||
| Age-adjusted results | 1.00 | 0.60 | 0.41, 0.89 | 0.83 | 0.57, 1.19 | 0.75 | 0.52, 1.08 | 0.26 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.61 | 0.41, 0.91 | 0.84 | 0.58, 1.21 | 0.76 | 0.52, 1.10 | 0.30 | ||||||||
| Allium vegetablesd | 55 | ≤2.9 | 57 | >2.9–5.5 | 44 | >5.5–10.2 | 50 | >10.2 | ||||||||
| Age-adjusted results | 1.00 | 1.12 | 0.78, 1.63 | 0.92 | 0.62, 1.37 | 1.06 | 0.72, 1.56 | 0.98 | ||||||||
| Fully adjusted resultsa | 1.00 | 1.13 | 0.78, 1.64 | 0.94 | 0.63, 1.41 | 1.10 | 0.74, 1.63 | 0.88 | ||||||||
| Legumese | 58 | ≤13.8 | 41 | >13.8–23.9 | 51 | >23.9–39.5 | 56 | >39.5 | ||||||||
| Age-adjusted results | 1.00 | 0.76 | 0.51, 1.14 | 0.95 | 0.65, 1.39 | 1.04 | 0.72, 1.50 | 0.63 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.80 | 0.53, 1.20 | 1.02 | 0.69, 1.50 | 1.14 | 0.77, 1.69 | 0.35 | ||||||||
| Other vegetablesf | 66 | ≤84.2 | 50 | >84.2–134.1 | 46 | >134.1–205.9 | 44 | >205.9 | ||||||||
| Age-adjusted results | 1.00 | 0.87 | 0.60, 1.26 | 0.85 | 0.57, 1.24 | 0.84 | 0.57, 1.24 | 0.36 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.90 | 63, 1.33 | 0. | 0.91 | 0.61, 1.34 | 0.92 | 0.61, 1.40 | 0.68 | |||||||
| Men | ||||||||||||||||
| All vegetables | 35 | ≤212.9 | 35 | >212.9–307.2 | 30 | >307.2–429.3 | 32 | >429.3 | ||||||||
| Age-adjusted results | 1.00 | 1.11 | 0.69, 1.77 | 0.97 | 0.60, 1.58 | 1.01 | 0.62, 1.63 | 0.90 | ||||||||
| Fully adjusted resultsa | 1.00 | 1.13 | 0.70, 1.82 | 0.99 | 0.60, 1.63 | 1.00 | 0.59, 1.68 | 0.87 | ||||||||
| Cruciferous vegetablesb | 40 | ≤71.4 | 25 | >71.4–110.8 | 28 | >110.8–164.9 | 39 | >164.9 | ||||||||
| Age-adjusted results | 1.00 | 0.68 | 0.41, 1.12 | 0.77 | 0.48, 1.25 | 1.05 | 0.67, 1.63 | 0.76 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.70 | 0.42, 1.15 | 0.79 | 0.49, 1.30 | 1.05 | 0.66, 1.66 | 0.77 | ||||||||
| Green leafy vegetablesc | 42 | ≤38.3 | 28 | >38.3–65.8 | 26 | >65.8–117.1 | 36 | >117.1 | ||||||||
| Age-adjusted results | 1.00 | 0.71 | 0.44, 1.15 | 0.67 | 0.41, 1.09 | 0.92 | 0.59, 1.43 | 0.62 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.73 | 0.45, 1.17 | 0.67 | 0.41, 1.10 | 0.90 | 0.57, 1.42 | 0.58 | ||||||||
| Allium vegetablesd | 41 | ≤6.7 | 31 | >6.7–11.7 | 24 | >11.7–20.0 | 36 | >20.0 | ||||||||
| Age-adjusted results | 1.00 | 0.84 | 0.53, 1.34 | 0.68 | 0.41, 1.12 | 1.02 | 0.65, 1.60 | 0.84 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.82 | 0.51, 1.31 | 0.64 | 0.39, 1.07 | 0.92 | 0.58, 1.46 | 0.51 | ||||||||
| Legumese | 37 | ≤19.7 | 29 | >19.7–33.2 | 32 | >33.2–52.0 | 34 | >52.0 | ||||||||
