Abstract
Most epidemiological studies evaluating the association of fruit and vegetable intakes on lung cancer risk were conducted in North American and European countries. We investigated the association of intakes of fruits, vegetables, dietary vitamins A and C, and folate with lung cancer risk among 61,491 Chinese adult men who were recruited to the Shanghai Men's Health Study, a population-based, prospective cohort study. Baseline dietary intake was assessed through a validated food frequency questionnaire during in-home visits. Multivariate Cox regression was used to estimate hazard ratios (HR) and 95% confidence intervals (CI) of lung cancer risk associated with dietary intakes. During a median follow-up of 5.5 years, 359 incident lung cancer cases accrued after the first year of follow-up and 68.8% of them were current smokers. Intakes of green leafy vegetables, β-carotene-rich vegetables, watermelon, vitamin A, and carotenoids were inversely associated with lung cancer risk; the corresponding HR (95% CI) comparing the highest with the lowest quartiles were 0.72 (0.53–0.98), 0.69 (0.51–0.94), 0.65 (0.47–0.90), 0.63 (0.44–0.88), and 0.64 (0.46–0.88). Intake of all fruits and vegetables combined was marginally associated with lower risk. Our study suggests that the consumption of carotenoid-rich vegetables is inversely associated with lung cancer risk.
Keywords: fruits, vegetables, carotenoids, dietary intake, lung cancer, epidemiological
Introduction
Evidence from observational studies has supported a protective effect of fruits and vegetables on lung cancer risk (1–3). In prospective studies, fruit intake has a more consistent inverse association with lung cancer risk than vegetable intake (2,3). The protective effect of antioxidant-rich carotenoids may explain this inverse association (4). Several prospective studies found an inverse association between lung cancer risk and dietary intakes of β-carotene, lycopene, β-cryptoxanthin, and lutein (4–6). In contrast, two clinical trials of high-dose β-carotene supplementation among smokers and/or asbestos-exposed workers found a higher risk of lung cancer among the supplement group compared with the placebo group (7,8). For both trials, six years after the supplementation was discontinued, lung cancer risk was reassessed and the higher risk initially observed among the supplement group was attenuated and no longer statistically significant (9,10). Other trials of lower dose β-carotene supplementation, along with aspirin or other micronutrients, reported no increase in risk of lung cancer (11,12). Similarly, a pooled analysis of prospective studies, primarily conducted in North America and Europe, found no association between lung cancer risk and vitamin C or folate intake from either foods or supplements (13).
The vast majority of current evidence on associations of diet with lung cancer is derived from populations in North America and Europe. The role of dietary intake on lung cancer in other regions, such as China, has not been well investigated. The prevalence of smoking among men in China is one of the highest in the world (59.5%) (14), yet the lung cancer incidence rate was only 45.9 per 100,000 person-years in 2008. This is in the middle range compared with rates in many developed countries and lower than the rate of 71.2 per 100,000 person-years in 2008 in the United States (15), which had a lower smoking prevalence among men (21.3%) (16) than China. Furthermore, the traditional Chinese diet is plant-based, and dietary and nutrient intake patterns in Chinese populations differ from those in other populations. Hence, we investigated the association of dietary intakes of fruits, vegetables, vitamin A, vitamin C, and folate with lung cancer risk in the Shanghai Men's Health Study (SMHS), a population-based, prospective cohort study of 61,491 middle-aged and elderly men.
Materials and methods
Study population
The SMHS included 61,491 men who lived in eight communities of urban Shanghai and were aged 40 to 74 years at baseline between 2002 and 2006 (17). Individuals who had been diagnosed with any type of cancer at the time of recruitment were excluded from this study. Questionnaires were administered in person by trained interviewers during in-home visits. Anthropometric measurements, including body weight, height, and hip and waist circumferences, were taken by the interviewers during the visits. All participants provided written, informed consent, and the study was approved by the Institutional Review Boards of all participating institutes.
Dietary assessment
The food frequency questionnaire (FFQ) consisted of 81 items, including 38 items for vegetables and eight items for fruits, and asked about typical intake during the year before the interview in terms of the frequency (five categories ranging from never to every day) and the amount of consumption in liang (1 liang ≈ 50 grams). The FFQ was validated against twelve monthly, unannounced 24-hour dietary recalls, which covered roughly the same one-year period as the FFQ, of 195 randomly selected participants of the same study (18). The Spearman correlation coefficients for carotenoids, vitamin C, total vegetables, and total fruits were 0.38, 0.42, 0.42, and 0.72, respectively. The reproducibility of reported dietary intakes was assessed by administering the same FFQ twice about one year apart. The corresponding Spearman correlation coefficients for carotenoids, vitamin C, total vegetables, and total fruits were 0.40, 0.38, 0.43, and 0.64. Misclassifications of participants to opposite quartiles of intake between the two assessments occurred only rarely. Nutrient intakes were calculated using the Chinese Food Composition Tables (19).
Follow-up and cancer ascertainment
Follow up is conducted every 2–3 years by in-home visits to determine cancer incidence, occurrence of other chronic diseases, and vital status for all cohort members. This in-person follow-up is supplemented by annual linkage of our participants' files to the Shanghai Cancer Registry and the Shanghai Vital Statistics Registry databases. Initial reports from the cancer registry of a diagnosis of primary lung cancer [defined using International Classification of Disease, Ninth revision (ICD-9) code 162] or from a self-report by participants during a follow-up visit were verified by medical chart review. We obtained the date of diagnosis and primary site of cancer for all cases. Confirmed cases diagnosed with primary lung cancer between October 2002 and December 2009 were included in this analysis.
Statistical analyses
Participants who were excluded from this analysis include: 1) those who were lost during the first follow-up, which occurred two years after the baseline survey (n =14); 2) those who were censored during first year of the follow-up due to a lung cancer diagnosis (n = 66) or death (n = 227); and 3) those who reported total caloric intake outside the range of 500–4,000 kcal/day (n = 91). In addition, one participant was excluded due to incomplete cigarette smoking history. Hence, 61,092 participants were included in the current analysis. Body mass index (BMI) values were missing for 37 participants and they were categorized into the most frequent category in the cohort (23 ≤ BMI< 25 kg/m2). Similarly, missing values for education (n = 860) and occupation (n = 70) were replaced with the most common categories (high school and manual laborers, respectively). The results of the analyses conducted as above (with replaced values) and analyses conducted after excluding missing responses were materially similar. Characteristics of the study population were compared between cases and non-cases after adjusting for age at baseline. Intakes of nutrients, fruits, and vegetables were compared between the two groups after further adjusting for total caloric intake.
