Plant Foods, Antioxidants, and Prostate Cancer Risk

share page

Share this with your friends
Tagged:  

Abstract: Epidemiological data on most cancer sites suggest that consumption of plant foods, which contain high levels of antioxidants, might slow or prevent the appearance of cancer. We used data from three case-control studies to test this hypothesis. The total study population consisted of 617 incident cases of prostate cancer and 636 population controls from Ontario, Quebec, and British Columbia. Dietary information was collected by an in-person interview with a detailed quantitative dietary history. Unconditional logistic regression analyses were performed to estimate odds ratios (ORs) and 95% confidence intervals (Cls). A decreasing, statistically significant association was found with increasing intakes of green vegetables (OR = 0.54, 95% Cl = 0.400.71 for 4th quartile), tomatoes (OR = 0.64, 95% Cl = 0.45-0.91), beans/lentils/nuts (OR = 0.69, 95% Cl = 0.530.91), and cruciferous vegetables (OR = 0.69, 95% Cl = 0.52-0.91 for 3rd quartile). Higher intakes of fruit were associated with higher ORs (OR = 1.51, 95% Cl = 1.14-2. 01 for 4th quartile), an effect that was seen for total fruit and citrus fruit, as well as for all other noncitrus fruits. Among the grains, refined-grain bread intake was associated with a decrease in risk (OR = 0.65for 4th quartile), whereas wholegrain breakfast cereals were associated with a higher risk for prostate cancer. Of all the antioxidant nutrients studied, the ORs were higher with higher intakes of cryptoxanthin (OR = 1.44, 95% Cl = 1.09-1.89 for 4th quartile). Exposure to certain dietary components of plant origin, which are potentially modifiable, indicates the theoretical scope for reducing the risk from prostate cancer. Future experimental studies or trials are warranted for further understanding.

Introduction
Epidemiological data on most cancer sites suggest that consumption of plant foods might slow or prevent the appearance of cancer. There are many biologically plausible reasons for this. These include the presence in plant foods of such potentially active anticarcinogenic substances as carotenoids, vitamin C, vitamin E, selenium, dietary fiber (and fractions), dithiolthiones, isothiocyanates, indoles, phenols, protease inhibitors, allium compounds, sterols, phytoestrogens, and limonene ( 1). However, the available data regarding their effects on prostate cancer are equivocal. Findings of case-control and cohort studies, as summarized in recent reviews ( 1-4), suggest that the pattern of association between prostate cancer and fruits and vegetables is not clear. Most studies found no association or even increased risk with some fruit categories and a marginal decrease in risk with vegetables. The bioactive constituents in fruits and vegetables may interact in complex ways to increase or protect against carcinogenesis in the prostate. Among the other plant foods, whole-grain intake was reported to be marginally protective in the only study ( 5) reported in a recent review ( 6). Similarly, weakly protective effects of legumes and nuts have also been reported ( 3).

The present analysis was undertaken to study whether vegetables, fruits, or grains decrease the risk of prostate cancer in data from a multicenter collaborative study in Canada and also to examine the effect of various known and emerging antioxidants simultaneously. Because large intervention trials have shown no effect of beta-carotene on prostate cancer ( 7, 8) and laboratory studies have shown that vitamin A and carotenoids can increase and decrease carcinogenesis under various experimental conditions ( 2), it was of interest to examine the relative contribution of various types of carotenoids, retinol, and other known antioxidants. It was thought possible that some of the discrepancies in previous studies might have been related to a failure to consider interactions between dietary energy or fat and potential protective constituents in plant foods. The present study permitted us to consider these interactions.

Materials and Methods
This case-control study was conducted in three geographical areas of Canada: Metropolitan Toronto (Ontario), Montreal (Quebec), and Vancouver (British Columbia). In each center, the study was conducted according to a common protocol, which included the use of questionnaires with a core set of identical questions. The sampling frames for the controls differed in the three centers, but all were essentially population based. The total study population consisted of 617 cases (187 in Ontario, 229 in Quebec, and 201 in British Columbia) and 636 controls (207 in Ontario, 230 in Quebec, and 199 in British Columbia).

Cases
Men were eligible for inclusion in the study as cases if they had a recent, histologically confirmed diagnosis of adenocarcinoma of the prostate.

Ontario: All cases referred to the Ontario Cancer Registry between April 1990 and April 1992 and resident in metropolitan Toronto and surrounding areas (York, Peel, and Halton) were included in the study. Pathology reports were received at the registry within six months of diagnosis for >90% of cases occurring in Ontario. A random sample of 750 was selected from a total of 4,227 cases referred to the registry during this period. Of the 367 eligible cases from these for whom doctor's permission was obtained, 86 cases were not interviewed for various reasons, 207 were interviewed (74%), and 74 refused. For this analysis, 17 more cases were deleted, inasmuch as their diagnosis date was before the study period, and another 3 reported extremely high caloric intake of > 10,000 kcal/day.

Quebec: During 1989-1993, 881 newly diagnosed patients were identified through admission offices of five major francophone teaching hospitals in Montreal. Of the 362 eligible cases from these, 232 (81%) were interviewed for the study, 3 interviewed cases were later excluded because of incorrect diagnosis and lack of sufficient data, 56 declined to be interviewed, 31 were not contacted because of incorrect address, and 40 were not interviewed because of termination of the study.

British Columbia: A total of 6,183 cases were referred to the British Columbia Cancer Registry between 1 January 1989 and 31 December 1991, and a random sample of those eligible was recruited each month by means of random number tables. About one-half of those selected were <70 years of age, and one-half were >70 years of age. Of the 229 cases approached, 201 (88%) were interviewed, 5 had died, 3 were too ill to participate, 4 had relocated, and 16 refused to participate.

Controls
Controls were frequency matched on five-year age group distribution of the cases and residence. Attempts were made to recruit one control per case.

Ontario: Potential controls were identified by randomly selecting men from assessment lists maintained by the Ontario Ministry of Finance. These lists contain the age, gender, and address of all residents and are updated twice a year. For the purposes of the present study, the Ministry of Finance provided a list consisting of the names of a sample of men residing in the geographical areas of cases, stratified by five-year age group. In Ontario, 733 subjects were approached, of whom 109 did not speak English, 58 had moved out of the study area, and 41 had died. Of the remaining 525 subjects, 207 (52%) were interviewed, 194 refused to participate, 27 were too ill or senile, and 97 could not be contacted.

Quebec: Population-based controls were selected by a modified random-digit-dialing method. A page from the telephone directory was selected randomly from the sampling frame, and the names and addresses of 10 individuals were chosen with the same first three telephone digits as the patient. Of the 711 controls selected, 462 (65%) were eligible and the remaining 249 (35%) were noneligible (e.g., did not answer, refused to participate before the study was explained, wrong age, non-French-Canadians). Therefore, a total of 231 (50%) of the eligible controls were interviewed (1 was later deleted because of insufficient information).

British Columbia: Potential controls were identified by random selection of men from rosters maintained by the Medical Services Plan of British Columbia, which covers almost the entire population of the province. A total of 262 subjects were approached, of whom 9 had moved and 14 had died. Of the remaining 239 subjects, 199 (87%) were interviewed, 36 refused, 1 provided incomplete information, and 3 were too ill to participate.