| Age-adjusted results | 1.00 | 0.84 | 0.51, 1.36 | 0.94 | 0.59, 1.51 | 0.97 | 0.61, 1.54 | 0.99 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.84 | 0.51, 1.37 | 0.94 | 0.58, 1.51 | 0.94 | 0.57, 1.51 | 0.86 | ||||||||
| Other vegetablesf | 37 | ≤82.1 | 34 | >82.1–129.9 | 32 | >129.9–198.6 | 29 | >198.6 | ||||||||
| Age-adjusted results | 1.00 | 1.02 | 0.64, 1.63 | 0.99 | 0.61, 1.59 | 0.86 | 0.53, 1.39 | 0.53 | ||||||||
| Fully adjusted resultsa | 1.00 | 1.05 | 0.66, 1.68 | 1.02 | 0.63, 1.66 | 0.86 | 0.51, 1.46 | 0.60 | ||||||||
Abbreviations: CI, confidence interval; Q, quartile.
Adjusted for age, education, smoking, and total energy intake.
Includes bok choy, cabbage, Napa cabbage, cauliflower, white turnip, garland chrysanthemum, shepherd's purse, clover, and amaranth.
Includes bok choy and spinach.
Includes garlic, garlic shoots, heads of garlic, onions, green onions, and Chinese chives.
Includes baby soybeans, fresh peas, fresh broad beans, yard long beans, green beans, snowpeas, snowpea shoots, and peanuts.
Includes soybean sprouts, mung bean sprouts, black and white edible tree fungi, dried xianggu mushroom, celery, eggplant, wild rice stems, asparagus lettuce, potatoes, sweet potatoes, wax gourds, cucumbers, luffa, fresh mushrooms, red and green peppers, bamboo shoots, lotus root, carrots, and tomatoes.
None of the dietary nutrients examined—vitamin A, vitamin C, vitamin E, carotene, retinol, selenium, or folic acid—was associated with distal gastric cancer risk among the men or women in our population (Table 4).
Table 4.
Hazard Ratios of Distal Gastric Cancer by Quartiles of Dietary Intake of Nutrients in the Shanghai Women's and Men's Health Studies, 1996–2008
| Q1 (Referent) |
Q2 |
Q3 |
Q4 |
Ptrend | ||||||||
| Range, μg/day | Hazard Ratio | Range, μg/day | Hazard Ratio | 95% CI | Range, μg/day | Hazard Ratio | 95% CI | Range, μg/day | Hazard Ratio | 95% CI | ||
| Women | ||||||||||||
| Vitamin A | ≤444.5 | >444.5–608.9 | >608.9–815.6 | >815.6 | ||||||||
| Fully adjusted resultsa | 1.00 | 1.35 | 0.93, 1.97 | 1.14 | 0.76, 1.72 | 1.30 | 0.84, 2.0 | 0.38 | ||||
| Vitamin C | ≤57.0 | >57.0–81.5 | >81.5–113.8 | >113.8 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.75 | 0.51, 1.11 | 0.96 | 0.65, 1.41 | 0.97 | 0.64, 1.48 | 0.93 | ||||
| Vitamin E | ≤9.5 | >9.5–12.5 | >12.5–16.2 | >16.2 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.94 | 0.64, 1.38 | 0.95 | 0.63, 1.43 | 0.81 | 0.51, 1.31 | 0.45 | ||||
| Carotene | ≤1,845.8 | >1,845.8–2,640.9 | >2,640.9–3,657.4 | >3,657.4 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.86 | 0.58, 1.26 | 1.04 | 0.71, 1.52 | 0.97 | 0.64, 1.46 | 0.92 | ||||
