NUTRITION AND CANCER, 55(1), 71–77 Copyright © 2006, Lawrence Erlbaum Associates, Inc.

Antioxidant Vitamins and Risk of Gastric Cancer: A Case-Control Study in Portugal
Nuno Lunet, Carmen Valbuena, Fátima Carneiro, Carlos Lopes, and Henrique Barros

Abstract: We quantified the effect of antioxidant vitamins in gastric cancer risk, taking into account Helicobacter pylori seropositivity and overall fruit and vegetable intake. Incident cases were identified in two large hospitals in Porto, Portugal, and controls were randomly sampled among city dwellers. Food intake was assessed with a previously validated semiquantitative food-frequency questionnaire. A commercially available chromatographic immunoassay was used for the detection of immunoglobulin G antibodies. Complete questionnaire information and serum samples were available for 233 cases and 311 controls. Compared with subjects in the lowest tertile of dietary intake, the odds ratios (ORs) for those in the highest were 0.85 (95% confidence interval, CI = 0.45–1.60) for vitamin C, 1.04 (95% CI = 0.60–1.80) for vitamin E, and 1.33 (95% CI = 0.77–2.30) for provitamin A carotenoids after further adjusting for fruit and vegetable consumption. Fruit and vegetables remained an independent protective factor (OR = 0.45; 95% CI = 0.23–0.89) after further adjustment for the intake of antioxidant vitamins. H. pylori status had no significant interaction with dietary items. Factors other than H. pylori infection and intake of vitamin C and provitamin A carotenoids seem to account for the inverse association between fruit and vegetable consumption and gastric cancer.

At an ecological level, the availability of fruit and vegetables is poorly correlated with gastric cancer frequency (7), and countries such as Portugal, Spain, or Italy present high intakes of fruit and vegetables but also the highest stomach cancer rates in Western Europe. In addition, the marked differences in gastric cancer incidence across regions with a similar prevalence of Helicobacter pylori infection suggest that outcomes depend on the combined effect of H. pylori pathogenicity, host susceptibility, and environmental exposures such as diet. Epidemiological studies have demonstrated a causal link between H. pylori infection and gastric carcinogenesis (8) but rarely addressed the potential interaction between H. pylori and dietary factors (9–11). Therefore, we aimed to quantify the effect of dietary antioxidant vitamins on gastric cancer risk, taking into account the overall contribution of fruit and vegetable consumption and H. pylori infection.

Materials and Methods From June 2001 to May 2004, we evaluated incident cases of gastric cancer and suitable community controls in Porto, the largest town in the northern region of Portugal. Complete questionnaire information and serum samples were available for 233 cases and 311 controls. Cases During the study period, we identified 429 cancer cases of stomach adenocarcinoma admitted to the surgery wards of the two major public hospitals caring for cancer patients in the north of Portugal, Hospital de S. João and Instituto Português de Oncologia Francisco Gentil, both in Porto. To be eligible, patients had to be diagnosed with an incident gastric cancer, have no previous cancer diagnosis except skin non-melanoma and no subtotal gastrectomy for benign conditions, and be able to provide informed consent. Cases were evaluated during their hospital stay, as soon as possible after admission, and

Introduction There is a convincing body of evidence linking high fruit and vegetable intake with a reduced risk of gastric cancer (1,2). Vitamin C and carotenoids (1) have been considered the most likely preventative candidates, but other compounds present in fruit and vegetables may also be involved (3,4). The strength of previously reported associations was challenged by the results of prospective observational studies (5) and the experimental investigations (6) that failed to show a decreased risk of gastric cancer when subjects were exposed to nondietary sources of antioxidants.