Procedures
Potential cases and controls in all three study areas were initially contacted by a letter and subsequently by telephone. Those with unlisted telephone numbers were not included, since participation rates are generally very low among such subjects. Several telephone calls were made on different days and times, until a contact was made or a failure to contact was established by the interviewer. All interviews were conducted at the subject's home by trained interviewers, except in Quebec, where interviews for hospitalized patients and controls were also conducted in hospitals. The interviewers used a structured questionnaire on dietary and nondietary factors. The nondietary component of the questionnaire elicited information on demographic characteristics, height, weight at various times during life, physical activity patterns, personal and medical history, smoking, frequency of use of the medical system, and history of rectal examination.

A validated, quantitative diet history was used to obtain estimates of daily intake of alcohol and a great majority of foods found in the Canadian diet ( 9). The diet history contained questions concerning seasonality, usual frequency, and usual amount of consumption of various foods. Quantification was achieved by the use of physical volumetric food models. In reporting intake, cases were asked to address their usual intake to the one-year prediagnosis period, whereas controls were asked to address it to the one year before the interview date.

A total of 107,000 food entries recorded by all subjects encompassed 1,129 unique food items. These were first grouped into 142 food groups according to the food-grouping criteria used by the National Food Consumption Survey classification system ( 10) and a few groups based on the classification described by Smith and co-workers ( 11). These food groups were further collapsed together to obtain gram intakes per day of approximately 29 food groups for these analyses. These food groups were grains (including breads, pasta, breakfast cereals, muffins, pies, cakes, crackers, pancakes), fruit (including citrus fruit, berries, and all other fruits), vegetables (including green, leafy, root vegetables, tomatoes, tomato or vegetable soup, other vegetables), all plant foods (including all grains, all fruits, all vegetables), milk (all kinds), yogurt (all kinds), cream, cheeses, red meat (all beef, pork, veal, lamb, game, meat stews, meat soups), chicken, fish, casseroles (TV dinners, mixed-food-group items), eggs, beans (beans, lentils, nuts, seeds), refined-grain breads (not including breakfast cereals), whole-grain breads (not including breakfast cereals), refined-grain breakfast cereals, whole-grain breakfast cereals, pasta, table or cooking fat (butter, lard, beef tallow, drippings, not margarine), cooking oils (including margarines, salad dressings), table sugar, and jams. Subsets of some of the above food groups were also constructed as follows: citrus fruits, other fruits (including berries), leafy vegetables, green vegetables, tomato (including tomato soups, sauces), other vegetables (not including leafy, green, and tomato), and cruciferous vegetables.

Daily intakes of most nutrients were calculated using Handbook No. 8 of the US Department of Agriculture (USDA) ( 12), which we expanded for Canadian foods. Specific estimates of carotenoids (alpha-carotene, beta-carotene, lycopene, cryptoxanthin, lutein), folic acid, and an antioxidant conjugated linoleic acid were obtained on the basis of USDA-National Cancer Institute carotenoid food composition database on >2,400 fruits and vegetables and multi-ingredient foods containing fruits and vegetables ( 13, 14). The estimated total beta-carotene reported by us previously (15, 16) is not directly related to estimates reported here. It was based on certain assumed fractions applied to the amount of total vitamin A reported in Handbook No. 8 in the absence of published values when our previous nutrient database was compiled and had not been revised at the time those analyses were undertaken. Folic acid values were extracted from the Canadian Nutrient File (1997 version) ( 17), a database essentially derived from Handbook No. 8, whereas conjugated linoleic acid values were derived from various sources (1821). In all, 19 nutrients were examined for this analysis: total calories, total fat, saturated fat, oleic acid, linoleic acid, dietary fiber, total carotenoids, alpha-carotene, beta-carotene, lycopene, cryptoxanthin, lutein, folic acid, conjugated linoleic acid, vitamin E, vitamin C, retinol, selenium, and total sugar.

Statistical Analysis
Foods and nutrients of interest were classified into quartiles or other suitable categories on the basis of their distribution among controls and were treated as separate models. A value of 1.0 or 0.1 was assigned to 0 values for foods or nutrients, respectively, before their conversion to logarithmic values. Odds ratios (ORs) and 95% confidence intervals (CIs) for the association between foods and risk of prostate cancer were obtained from unconditional logistic regression models ( 22). All models were first adjusted for age and logarithmically converted total energy intake, since all foods were positively associated with total energy. We used this standard method of energy adjustment, since our interest was mainly in analyzing food groups rather than nutrients ( 23). In the second set of analyses, using the forward stepwise regression procedure in Statistical Analysis Software, all models included the food/nutrient of interest (forced into the model), logarithmically converted energy intake (forced into the models), age (in years), ever/never smoked filter cigarettes, marital status (married at the time of diagnosis/interview vs. not married), study area (Toronto, Vancouver, Montreal), vasectomy (yes/no), body mass index, education (total years of education), dietary fiber (g), and use of multivitamin supplements in the one year before diagnosis or interview (yes/no). The analyses were conducted for the three centers combined, since any differences between the centers were largely due to differences in energy intake, and the two previous publications showed similar results for Toronto and Montreal areas (15, 16, 24). The ORs from the second set of analyses did not differ appreciably from the ORs obtained by the simpler models.

In the third set of analyses, the stepwise regression procedure, all models were further adjusted for logarithmically converted intakes (to avoid overweighing for extreme values) of total grains, total fruit, total vegetables, total plant foods, total carotenoids, folic acid, conjugated linoleic acid, vitamin E, vitamin C, retinol, total fat, and linoleic acid (avoiding any duplication with the main food of interest). Additional adjustment by saturated fat and oleic acid had no impact on the results. The vitamin estimates were based on contribution from food sources only, not from vitamin supplements. Some of these foods and nutrients were included to adjust for the effect of antioxidants. This approach for analysis changed the point estimates and their precision slightly, and therefore results of this all-inclusive model are presented as the multivariate OR. Category 1, with an OR of 1.00 is the reference category throughout the analysis. Tests for linear trend over categories were performed.

Results
The analyses included 1,253 subjects, 617 cases and 636 controls, from all three centers. The means for various characteristics for cases were 69.8 yr of age, 79.3 kg body wt, and 2,951 kcal daily energy intake. The values for controls were 69.9 yr of age, 78.5 kg body wt, and 2,731 kcal daily energy intake. Statistically significant positive associations were observed in the present data set with a family history of prostate cancer in first-degree relatives and ever-smoking of filter cigarettes, whereas a history of rectal examination in the last five years and vasectomy ware negatively associated with prostate cancer risk. High intakes of total energy and polyunsaturated fats were positively associated. These have been described in detail in previous publications (15, 16) and were considered as confounders in the present analyses.

The relation of intake of various plant foods with these putative factors was evaluated by examining the percent frequency distribution of control subjects among quartile categories of food groups (data not shown). There were no significant differences in the distribution of the following characteristics between quartiles of grain, fruit, and vegetable food groups: age (<70 vs. >70), education (grade 12 vs. >12), history of benign prostate hypertrophy, and family history of cancer. Higher grain intake was more likely among single men than among married men and men who did not report having a vasectomy. People who ever smoked filter cigarettes were more likely to consume less fruits and less vegetables than never-smokers. Grain intake or vegetable intake was positively correlated (p < 0.01).with total energy intake, saturated fat, linoleic acid, retinol, dietary fiber, fruit, and vegetable intake, reflecting a general effect of the food quantity consumed. This effect of total energy also accounted for any differences observed between study areas in their grain, fruit, or vegetable intake. These associations were considered when testing for confounders of the main effect.