| Retinol | ≤86.2 | >86.2–144.1 | >144.1–213.7 | >213.7 | ||||||||
| Fully adjusted resultsa | 1.00 | 1.99 | 1.37, 2.87 | 1.30 | 0.84, 2.01 | 1.47 | 0.93, 2.32 | 0.30 | ||||
| Selenium | ≤30.9 | >30.9–41.4 | >41.4–55.0 | >55.0 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.92 | 0.63, 1.34 | 0.87 | 0.56, 1.34 | 0.92 | 0.56, 1.52 | 0.67 | ||||
| Folic acid | ≤218.7 | >218.7–276.1 | >276.1–346.5 | >346.5 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.93 | 0.62, 1.41 | 1.28 | 0.86, 1.93 | 1.39 | 0.88, 2.19 | 0.08 | ||||
| Men | ||||||||||||
| Vitamin A | ≤462.0 | >462.0–633.5 | >633.5–851.8 | >851.8 | ||||||||
| Fully adjusted resultsa | 1.00 | 1.20 | 0.74, 1.95 | 1.26 | 0.77, 2.06 | 0.84 | 0.48, 1.48 | 0.67 | ||||
| Vitamin C | ≤61.0 | >61.0–86.6 | >86.6–119.6 | >119.6 | ||||||||
| Fully adjusted resultsa | 1.00 | 1.02 | 0.64, 1.63 | 0.89 | 0.54, 1.47 | 0.92 | 0.55, 1.55 | 0.65 | ||||
| Vitamin E | ≤10.5 | >10.5–13.9 | >13.9–17.9 | >17.9 | ||||||||
| Fully adjusted resultsa | 1.00 | 1.08 | 0.65, 1.79 | 1.27 | 0.76, 2.11 | 1.08 | 0.60, 1.92 | 0.66 | ||||
| Carotene | ≤1,958.4 | >1,958.4–2,848.1 | >2,848.1–3,981.4 | >3,981.4 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.97 | 0.61, 1.54 | 0.84 | 0.51, 1.38 | 0.82 | 0.49, 1.36 | 0.37 | ||||
| Retinol | ≤91.0 | >91.0–142.0 | >142.0–200.9 | >200.9 | ||||||||
| Fully adjusted resultsa | 1.00 | 1.29 | 0.77, 2.18 | 1.46 | 0.87, 2.45 | 1.69 | 0.99, 2.88 | 0.05 | ||||
| Selenium | ≤35.8 | >35.8–46.8 | >46.8–61.4 | >61.4 | ||||||||
| Fully adjusted resultsa | 1.00 | 1.24 | 0.75, 2.03 | 1.20 | 0.70, 2.06 | 1.38 | 0.76, 2.50 | 0.34 | ||||
| Folic acid | ≤258.2 | >258.2–324.1 | >324.1–404.3 | >404.3 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.67 | 0.38, 1.17 | 1.39 | 0.86, 2.27 | 0.94 | 0.52, 1.68 | 0.54 | ||||
Abbreviations: CI, confidence interval; Q, quartile.
Adjusted for age, education, smoking, and total energy intake.
Additional analyses were conducted to evaluate the potential modifying effect of cigarette smoking on the association between fruit intake and distal gastric cancer risk among men. A reduction in risk of distal gastric cancer from increased fruit intake was seen among ever smokers only (for the highest vs. the lowest quartile of intake of all fruit except watermelon, HR for ever smokers = 0.43, 95% CI: 0.23, 0.79; Ptrend = 0.001). For never smokers, there was a suggestion of a smaller reduced risk from a high intake of fruits (HR for never smokers = 0.77, 95% CI: 0.27, 2.23; Ptrend = 0.67) (Table 5). However, the interaction by smoking status was not statistically significant (Pinteraction = 0.27).
Table 5.