N. Lunet and H. Barros are affiliated with the Department of Hygiene and Epidemiology, University of Porto Medical School, Porto, Portugal. C. Valbuena and F. Carneiro are affiliated with the Department of Pathology, Hospital S. João, Porto, Portugal. F. Carneiro is also affiliated with the Institute of Molecular Pathology, University of Porto (IPATIMUP) and the University of Porto Medical School, Porto, Portugal. C. Lopes is affiliated with the Department of Pathology, Portuguese Institute of Oncology, Porto, Portugal.

mostly before treatment. We could not interview 66 cases due to early discharge or premature death (n = 27), incapacity to understand the studypurposes (n = 26), or refusal to participate (n = 13). Patients not interviewed were significantlyolder (median age = 73.5 vs. 65.0; P < 0.001) and were more frequently males (72.7% vs. 57.6%; P = 0.02). To assess cognitive function, all individuals older than 64 yr took a Mini-Mental State Examination. This resulted in the exclusion from analysis of 28 cases scoring less than 18 (12). Controls A representative sample of the adult Porto population assembled as part of an ongoing health and nutrition survey was used as source of controls. A detailed description of the general selection procedures and subject characteristics was made previously (13). In brief, subjects were recruited by random digit dialing using households as the sampling frame followed by simple random sampling to select one eligible person among permanent residents in each household that was invited to visit our department for interview and blood collection. Serum was kept frozen at –20°C until analysis. Twenty-five hundred community controls aged 18 to 93 were finally evaluated corresponding to a participation proportion of 70%. For each case, we selected one age and sex frequency-matched control. Subjects aged above 64 yr scoring less than 18 in the Mini-Mental State Examination (12) were not eligible. As more than one control was available for every case, we opted for those more recently evaluated. Evaluation of Subjects Cases and controls completed the same comprehensive structured questionnaire, providing information on demographic, social, behavioral, and medical characteristics, namely, socioeconomic status, diet, and other lifestyles, and a blood sample was drawn. Dietary habits were recorded using a semiquantitative food-frequency questionnaire (FFQ) comprising 82 food items or beverage categories. It was designed according to Willett (14) and adapted by inclusion of a variety of typical Portuguese food items. The choice of relevant food items to be included in the FFQ was based on their contribution to the between-person variance of intake of total energy, proteins, fat, carbohydrates, cholesterol, dietary fiber, vitamin A, provitamin A carotenoids, vitamins C and E, calcium, alcohol, and caffeine. Certain foods with similar nutrient composition were grouped together as a single item. The FFQ was validated by two different methods, with a 7-day food record in 75 female and 71 male community subjects and, regarding the fatty acid composition, with the composition of subcutaneous adipose tissue in 64 female and 52 male subjects (15). Fruits considered in the analysis were apple or pear, orange or tangerine, banana, kiwi, strawberry, cherry, peach or plum, melon or watermelon, persimmon, fig or loquat or apricot, and grapes. The vegetables were lettuce, watercress, tomato, cucumber, green and white cabbages, broccoli, cau72

liflower or Brussels sprout, spinach or spring greens or turnip greens, spinach, bean pod, carrot, turnip, green pepper, and onion. Vegetables from mixed dishes were not included. For each FFQ item, subjects were asked the average frequency of consumption (nine possible responses ranging from never to six or more times per day), the portion size usually consumed (based on a manual of photographs with small, medium, and large portion sizes), and the number of months during which it had been consumed in the previous year. This information was used to estimate, for each FFQ item, the frequency of consumption of medium servings corrected for seasonality by multiplying the reported frequency by the ratio between the weight of the serving more frequently consumed and the weight of the medium serving and by the ratio between the number of months during which the food item was reported to be consumed and 12 [for example, for a subject eating four large servings of watermelon per week but only in the three summer months, we considered an average consumption of two times per week during the previous 12 mo, that is (four times per week) × (3 mo/12 mo) × (300 g/150 g) = two times per week]. In the analysis, the overall food group intake was established by adding up the amounts of single items or groups consumed per day. Food Processor Plus®, version 5.0 (ESHA Research, Salem, OR), was used to obtain nutrient data, and energy-adjusted nutrient intakes were computed using the Willet and Stampfer method (16). The FFQ evaluated dietary intake in the previous year or the year before onset of symptoms when appropriate for cases. In addition to dietary information, subjects were asked about the use of vitamin and mineral supplements in the same period. Complete information on dietary habits was obtained for 320 of 335 cases, from which 15 declared to have changed their food intake 12 or more months before interview due to gastrointestinal symptoms and were excluded from this analysis. Subjects that modified their habits during the previous year were not excluded but were asked to recall dietary intake in the year before the change in habits. Two hundred thirty-three of 305 eligible cases provided a venous blood sample for the study purposes, and serum was frozen at –20°C until analysis. A commercially available chromatographic immunoassay (H. pylori Rapid Test Device, SureScreen Diagnostics, Derby, UK) was used for the detection of IgG antibodies, with results interpreted according to the manufacturer’s instructions. Cases from whom a blood sample was not available were not significantly different concerning age (median = 64.5 yr vs. 63 yr; P = 0.36), sex (males = 63.5% vs. 59.8%; P = 0.54), or intake of fruit and vegetables (median number of servings per day = 3.2 vs. 3.1; P = 0.44). The proportion of controls providing a blood sample was 82.3%. Controls that did not have their anti–H. pylori IgG assessed, because blood was not available or controls were not selected for matching with cases, were not significantly different concerning age (median = 61 yr vs. 61 yr; P = 0.28) or intake of fruit and vegetables (median number of servings per Nutrition and Cancer 2006