Categorical Plant Food Intake
The overall risk of prostate cancer with consumption of various plant foods is shown in Table 1. Category formation was based on the distribution of control subjects in quartiles or other categories of intake. When risks were examined by level of daily vegetable consumption in the multivariate analysis using stepwise regression, a decreasing, statistically significant association was found with increasing intakes of green vegetables (OR = 0.54, 95% Cl = 0.40-0.71 in the 4th quartile), tomatoes (OR = 0.64, 95% Cl = 0.45-0.91), and cruciferous vegetables (OR = 0.69, 95 % Cl = 0.52-0.91) in the third quartile and with the beans/lentils group (OR = 0.69, 95% Cl = 0.53-0.91). Higher intakes of fruit were associated with ORs > 1.00 for risk of prostate cancer, an effect that was seen for total fruit, citrus fruit, and all other noncitrus fruits. Among the grains, refined-grain bread intake was associated with a decrease in risk, whereas wholegrain breakfast cereals were associated with a higher risk for prostate cancer. A separate analysis with quartiles of dietary fiber did not show any association with prostate cancer in this data set (OR = 0.96, 95% Cl = 0.67-1.40 for the 4th quartile compared with OR = 1.00 for the 1st quartile, data not shown). No particular pattern was observed with total vegetables, total grains, total plant foods, leafy vegetables, vegetables other than leafy, green, or tomatoes, wholegrain breads, refined-grain breakfast cereals, and pasta. When a subgroup of all whole grains (whole-grain bread + whole-grain cereals) was formed, higher intakes of all whole grains showed a nonsignificant increase in risk (OR = 1.13, Cl = 0.82-1.56 in the 4th quartile, data not shown). Higher intakes of a subgroup of all refined grains (refined-grain breads, refined-grain breakfast cereals, pasta, flour) showed a decreased risk (OR = 0.65, Cl = 0.45-0.94 in the 4th quartile, data not shown). In all the models shown in Table 1, smoking and higher total energy intakes or fruits consistently remained a significant risk factor, whereas having undergone a vasectomy or higher total plant food intake lowered the risk.

Categorical Intake of Selected Nutrients
Table 2 shows the effect of various carotenoids and selected antioxidants on the risk of developing prostate cancer. The ORs increased with higher intakes of cryptoxanthin (OR = 1.44, 95% Cl = 1.09-1.89 for the 4th quartile). Intakes of other carotenoids, including alpha-carotene, beta-carotene, lutein, and lycopene, and folic acid or conjugated linoleic acid did not appreciably change in the ORs. Again, in all the models shown in Table 3, smoking and higher total energy intakes or fruits consistently remained a significant risk factor, whereas having undergone a vasectomy or higher total plant food intake lowered the risk. Retinol or vitamin E intakes did not show any significant association with risk, but a slightly protective effect of vitamin C was observed with the multivariate analysis (OR = 0.69, 95% Cl = 0.500.97 for the 3rd quartile). Intakes of selenium were associated with a decrease in risk in the second quartile (OR = 0.68, 95% Cl = 0.52-0.92). Because the intake of carotenoids from supplements was not very prevalent at the time of this study, no attempt was made to study the effect of supplemental sources for this study.

Categorical Intake of Foods Other Than Plant Foods
In addition, the effect of intakes of various other foods on risk of prostate cancer was examined (Table 3). No appreciable change in ORs was observed with higher intakes of yogurt, cream, cheese, red meats, chicken, eggs, cooking oils, and table sugars. Higher ORs were observed with the highest quartile of intakes of milk (OR = 1.47, 95% Cl = 1.11-1.94) and ever-use of table fats (OR = 1.39, 95% Cl = 1.14-1.69). The direction of effect of milk was consistent with intakes of yogurt or table fats but not with creams and cheese intakes, making interpretations difficult. A lower OR was observed with the highest quartile of intake of fish (OR = 0.66, 95% Cl = 0.50-0.89) and mixed-dish casseroles (OR = 0.73, 95% Cl = 0.56-0.95).

Simultaneous and Joint Intakes of Nutrients and Plant Foods: Continuous Variable Analysis
To further examine whether the effects of dietary intakes of various plant foods and antioxidant nutrients in relation to prostate cancer were independent of each other, we also conducted analyses with all nutrients and foods as continuous variables in the model simultaneously (data not shown). In a forward stepwise regression model, all 29 food groups, 18 nutrients, and 8 nondietary variables (to control for confounders) were included. All dietary variables were converted to logarithmic values, and thus the results obtained were for 1 log unit of food in grams or 1 log unit of nutrient, e.g., 1 log kcal for total energy. The ORs obtained from these results were therefore difficult to translate into actual consumption units but were essentially obtained as information for interpretation. Total energy intake was forced into the model. The variables that were retained in the final model (p for estimates

We also examined the trend across categories of intake of various foods and nutrients. A somewhat linear trend was observed with intakes of green vegetables, cruciferous vegetables, fruits other than citrus, beans and lentils, milk, fish, table fat, total energy, and cryptoxanthin (p for trend = 0.06, 0.05, 0.04, 0.03, 0.04, 0.05, 0.02, 0.01, and 0.02, respectively). With most variables including total fruit, citrus fruit, lutein, vitamin C, vitamin E, folic acid, and selenium, we did not observe a linear effect but either a J- or a U-shaped effect (or inverse).

Discussion
Results of the present study suggest the presence of an inverse association between green vegetable intake and subsequent prostate cancer incidence. The effect of fruits was in the opposite direction. In addition, there was a marginal decrease in risk with higher intakes of tomatoes, but, unlike other reports in the literature, a parallel effect was not seen with intakes of lycopene as a major carotenoid from tomatoes ( 25). Of the various other antioxidant vitamins studied in this analysis, higher intakes of cryptoxanthin appear to marginally increase the risk of prostate cancer. In this study and in previous analyses of data from the present study population (15, 16), no change in risk of prostate cancer was found with intakes of dietary fiber, vitamin E, and other known antioxidant substances in fruits, vegetables, and grains.

The assessment of various nutrients considered in this analysis is subject to methodological problems because of their wide distribution in foods and limitation of values in the available food composition database. However, our estimates of carotenoids compared well with intakes reported for US populations ( 25, 26). The third-quintile ranges (i.e., 40th-60th percentile) of carotenoid intakes for men reported by Giovannucci and co-workers ( 25), using a food frequency questionnaire, were (in mug/day) 523-722 for alpha-carotene, 3,902-5,189 for beta-carotene, 41-67 for beta-cryptoxanthin, 2,666-3,620 for lutein, and 3,367-4,591 for lycopene. Our median values were (in mug/day) 1,515 for alpha-carotene, 5,188 for beta-carotene, 49.5 for -cryptoxanthin, 1,654 for lutein, and 5,252 for lycopene. The validity and reliability of the nutrient estimates obtained by the dietary questionnaire used by us have been comparable to those obtained from other similar instruments ( 9). Because we used a very detailed personal interview for diet assessment, we obtained reports of >1,129 unique foods that contributed to these nutrients. It was therefore necessary to collapse them into workable food groups to obtain a suitable range of intake for food group analyses. This may, at times, result in misclassification of certain types of foods. However, the comparability of risk estimates obtained in this study to reports from other studies in the literature does not suggest any appreciable problem from this source.

