Dietary Carotenoid Intake and Colorectal Cancer Risk

Transcription

1 NUTRITION AND CANCER, 42(2), Copyright 2002, Lawrence Erlbaum Associates, Inc. Dietary Carotenoid Intake and Colorectal Cancer Risk Paul Terry, Meera Jain, Anthony B. Miller, Geoffrey R. Howe, and Thomas E. Rohan Abstract: Several studies have found inverse associations between fruit and vegetable consumption and colorectal cancer risk, suggesting the potential etiological importance of carotenoids (and other phytochemicals) contained in these foods. However, only one study (a case-control study) has examined the association between dietary carotenoids other than -carotene and colorectal cancer risk. In the study reported here, we examined the relationships between dietary intakes of carotene, -carotene, lycopene, lutein, and -cryptoxanthin and colorectal cancer risk in a large cohort study of Canadian women. A case-cohort analysis was undertaken within the cohort of 56,837 women who were enrolled in the Canadian National Breast Screening Study and who completed a self-administered dietary questionnaire. During follow-up to the end of 1993, a total of 388 women were diagnosed with colorectal cancer. For comparative purposes, a subcohort of 5,681 women was randomly selected. After exclusions for various reasons, the analyses were based on 295 cases and 5,334 noncases. We did not find any clear association between intake of any of the studied carotenoids and colorectal cancer risk in the study population as a whole or in subgroups defined by smoking status, relative body weight (body mass index), intakes of total fat, energy, alcohol, and folic acid, or menopausal status. Our data do not support any association between dietary intakes of the studied carotenoids and colorectal cancer risk. However, given that this is the first prospective cohort study of carotenoids in relation to colorectal cancer, further studies are warranted. Introduction Colorectal cancer is the second most commonly occurring cancer among women and the third most commonly occurring cancer among men worldwide (1). The incidence of colorectal cancer is highest in developed countries, where incidence rates are generally per 100,000 per year (1). -Carotene, an antioxidant pigment found primarily in plants, has attracted considerable attention in previous studies of cancer at various sites, including colorectal cancer. However, an association between dietary -carotene intake and colorectal cancer risk has not been clearly established. Although case-control studies have mostly found inverse associations with risk of colorectal cancer (2 5) and adenomas (6,7), clinical trials of -carotene supplementation have found no clear benefits (8 12). Although the conflicting results according to study design suggest the possibility that selective recall of fruit and vegetable consumption in the case-control studies might have resulted in spurious inverse associations, it has also been speculated that -carotene might be less important than other carotenoids in cancer prevention (13,14) and that total carotenoids might be more important than any single carotenoid studied in isolation (14). However, information on the individual carotenoid content of various foods has become available only in the past few years (15), and, consequently, studies examining dietary carotenoids other than -carotene are few. A recent case-control study, apparently the only study to examine the associations between the intakes of several individual carotenoids and colon cancer risk (13), found a statistically significant inverse association with lutein intake but no association with intakes of -carotene, -carotene, cryptoxanthin, or lycopene (among the major carotenoids in the human diet) (13,15). However, that study was based on retrospectively collected data, which are potentially vulnerable to bias related to the recall of diet and the selection of study participants (16). Furthermore, the association between the intake of carotenoids and rectal cancer risk has not been examined. Therefore, in the study reported here, we examined dietary intake of carotenoids in relation to colon and rectal cancer risk in a large cohort of Canadian women. Overview Subjects and Methods The investigation was conducted by performing a casecohort analysis using data from a cohort of 56,837 women in P. Terry and T. E. Rohan are affiliated with the Department of Epidemiology and Social Medicine, Albert Einstein College of Medicine, Bronx, NY M. Jain and A. B. Miller are affiliated with the Department of Public Health Sciences, University of Toronto, Toronto, ON, Canada. M. Jain is also affiliated with Toronto Public Health, City of Toronto, Toronto, ON, Canada, and A. B. Miller with the Division of Clinical Epidemiology, Deutsches Krebsforschungszentrum, Heidelberg, Germany. G. R. Howe is affiliated with the Division of Epidemiology, School of Public Health, Columbia University, New York, NY.