Hazard Ratios of Distal Gastric Cancer by Quartiles of Intake of Fruit, Excluding Watermelon, Stratified by Smoking Status,* in the Shanghai Men's Health Study, 2002–2008
| Q1 (Referent) |
Q2 |
Q3 |
Q4 |
Ptrend | ||||||||
| No. of Cases | Hazard Ratio | No. of Cases | Hazard Ratio | 95% CI | No. of Cases | Hazard Ratio | 95% CI | No. of Cases | Hazard Ratio | 95% CI | ||
| Fruit intake among ever smokers | 40 | 27 | 15 | 15 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.72 | 0.44, 1.19 | 0.43 | 0.23, 0.78 | 0.43 | 0.23, 0.79 | 0.001 | ||||
| Fruit intake among never smokers | 6 | 8 | 11 | 10 | ||||||||
| Fully adjusted resultsa | 1.00 | 0.90 | 0.31, 2.62 | 1.01 | 0.36, 2.79 | 0.77 | 0.27, 2.23 | 0.67 | ||||
Abbreviations: CI, confidence interval; Q, quartile.
Pinteraction = 0.27.
Adjusted for age, education, and total energy intake.
When the analyses were then repeated excluding those cases whose histologic codes did not identify them as pure distal gastric cancer, the results were qualitatively the same as those found in all the primary analyses. Again, the only significant finding was that of increasing intake of all fruit except watermelon being associated with a decrease in risk of distal gastric cancer for men only (Ptrend = 0.02), with a 46% (95% CI: 0.30, 0.90) reduction observed for men in the fourth quartile, compared with men in the lowest, first quartile of fruit intake. When stratified by smoking status, again only male smokers experienced a reduction in risk from fruit intake (for the highest vs. the lowest quartile of intake, HR = 0.51, 95% CI: 0.26, 0.99; Ptrend = 0.02), but the interaction with smoking was not significant (Pinteraction = 0.27).
DISCUSSION
In these Shanghai cohorts, men with the highest quartile of fruit intake had an approximately 50% reduction in risk of distal gastric cancer compared with men in the lowest quartile of fruit intake. This association was especially strong among male ever smokers, for whom those in the highest quartile of fruit intake had a nearly 60% reduced risk of distal gastric cancer. No association between distal gastric cancer risk and fruit was found for women, most of whom were lifetime nonsmokers. Additionally, no association with distal gastric cancer risk was found for men or women with either vegetable intake or dietary intake of micronutrients.
Of the 10 comparable, previously published, prospective studies of fruit intake and gastric cancer risk, only 3 observed significant decreases in risk for men who are high consumers of fruit, as in the present study (11, 19, 20). The magnitude of the reduction in risk from those 3 studies ranged from 34% to 40%, weaker than the 50% reduction in risk found in the present study. These studies were also carried out among high-risk populations: 2 were among Japanese in Hawaii and 1 was a study of Finnish male smokers. Two Japanese studies found nonsignificant suggestions of an inverse association with fruit (21, 22). Of the 5 other studies that found no association with fruit intake, 4 were conducted in European or American populations (8, 9, 10, 23) and 1 from a very poor region of China, in a population with an overall very low fruit intake (the median intake of fruit was described as 5 times a year) (12). Additionally, compared with these previous cohort studies that found no association between gastric cancer risk and fruit intake, the current study had by far the largest number of female cases, allowing for sufficient power to detect associations separately by sex.