Our results on vegetables and fruits are consistent with the various case-control and cohort studies reported and reviewed ( 1-3). In contrast to most other cancer sites, where vegetables show consistently protective associations, the available data on the risk of prostate cancer are much less clear. Several studies have shown a protective association with higher consumption of green and yellow vegetables ( 27-29), but some did not find any relation between overall intake of vegetables and risk of prostate cancer ( 5, 25, 30-34). Some of the discrepancies may be due to differences in the food groupings used in various studies. Our study shows a protective role only for vegetables defined as green (including leafy and nonleafy) and a borderline effect of cruciferous vegetables. Protective effects have been reported with consumption of carrots ( 35), tomato-based foods ( 25, 30), cabbage and spinach ( 36), and beans, peas, and lentils ( 30, 34, 35). Contrary to the general belief that fruits also have a chemopreventive role because of their high vitamin content, we did not find a protective role for fruit intake; instead a positive association was observed. Higher intakes of fruits have been reported to increase risk for prostate cancer in several other studies (OR = 1.3-2.5 for highest vs. lowest category) ( 5, 28, 34, 37-39) and not change in risk in two other studies ( 29, 30). This raises the possibility that some of the dietary risk factors for prostate (hormone-mediated) cancer are different from those for many other cancers.

Although the active ingredients in fruits and vegetables (whether increasing or decreasing risk) can only be speculated on at present, carotenoids have been the most extensively studied components, mainly because of their antioxidant capabilities. Besides beta-carotene, other main carotenoids include alpha-carotene, lutein, lycopene, and cryptoxanthin. Epidemiological investigations on the relative contribution of the various sources of carotenoids to cancers have been limited, primarily because of the unavailability of food tables on these nutrients until very recently. Findings on beta-carotene and prostate cancer are conflicting. Some case-control studies have reported a positive association with beta-carotene intake, particularly among men >70 years of age, whereas others have found no association or even protection ( 3). Of interest for this study is a case-control study in Hawaii by Le Marchand and colleagues ( 39). They noticed that among all the fruits and vegetables they examined, intakes of papaya resulted in a higher risk of prostate cancer. Their study failed to support an association with main carotenoids, such as beta-carotene, lutein, and lycopene, or with other cancer inhibitors particularly abundant in cruciferous vegetables (e.g., phenols, indoles, flavones), since consumption of tomatoes, dark green vegetables, and cruciferous vegetables was similar among cases and controls. Papaya contains significant amounts of the carotenoid beta-cryptoxanthin, and it was speculated that this may be one of the main active ingredients for prostate cancer. Our results from this study tend to support this observation when examined as fruit and as cryptoxanthin. Findings of a recent cohort study that has reported a marginal but significant positive association of prostate cancer risk with consumption of oranges (relative risk = 1.07) ( 34) also suggest some possible role of cryptoxanthin, oranges and papaya being the most concentrated sources of this carotenoid. If citrus fruits and fruit juices could be considered an indicator of oranges/tangerines, then our data tend to support a role for this carotenoid.

Among the other antioxidants, Giovannucci and co-workers ( 25) observed a reduced risk of prostate cancer with higher intakes of lycopene (OR = 0.79 in the 4th quartile) and tomato-based foods as the major dietary source of lycopene in a cohort of 47,894 men. They did not find any association with other carotenoids. The possible antioxidant mechanisms for the action of lycopene and lutein have been recently proposed ( 40). Although a similar protective effect of tomatoes is evident in our study, intakes of lutein, but not lycopene, paralleled the findings of tomatoes, suggesting that factors in tomatoes other than lycopene may be of significance.

Our study shows an overall decrease in risk with wholegrain breads but not with whole-grain cereals. One can only speculate whether there are any confounding factors that affect the intake of these two sources of whole grains. A weak protective effect of whole-meal bread has been reported in one study from Italy ( 5). The effect was observed only for <70-year-old men, and not for >70-year-old men. We also found a similar nonsignificant decrease in risk ratios with "whole-grain breads" in the younger age group only. When we examined "all whole grains combined" or "wholegrain cereal" by age group, there was no difference in effect between <70- and >70-year-old men (data not shown). We also found a protective effect with refined-grain breads and beans/lentils/nuts/seeds. Dietary fiber as a surrogate of whole grains or beans did not show any significant protective effect. The decrease in risk observed with higher intakes of refined foods may partly be a reflection of overall higher food intakes and thus higher amounts of vegetables as well, but it could also be a chance observation. The inverse effect of beans/lentils/nuts/seeds could be due to the phytoestrogens present in them ( 1, 41). Similar to our study, others have reported a relationship of decreased risk with high intakes of fish ( 29, 30, 42) and increased risk with milk ( 5, 29, 37, 43, 44) and animal fat ( 29, 43). Our study only marginally supports the findings from a recently reported study that found a reduced risk of prostate cancer with higher levels of selenium ( 45). Selenium is a mineral that acts as an antioxidant, maintains the elasticity of body tissues, and improves the oxygen supply to the heart. However, literature is inconsistent with regard to its role in prostate cancer. Although no association between baseline levels of selenium and prostate cancer was found in a beta-carotene alpha-tocopherol trial-based cohort of men ( 46), the incidence was reported to be reduced in a double-blind trial with selenium supplementation ( 47).

We recognize the problems associated with multiple comparisons and that some findings will occur as significant just by chance. However, we tried to focus on our a priori hypothesis that intakes of vegetables, other plant foods, and antioxidants are associated with a decreased risk of prostate cancer. This was evaluated in light of various nutrients that may be the principal components of these foods such as vitamins and phytoestrogens. Our results support this assumption that vegetable intake, particularly green vegetables and cruciferous vegetables, tomatoes, and beans, is associated with a reduced risk, perhaps mediated by an effect of lutein, lycopene, and phytoestrogens. The fruits do not contribute to this protection; their effect is that of enhancing risk, independent of antioxidant vitamin intake and perhaps mediated by beta-cryptoxanthin. Exposure to certain dietary components of plant origin, which are potentially modifiable, indicates the theoretical scope for reducing the risk from prostate cancer. Future experimental studies or trials are warranted for further understanding.

Acknowledgments and Notes
The authors thank the various study coordinators, Dr. Chong Teh, David Butch, and Chantal Perret, who helped in the data collection phase of the study and acknowledge the important contribution made by our interviewers and study participants. The financial support for this project was partly provided by the National Cancer Institute of Canada. Address reprint requests to Meera Jain, Cancer Epidemiology Unit, Dept. of Public Health Sciences, 308-12 Queen's Park Crescent West, University of Toronto, Toronto, ON, Canada M5S IA8.

Submitted 11 January 1999; accepted in final form 1 April 1999.