day = 3.9 vs. 4.0; P = 0.46). The proportion of males was lower in controls with unknown H. pylori infection status (46.9% vs. 57.9%; P < 0.001). Statistical Analysis Sociodemographic and lifestyle characteristics of cases and controls were compared using the χ2 and the Kruskal-Wallis tests, respectively, for proportions and quantitative variables. The association between energy-adjusted nutrient intakes or fruit and vegetable consumption and cancer was quantified with odds ratios (ORs) and the corresponding 95% confidence intervals (CIs). Risk estimates for tertiles of intake of each evaluated item were derived from unconditional multiple logistic regression models computed with STATA®, version 9.0 (Stata Corp., College Station, TX). ORs for the association between antioxidant vitamins or fruit and vegetables and gastric cancer, according to H. pylori infection status, were obtained with models including interaction terms between infection and each of the dietary variables. The coefficients for the dietary items originate the OR for the H. pylori-noninfected cases and controls, and, when added to the different coefficients of interaction, the risk estimates for the H. pylori-infected cases and controls are obtained. All the analyses in this article were performed using 233 cases (194 H. pylori positive and 39 H. pylori negative) and 311 controls (232 H. pylori positive and 79 H. pylori negative). Ethics The ethics committee of each involved hospital approved the study, and all subjects provided written informed consent.

ber of siblings and were more likely to be infected with H. pylori and less likely to use vitamin and mineral supplements. The median consumption of fruit and vegetables and the dietary intake of vitamins C and E were significantly lower in cases. All the variables presented in Table 1 were considered in the multivariate analysis of dietary data. In multivariate analysis (Table 2), using the lowest tertile of dietary intake as the reference category, the OR for the highest tertile was 0.56 (95% CI = 0.33–0.95) for vitamin C, 0.76 (95% CI = 0.46–1.24) for vitamin E, and 0.94 (95% CI = 0.57–1.54) for provitamin A carotenoids. The risk estimates were, respectively, 0.85, 1.04, and 1.33 for vitamin C, vitamin E, and provitamin A carotenoids when fruit and vegetable consumption was added to the multivariate model, but the magnitude of the association between gastric cancer and the consumption of fruit and vegetables remained similar when further adjustments were performed accounting for the specific intake of antioxidant vitamins. No statistically significant interaction was observed between vitamin intake and H. pylori infection status (Table 3). For fruit and vegetables, the ORs favoring stomach cancer were 0.53 (95% CI = 0.30–0.94) in H. pylori–positive subjects and 0.21 (95% CI = 0.06–0.75) among those without serological evidence of infection (P for interaction = 0.28).

Discussion Our results confirm previous findings of case-control studies, showing an inverse association between fruit and vegetable consumption and risk of gastric cancer, but do not support the hypothesis of vitamin C, vitamin E, or provitamin A carotenoids as major contributors for this effect. No significant association was found between vitamin intake and gastric cancer after inclusion of fruit and vegetable consumption in the statistical model, both in H. pylori–positive and –negative subjects. Methodological shortcomings of case-control studies, and specifically the design options of the present investigation, should be discussed. Cancer registry data could not be used to identify cases because of the local delay in cancer notifica-