As we detected an association only for men, whose average daily fruit intake is much lower than that for the women in our study, and primarily for men who are ever smokers, our findings suggest that it is in a high-risk group that fruit intake confers a protective effect against distal gastric cancer. The potential anticarcinogenic properties of fruit are believed to be related to the presence of antioxidant carotenoids or other phytochemicals (24). It is possible that there is a threshold effect, whereby individuals consuming above a certain level of phytochemicals receive no further benefit. Data from our study support this hypothesis, as the amount of reduction in risk for men associated with intake of all fruits except watermelon flattens out at the third quartile. Additionally, the range of the amount of fruit eaten by men in the male third quartile of intake is the same as that eaten by women in the female first and second quartiles, and correspondingly we found no association with increased intake of fruit and decreased risk of gastric cancer among the women in our study. As for the suggested difference in effect of fruit intake by smoking status, studies (including 1 in Shanghai) have consistently observed that smokers have lower baseline levels of serum carotenoids than do nonsmokers (25, 26). In fact, even when consuming similar levels of dietary carotenoids, smokers have much lower serum carotenoid levels than do nonsmokers, suggesting that smoking itself reduces blood nutrient levels (27). In relation to the findings of the present study, this further supports the possibility that smokers benefit more greatly from fruit intake because they are of even higher risk for gastric cancer because of even lower levels of carotenoids. Our finding of no association of gastric cancer risk with vegetable intake reflects the general results from other prospective studies (7), as only 2 comparable studies have found significant inverse associations (10, 20). Within the SWHS, a vegetable-rich diet was not found to be related to women's overall mortality, whereas a fruit-rich diet was found to be protective (28). In the SMHS, a vegetable-rich diet was more common among men with chronic disease than among men without, which suggests that our analysis of vegetable intake with gastric cancer risk could be biased by residual confounding by factors related to chronic disease prevalence (15).
In this study, dietary intake of selected micronutrients—vitamin A, vitamin C, vitamin E, carotene, retinol, selenium, and retinol—was not associated with distal gastric cancer risk among men or women, or when stratified by smoking status. Few prospective studies have investigated the relation between dietary nutrient intake and gastric cancer risk, and the results have generally been conflicting, although a number have found significant relations with vitamin C and selenium (5). There has also been inconsistency in the findings of studies examining micronutrients and gastric cancer risk using prediagnostic blood, although several recent studies have found an inverse association with serum or plasma levels of some carotenoids (29–31).
There are a number of limitations to the present analyses. Although the validity of vegetable and nutrient intakes was fairly good for both men and women (correlations between food frequency questionnaire and 24-hour recalls ranging from 0.38 to 0.42), the correlations of 0.72 and 0.51 for men and women, respectively, were better for fruit intake, particularly for men (16, 17). Ideally, we would have H. pylori status for all cohort members and could then both adjust for this main distal gastric cancer risk factor and determine whether it is an effect modifier for the relation of fruits and vegetables and gastric cancer. Although we did not have this information for all cohort members, we do know that, in our population, the majority of individuals (e.g., 95% in the SWHS) are infected with H. pylori (32). Studies that have stratified by H. pylori status have generally found no evidence of effect modification (33) or have found the suggestion that fruits and vegetables are more protective for individuals positive for the bacteria (9). Thus, by including H. pylori-negative individuals in our study, our results may have been biased toward the null. Additionally, regular use of nonsteroidal antiinflammatory drugs has been consistently identified as having an inverse association with gastric cancer risk (34, 35). However, only 2% of women and approximately 7% of men in our population took these drugs regularly, and thus potential confounding by this variable is likely to be small. Although excluding from the analyses individuals with a history of gastrectomy (0.3% of women and 2.2% of men), which was necessary to define the at-risk group, may create concern of bias, the numbers were very small, and the sex difference was not unanticipated as men have much higher rates of ulcers than women do, ulcers being the main reason after gastric cancer for a gastrectomy. Finally, the follow-up time for the subjects in this study, particularly the men, was rather short (median follow-up times, SWHS = 9.2 years, SMHS = 3.6 years). As reported above, a secondary analysis excluding cases diagnosed within 1 year of enrollment did not change the results, suggesting that a change in diet due to disease or symptoms of disease did not create the associations found. Extended follow-up of these cohorts would allow us to evaluate this issue further in the future.
A strength of the present study is the use of a comprehensive, validated food frequency questionnaire for ascertainment of the exposures of interest. Additionally, we restricted our case definition to individuals with distal gastric cancer, as gastric cardia cancer is rare in China and appears to have a different etiology (36). Furthermore, we utilized 2 well-established cohorts from a high-risk population that includes the largest number of female cases of distal gastric cancer yet to be analyzed in terms of an association with intakes of fruits and vegetables in a prospective study. Our finding—that fruit intake is inversely associated with distal gastric cancer risk in men, particularly in men who have ever smoked—provides further evidence that future research on fruits and gastric cancer in high-risk populations is warranted.