Table 1. ORs and 95% Cls for Prostate Cancer by Categories of Food Consumption in Canada, 1989-93[a,b]

Legend for Chart:

A - Food Item
B - Category 1
C - Category 2
D - Category 3
E - Category 4

A B
C
D
E

All plant foods (vegetables + fruit + grains, no beans)

g/day 0-825
826-1,098
1,099-1,403
> 1,404

Cases/controls 144/158
141/159
147/160
185/159

OR 1
0.89 (0.64-1.24)
0.87 (0.62-1.22)
0.99 (0.69-1.42)

Multivariate OR 1
0.80 (0.59-1.08)
0.76 (0.54-1.06)
0.78 (0.52-1.18)

Total fruit

g/day 183.3
183.3-342.7
342.7-514.4
>514.4

Cases/controls 120/159
171/159
146/159
180/159

OR 1
1.42 (1.03-1.96)[c]
1.18 (0.85-1.64)[d]
1.37 (0.99-1.91)[e]

Multivariate OR 1
1.52 (1.15-2.00)[f]
1.28 (0.97-1.70)[g]
1.51 (1.1 4-2.01)[f]

Total vegetables

g/day <286.5
286.5-428.7
428.8-594.6
>594.6

Cases/controls 138/159
183/159
138/159
158/159

OR 1
1.21 (0.88-1.66)
0.85 (0.61-1.20)
0.90 (0.63-1.27)

Multivariate OR 1
1.26 (0.95-1.66)
0.89 (0.66-1.21)
0.95 (0.68-1.33)

Total grains

g/day <194.2
194.2-275.8
275.9-396.1
>396.1

Cases/controls 138/156
149/162
170/159
160/159

OR 1
0.92 (0.67-1.29)
0.99 (0.70-1.39)
0.84 (0.58-1.22)

Multivariate OR 1
0.94 (0.72-1.25)
1.00 (0.75-1.33)
0.83 (0.61-1.14)

Fruit, citrus

g/day <34.3
34.3-145.3
145.4-265.0
>265

Cases/controls 117/159
193/159
142/159
165/159

OR 1
1.66 (1.21-2.29)[h]
1.22 (0.88-1.70)[i]
1.38 (1.00-1.91)[j]

Multivariate OR 1
1.78 (1.35-2.33)[k]
1.35 (1.01-1.79)[l]
1.48 (1.12-1.96)[m]

Fruit, other than citrus

g/day <101.3
101.3-195.1
195.2-315.8
>315.8

Cases/controls 118/159
154/159
156/159
189/159

OR 1
1.28 (0.92-1.77)
1.28 (0.92-1.78)
1.44 (1.04-20.1)[n]

Multivariate OR 1
1.28 (0.94-1.73)
1.23 (0.89-1.71)
1.54 (1.07-2.22)[j]

Vegetables, leafy 0
>0
--
--

Cases/controls 292/298
325/338
--
--

OR 1
0.95 (0.76-1.18)
--
--

Multivariate OR 1
0.94 (0.78-1.14)
--
--

Vegetables, green

g/day <24.0
24.0-42.6
42.7-72.7
>72.7

Cases/controls 191/158
128/159
165/159
133/160

OR 1
0.62 (0.45-0.85)[o]
0.76 (0.56-1.04)[p]
0.58 (0.42-0.80)[q]

Multivariate OR 1
0.59 (0.55-0.77)[q]
0.72 (0.55-0.95)[j]
0.54 (0.40-0.71)[r]

Tomatoes

g/day <9.3
9.3-45.5
45.6-109.6
>109.6

Cases/controls 161/158
139/159
161/160
156/159

OR 1
0.84 (0.61-1.16)
0.93 (0.68-1.28)
0.85 (0.62-1.18)

Multivariate OR 1
0.77 (0.58-1.01)
0.78 (0.58-1.04)
0.64 (0.45-0.91)[n]

Vegetables, other than leafy, green and tomatoes

g/day <246.9
246.9-382.7
382.8-541.7
>541.7

Cases/controls 139/159
172/157
149/161
157/159

OR 1
1.16 (0.84-1.59)
0.94 (0.68-1.31)
0.93 (0.66-1.30)

Multivariate OR 1
1.20 (0.91-1.59)
1.03 (0.77-1.39)
1.05 (0.75-1.47)

Vegetables, cruciferous

g/day <8.7
8.7-24.0
24.1-44.6
>44.6

Cases/controls 176/159
166/159
122/159
153/159

OR 1
0.93 (0.69-1.27)
0.67 (0.48-0.92)[f]
0.80 (0.59-1.10)

Multivariate OR 1
0.95 (0.73-1.24)
0.69 (0.52-0.91)[c]
0.85 (0.64-1.13)

Beans, lentils, nuts, seeds

g/day <3.6
3.6-12.9
13.0-30.7
>30.7

Cases/controls 179/159
144/156
143/161
151/160

OR 1
0.79 (0.57-1.08)
0.73 (0.54-1.01)[j]
0.70 (0.51-0.97)[c]

Multivariate OR 1
0.79 (0.61-1.04)
0.72 (0.55-0.95)[j]
0.69 (0.53-0.91)[c]

Breads, refined grain

g/day <56.1
56.1-101.2
101.3-169.3
>169.3

Cases/controls 156/159
145/159
169/159
147/159

OR 1
0.86 (0.62-1.18)
0.93 (0.68-1.29)
0.71 (0.50-1.01)[e]

OR 1
0.84 (0.64-1.10)
0.89 (0.67-1.16)
0.65 (0.48-0.88)[m]

Breads, whole grain

g/day 0
>0-69
>69
--

Cases/controls 205/199
253/280
159/157
--

OR 1
0.91 (0.70-1.19)
0.96 (0.71-1.29)
--

Multivariate OR 1
0.92 (0.73-1.15)
0.99 (0.76-1.28)
--

Pasta

g/day 0
>0-20.6
>20.6
--

Cases/controls 266/304
176/180
175/152
--

OR 1
1.10 (0.84-1.44)
1.21 (0.91-1.60)
--

Multivariate OR 1
1.06 (0.85-1.33)
1.20 (0.94-1.52)
--

Breakfast cereals, refined grain

g/day 0
>0
--
--

Cases/controls 354/376
263/260
--
--

OR 1
1.06 (0.84-1.33)
--
--

Multivariate OR 1
1.00 (0.83-1.21)
--
--

Breakfast cereals, whole grain

g/day <29.1
29.1-74.2
74.3-152.0
>152.0

Cases/controls 131/159
164/158
153/160
169/159

OR 1
1.23 (0.89-1.69)
1.11 (0.80-1.53)
1.19 (0.86-1.64)

Multivariate OR 1
1.27 (0.96-1.66)
1.19 (0.90-1.57)
1.41 (1.05-1.88)[j]
a: Odds ratios (ORs) were adjusted for age and log total energy intake; values in parentheses are 95% confidence intervals (Cls). Multivariate ORs were obtained by forward stepwise regression model adjusted for log total energy, vasectomy (yes/no), age, ever-smoked (yes/no), marital status (married at interview/others), study area, body mass index, education, ever-used multivitamin supplements in 1 yr before diagnosis/interview, area of study, and log-converted amounts for grains, fruit, vegetables, total plants, total carotenoids, folic acid, dietary fiber, conjugated linoleic acid, vitamin E, vitamin C, retinol, total fat, and linoleic acid; n = 1,253 (617 cases and 636 controls).

b: Statistical significance (for comparison with Category 1) is as follows: c, p = 0.03; d, p = 0.32; e, p = 0.06; f, p = 0.01; g, p = 0.14; h, p = 0.002; i, p = 0.24; j, p = 0.05; k, p = 0.0005; l, p = 0.08; m, p = 0.02; n, p = 0.04; o, p = 0.003; p, P = 0.09; q, p = 0.001; r, p = 0.0003.