Cox regression analysis was performed with age as the time scale. Entry time was defined as age at study enrollment and exit time was defined as age at lung cancer diagnosis or censoring (either December 31, 2009, the date of the latest record linkage, or the date of death), whichever came first. All analyses were adjusted for age (years, continuous), smoking [years of smoking (four categories) and the number of cigarettes smoked per day (four categories)], education (four categories ranging from elementary school or less to college or more), BMI (five categories ranging from <20 to >27.5 kg/m2), regular tea consumption (at least three times per week continuously for at least six months), history of chronic bronchitis (ICD-9 code 491, yes or no), family history of lung cancer among first-degree relatives (yes or no), and total caloric intake (Kcal/day, continuous). The following variables were assessed as potential confounders or covariates, but not included due to their small effect on the risk of lung cancer: family history of any cancer (yes or no); history of asthma (ICD-9 code 493, yes or no), emphysema (ICD-9 code 492, yes or no), or pulmonary tuberculosis (ICD-9 code 011, yes or no); occupation (office workers, clerical workers, or manual laborers); household income (four categories ranging from <500 to >2000 yuan/capita); regular exercise (at least once per week continuously for at least three months); waist-to-hip ratio (continuous); and alcohol consumption (at least three times a week continuously for at least six months).
Quartiles of intake of each dietary variable were created according to the distribution of the intake at baseline in the entire cohort. Hazard ratio (HR) and 95% confidence intervals (CI) for each quartile were calculated using the first quartile as reference. The linear trend of the association between dietary intake and lung cancer risk was assessed by assigning the median value of each quartile and treating it as continuous in a model. The current analyses did not include soy intake, which will be investigated in a separate report. In addition to total fruit and/or vegetable intakes, five vegetable subgroups (cruciferous, allium, green leafy, legumes, and other) and one fruit subgroup (citrus) were evaluated in the current analyses, given that the fruits and vegetables within these subgroups share a more similar nutritional profile compared with all fruits or vegetables considered together. To further elucidate the effect of subtypes of carotenoids, the following five vegetable or fruit groups were also evaluated: carotenoid-rich vegetables, β-carotene-rich vegetables, tomatoes (a lycopene-rich vegetable), lutein/zeaxanthin-rich vegetables, and watermelon (a lycopene-rich fruit) (20).
Subgroup analyses were conducted by smoking status (never, former, or current smokers at baseline), pack-years of smoking [never smokers, light smokers (<30 pack-years), or heavy smokers (≥30 pack-years)], alcohol consumption (never or ever), history of lung disease (chronic bronchitis, asthma, emphysema, or pulmonary tuberculosis), any vitamin supplement use (any use of single vitamin A, B, C, D, or E supplements; vitamins A/D; or multivitamins), and vitamin A or vitamin C-containing supplement use. The statistical significance of multiplicative interaction effects between smoking history and dietary intakes was tested by including the main effects and their cross-product term in the model and assessing the P-value. All statistical analyses were performed using SAS 9.3 software (SAS Institute Inc., Carey, NC). All P values were two-sided.
Results
During the median follow-up of 5.5 years, 359 lung cancer cases were documented after the first year of follow-up. Lung cancer cases were older than non-cases and had lower socioeconomic status than non-cases based on education and household income (Table 1). Family history of lung cancer among first-degree relatives was more common in cases than in non-cases, as was personal history of emphysema, chronic bronchitis, and asthma. More cases had smoked cigarettes than non-cases and 68.8% of the cases were current smokers. The incidence of lung cancer was highest in former smokers (1.8 per 1000 person-years), followed by current smokers (1.3 per 1000 person-years) and never smokers (0.4 per 1000 person-years). The mean BMI was lower in cases than in non-cases. Alcohol and tea consumption were more common in cases than non-cases. The use of any vitamin supplement (vitamin A, B, C, D/AD, E, or multivitamins) was less common in cases than non-cases (10.7% versus 15.3%). Total caloric intake was lower in cases than non-cases (Table 2). Overall, cases had lower mean consumption of vitamins A and C, carotenoids, vegetables (total vegetables and the cruciferous, green leafy, carotenoid-rich, and β-carotene-rich vegetable groups), and fruits (total fruits and the watermelon and citrus fruit groups) than non-cases.
Table 1.
Baseline Characteristics of the Study Population1
| Non-Cases | Cases | P-difference | |
|---|---|---|---|
| Number | 60,733 | 359 | - |
|
| |||
| Age (years) | 55.3 (0.04) | 63.4 (0.49) | <.0001 |
|
| |||
| Education | |||
| Elementary school or less | 6.5% | 10.4% | |
| Middle school | 33.0% | 44.2% | |
| High school | 37.0% | 32.6% | |
| Professional education/college or more | 23.5% | 12.8% | <.0001 |
|
| |||
| Income (per capita) | |||
| <500 yuan | 12.5% | 19.4% | |
| 500–<1000 yuan | 42.5% | 45.3% | |
| 1000–<2000 yuan | 35.2% | 29.0% | |
| ≥2000 yuan | 9.8% | 6.3% | 0.06 |
|
| |||
| Occupation | |||
| Office workers | 26.6% | 18.7% | |
| Clerical workers | 21.9% | 23.7% | |
| Manual laborers | 51.5% | 57.6% | 0.002 |
|
| |||
| Family history of lung cancer | 5.9% | 9.8% | 0.006 |
|
| |||
| Personal history of lung disease | |||
| Emphysema | 1.5% | 2.2% | 0.008 |
| Pulmonary tuberculosis | 5.2% | 3.5% | 0.12 |
| Chronic bronchitis | 4.6% | 8.7% | <.0001 |
| Asthma | 2.5% | 3.9% | 0.05 |
|
| |||
| Smoking | |||
| Ever smoked | 69.5% | 90.3% | <.0001 |
| Age started smoking (years)2 | 23.3 (0.03) | 20.3 (0.38) | <.0001 |
| Cigarettes (day)2 | 16.5 (0.04) | 19.7 (0.50) | <.0001 |
| Smoking (pack-years)2 | 24.4 (0.08) | 33.7 (0.88) | <.0001 |
|
| |||
| Alcohol consumption | |||
| Ever consumed | 33.6% | 46.8% | <.0001 |
| Drinks (day)2 | 0.83 (0.01) | 1.58 (0.09) | <.0001 |
|
| |||
| Regular tea consumption | 67.1% | 75.5% | <.0001 |
|
| |||
| Vitamin supplement use3 | 15.3% | 10.7% | 0.07 |
|
| |||
| Vitamin A-containing supplement use | 8.6% | 6.1% | 0.20 |
|
| |||
| Vitamin C-containing supplement use | 12.0% | 7.8% | 0.07 |
|
| |||
| Multivitamin supplement use | 7.6% | 5.0% | 0.18 |
|
| |||
| BMI (kg/m2) | 23.7 (0.01) | 22.8 (0.16) | <.0001 |
|
| |||
| Regular exercise4 | 35.6% | 25.7% | 0.0003 |
For all characteristics except age, means (standard errors) or percentages were adjusted for age at baseline.