Acknowledgments
Author affiliations: Division of Epidemiology, Department of Medicine, Vanderbilt-Ingram Cancer Center, Vanderbilt University School of Medicine, Nashville, Tennessee (Meira Epplein, Xiao-Ou Shu, Gong Yang, Hui Cai, Wei Zheng); Department of Epidemiology, Shanghai Cancer Institute, Shanghai, People's Republic of China (Yong-Bing Xiang, Hong-Lan Li, Yu-Tang Gao); and Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, Maryland (Wong-Ho Chow, Bu-Tian Ji).
Supported by National Cancer Institute grants R37 CA70867-12 and R01 CA82729-09.
Conflict of interest: none declared.
Glossary
Abbreviations
- CI
confidence interval
- HR
hazard ratio
- ICD-O
International Classification of Diseases for Oncology
- SMHS
Shanghai Men's Health Study
- SWHS
Shanghai Women's Health Study
References
- 1.Ferlay J, Bray F, Pisani P, et al. GLOBOCAN 2002. Cancer Incidence, Mortality and Prevalence Worldwide. IARC Cancer Base No. 5, Version 2.0. Lyon, France: IARC Press; 2004. [Google Scholar]
- 2.Parkin DM, Bray F, Ferlay J, et al. Global cancer statistics, 2002. CA Cancer J Clin. 2005;55(2):74–108. doi: 10.3322/canjclin.55.2.74. [DOI] [PubMed] [Google Scholar]
- 3.Suerbaum S, Michetti P. Helicobacter pylori infection. N Engl J Med. 2002;347(15):1175–1186. doi: 10.1056/NEJMra020542. [DOI] [PubMed] [Google Scholar]
- 4.Lunet N, Valbuena C, Vieira AL, et al. Fruit and vegetable consumption and gastric cancer by location and histological type: case-control and meta-analysis. Eur J Cancer Prev. 2007;16(4):312–327. doi: 10.1097/01.cej.0000236255.95769.22. [DOI] [PubMed] [Google Scholar]
- 5.Liu C, Russell RM. Nutrition and gastric cancer risk: an update. Nutr Rev. 2008;66(5):237–249. doi: 10.1111/j.1753-4887.2008.00029.x. [DOI] [PubMed] [Google Scholar]
- 6.IARC Handbooks of Cancer Prevention. Vol. 8. Fruits and Vegetables. Lyon, France: IARC Press; 2003. [Google Scholar]
- 7.Lunet N, Lacerda-Vieira A, Barros H. Fruit and vegetables consumption and gastric cancer: a systematic review and meta-analysis of cohort studies. Nutr Cancer. 2005;53(1):1–10. doi: 10.1207/s15327914nc5301_1. [DOI] [PubMed] [Google Scholar]
- 8.Freedman ND, Subar AF, Hollenbeck AR, et al. Fruit and vegetable intake and gastric cancer risk in a large United States prospective cohort study. Cancer Causes Control. 2008;19(5):459–467. doi: 10.1007/s10552-007-9107-4. [DOI] [PubMed] [Google Scholar]
- 9.González CA, Pera G, Agudo A, et al. Fruit and vegetable intake and the risk of stomach and oesophagus adenocarcinoma in the European Prospective Investigation into Cancer and Nutrition (EPIC-EURGAST) Int J Cancer. 2006;118(10):2559–2566. doi: 10.1002/ijc.21678. [DOI] [PubMed] [Google Scholar]
- 10.Larsson SC, Bergkvist L, Wolk A. Fruit and vegetable consumption and incidence of gastric cancer: a prospective study. Cancer Epidemiol Biomarkers Prev. 2006;15(10):1998–2001. doi: 10.1158/1055-9965.EPI-06-0402. [DOI] [PubMed] [Google Scholar]
- 11.Nouraie M, Pietinen P, Kamangar F, et al. Fruits, vegetables, and antioxidants and risk of gastric cancer among male smokers. Cancer Epidemiol Biomarkers Prev. 2005;14(9):2087–2092. doi: 10.1158/1055-9965.EPI-05-0038. [DOI] [PubMed] [Google Scholar]