Table 2. ORs and 95% Cls for Prostate Cancer by Categories of Nutrient Intake in Canada, 1989-93[a,b]

Legend for Chart:

A - Food Item
B - Category 1
C - Category 2
D - Category 3
E - Category 4

A B
C
D
E

alpha-Carotene

mug/day <839
839-1,514
1,515-2,187
>2,158

Cases/controls 147/159
172/159
131/158
167/160

OR 1
1.15 (0.84-1.57)
0.83 (0.60-1.15)
1.00 (0.72-1.38)

Multivariate OR 1
1.16 (0.89-1.53)
0.87 (0.65-1.16)
1.06 (0.79-1.43)

13-Carotene

mug/day <3,002
3,002-5,187
5,188-7,829
>7,829

Cases/controls 131/158
174/159
156/160
156/159

OR 1
1.24 (0.90-1.71)
1.06 (0.76-1.47)
1.01 (0.72-1.42)

Multivariate OR 1
1.25 (0.94-1.65)
1.07 (0.80-1.44)
1.06 (0.74-1.42)

Lutein

mug/day <1,019
1,018-1,653
1,654-2,684
>2,684

Cases/controls 166/158
131/159
149/160
171/159

OR 1
0.70 (0.51-0.97)[c]
0.78 (0.56-1.07)
0.85 (0.61-1.17)

Multivariate OR 1
0.71 (0.53-0.94)[d]
0.80 (0.60-1.06)[e]
0.81 (0.65-1.18)[f]

Lycopene

mug/day <2,103
2,103-5,251
5,252-12,681
>12,681

Cases/controls 152/159
146/159
151/159
168/159

OR 1
0.91 (0.66-1.25)
0.92 (0.67-1.27)
0.99 (0.72-1.36)

Multivariate OR 1
0.90 (0.68-1.19)
0.90 (0.68-1.19)
1.01 (0.76-1.35)

Cryptoxanthin

mug/day <17.9
17.9-49.4
49.5-100
>100

Cases/controls 129-158
127/159
176/160
185/159

OR 1
0.98 (0.70-1.36)
1.32 (0.96-1.82)[g]
1.34 (0.98-1.85)[h]

Multivariate OR 1
1.02 (0.77-1.35)
1.45 (1.11-1.90)[i]
1.44 (1.09-1.89)[c]

All carotenoids (alpha, beta, cryptoxanthin, lutein, lycopene)

mug/day 0-8,942
8,943-15,470
15,471-24,734
>24,735

Cases/controls 145/159
168/159
133/159
171/159

OR 1
1.10 (0.81-1.51)
0.82 (0.59-1.14)
1.02 (0.74-1.41)

Multivariate OR 1
1.12 (0.85-1.48)
0.84 (0.62-1.13)
1.06 (0.77-1.45)

Folic acid

mug/day <258.6
258.6-334.9
335-431.6
>431.6

Cases/controls 137/158
141/160
153/159
186/159

OR 1
0.91 (0.65-1.27)
0.93 (0.66-1.32)
1.01 (0.69-1.49)

Multivariate OR 1
0.89 (0.66-1.21)
0.95 (0.68-1.33)
1.04 (0.70-1.56)

Vitamin E

mg/day <17.17
17.17-25.30
25.31-37.25
>37.25

Cases/controls 135/159
129/160
180/158
173/159

OR 1
0.87 (0.62-1.21)
1.16 (0.83-1.62)
0.99 (0.69-1.44)

Multivariate OR 1
0.89 (0.67-1.19)
1.27 (0.95-1.70)
1.12 (0.81-1.55)

Vitamin C

mg/day <121.08
121.08-178.21
178.22-243.70
>243.70

Cases/controls 156/159
137/159
147/159
177/159

OR 1
0.84 (0.61-1.16)
0.86 (0.62-1.18)
0.97 (0.70-1.35)

Multivariate OR 1
0.71 (0.53-0.97)[h]
0.69 (0.50-0.97)[h]
0.75 (0.50-1.22)[j]

Retinol

IU/day <1,100
1,100-1,651
1,652-2,495
>2,495

Cases/controls 142/159
147/158
136/160
192/159

OR 1
0.93 (0.67-1.30)
0.80 (0.56-1.13)
1.03 (0.71-1.49)

Multivariate OR 1
0.95 (0.72-1.26)
0.80 (0.60-1.08)
1.04 (0.76-1.42)

Selenium

mug/day 157/159
88.38-118.70
118.71-158.23
>158.23

Cases/controls <88.38
116/159
167/159
177/159

OR 1
0.68 (0.49-0.95)[i]
0.91 (0.65-1.26)
0.84 (0.58-1.22)

Multivariate OR 1
0.69 (0.52-0.92)[c]
0.98 (0.74-1.31)
0.93 (0.68-1.28)

Conjugated linoleic acid

g/day <81.2
81.2-133.8
133.9-203.7
>203.7

Cases/controls 131/159
155/158
152/159
179/160

OR 1
1.10 (0.79-1.52)
0.98 (0.69-1.38)
1.03 (0.71-1.50)

Multivariate OR 1
1.05 (0.80-1.39)
0.96 (0.71-1.29)
1.00 (0.72-1.39)
a: ORs were adjusted for age and log total energy intake; values in parentheses are 95% CIs. Multivariate ORs were obtained by forward stepwise regression model adjusted for log total energy, vasectomy (yes/no), age, ever-smoked (yes/no), marital status (married at interview/others), study area, body mass index, education, ever-used multivitamin supplements in 1 yr before diagnosis/interview, area of study, and log-converted amounts for grains, fruit, vegetables, total plants, total carotenoids, folic acid, dietary fiber, conjugated linoleic acid, vitamin E, vitamin C, retinol, total fat, and linoleic acid; n = 1,253 (617 cases and 636 controls).

b: Statistical significance (for comparison with Category 1) is as follows: c, p = 0.03; d, p = 0.04; e, P = 0.18; f, p = 0.46; g, p = 0.09; h, p = 0.07; i, p = 0.02; j, p = 0.24.

Table 3. ORs and 95% CIs for Prostate Cancer by Categories of Food Consumption in Canada, 1989-93[a,b]

Legend for Chart:

A - Food Item
B - Category 1
C - Category 2
D - Category 3
E - Category 4

A B
C
D
E

Red meat

g/day <15.9
15.9-29.4
29.5-55.1
>55.1

Cases/controls 151/158
130/159
152/160
184/159

OR 1
0.82 (0.59-1.13)
0.91 (0.66-1.26)
1.02 (0.73-1.43)

Multivariate OR 1
0.85 (0.65-1.12)
0.93 (0.71-1.22)
1.02 (0.77-1.35)

Chicken

g/day <9.9
9.9-23.1
23.2-44.6
>44.6

Cases/controls 150/159
142/158
162/161
163/158

OR 1
0.94 (0.68-1.30)
1.02 (0.74-1.40)
0.98 (0.71-1.35)

Multivariate OR 1
0.95 (0.72-1.24)
1.07 (0.82-1.40)
1.02 (0.77-1.34)

Fish

g/day <9.5
9.5-24.4
24.5-45.1
>45.1

Cases/controls 185/156
140/159
143/162
149/159

OR 1
0.74 (0.54-1.02)[c]
0.72 (0.53-0.98)[d]
0.73 (0.54-1.00)[e]