Only smokers or alcohol consumers were included.
Any vitamin A, C, D, A/D, or E supplement, or multivitamins.
At least once per week continuously for at least three months
Table 2.
Baseline Nutrient, Vegetable, and Fruit Intakes of the Study Population1
| Non-cases | Cases | P-difference | |
|---|---|---|---|
| Number | 60,733 | 359 | - |
|
| |||
| Total calories (kcal/day) | 1,907 (1.92) | 1,833 (25.0) | 0.003 |
|
| |||
| Nutrients | |||
| Fat (g/day) | 34.5 (0.05) | 35.1 (0.62) | 0.35 |
| Protein (g/day) | 78.2 (0.05) | 78.5 (0.70) | 0.74 |
| Carbohydrates (g/day) | 320.6 (0.15) | 319.1 (1.91) | 0.42 |
| Fiber (g/day) | 11.5 (0.01) | 11.2 (0.18) | 0.08 |
| Vitamin A (μg RE/day) | 693 (1.26) | 628 (16.4) | <.0001 |
| Carotenoids (μg/day) | 3164 (6.6) | 2809 (86.6) | <.0001 |
| Retinol (μg/day) | 165 (0.5) | 160 (7.1) | 0.48 |
| Vitamin C (mg/day) | 96.2 (0.19) | 87.5 (2.51) | 0.0005 |
| Vitamin E (mg/day) | 14.9 (0.02) | 14.9 (0.29) | 0.80 |
| Folate (μg/day) | 342.1 (0.39) | 333.1 (5.15) | 0.08 |
|
| |||
| Vegetables2 | |||
| Total vegetables (g/day) | 343.8 (0.72) | 318.3 (9.39) | 0.007 |
| Allium vegetables (g/day) | 16.4 (0.07) | 16.9 (0.89) | 0.58 |
| Cruciferous vegetables (g/day) | 127.6 (0.32) | 114.4 (4.12) | 0.001 |
| Green leafy vegetables (g/day) | 100.4 (0.27) | 90.5 (3.47) | 0.005 |
| Legumes (g/day) | 40.8 (0.13) | 39.3 (1.65) | 0.36 |
| Tomatoes (g/day) | 29.2 (0.14) | 26.4 (1.88) | 0.14 |
| Carotenoid-rich vegetables (g/day) | 147.6 (0.35) | 133.7 (4.57) | 0.002 |
| β-carotene-rich vegetables (g/day) | 104.7 (0.27) | 93.4 (3.53) | 0.001 |
| Lutein/zeaxanthin-rich vegetables (g/day) | 12.7 (0.06) | 13.9 (0.73) | 0.11 |
| Other vegetables (g/day) | 124.1 (0.33) | 116.2 (4.30) | 0.07 |
| Fruit2 | |||
| Total fruit (g/day) | 151.9 (0.49) | 121.7 (6.40) | <.0001 |
| Watermelon (g/day) | 79.2 (0.31) | 62.9 (4.09) | <.0001 |
| Citrus fruits (g/day) | 12.6 (0.07) | 9.5 (0.96) | 0.001 |
Means (standard errors) for all nutrients and food items except for total caloric intake were adjusted for age at baseline.
Allium vegetables include garlic, garlic shoots, heads of garlic, onions, green onions, and Chinese chives; cruciferous vegetables include Chinese greens, cabbage, Napa cabbage, cauliflower, white turnip, garland chrysanthemum, shepherd's purse, watercress, and amaranth; green leafy vegetables include spinach, Chinese greens, garland chrysanthemum, shepherd's purse, watercress, and amaranth; legumes include baby soybeans, fresh peas, fresh broad beans, yard long beans, green beans, snow peas, snow pea shoots, and peanuts; carotenoid-rich vegetables include Chinese greens, spinach, snow pea shoots, tomatoes, Chinese chives, green onions, carrot, (green) peas, garland chrysanthemum, shepherd's purse, watercress, and amaranth; β-carotene-rich vegetables include Chinese greens, spinach, Chinese chives, green onions, carrot and garland chrysanthemum; lutein/zeaxanthin-rich vegetables include spinach, (green) peas, and garland chrysanthemum; other vegetables include soybean sprouts, mung bean sprouts, celery, eggplant, wild rice stem, lettuce, potato, winter melon, cucumber, luffa, fresh mushroom, red/green pepper, bamboo shoots, lotus shoots, black/white tree fungi, and dried xianggu mushroom; citrus fruits includes tangerines, oranges, and grapefruit.
Categorical analyses showed a marginally significant inverse association between lung cancer risk and combined intake of total fruits and vegetables overall (HR = 0.76; 95% CI = 0.55–1.07 for comparing the fourth with the first quartile of intake; P-trend = 0.07) (Table 3). There was no association for total vegetable intake (HR = 0.88, 95% CI = 0.64–1.22; P-trend = 0.49). Within vegetable groups, men in the fourth quartile of green leafy vegetable intake had a 28% lower risk of lung cancer than those in the first (HR = 0.72, 95% CI = 0.53–0.98) and the linear trend for this association was marginally significant (P-trend = 0.08). Specifically, intake of Chinese greens, which constituted 77.5% of green leafy vegetable intake, was inversely associated with the risk of lung cancer (HR = 0.70; 95% CI = 0.51–0.98; P-trend = 0.05), but not other green leafy vegetables, when assessed individually. In addition, high intake of the β-carotene-rich vegetable group was also associated with lower risk of lung cancer (HR = 0.69; 95% CI = 0.51–0.94; P-trend = 0.03). No other vegetable subgroups were associated with the risk of lung cancer.
Table 3.