- 12.Tran GD, Sun XD, Abnet CC, et al. Prospective study of risk factors for esophageal and gastric cancers in the Linxian general population trial cohort in China. Int J Cancer. 2005;113(3):456–463. doi: 10.1002/ijc.20616. [DOI] [PubMed] [Google Scholar]
- 13.World Cancer Research Fund/American Institute for Cancer Research. Food, Nutrition, Physical Activity, and the Prevention of Cancer: A Global Perspective. Washington, DC: American Institute for Cancer Research; 2007. [Google Scholar]
- 14.Zheng W, Chow WH, Yang G, et al. The Shanghai Women's Health Study: rationale, study design, and baseline characteristics. Am J Epidemiol. 2005;162(11):1123–1131. doi: 10.1093/aje/kwi322. [DOI] [PubMed] [Google Scholar]
- 15.Cai H, Zheng W, Xiang YB, et al. Dietary patterns and their correlates among middle-aged and elderly Chinese men: a report from the Shanghai Men's Health Study. Br J Nutr. 2007;98(5):1006–1013. doi: 10.1017/S0007114507750900. [DOI] [PubMed] [Google Scholar]
- 16.Shu XO, Yang G, Jin F, et al. Validity and reproducibility of the food frequency questionnaire used in the Shanghai Women's Health Study. Eur J Clin Nutr. 2004;58(1):17–23. doi: 10.1038/sj.ejcn.1601738. [DOI] [PubMed] [Google Scholar]
- 17.Villegas R, Yang G, Liu D, et al. Validity and reproducibility of the food-frequency questionnaire used in the Shanghai Men's Health Study. Br J Nutr. 2007;97(5):993–1000. doi: 10.1017/S0007114507669189. [DOI] [PubMed] [Google Scholar]
- 18.Yang YX, Wang GY, Pan XC, editors. China Food Composition Tables 2002. Beijing, China: Beijing University Medical Press; 2002. [Google Scholar]
- 19.Galanis DJ, Kolonel LN, Lee J, et al. Intakes of selected foods and beverages and the incidence of gastric cancer among the Japanese residents of Hawaii: a prospective study. Int J Epidemiol. 1998;27(2):173–180. doi: 10.1093/ije/27.2.173. [DOI] [PubMed] [Google Scholar]
- 20.Nomura AM, Stemmermann GN, Chyou PH. Gastric cancer among the Japanese in Hawaii. Jpn J Cancer Res. 1995;86(10):916–923. doi: 10.1111/j.1349-7006.1995.tb03001.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Inoue M, Tajima K, Kobayashi S, et al. Protective factor against progression from atrophic gastritis to gastric cancer—data from a cohort study in Japan. Int J Cancer. 1996;66(3):309–314. doi: 10.1002/(SICI)1097-0215(19960503)66:3<309::AID-IJC7>3.0.CO;2-2. [DOI] [PubMed] [Google Scholar]
- 22.Kobayashi M, Tsubono Y, Sasazuki S, et al. Vegetables, fruit and risk of gastric cancer in Japan: a 10-year follow-up of the JPHC Study Cohort I. Int J Cancer. 2002;102(1):39–44. doi: 10.1002/ijc.10659. [DOI] [PubMed] [Google Scholar]
- 23.Botterweck AA, van den Brandt PA, Goldbohm RA. A prospective cohort study on vegetable and fruit consumption and stomach cancer risk in the Netherlands. Am J Epidemiol. 1998;148(9):842–853. doi: 10.1093/oxfordjournals.aje.a009709. [DOI] [PubMed] [Google Scholar]
- 24.Steinmetz KA, Potter JD. Vegetables, fruit, and cancer. II. Mechanisms. Cancer Causes Control. 1991;2(6):427–442. doi: 10.1007/BF00054304. [DOI] [PubMed] [Google Scholar]