Multivariate OR 1
0.73 (0.56-0.95)[e]
0.66 (0.50-0.87)[f]
0.66 (0.50-0.89)[f]

Casseroles

g/day 0
>0-87.4
>87.4
--

Cases/controls 193/172
272/305
152/159
--

OR 1
0.77 (0.59-1.00)[e]
0.76 (0.56-1.04)[d]
--

Multivariate OR 1
0.75 (0.60-0.94)[d]
0.73 (0.56-0.95)[e]
--

Eggs

g/day <7.2
7.2-15.1
>15.1-28.5
>28.5

Cases/controls 128/144
137/172
139/141
213/179

OR 1
0.88 (0.63-1.22)
1.04 (0.74-1.46)
1.20 (0.89-1.68)

Multivariate OR 1
0.88 (0.67-1.16)
1.04 (0.78-1.39)
1.26 (0.96-1.65)

Milk, all kinds

g/day <131
131-257
258-488
>488

Cases/controls 130/159
131/159
157/158
199/160

OR 1
0.98 (0.70-1.36)
1.15 (0.83-1.59)
1.36 (0.98-1.89)[c]

Multivariate OR 1
1.00 (0.76-1.33)
1.18 (0.90-1.56)
1.47 (1.11-1.94)[g]

Yogurt

g/day 0
>0
--
--

Cases/controls 392/414
225/222
--
--

OR 1
1.06 (0.84-1.34)
--
--

Multivariate OR 1
1.05 (0.86-1.28)
--
--

Cream

g/day 0
>0
--
--

Cases/controls 379/369
238/267
--
--

OR 1
0.83 (0.66-1.04)
--
--

Multivariate OR 1
0.82 (0.68-1.00)[h]
--
--

Cheese

g/day <3.2
3.2-12.5
12.6-32.3
>32.3

Cases/controls 154/154
150/164
141/159
172/159

OR 1
0.91 (0.67-1.25)
0.87 (0.62-1.20)
0.98 (0.71-1.35)

Multivariate OR 1
0.91 (0.70-1.19)
0.86 (0.65-1.12)
0.92 (0.70-1.21)

Table and cooking fat (butter, lard)

g/day 0
>0
--
--

Cases/controls 210/274
407/362
--
--

OR 1
1.43
(1.14-1.80)[i]
--

Multivariate OR 1
1.39
(1.14-1.69)[j]
--

Cooking oil (includes margarine, salad dressing)

g/day <4.0
4.0-14.2
14.3-29.4
>29.4

Cases/controls 150/157
175/160
153/160
139/159

OR 1
1.11 (0.81-1.52)
0.98 (0.71-1.34)
0.82 (0.59-1.13)

Multivariate OR 1
1.16 (0.89-1.51)
1.01 (0.77-1.32)
0.85 (0.64-1.13)

Table sugar, syrup, jams, candy

g/day <10.7
10.7-28.6
28.7-57.9
>57.9

Cases/controls 154/158
132/160
144/158
187/160

OR 1
0.81 (0.59-1.12)
0.85 (0.62-1.18)
1.03 (0.74-1.42)

Multivariate OR 1
0.81 (0.61-1.06)
0.88 (0.67-1.16)
1.03 (0.78-1.36)
a: ORs were adjusted for age and log total energy intake; values in parentheses are 95% Cls. Multivariate ORs were obtained by forward stepwise regression model adjusted for log total energy, vasectomy (yes/no), age, ever-smoked (yes/no), marital status (married at interview/others), study area, body mass index, education, ever-used multivitamin supplements in 1 yr before diagnosis/interview, area of study, and log-converted amounts for grains, fruit, vegetables, total plants, total carotenoids, folic acid, dietary fiber, conjugated linoleic acid, vitamin E, vitamin C, retinol, total fat, and linoleic acid; n = 1,253 (617 cases and 636 controls).

b: Statistical significance (for comparison with Category 1) is as follows: c, p = 0.06; d p = 0.04; e, p = 0.05; f, p = 0.01; g, p = 0.02; h, p = 0.10; i, p = 0.002; j, p = 0.006; k, p = 0.08.

References
1. Potter, JD, and Steinmetz, K: Vegetables, fruit and phytoestrogens as preventive agents. In Principles of Chemoprevention, BW Stewart, D McGregor, and P Kleihues (eds). Lyon, France: Int Agency Res Cancer, 1996, pp 61-90. (IARC Sci Publ 139)

2. Kolonel, LN: Nutrition and prostate cancer. Cancer Causes Control 7, 83-94, 1996.

3. American Institute of Cancer Research: Prostate. In Food, Nutrition and the Prevention of Cancer: A Global Perspective. Washington, DC: Am Inst Cancer Res, 1997, pp 310-323.

4. Ellison, LF, Stokes, J, Gibbons, L, Lindsay, J, Levy, et al.: Monograph series on aging-related diseases. X. Prostate cancer. Chron Dis Can 19, 1-18, 1998.

5. Talamini, R, Franceschi, S, La Vecchia, C, Serraino, D, Barra, S, et al.: Diet and prostate cancer: a case-control study in Northern Italy. Nutr Cancer 18, 277-286, 1992.

6. Jacobs, DR, Jr, Marquart, L, Slavin, J, and Kushi, LH: Whole-grain intake and cancer: an expanded review and meta-analysis. Nutr Cancer 30, 85-96, 1998.

7. Omenn, GS, Goodman, GE, Thornquist, MD, Balmes, J, Cullen, MR, et al.: Effects of a combination of beta-carotene and vitamin A on lung cancer and cardiovascular disease. New Engl J Med 334, 1150-1155, 1996.

8. alpha-Tocopherol, beta-Carotene Cancer Prevention Study Group: The effect of vitamin E and beta-carotene on the incidence of lung cancer and other cancers in male smokers. New Engl J Med 330, 1029-1035, 1994.

9. Jain, M, Howe, GR, and Rohan, T: Dietary assessment in epidemiology: comparison of food frequency and a diet history questionnaire with a 7-day food record. Am J Epidemiol 143, 953-960, 1996.

10. National Food Consumption Survey Classification System. Washington, DC: US Dept of Agriculture-National Institutes of Health, 1993.

11. Smith, SA, Campbell, DR, Elmer, PJ, Martini, MC, Slavin, JL, et al.: The University of Minnesota Cancer Prevention Research Unit vegetable and fruit classification scheme (United States). Cancer Causes Control 6, 292-302, 1995.

12. Watt, BK, and Merril, AL: Composition of Foods: Raw, Processed, Prepared. Washington, DC: US Dept of Agriculture, Agricultural Research Service, 1968, expansion 1972. (Handbook No. 8)

13. Mangels, AR, Holden, JM, Beecher, GR, Forman, MR, and Lanza, E: The carotenoid content of fruits and vegetables: an evaluation of analytical data. J Am Diet Assoc 93, 284-296, 1993.

14. Chug-Ahuja, JK, Holden, JM, Forman, MR, Mangels, AR, Beecher, GR, et al.: The development and application of a carotenoid database for fruits, vegetables, and multicomponent foods containing fruits and vegetables. J Am Diet Assoc 93, 318-323, 1993. 15. Rohan, TE, Howe, GR, Burch, D, and Jain, M: Dietary factors and risk of prostate cancer: a case-control study in Ontario, Canada. Cancer Causes Control 6, 145-154, 1995.