Risk of Lung Cancer by Vegetable and Fruit Intakes (g/day)1
| Vegetable and/or fruit | Quartiles | Median | Cases | HR | 95% CI | P-trend2 | |
|---|---|---|---|---|---|---|---|
| Fruits and Vegetables | 1 | 233.4 | 138 | Reference | |||
| 2 | 381.8 | 93 | 0.92 | 0.70, | 1.20 | ||
| 3 | 527.0 | 68 | 0.77 | 0.57, | 1.04 | ||
| 4 | 779.7 | 60 | 0.76 | 0.55, | 1.07 | 0.07 | |
|
| |||||||
| Total vegetables | 1 | 158.0 | 131 | Reference | |||
| 2 | 260.0 | 84 | 0.88 | 0.66, | 1.16 | ||
| 3 | 360.1 | 75 | 0.90 | 0.67, | 1.21 | ||
| 4 | 545.5 | 69 | 0.88 | 0.64, | 1.22 | 0.49 | |
|
| |||||||
| Allium vegetables | 1 | 4.6 | 86 | Reference | |||
| 2 | 9.1 | 103 | 1.41 | 1.06, | 1.88 | ||
| 3 | 15.0 | 81 | 1.17 | 0.86, | 1.60 | ||
| 4 | 29.6 | 89 | 1.37 | 1.00, | 1.86 | 0.17 | |
|
| |||||||
| Cruciferous vegetables | 1 | 48.1 | 122 | Reference | |||
| 2 | 90.2 | 85 | 0.95 | 0.72, | 1.26 | ||
| 3 | 136.2 | 84 | 0.97 | 0.73, | 1.29 | ||
| 4 | 216.7 | 68 | 0.80 | 0.59, | 1.10 | 0.20 | |
|
| |||||||
| Green leafy vegetables | 1 | 34.6 | 125 | Reference | |||
| 2 | 68.7 | 78 | 0.82 | 0.61, | 1.09 | ||
| 3 | 107.0 | 90 | 0.98 | 0.74, | 1.29 | ||
| 4 | 176.3 | 66 | 0.72 | 0.53, | 0.98 | 0.08 | |
|
| |||||||
| Legumes | 1 | 12.5 | 108 | Reference | |||
| 2 | 26.3 | 83 | 0.96 | 0.72, | 1.28 | ||
| 3 | 41.1 | 91 | 1.15 | 0.86, | 1.53 | ||
| 4 | 72.3 | 77 | 0.97 | 0.71, | 1.33 | 0.98 | |
|
| |||||||
| Other vegetables | 1 | 44.9 | 126 | Reference | |||
| 2 | 83.1 | 86 | 0.98 | 0.74, | 1.30 | ||
| 3 | 124.8 | 79 | 1.04 | 0.78, | 1.40 | ||
| 4 | 207.6 | 68 | 0.99 | 0.72, | 1.36 | >0.99 | |
|
| |||||||
| Carotenoid-rich vegetables | 1 | 59.1 | 133 | Reference | |||
| 2 | 106.3 | 75 | 0.75 | 0.57, | 1.00 | ||
| 3 | 155.6 | 82 | 0.88 | 0.66, | 1.16 | ||
| 4 | 245.5 | 69 | 0.77 | 0.56, | 1.05 | 0.16 | |
|
| |||||||
| β-carotene-rich vegetables | 1 | 37.6 | 129 | Reference | |||
| 2 | 72.2 | 83 | 0.82 | 0.62, | 1.08 | ||
| 3 | 111.6 | 81 | 0.84 | 0.63, | 1.12 | ||
| 4 | 181.9 | 66 | 0.69 | 0.51, | 0.94 | 0.03 | |
|
| |||||||
| Tomatoes | 1 | 3.5 | 141 | Reference | |||
| 2 | 13.9 | 84 | 0.85 | 0.65, | 1.12 | ||
| 3 | 28.9 | 62 | 0.79 | 0.58, | 1.08 | ||
| 4 | 69.3 | 72 | 0.99 | 0.73, | 1.34 | 0.90 | |
|
| |||||||
| Lutein/zeaxanthin-rich vegetables | 1 | 2.5 | 113 | Reference | |||
| 2 | 6.5 | 86 | 1.03 | 0.78, | 1.37 | ||
| 3 | 11.8 | 84 | 1.13 | 0.84, | 1.50 | ||
| 4 | 23.6 | 76 | 1.16 | 0.86, | 1.58 | 0.30 | |
|
| |||||||
| Total fruit | 1 | 21.1 | 142 | Reference | |||
| 2 | 92.9 | 84 | 0.75 | 0.57, | 0.99 | ||
| 3 | 168.6 | 70 | 0.76 | 0.57, | 1.03 | ||
| 4 | 286.3 | 63 | 0.75 | 0.54, | 1.04 | 0.09 | |
|
| |||||||
| Watermelon | 1 | 4.6 | 144 | Reference | |||
| 2 | 46.7 | 120 | 0.78 | 0.61, | 1.00 | ||
| 3 | 95.0 | 41 | 0.88 | 0.62, | 1.25 | ||
| 4 | 163.3 | 54 | 0.65 | 0.47, | 0.90 | 0.02 | |
|
| |||||||
| Citrus fruits | 1 | 0.0 | 132 | Reference | |||
| 2 | 3.6 | 89 | 0.84 | 0.64, | 1.10 | ||
| 3 | 10.8 | 78 | 0.80 | 0.60, | 1.07 | ||
| 4 | 27.0 | 60 | 0.72 | 0.53, | 1.00 | 0.07 | |
Adjusted for age, years of smoking, the number of cigarettes smoked per day, current smoking status, total caloric intake, education, BMI category, ever consumption of tea, history of chronic bronchitis, family history of lung cancer among first-degree relatives, total fruit intake for each vegetable, and total vegetable intake for each fruit.
Statistically significant P-trend are presented in boldface type.
The inverse association for total fruit intake was marginally significant (HR = 0.75; 95% CI = 0.54–1.04 for comparing the fourth with the first quartile of intake; P-trend = 0.09). Watermelon intake was inversely associated with lung cancer risk (HR = 0.65; 95% CI = 0.47–0.90; P-trend = 0.02). The inverse association between citrus fruit intake and lung cancer risk was borderline significant (HR = 0.72; 95% CI = 0.53–1.00; P-trend = 0.07).
Dietary vitamin A and carotenoid intakes were each inversely associated with lung cancer risk; the HR (95% CI) for the fourth quartile compared with the first and P-trend were 0.63 (0.44–0.88) and 0.01 for vitamin A and 0.64 (0.46–0.88) and 0.008 for carotenoids, respectively (Table 4). The Spearman correlations of carotenoid intake with intakes of vitamin A (r = 0.93; P <0.0001), green leafy vegetables (r = 0.86; P <0.0001), and β-carotene-rich vegetables (r = 0.88; P <0.0001) were high, whereas the correlations with intakes of citrus fruit (r = 0.32; P <0.0001) and watermelon (r = 0.40; P <0.0001) were relatively low. When both carotenoids and watermelon were entered in a model, the inverse association for carotenoids was strengthened (HR = 0.55; 95% CI = 0.33–0.91; P-trend= 0.02), whereas that for watermelon was attenuated and became non-significant (HR = 0.73; 95% CI = 0.52–1.02; P-trend= 0.10). No association was observed with vitamin C (HR = 0.84; 95% CI = 0.61–1.16; P-trend = 0.17) or folate intake (HR = 0.99; 95% CI = 0.70–1.40; P-trend = 0.92).