Results Table 1 presents selected sociodemographic and lifestyle characteristics of the study subjects as well as their H. pylori infection status. Compared with controls, cancer patients had a lower median number of education years and a higher numTable 1. Characteristics of the Subjectsa
Controls (n = 311) Males (%) Age (yr) Education (yr) Number of siblings Helicobacter pylori infected (%) Calories (kcal/day) Fruit and vegetables (servings/day) Vitamin C (mg/day) Vitamin E (mg/day) Carotenoids (retinol equivalents/day)b Use of vitamin/mineral supplements 57.9 61 (51–71) 6 (4–11) 3 (2–5) 74.6 2,249 (1,865–2,693) 4.0 (2.8–5.6) 118.8 (85.1–163.5) 8.0 (6.1–9.6) 994 (633–1,395) 27.6

Cases (n = 233) 57.8 63 (50–72) 4 (3–4) 5 (3–7) 83.3 2,163 (1,743–2,654) 3.1 (1.9–4.5) 100.9 (66.0–137.4) 7.2 (5.5–9.1) 921 (646–1,291) 19.2

P 0.97 0.52 <0.001 <0.001 0.02 0.30 <0.001 <0.001 <0.005 0.55 0.02

a: Results are presented as a proportion or as a median (interquartile range) as appropriate. b: Provitamin A carotenoids (1 retinol equivalent = 10 International Units of vitamin A activity from β-carotene).

Vol. 55, No. 1

73

Table 2. Fruit, Vegetable, and Antioxidant Vitamin Intake and Risk of Gastric Cancer (233 cases and 311 controls)a
Controls n (%) Fruit and vegetables (servings/day) 1st tertile (≤3.4)d 2nd tertile (3.5–5.1)f 3rd tertile (≥5.2)g Vitamin C (mg/day) 1st tertile (≤100.3) 2nd tertile (100.4–144.6) 3rd tertile (≥144.7) Vitamin E (mg/day) 1st tertile (≤7.2) 2nd tertile (7.3–8.5) 3rd tertile (≥8.6) Carotenoids (retinol equivalents/day)h 1st tertile (≤756.0) 2nd tertile (756.1–1,265.0) 3rd tertile (≥1,265.1) Cases n (%) OR (95% CI)
b

OR (95% CI)

c

104 (33.4) 104 (33.4) 103 (33.2) 104 (33.4) 104 (33.4) 103 (33.2) 104 (33.4) 104 (33.4) 103 (33.2) 104 (33.4) 104 (33.4) 103 (33.2)

130 (55.8) 58 (24.9) 45 (19.3) 116 (49.8) 80 (34.3) 37 (15.9) 108 (46.4) 65 (27.9) 60 (25.7) 78 (33.5) 86 (36.9) 69 (29.6)

1 0.44 (0.27–0.71) 0.45 (0.27–0.77) 1 0.90 (0.58–1.42) 0.56 (0.33–0.95) 1 0.68 (0.43–1.08) 0.76 (0.46–1.24) 1 1.01 (0.63–1.63) 0.94 (0.57–1.54)

1e 0.43 (0.25–0.74) 0.45 (0.23–0.89) 1 1.20 (0.73–1.97) 0.85 (0.45–1.60) 1 0.89 (0.54–1.47) 1.04 (0.60–1.80) 1 1.25 (0.76–2.05) 1.33 (0.77–2.30)

a: Abbreviations are as follows: OR, odds ratio; CI, confidence interval. b: Adjusted for age (continuous), sex, education (0–3, 4, and ≥5 yr), number of siblings (0, 1–2, 3–4, and ≥5 siblings), Helicobacter pylori infection, vitamin and mineral supplement use, and total caloric intake (continuous). c: Adjusted as in b and for fruit and vegetable dietary intake (tertiles), except when fruit and vegetable dietary intake is evaluated. d: ≤295 g. e: Adjusted as in b and for vitamin C, vitamin E, and carotenoid dietary intake (tertiles). f: 296–458 g. g: ≥459 g. h: Provitamin A carotenoids (1 retinol equivalent = 10 International Units of vitamin A activity from β-carotene).