16. Ghadirian, P, Lacroix, P, Maisonneuve, C, Perret, G, Drouin, G, et al.: Nutritional factors and prostate cancer: a case-control study of French Canadians in Montreal. Cancer Causes Control 7, 428-436, 1996.

17. Canadian Nutrient File. Ottawa, ON, Canada: Health Canada, 1997.

18. Chin, SF, Liu W, Storkson JM, Ha, YL, and Pariza, MW: Dietary sources of conjugated dienoic isomers of linoleic acid, a newly recognized class of anticarcinogens. J Food Comp Anal 5, 185-197, 1992.

19. Shantha, NC, and Decker, EA: Conjugated linoleic acid concentrations in processed cheese containing hydrogen donors, iron and dairy-based additives. Food Chem 47, 257-261, 1993.

20. Shantha, NC: Conjugated linoleic acid concentrations in dairy products as affected by processing and stomge. J Food Sci 60, 695, 1995.

21. Lin, H, Boylston, TD, Chang, MJ, Luedecke, LO, and Shultz, TD: Survey of the conjugated linoleic acid content of dairy products. J Dairy Sci 78, 2358-2365, 1995.

22. Breslow, NE, and Day, NE: Statistical Methods in Cancer Research. The Analysis of Case-Control Studies. Lyon, France: Int Agency Res Cancer, 1980. (IARC Sci Publ 32)

23. Wiilett, WC, Howe, GR, and Kushi, LH: Adjustment for total energy intake in epidemiologic studies. Am J Clin Nutr 65 Suppl, 1220s-1228s, 1997.

24. Rohan, TE, Hislop, TG, Howe, GR, Gallagher, RP, Teh, CZ, et al.: Cigarette smoking and risk of prostate cancer: a population-based case-control study in Ontario and British Columbia, Canada. Eur J Cancer Prev 6, 382-388, 1997.

25. Giovannucci, E, Ascherio, A, Rimm, EB, Stampfer, MJ, Colditz, GA, et al.: Intake of carotenoids and refinol in relation to risk of prostate cancer. JNCI 87, 1767-1776, 1995.

26. Yong, LC, Forman, MR, Beecher, GR, Graubard, BI, Campbel, WS, et al.: Relationship between dietary intake and plasma concentrations of carotenoids in premenopausal women: application of the USDANCI carotenoid food-composition database. Am J Clin Nutr 60, 223230, 1994.

27. Hirayama, T: A large-scale cohort study on cancer risks by diet--with special reference to the risk-reducing effects of green-yellow vegetable consumption. Proc 16th Int Syrup Princess Takamatsu Cancer Res Fund, 1985, Tololo, Japan, pp 41-53.

28. Ohno, Y, Yoshida, O, Oishi, K, Okada, K, Yamabe, H, et al.: Dietary beta-carotene and cancer of the prostate: a case-control study in Kyoto, Japan. Cancer Res 48, 1331-1336, 1988.

29. Hebert, JR, Hurley, TG, Olendzki, BC, Teas, J, Ma, Y, et al.: Nutritional and socioeconomic factors in relation to prostate cancer mortality: a cross-national study. JNCI 90, 1637-1647, 1998.

30. Mills, PK, Beeson, WL, Phillips, RL, and Fraser, GE: Cohort study of diet, lifestyle, and prostate cancer in Adventist men. Cancer 64, 598-604, 1989.

31. Snowdon, DA, Phillips, RL, and Choi, W: Diet, obesity, and risk of fatal prostate cancer. Am J Epidemiol 120, 244-250, 1984.

32. Shibata, A, Paganini-Hill, A, Ross, RK, and Henderson, BE: Intake of vegetables, fruits, beta-carotene, vitamin C and vitamin supplements and cancer incidence among the elderiy: a prospective study. Br J Cancer 66, 673-679, 1992.

33. Hsing, AW, McLaughlin, JK, Schuman, LM, Bjelke, E, Gridley G, et al.: Diet, tobacco use, and fatal prostate cancer: results from the Lutheran Brotherhood Cohort Study. Cancer Res 50, 6836-6840, 1990.

34. Schuurman, AG, Goldbohm, RA, Dorant, E, and vanden Brandt, PA: Vegetable and fruit consumption and prostate cancer risk: a cohort study in the Netherlands. Cancer Epidemiol Biomarkers Prev 7, 673680, 1998.

35. Schuman, LM, Mandel, JS, Radke, A, Seal, U, and Halberg, F: Some selected features of the epidemiology of prostatic cancer: Minneapolis-St Paul, Minnesota case-control study, 1976-1979. In Trends in Cancer Incidence: Causes and Implications, K Magnus (ed). Washington, DC: Hemisphere, 1982, pp 345-354.

36. Walker, ARP, Walker, BF, Tsotetsi, NG, Sebitso, C, Siwedi, D, et al.: Case-control study of prostate cancer in black patients in Soweto, South Africa. Br J Cancer 65, 438-441, 1992.

37. De Stefani, E, Fierro, L, Barrios, E, and Ronco, A: Tobacco, alcohol, diet and risk of prostate cancer. Tumori 81, 315-320, 1995.

38. Severson, RK, Nomura, AMY, Grove, JS, and Stemmermann, GN: A prospective study of demographics, diet, and prostate cancer among men of Japanese ancestry in Hawaii. Cancer Res 49, 1857-1860, 1989.

39. Le Marchand, L, Hankin, JH, Kolonel, LN, and Wilkens, LR: Vegetable and fruit consumption in relation to prostate cancer risk in Hawaii: a reevaluation of the effect of dietary beta-carotene. Am J Epidemiol 133, 215-219, 1991.

40. Khachik, F, Beecher, GR, and Smith, JC, Jr: Lutein, lycopene, and their oxidative metabolites in chemoprevention of cancer. J Cell Biochem Suppl 22, 236-246, 1995.

41. Key, TJ, Silcocks, PB, Davey, GK, Appleby, PN, and Bishop, DT: A case-control study of diet and prostate cancer. Br J Cancer 76, 678687, 1997.

42. Pawlega, J, Rachtan, J, and Dyba, T: Dietary factors and risk of prostate cancer in Poland. Results of case-control study. Neoplasma 43, 61-63, 1996.

43. Le Marchand, L, Kolonel, LN, Wilkens, LR, Myers, BC, and Hirohata, T: Animal fat consumption and prostate cancer: a prospective study in Hawaii. Epidemiology 5, 276-282, 1994.

44. La Vecchia, C, Negri, E, D'Avanzo, B, Franceschi, S, and Boyle, P: Dairy products and the risk of prostatic cancer. Oncology 48, 406-410, 1991.

45. Giovannucci, E: Selenium and risk of prostate cancer. Lancet 352, 755-756, 1998.

46. Hartman, TJ, Albanes, D, Pietinen, P, and Hartman, AM: The association between baseline vitamin E, selenium, and prostate cancer in the alpha-Tocopherol, beta-Carotene Cancer Prevention Study. Cancer Epidemiol Biomarkers Prev 7, 335-340, 1998.

47. Combs, GF, Jr, Clark, LC, and Turnbull, BW: Reduction of cancer mortality and incidence by selenium supplementation. Med Klin 92 Suppl 3, 42-45, 1997.

~~~~~~~~

By Meera G. Jain; Gregory T. Hislop; Geoffrey R. Howe and Parviz Ghadirian

Share this with your friends