Table 4.
Risk of Lung Cancer by Intakes of Dietary Vitamins A and C and Folate1
| Nutrient | Quartiles | Median | Cases | HR | 95% CI | P-trend2 | |
|---|---|---|---|---|---|---|---|
| Vitamin A (μg RE/day) | 1 | 359.4 | 137 | Reference | |||
| 2 | 549.8 | 89 | 0.86 | 0.65, | 1.13 | ||
| 3 | 729.2 | 79 | 0.85 | 0.64, | 1.14 | ||
| 4 | 1046.1 | 54 | 0.63 | 0.44, | 0.88 | 0.01 | |
|
| |||||||
| Carotenoids (μg/day) | 1 | 1449.8 | 138 | Reference | |||
| 2 | 2045.8 | 86 | 0.83 | 0.63, | 1.09 | ||
| 3 | 3346.9 | 78 | 0.82 | 0.62, | 1.10 | ||
| 4 | 5025.5 | 57 | 0.64 | 0.46, | 0.88 | 0.008 | |
|
| |||||||
| Retinol (μg/day) | 1 | 63.4 | 128 | Reference | |||
| 2 | 117.4 | 79 | 0.79 | 0.59, | 1.05 | ||
| 3 | 170.0 | 77 | 0.80 | 0.60, | 1.09 | ||
| 4 | 251.7 | 75 | 0.85 | 0.62, | 1.16 | 0.32 | |
|
| |||||||
| Vitamin C (mg/day) | 1 | 46.2 | 133 | Reference | |||
| 2 | 73.8 | 95 | 1.00 | 0.76, | 1.30 | ||
| 3 | 100.9 | 66 | 0.77 | 0.57, | 1.05 | ||
| 4 | 150.4 | 65 | 0.84 | 0.61, | 1.16 | 0.17 | |
|
| |||||||
| Folate (μg/day) | 1 | 217.9 | 125 | Reference | |||
| 2 | 291.3 | 83 | 0.90 | 0.67, | 1.20 | ||
| 3 | 360.2 | 75 | 0.87 | 0.63, | 1.19 | ||
| 4 | 474.4 | 76 | 0.99 | 0.70, | 1.40 | 0.92 | |
Adjusted for age, years of smoking, the number of cigarettes smoked per day, current smoking status, total caloric intake, education, BMI category, ever consumption of tea, history of chronic bronchitis, and family history of lung cancer among first-degree relatives.
Statistically significant P-trend are presented in boldface type.
We observed no association of lung cancer risk with the use of any vitamin supplement (HR = 1.11; 95% CI = 0.82–1.51), multivitamin supplements (HR = 1.08; 95% CI = 0.71–1.63), or vitamin A- or vitamin C-containing supplements (Table 5). Adjustment of supplemental vitamin intake for the dietary intake of the respective vitamin (e.g., adjusting for vitamin C-containing supplement use for dietary vitamin C intake) did not change the associations (data not shown). In our stratified analyses, some of the associations became stronger in selected subgroups, while the directions of the associations were unchanged; there was no significant interaction effect (data not shown). For instance, while there was no interaction effect by smoking, the inverse associations of lung cancer risk with intakes of β-carotene-rich vegetables and carotenoids were stronger in heavy smokers than never or light smokers (Table 6). Similar patterns were observed for associations with green leafy vegetable, carotenoid, and vitamin A intakes.
Table 5.
Risk of Lung Cancer by Vitamin Supplement Use1
| Supplement use | Cases | HR | 95% CI | P-difference | ||
|---|---|---|---|---|---|---|
| Any vitamin supplement | No | 301 | Reference | |||
| Yes | 58 | 1.11 | 0.82, | 1.51 | 0.50 | |
|
| ||||||
| Vitamin A-containing supplements | No | 329 | Reference | |||
| Yes | 30 | 1.13 | 0.77, | 1.66 | 0.54 | |
|
| ||||||
| Vitamin C-containing supplements | No | 317 | Reference | |||
| Yes | 42 | 1.05 | 0.75, | 1.47 | 0.77 | |
|
| ||||||
| Multivitamin supplements | No | 334 | Reference | |||
| Yes | 25 | 1.08 | 0.71, | 1.63 | 0.72 | |
Adjusted for age, years of smoking, the number of cigarettes smoked per day, current smoking status, total caloric intake, education, BMI category, ever consumption of tea, history of chronic bronchitis, and family history of lung cancer among first-degree relatives.
Table 6.