tion and data handling that precluded any attempt to interview cases in due time. Therefore, we opted to approach every case admitted to Hospital de S. João, the largest general hospital serving the northern region of Portugal and an academic institution with a long oncological tradition (17), and to Instituto Português de Oncologia Francisco Gentil, Porto, the cancer hospital serving the region, and to compare them with referents from the same town. The sex and age distribution of gastric cancer cases identified in our study is similar to that observed in cases from Porto. The patients’ social class, however, may be different in the hospitals involved in our study and in institutions that were not included, especially private clinics. The proportion of refusals was low, and we were able to interview 84.6% of eligible cases. Those not interviewed were older and therefore more likely to be not eligible due to cognitive impairment. The exposures of interest, however, are probably not related to the reasons that some patients were not interviewed. Serum samples were available only for 77.1% of cases. This was due to the fact that blood collection was performed when patients had blood drawn for routine analysis in the hospital, allowing for an extra vial to be collected. This procedure ensures that the patients are not unnecessarily bothered but makes the logistics of blood collection much more difficult, reflected in the loss of a quarter of the subjects for this evaluation. However, no statistically significant differences were observed in those providing blood samples compared with subjects for whom it was not possible to collect blood regarding gender, age, education, or fruit and vegetable consumption. The same was observed for controls, except for 74

gender, and no consequences are expected for our conclusions from the lack of H. pylori data for these subjects. It was impossible to blind interviewers for the case or control status of subjects, and information bias may occur with a case-control design. However, the team of interviewers was thoroughly trained to treat cases and controls uniformly, the questionnaires were precoded, and interviewers evaluated both cases and controls. Differential recall bias due to knowledge of the disease status is a major concern in case-control studies. Furthermore, cases might have changed diet due to their disease or its precursors, and the recall of past dietary intake can be influenced by current habits (18). Some gastric cancer patients might have decreased their consumption of fruit and vegetables before diagnosis because of gastrointestinal symptoms. In practice, this change may bias the recall of past intake, underestimating true consumption and resulting in spurious inverse associations. Because illness duration may tend to modify the accuracy of dietary reporting, our study was limited to incident cases, we obtained data reporting to the year preceding the diagnosis, and we excluded from analysis cases declaring to have changed dietary habits more than 1 yr before interview to improve the recall of food intake. Nevertheless, the stability of dietary habits in general does not exclude the hypothesis that subjects with subclinical cancer, in particular, gastrointestinal cancer, may have changed their diet without perceiving it because of symptoms associated with the not-yet-diagnosed disease. A validated FFQ (15) was used to estimate usual food intake, and seasonal variations in intake of specific food items were considered for a more accurate evaluation of the intake Nutrition and Cancer 2006

Vol. 55, No. 1
n (%) H. pylori Positive H. pylori Negative H. pylori Positive H. pylori Negative P Interaction H. pylori Positive OR (95% CI)b OR (95% CI)c H. pylori Negative P Interaction 178 (41.8) 135 (31.7) 113 (26.5) 0.28 178 (41.8) 147 (34.5) 101 (23.7) 0.68 174 (40.8) 127 (29.8) 125 (29.3) 141 (33.1) 150 (35.2) 135 (31.7) 41 (34.8) 40 (33.9) 37 (31.4) 1 1.01 (0.59–1.71) 1.07 (0.62–1.85) 1 1.02 (0.36–2.88) 0.54 (0.18–1.65) 38 (32.2) 42 (35.6) 38 (32.2) 1 0.68 (0.40–1.14) 0.87 (0.51–1.48) 1 0.66 (0.24–1.79) 0.40 (0.12–1.29) 42 (35.6) 37 (31.4) 39 (33.0) 1 0.96 (0.58–1.58) 0.53 (0.29–0.95) 1 0.71 (0.26–1.98) 0.69 (0.23–2.07) 56 (47.5) 27 (22.9) 35 (29.7) 1 0.42 (0.25–0.72) 0.53 (0.30–0.94) 1 0.58 (0.20–1.70) 0.21 (0.06–0.75) 1e 0.42 (0.24–0.74) 0.54 (0.26–1.10) 1 1.31 (0.76–2.28) 0.80 (0.40–1.59) 1e 0.58 (0.19–1.81) 0.21 (0.06–0.81) 1 0.82 (0.30–2.42) 1.11 (0.35–3.50) 0.25 0.51 0.43 1 0.89 (0.51–1.56) 1.15 (0.64–2.07) 1 0.82 (0.30–2.28) 0.59 (0.17–2.01) 0.57 0.45 1 1.22 (0.71–2.13) 1.48 (0.82–2.69) 1 1.30 (0.45–3.76) 0.84 (0.26–2.68) 0.56