- 25.Sugiura M, Nakamura M, Ogawa K, et al. Synergistic interaction of cigarette smoking and alcohol drinking with serum carotenoid concentrations: findings from a middle-aged Japanese population. Br J Nutr. 2009;102(8):1211–1219. doi: 10.1017/S0007114509382124. [DOI] [PubMed] [Google Scholar]
- 26.Yuan JM, Ross RK, Chu XD, et al. Prediagnostic levels of serum beta-cryptoxanthin and retinol predict smoking-related lung cancer risk in Shanghai, China. Cancer Epidemiol Biomarkers Prev. 2001;10(7):767–773. [PubMed] [Google Scholar]
- 27.Stryker WS, Kaplan LA, Stein EA, et al. The relation of diet, cigarette smoking, and alcohol consumption to plasma beta-carotene and alpha-tocopherol levels. Am J Epidemiol. 1988;127(2):283–296. doi: 10.1093/oxfordjournals.aje.a114804. [DOI] [PubMed] [Google Scholar]
- 28.Cai H, Shu XO, Gao YT, et al. A prospective study of dietary patterns and mortality in Chinese women. Epidemiology. 2007;18(3):393–401. doi: 10.1097/01.ede.0000259967.21114.45. [DOI] [PubMed] [Google Scholar]
- 29.Jenab M, Riboli E, Ferrari P, et al. Plasma and dietary carotenoid, retinol and tocopherol levels and the risk of gastric adenocarcinomas in the European Prospective Investigation into Cancer and Nutrition. Br J Cancer. 2006;95(3):406–415. doi: 10.1038/sj.bjc.6603266. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Persson C, Sasazuki S, Inoue M, et al. Plasma levels of carotenoids, retinol and tocopherol and the risk of gastric cancer in Japan: a nested case-control study. Carcinogenesis. 2008;29(5):1042–1048. doi: 10.1093/carcin/bgn072. [DOI] [PubMed] [Google Scholar]
- 31.Yuan JM, Ross RK, Gao YT, et al. Prediagnostic levels of serum micronutrients in relation to risk of gastric cancer in Shanghai, China. Cancer Epidemiol Biomarkers Prev. 2004;13(11 pt 1):1772–1780. [PubMed] [Google Scholar]
- 32.Dong LM, Shu XO, Gao YT, et al. Urinary prostaglandin E2metabolite and gastric cancer risk in the Shanghai Women's Health Study. Cancer Epidemiol Biomarkers Prev. 2009;18(11):3075–3078. doi: 10.1158/1055-9965.EPI-09-0680. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Epplein M, Nomura AM, Hankin JH, et al. Association of Helicobacter pylori infection and diet on the risk of gastric cancer: a case-control study in Hawaii. Cancer Causes Control. 2008;19(8):869–877. doi: 10.1007/s10552-008-9149-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Dai Y, Wang WH. Non-steroidal anti-inflammatory drugs in prevention of gastric cancer. World J Gastroenterol. 2006;12(18):2884–2889. doi: 10.3748/wjg.v12.i18.2884. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Epplein M, Nomura AM, Wilkens LR, et al. Nonsteroidal antiinflammatory drugs and risk of gastric adenocarcinoma: the Multiethnic Cohort Study. Am J Epidemiol. 2009;170(4):507–514. doi: 10.1093/aje/kwp162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Brown LM, Devesa SS. Epidemiologic trends in esophageal and gastric cancer in the United States. Surg Oncol Clin N Am. 2002;11(2):235–256. doi: 10.1016/s1055-3207(02)00002-9. [DOI] [PubMed] [Google Scholar]