Risk of Lung Cancer by Selected Vegetable Groups, Watermelon, and Vitamin Intakes by Smoking History1
| Smoking | Never smokers | Light smokers | Heavy smokers | P-interaction2 | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||||||
| Intake (g/day) | quartile | cases | HR | 95% CI | cases | HR | 95% CI | cases | HR | 95% CI | ||||
| Green leafy vegetables | 1 | 13 | Reference | 35 | Reference | 77 | Reference | |||||||
| 2 | 10 | 0.79 | 0.34, | 1.81 | 25 | 0.85 | 0.50, | 1.43 | 43 | 0.80 | 0.55, | 1.17 | ||
| 3 | 12 | 0.95 | 0.43, | 2.10 | 34 | 1.22 | 0.75, | 1.97 | 44 | 0.87 | 0.60, | 1.27 | ||
| 4 | 11 | 0.91 | 0.40, | 2.08 | 22 | 0.82 | 0.47, | 1.43 | 33 | 0.63 | 0.42, | 0.96 | ||
|
| ||||||||||||||
| P-trend | 0.95 | 0.72 | 0.05 | 0.97 | ||||||||||
|
| ||||||||||||||
| β-carotene-vegetables | 1 | 12 | Reference | 38 | Reference | 79 | Reference | |||||||
| 2 | 13 | 1.12 | 0.51, | 2.46 | 27 | 0.87 | 0.53, | 1.44 | 43 | 0.73 | 0.50, | 1.07 | ||
| 3 | 10 | 0.83 | 0.36, | 1.95 | 30 | 1.02 | 0.62, | 1.66 | 41 | 0.76 | 0.52, | 1.12 | ||
| 4 | 11 | 0.93 | 0.40, | 2.17 | 21 | 0.73 | 0.42, | 1.26 | 34 | 0.64 | 0.42, | 0.97 | ||
|
| ||||||||||||||
| P-trend | 0.73 | 0.33 | 0.045 | 0.95 | ||||||||||
|
| ||||||||||||||
| Watermelon | 1 | 10 | Reference | 48 | Reference | 86 | Reference | |||||||
| 2 | 20 | 1.16 | 0.54, | 2.49 | 36 | 0.63 | 0.41, | 0.98 | 64 | 0.82 | 0.59, | 1.13 | ||
| 3 | 9 | 1.42 | 0.57, | 3.55 | 11 | 0.59 | 0.30, | 1.14 | 21 | 0.97 | 0.60, | 1.59 | ||
| 4 | 7 | 0.60 | 0.22, | 1.64 | 21 | 0.64 | 0.38, | 1.10 | 26 | 0.70 | 0.44, | 1.10 | ||
|
| ||||||||||||||
| P-trend | 0.30 | 0.11 | 0.17 | 0.56 | ||||||||||
|
| ||||||||||||||
| Carotenoids | 1 | 10 | Reference | 40 | Reference | 88 | Reference | |||||||
| 2 | 15 | 1.43 | 0.64, | 3.21 | 30 | 0.93 | 0.57, | 1.50 | 41 | 0.69 | 0.47, | 1.01 | ||
| 3 | 10 | 0.98 | 0.40, | 2.39 | 27 | 0.93 | 0.56, | 1.55 | 41 | 0.78 | 0.53, | 1.14 | ||
| 4 | 11 | 1.03 | 0.42, | 2.53 | 19 | 0.70 | 0.39, | 1.25 | 27 | 0.56 | 0.36, | 0.89 | ||
|
| ||||||||||||||
| P-trend | 0.77 | 0.24 | 0.02 | 0.80 | ||||||||||
|
| ||||||||||||||
| Vitamin A | 1 | 9 | Reference | 45 | Reference | 83 | Reference | |||||||
| 2 | 16 | 1.70 | 0.67, | 4.31 | 28 | 0.76 | 0.47, | 1.23 | 45 | 0.80 | 0.55, | 1.17 | ||
| 3 | 12 | 1.09 | 0.35, | 3.35 | 23 | 0.71 | 0.42, | 1.20 | 44 | 0.93 | 0.63, | 1.36 | ||
| 4 | 9 | 0.65 | 0.17, | 2.49 | 20 | 0.65 | 0.36, | 1.17 | 25 | 0.58 | 0.36, | 0.94 | ||
|
| ||||||||||||||
| P-trend | 0.28 | 0.14 | 0.05 | 0.63 | ||||||||||
Adjusted for age, years of smoking, the number of cigarettes smoked per day, current smoking status, total caloric intake, education, BMI category, ever consumption of tea, history of chronic bronchitis, and family history of lung cancer among first-degree relatives where applicable; light smokers were those who smoked <30 pack-years and heavy smokers were those who smoked ≥30 pack-years.
P-interaction was tested by including the main effects and the products terms and by assessing the P-value.
Discussion
In this prospective study of adult men in Shanghai aged 40 to 74 years at baseline with a mean BMI of 23.7 kg/m2, intakes of green leafy vegetables, β-carotene-rich vegetables, watermelon, vitamin A, and carotenoids were inversely associated with lung cancer risk. The associations were stronger among smokers, although there was no significant interaction with smoking status. Total vegetable and fruit intake was marginally associated with lower risk.
The inverse association between dietary carotenoids and lung cancer risk observed in this prospective study is in agreement with previous prospective studies (21,22). However, it is in contrast to findings from two prospective studies of dietary supplements (23,24) and two clinical trials of β-carotene supplementation among smokers and/or asbestos-exposed workers (7,12). These previous studies assessed carotenoid intake from supplements, not foods, suggesting that the effect of carotenoids on lung cancer may differ by source (i.e., foods or supplements). In vitro studies, animal models, and prospective cohort studies have reported potential protective effects on lung cancer for certain carotenoid subtypes (e.g., lycopene and β-cryptoxanthin) (4,25–27). These nutrients are also found in foods high in β-carotene. In our study, vitamin A-containing supplement use (no information on single retinol or carotenoid supplement use was available) was not associated with lung cancer risk, suggesting that subtypes of dietary carotenoids or other bioactive compounds found in carotenoid-rich foods, individually or jointly, may be responsible for the observed inverse association. Unfortunately, the specific carotenoid content of foods is not available in the Chinese Food Composition Tables (19). However, our food group analyses showed inverse associations for β-carotene-rich vegetables and watermelon, a lycopene-rich fruit. The inverse association for carotenoids was strengthened after adjusting for watermelon. No significant associations were found for tomatoes, which are rich in lycopene, or for lutein/zeaxanthin-rich vegetables. These observations suggest that among carotenoid subtypes, β-carotene might have made the largest contribution to the inverse association found for carotenoid intake in our study. Nonetheless, follow-up studies that include biomarker measurements would help clarify our findings.
The association between dietary vitamin A and lung cancer risk remains unclear. While our finding of an inverse association is in corroboration with a pooled analysis of prospective studies (13), other dietary and supplemental intake studies have not supported this finding (23,28). Due to the lack of an association for retinol intake and the relatively small contribution of retinol to vitamin A intake (24%) in our study, the observed inverse association for vitamin A may be attributed more to carotenoids than retinol.
The lack of association between total vegetable intake and lung cancer risk reported in our study is consistent with most previous prospective studies (3,29–32). Within subgroups of vegetables, the observed inverse association for green leafy vegetables (i.e., Chinese greens, spinach, garland chrysanthemum, shepherd's purse, watercress, and amaranth) is somewhat consistent with the finding in one previous study on leafy vegetables (spinach, borage, chard, endive, lettuce, and thistle) (29), but not with two other analyses (3,32). Since there was a significant inverse association for intake of Chinese greens, but not other green leafy vegetables, our finding on green leafy vegetables was likely driven by this vegetable. Although green leafy vegetables are high in carotenoids and folate, there was no association with folate intake is our study. Thus, the green leafy vegetable association is more consistent with the proposed antioxidant property of carotenoids. Alternatively, green leafy vegetable intake might have also served as a surrogate for unmeasured protective factors or bioactive constituents or other non-dietary protective factors for lung cancer.