Table 3. Fruit, Vegetable, and Antioxidant Vitamin Intake and Risk of Gastric Cancer According to Helicobacter pylori Infection Status (233 cases and 311 controls)a

Fruit and vegetables (servings/day) 1st tertile (≤3.4)d 2nd tertile (3.5–5.1)f 3rd tertile (≥5.2)g Vitamin C (mg/day) 1st tertile (≤100.3) 2nd tertile (100.4–144.6) 3rd tertile (≥144.7) Vitamin E (mg/day) 1st tertile (≤7.2) 2nd tertile (7.3–8.5) 3rd tertile (≥8.6) Carotenoids (retinol equivalents/day)h 1st tertile (≤756.0) 2nd tertile (756.1–1265.0) 3rd tertile (≥1,265.1)

a: b: c: d: e: f: g: h:

Abbreviations are as follows: OR, odds ratio; CI, confidence interval. Adjusted for age (continuous), sex, education (0–3, 4, and ≥5 yr), number of siblings (0, 1–2, 3–4, and ≥5 siblings), vitamin and mineral supplement use, and total caloric intake (continuous). Adjusted as in b and for fruit and vegetable dietary intake (tertiles), except when fruit and vegetable dietary intake is evaluated. ≤295 g. Adjusted as in b and for vitamin C, vitamin E, and carotenoid dietary intake (tertiles). 296–458 g. ≥459 g. Provitamin A carotenoids (1 retinol equivalent = 10 International Units of vitamin A activity from β-carotene).

75

of food items rich in antioxidant vitamins, more often consumed in specific times of the year. In addition to the discussion of internal validity, it should also be noted that there are limits to the generalizability of our results, as the premises for the hypothesis being tested vary across populations. The differences in the epidemiology of H. pylori infection, namely, the age at acquisition of infection or the virulence of the strains, and the distinct exposure to fruit and vegetables, quantitatively and qualitatively, may contribute to findings specific to each setting. In the Mediterranean diet, foods of plant origin are the main dietary sources of vitamin C and carotenoids and are also an important source of vitamin E (19). However, individual nutrients might mainly indicate the relative intake of specific foods, which provide several other compounds able to act synergistically or alone. As in the recent reanalysis of data from an earlier published Greek study on diet and gastric cancer (4,20), we found no association between dietary vitamin C and gastric cancer occurrence after controlling for fruit and vegetable consumption. The present study also focused on the role of H. pylori infection as a potential modifier of the effect of antioxidant vitamins on gastric cancer risk. Gastric inflammation induced by H. pylori colonization releases free radicals, increasing the likelihood of carcinogenesis, and the levels of vitamin C (21–23), vitamin E (24,25), and β-carotene (25) may be lower when infection is present, suggesting a direct action of these compounds in restraining mucosal damage promoted by H. pylori. Therefore, it may be hypothesized that the increase in gastric cancer risk by H. pylori infection is higher when the levels of antioxidants or the dietary intake of fruit and vegetables are lower, due to a more-limited capacity to protect the stomach from the deleterious effects of H. pylori. On the other hand, the protective effect of these dietary items may be more important when infection is present. In our investigation, the power to evaluate subtle H. pylori–diet interactions is somewhat low, although it is adequate to assess stronger effect modification such as that suggested by other studies on this topic. Four previous reports addressed the effect of dietary habits on gastric cancer accounting for the effect of H. pylori infection. Kim et al. (295 cases and 295 controls) (11) found a statistically significant interaction between infection and vitamin C intake, with H. pylori significantly increasing gastric cancer risk only when vitamin C intake was low (OR = 4.68; 95% CI = 1.97–11.1) but not when consumption was high (OR = 0.72; 95% CI = 0.32–1.65). Machida-Montani et al. (122 cases and 235 controls) (10) found stronger associations between H. pylori infection and gastric cancer when fruit (OR = 5.8; 95% CI = 2.0–16.9 vs. OR = 10.6; 95% CI = 3.3–33.9) and vegetable (OR = 7.6; 95% CI = 2.3–25.2 vs. OR = 8.5; 95% CI = 2.4–29.9) intake was low, but the interaction was not statistically significant (P = 0.32 for fruit and P = 0.60 for vegetables). A hospital-based study (211 cases and 454 controls) (9), with vitamin C roughly estimated as the frequency of consumption of a selected group of fruit and vegetables, 76