Strengths of this study are the large sample size and prospective measurements of various exposures. Because of this, we were able to comprehensively adjust for potential confounders in the analyses, particularly for smoking history, which enabled us to adjust for the number of cigarettes and duration of smoking rather than smoking status alone. Nevertheless, given the well-described, strong association of smoking with lung cancer risk and the high correlation of the fruit-based dietary pattern with high socioeconomic status and health consciousness (healthier lifestyle, including non-smoking) reported in our study population (33), residual confounding due to smoking and the dietary pattern correlates cannot be fully ruled out. In contrast to North American and European populations (29–31), intake of vegetables was relatively high in our study population. Nonetheless, dietary intakes of vitamins A and C in our study and serum concentrations of these vitamins in another study in Shanghai, China (34) were relatively lower than in previous studies of Western populations (5,22,31). The lower intake levels and circulating concentrations of vitamin A probably reflect the low consumption of animal food sources, while the lower levels and concentrations of vitamin C may reflect the typical cooking practices (e.g., pan-frying, boiling, and steaming) in China. This study also has several limitations. Due to the availability of nutrients considered in the Chinese Food Composition Tables (19), we could not evaluate subtypes of carotenoids, such as β-carotene, lycopene, and lutein/zeaxanthin. Alternatively, we grouped vegetables and fruits based on the content of each carotenoid subtype and examined their associations with lung cancer risk. The lack of the dosage information on supplement use did not allow us to compute intake of vitamins from supplements. Our FFQ was previously validated against multiple 24-hour dietary recalls and its reproducibility was assessed by administering the same FFQ twice about one year apart (18). The correlation coefficients between the FFQ and multiple dietary recalls varied from 0.38 to 0.42 for carotenoids, vitamin A, and vitamin C intakes, which are comparable to the FFQs used in other large cohort studies of adult men (35,36). Give the prospective nature of our study, measurement errors in assessing dietary intakes are independent of lung cancer status and are likely non-differential (e.g., weekday/weekend variation and seasonality) (37), which would result in an underestimation of the true effect. However, like other epidemiological studies, we cannot completely rule out other unknown differential errors inherent in self-reporting of food intakes (38) . In addition, information on histologic type and stage of lung cancer was unavailable for 42% and 50% of the cases, respectively, which precluded further analysis. Certain subgroup analyses included a small number of cases (e.g., never smokers) and were therefore not sufficiently powered. Reevaluation of subgroup associations is needed when we have accrued more lung cancer cases.
In conclusion, our study found that intakes of green leafy vegetables, β-carotene-rich vegetables, watermelon, vitamin A, and carotenoids were each inversely associated with lung cancer risk. Mutual adjustment for carotenoid and watermelon intakes resulted in stronger associations for carotenoid intake, suggesting the importance of dietary carotenoid intake for lung cancer prevention, which is consistent with the summary report issued by the WCRF/AICR in 2007. Future studies are needed to confirm whether the inverse association between carotenoid intake and lung cancer risk is due to specific carotenoids or other unidentified bioactive constituents.
Acknowledgements
This work was supported by a National Cancer Institute grant (R01 CA082729). The authors thank the research staff of the Shanghai Men's Health Study for their contributions.
Footnotes
No potential conflicts of interest were declared.
References
- 1.World Cancer Research Fund/ American Institute for Cancer Research . Food, Nutrition, Physical Activity, and the Prevention of Cancer: A Global Perspective. American Institute for Cancer Research; Washington, DC: 2007. [Google Scholar]
- 2.Riboli E, Norat T. Epidemiologic evidence of the protective effect of fruit and vegetables on cancer risk. Am J Clin Nutr. 2003;78:559S–69S. doi: 10.1093/ajcn/78.3.559S. [DOI] [PubMed] [Google Scholar]
- 3.Smith-Warner SA, Spiegelman D, Yaun SS, et al. Fruits, vegetables and lung cancer: a pooled analysis of cohort studies. Int J Cancer. 2003;107:1001–11. doi: 10.1002/ijc.11490. [DOI] [PubMed] [Google Scholar]
- 4.Gallicchio L, Boyd K, Matanoski G, et al. Carotenoids and the risk of developing lung cancer: a systematic review. Am J Clin Nutr. 2008;88:372–83. doi: 10.1093/ajcn/88.2.372. [DOI] [PubMed] [Google Scholar]
- 5.Michaud DS, Feskanich D, Rimm EB, et al. Intake of specific carotenoids and risk of lung cancer in 2 prospective US cohorts. Am J Clin Nutr. 2000;72:990–7. doi: 10.1093/ajcn/72.4.990. [DOI] [PubMed] [Google Scholar]
- 6.Knekt P, Jarvinen R, Teppo L, et al. Role of various carotenoids in lung cancer prevention. J Natl Cancer Inst. 1999;91:182–4. doi: 10.1093/jnci/91.2.182. [DOI] [PubMed] [Google Scholar]
- 7.Omenn GS, Goodman GE, Thornquist MD, et al. Effects of a combination of beta carotene and vitamin A on lung cancer and cardiovascular disease. N Engl J Med. 1996;334:1150–5. doi: 10.1056/NEJM199605023341802. [DOI] [PubMed] [Google Scholar]
- 8.The effect of vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers The Alpha-Tocopherol, Beta Carotene Cancer Prevention Study Group. N Engl J Med. 1994;330:1029–35. doi: 10.1056/NEJM199404143301501. [DOI] [PubMed] [Google Scholar]
- 9.Goodman GE, Thornquist MD, Balmes J, et al. The Beta-Carotene and Retinol Efficacy Trial: incidence of lung cancer and cardiovascular disease mortality during 6-year follow-up after stopping beta-carotene and retinol supplements. J Natl Cancer Inst. 2004;96:1743–50. doi: 10.1093/jnci/djh320. [DOI] [PubMed] [Google Scholar]
- 10.Virtamo J, Pietinen P, Huttunen JK, et al. Incidence of cancer and mortality following alpha-tocopherol and beta-carotene supplementation: a postintervention follow-up. JAMA. 2003;290:476–85. doi: 10.1001/jama.290.4.476. [DOI] [PubMed] [Google Scholar]