showed that vitamin C intake was associated with a nonstatistically significant reduction in gastric cancer risk and no interaction between H. pylori CagA status and vitamin C intake on cancer risk. The serologic test used in our study does not differentiate between H. pylori subtypes, but, in a previous study (26) in our region, 55.6% of subjects with gastritis and 87.9% of gastric cancer cases were CagA positive. More recently, in a report from the EPIC cohort (a nested case-control analysis with 241 cases and 1,141 controls) (27), small differences were observed for the association between gastric cancer and fresh fruit consumption in subjects H. pylori infected (OR = 0.98; 95% CI = 0.81–1.20) and not infected (OR = 0.72; 95% CI = 0.39–1.33; P for interaction = 0.013). In addition to the conflicting findings obtained in other investigations, in a previous analysis of our data (28) regarding the interaction between vitamin C intake and H. pylori infection, the point estimates did not vary consistently when different categories of exposure were considered. The elucidation of interactions between H. pylori and lifestyles is essential to understand individual risk and differences in gastric cancer frequency between regions with similar prevalence of infection. However, interactions such as the one anticipated for H. pylori and vitamin C clearly need larger samples to be disclosed. In conclusion, high fruit and vegetable consumption decreases gastric cancer risk, both in H. pylori–positive and in H. pylori–negative subjects, and factors other than dietary antioxidant vitamins C and E and provitamin A carotenoids seem to account for this protective effect.

Acknowledgments and Notes
Grants from Fundação para a Ciência e a Tecnologia (POCTI/SAUESP/56126/2004, POCI/SAU-ESP/61685/2004, SFRH/BD/3293/2000, POCTI/ESP/35769/99, and PRAXIS/2/2.1/SAU/ 1332/1995) are gratefully acknowledged. Address correspondence to N. Lunet, Serviço de Higiene e Epidemiologia da Faculdade de Medicina do Porto, Al. Prof. Hernâni Monteiro, 4200 - 319 Porto, Portugal. Phone: + 351 225513652. FAX: + 351 225513653. E-mail: nlunet@med.up.pt. Submitted 18 October 2005; accepted in final form 23 March 2006.

References
1. World Cancer Research Fund: Food, Nutrition and the Prevention of Cancer: A Global Perspective. Washington, DC: American Institute for Cancer Research, 1997. 2. Gonzalez CA and EPIC Working Group on Gastric Cancer: Vegetable, fruit and cereal consumption and gastric cancer risk. IARC Sci Publ 156, 79–83, 2002. 3. McCullough ML and Giovannucci EL: Diet and cancer prevention. Oncogene 23, 6349–6364, 2004. 4. Lagiou P, Samoli E, Lagiou A, Peterson J, Tzonou A, et al.: Flavonoids, vitamin C and adenocarcinoma of the stomach. Cancer Causes Control 15, 67–72, 2004.