- 11.Blot WJ, Li JY, Taylor PR, et al. Nutrition intervention trials in Linxian, China: supplementation with specific vitamin/mineral combinations, cancer incidence, and disease-specific mortality in the general population. J Natl Cancer Inst. 1993;85:1483–92. doi: 10.1093/jnci/85.18.1483. [DOI] [PubMed] [Google Scholar]
- 12.Hennekens CH, Buring JE, Manson JE, et al. Lack of effect of long-term supplementation with beta carotene on the incidence of malignant neoplasms and cardiovascular disease. N Engl J Med. 1996;334:1145–9. doi: 10.1056/NEJM199605023341801. [DOI] [PubMed] [Google Scholar]
- 13.Cho E, Hunter DJ, Spiegelman D, et al. Intakes of vitamins A, C and E and folate and multivitamins and lung cancer: a pooled analysis of 8 prospective studies. Int J Cancer. 2006;118:970–8. doi: 10.1002/ijc.21441. [DOI] [PubMed] [Google Scholar]
- 14.Mackay J, Eriksen M, Shafey O. The Tobacco Atlas. American Cancer Society; Atlanta, GA: 2006. [Google Scholar]
- 15.Howlader N, Noone AM, Krapcho M, et al. Surveillance Epidemiology and End Results Program Cancer Statistucs Review, 1975–2008. National Cancer Institute; Bethesda, MD: 2011. [Google Scholar]
- 16.Centers for Disease Control and Prevention State-specific prevalence and trends in adult cigarette smoking - United States, 1998–2007. MMWR. 2009;58:221–6. [PubMed] [Google Scholar]
- 17.Yang G, Xiang YB, Zheng W, et al. Body weight and weight change in relation to blood pressure in normotensive men. J Hum Hypertens. 2007;21:45–52. doi: 10.1038/sj.jhh.1002099. [DOI] [PubMed] [Google Scholar]
- 18.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;97:993–1000. doi: 10.1017/S0007114507669189. [DOI] [PubMed] [Google Scholar]
- 19.Yang G, Wang G, Pan X. China Food Composition Tables. Beijing University Medical Press; Beijing, China: 2002. [Google Scholar]
- 20.Perkins-Veazie P, Collins JK. Carotenoid changes of intact watermelons after storage. J Agric Food Chem. 2006;54:5868–74. doi: 10.1021/jf0532664. [DOI] [PubMed] [Google Scholar]
- 21.Bandera EV, Freudenheim JL, Marshall JR, et al. Diet and alcohol consumption and lung cancer risk in the New York State Cohort (United States) Cancer Causes Control. 1997;8:828–40. doi: 10.1023/a:1018456127018. [DOI] [PubMed] [Google Scholar]
- 22.Holick CN, Michaud DS, Stolzenberg-Solomon R, et al. Dietary carotenoids, serum beta-carotene, and retinol and risk of lung cancer in the alpha-tocopherol, beta-carotene cohort study. Am J Epidemiol. 2002;156:536–47. doi: 10.1093/aje/kwf072. [DOI] [PubMed] [Google Scholar]
- 23.Satia JA, Littman A, Slatore CG, et al. Long-term use of beta-carotene, retinol, lycopene, and lutein supplements and lung cancer risk: results from the VITamins And Lifestyle (VITAL) study. Am J Epidemiol. 2009;169:815–28. doi: 10.1093/aje/kwn409. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Roswall N, Olsen A, Christensen J, et al. Source-specific effects of micronutrients in lung cancer prevention. Lung Cancer. 2010;67:275–81. doi: 10.1016/j.lungcan.2009.11.010. [DOI] [PubMed] [Google Scholar]
- 25.Kohno H, Taima M, Sumida T, et al. Inhibitory effect of mandarin juice rich in beta-cryptoxanthin and hesperidin on 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone-induced pulmonary tumorigenesis in mice. Cancer Lett. 2001;174:141–50. doi: 10.1016/s0304-3835(01)00713-3. [DOI] [PubMed] [Google Scholar]
- 26.Lian F, Hu KQ, Russell RM, et al. Beta-cryptoxanthin suppresses the growth of immortalized human bronchial epithelial cells and non-small-cell lung cancer cells and up-regulates retinoic acid receptor beta expression. Int J Cancer. 2006;119:2084–9. doi: 10.1002/ijc.22111. [DOI] [PubMed] [Google Scholar]
- 27.Lian F, Smith DE, Ernst H, et al. Apo-10'-lycopenoic acid inhibits lung cancer cell growth in vitro, and suppresses lung tumorigenesis in the A/J mouse model in vivo. Carcinogenesis. 2007;28:1567–74. doi: 10.1093/carcin/bgm076. [DOI] [PubMed] [Google Scholar]
- 28.Chow WH, Schuman LM, McLaughlin JK, et al. A cohort study of tobacco use, diet, occupation, and lung cancer mortality. Cancer Causes Control. 1992;3:247–54. doi: 10.1007/BF00124258. [DOI] [PubMed] [Google Scholar]
- 29.Buchner FL, Bueno-de-Mesquita HB, Linseisen J, et al. Fruits and vegetables consumption and the risk of histological subtypes of lung cancer in the European Prospective Investigation into Cancer and Nutrition (EPIC) Cancer Causes Control. 2010;21:357–71. doi: 10.1007/s10552-009-9468-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Wright ME, Park Y, Subar AF, et al. Intakes of fruit, vegetables, and specific botanical groups in relation to lung cancer risk in the NIH-AARP Diet and Health Study. Am J Epidemiol. 2008;168:1024–34. doi: 10.1093/aje/kwn212. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Neuhouser ML, Patterson RE, Thornquist MD, et al. Fruits and vegetables are associated with lower lung cancer risk only in the placebo arm of the beta-carotene and retinol efficacy trial (CARET) Cancer Epidemiol Biomarkers Prev. 2003;12:350–8. [PubMed] [Google Scholar]
- 32.Shibata A, Paganini-Hill A, Ross RK, et al. Intake of vegetables, fruits, beta-carotene, vitamin C and vitamin supplements and cancer incidence among the elderly: a prospective study. Br J Cancer. 1992;66:673–9. doi: 10.1038/bjc.1992.336. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.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;98:1006–13. doi: 10.1017/S0007114507750900. [DOI] [PubMed] [Google Scholar]
- 34.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;10:767–73. [PubMed] [Google Scholar]
- 35.Rimm EB, Giovannucci EL, Stampfer MJ, et al. Reproducibility and validity of an expanded self-administered semiquantitative food frequency questionnaire among male health professionals. Am J Epidemiol. 1992;135:1114–26. doi: 10.1093/oxfordjournals.aje.a116211. [DOI] [PubMed] [Google Scholar]
- 36.Stram DO, Hankin JH, Wilkens LR, et al. Calibration of the dietary questionnaire for a multiethnic cohort in Hawaii and Los Angeles. Am J Epidemiol. 2000;151:358–70. doi: 10.1093/oxfordjournals.aje.a010214. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Cai H, Yang G, Xiang YB, et al. Sources of variation in nutrient intakes among men in Shanghai, China. Public Health Nutr. 2005;8:1293–9. doi: 10.1079/phn2005748. [DOI] [PubMed] [Google Scholar]
- 38.Kohlmeier L, Bellach B. Exposure assessment error and its handling in nutritional epidemiology. Annu Rev Public Health. 1995;16:43–59. doi: 10.1146/annurev.pu.16.050195.000355. [DOI] [PubMed] [Google Scholar]