Nutrition and Cancer 2006

5. Lunet N, Lacerda-Vieira A, and Barros H: Fruit and vegetables consumption and gastric cancer: a systematic review and meta-analysis of cohort studies. Nutr Cancer 53, 1–10, 2005. 6. Bjelakovic G, Nikolova D, Simonetti RG, and Gluud C: Antioxidant supplements for prevention of gastrointestinal cancers: a systematic review and meta-analysis. Lancet 364, 1219–1228, 2004. 7. Lunet N and Barros H: Helicobacter pylori infection and gastric cancer: facing the enigmas. Int J Cancer 106, 953–960, 2003. 8. International Agency for Research on Cancer: Schistosomes, Liver Flukes and Helicobacter pylori. Lyon, France: International Agency for Research on Cancer, 1994. (IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, vol 61) 9. Lopez-Carrillo L, Torres-Lopez J, Galvan-Portillo M, Munoz L, and Lopez-Cervantes M: Helicobacter pylori-CagA seropositivity and nitrite and ascorbic acid food intake as predictors for gastric cancer. Eur J Cancer 40, 1752–1759, 2004. 10. Machida-Montani A, Sasazuki S, Inoue M, Natsukawa S, Shaura K, et al.: Association of Helicobacter pylori infection and environmental factors in non-cardia gastric cancer in Japan. Gastric Cancer 7, 46–53, 2004. 11. Kim DS, Lee MS, Kim YS, Kim DH, Bae JM, et al.: Effect modification by vitamin C on the relation between gastric cancer and Helicobacter pylori. Eur J Epidemiol 20, 67–71, 2005. 12. Murden RA, McRae TD, Kaner S, and Bucknam ME: Mini-Mental State exam scores vary with education in blacks and whites. J Am Geriatr Soc 39, 149–155, 1991. 13. Ramos E, Lopes C, and Barros H: Investigating the effect of nonparticipation using a population-based case-control study on myocardial infarction. Ann Epidemiol 14, 437–441, 2004. 14. Willett W: Nutritional Epidemiology (2nd ed). New York: Oxford University Press, 1998. 15. Lopes C: [Reproducibility and validity of a food frequency questionnaire]. In [Diet and Acute Myocardial Infarction] (PhD thesis). Porto, Portugal: University of Porto Medical School, 2000, pp 79–115. 16. Willet WC and Stampfer MJ: Total energy intake: implications for epidemiologic analyses. Am J Epidemiol 124, 17–27, 1986. 17. Sobrinho-Simões M: “Re-makes,” “Re-visitas” e o Prof. Joaquim Bastos. Arq Med 3, 327–328, 1990. 18. Botterweck AM, van den Brandt PA, and Goldbohm RA: A prospective study cohort study on vegetables and fruit consumption and stomach cancer risk in Netherlands. Am J Epidemiol 148, 842–853, 1998.

19. Garcia-Closas R, Berenguer A, Jose Tormo M, Jose Sanchez M, Quiros JR, et al.: Dietary sources of vitamin C, vitamin E and specific carotenoids in Spain. Br J Nutr 91, 1005–1011, 2004. 20. Trichopoulos D, Ouranos G, Day NE, Tzonou A, Manousos O, et al.: Diet and cancer of the stomach: a case–control study in Greece. Int J Cancer 36, 291–297, 1985. 21. Ruiz B, Rood JC, Fontham ET, Malcom GT, Hunter FM, et al.: Vitamin C concentration in gastric juice before and after anti-Helicobacter pylori treatment. Am J Gastroenterol 89, 533–539, 1994. 22. Rokkas T, Papatheodorou G, Karameris A, Mavrogeorgis A, Kalogeropoulos N, et al.: Helicobacter pylori infection and gastric juice vitamin C levels. Impact of eradication. Dig Dis Sci 40, 615–621, 1995. 23. Zhang ZW, Patchett SE, Perrett D, Katelaris PH, Domizio P, et al.: The relation between gastric vitamin C concentrations, mucosal histology, and CagA seropositivity in the human stomach. Gut 43, 322–326, 1998. 24. Phull PS, Price AB, Thorniley MS, Green CJ, and Jacyna MR: Vitamin E concentrations in the human stomach and duodenum—correlation with Helicobacter pylori infection. Gut 39, 31–35, 1996. 25. Zhang ZW, Patchett SE, Perrett D, Domizio P, and Farthing MJ: Gastric alpha-tocopherol and beta-carotene concentrations in association with Helicobacter pylori infection. Eur J Gastroenterol Hepatol 12, 497–503, 2000. 26. Figueiredo C, van Doorn LJ, Nogueira C, Soares JM, Pinho P, et al.: Helicobacter pylori genotypes are associated with clinical outcome in Portuguese patients and show a high prevalence of infections with multiple strains. Scand J Gastroenterol 36, 128–135, 2001. 27. Gonzalez CA, Pera G, Agudo A, Bueno-de-Mesquita HB, Ceroti M, 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 118, 2559–2566, 2006. 28. Lunet N and Barros H: Does vitamin C dietary intake modify the association between Helicobacter pylori infection and gastric cancer? Eur J Epidemiol 19, 1061–1062, 2004.

Vol. 55, No. 1

77