2 the Canadian National Breast Screening Study (NBSS) who completed a self-administered quantitative food frequency questionnaire (in addition to an epidemiological questionnaire, which was completed by all NBSS participants). The NBSS is a multicenter randomized controlled trial of (primarily) mammographic screening for breast cancer in 89, 835 women aged yr at recruitment (17,18). Participants were recruited between 1980 and 1985 by various means, including personal invitation by letter, group mailings to employees of large institutions and members of professional associations, advertisements in newspapers, and public service announcements on radio and television. Dietary and Risk Factor Data On enrollment in the NBSS, all participants completed a questionnaire that sought data on demographic characteristics, menstrual and reproductive history, education, and cigarette smoking. Starting in 1982, a modified version of a previously validated self-administered quantitative food frequency questionnaire (19), which also elicited information on alcohol consumption and physical activity, was distributed to all new attendees at all screening centers and to women returning to the screening centers for rescreening. By the time the dietary questionnaire was introduced, some women had already been enrolled in the study and were not seen again at the screening centers. A total of 56,837 women returned completed dietary questionnaires, which asked about the frequency of consumption and usual portion size of 86 food items. Photographs of various portion sizes were included in the questionnaire to assist participants in quantifying intake. A comparison between the self-administered questionnaire and a full interviewer-administered questionnaire, which has been subjected to validity and reliability testing (20) and used in a number of epidemiological studies (21), revealed that the two methods gave estimates of intake of the major macronutrients and dietary fiber that were moderately to strongly correlated with each other (reported correlation coefficients of ) (22). Data from the completed self-administered questionnaires were used to estimate daily intake of carotenoids ( carotene, -carotene, lycopene, lutein + zeaxanthin, and -cryptoxanthin), total energy, various macronutrients, and other micronutrients using a nutrient database developed by modifying and extending food composition tables from the US Department of Agriculture to include typically Canadian foods (19). The values for carotenoid intake presented here are for intake from dietary sources alone (15), since supplemental carotenoids were not in general use in Canada when the data were collected. Case Definition and Ascertainment Cases were women who were diagnosed during follow-up with incident colorectal cancer. Outcome (incident colorectal cancer or death) was ascertained by means of computerized record linkage to the Canadian Cancer Database (a composite of data collected by the provincial populationbased cancer registries) and to the National Mortality Database, both of which are maintained by Statistics Canada. There is good evidence from the NBSS and from other sources that the use of record linkage to ascertain incident cancer cases and deaths in Canada is accurate and complete (23,24). Linkages to the databases yielded data on mortality and cancer incidence to 31 December In all, 388 cases were identified, of whom 48 were in the subcohort (see Construction of the Subcohort). Of these 388 cases, 41 were prevalent at recruitment and were excluded, as were an additional 52 because of missing dietary information, leaving 295 cases available for analysis. Construction of the Subcohort A subcohort was constructed by selecting a random sample of 5,681 women from the dietary cohort at baseline, which gave rise to 46 of the cases described above. Of those selected, 259 women whose dietary questionnaires were not available, 42 women with extreme energy values (±3 SD of the logarithm of energy intake), and 1 woman with prevalent colorectal cancer were excluded from the present analysis, leaving 5,379 women available for analysis. Statistical Analysis Analyses were based on the 295 incident cases (198 colon, 97 rectal) and 5,379 women (including 45 of the case patients) in the subcohort with dietary information and without extreme energy values. Cases contributed person-time to the study from their date of enrollment to the date of diagnosis of their colorectal cancer, and noncases contributed person-time from their date of enrollment to the termination of follow-up (31 December 1993) or death, whichever came first. Incidence rate ratios for the associations between dietary carotenoid intake and risk of colorectal cancer were estimated using Poisson regression, and robust standard errors were calculated (25), thereby yielding the appropriate confidence intervals for the rate ratios given the case-cohort sampling. For these analyses, estimates of nutrient intake were adjusted for total energy using the residual method (26). Individual carotenoids were categorized by quintiles and analyzed in models with and without mutual adjustment. The incidence rate ratios were adjusted for energy intake (fitted as a continuous variable) and for potential confounding by the following variables: age (in 5-yr age groups), smoking (never, past, current), body mass index (in quintiles), hours of vigorous physical activity (in quintiles), education (3 categories), multivitamin use (yes/no), and quintiles of alcohol, total fat, dietary fiber, folate, and calcium. In addition to analyzing risk in association with the individual carotenoids, we analyzed risk in association with a carotenoid index, which was created by summing the quintile scores for each of the five major carotenoids and then categorized by quartiles. For tests of trend in risk across successive levels of categorical variables, median values of each category were fitted in the 168 Nutrition and Cancer 2002

3 risk models as successive integers. Tests for interaction were based on likelihood ratio tests comparing models with and without product terms representing the variables of interest. The likelihood ratio test that all the interaction parameters were zero was performed by referring the differences between the deviances of models with and without interaction terms to the 2 distribution on degrees of freedom equal to the number of interaction parameters. Results In age- and multivariate-adjusted models (multivariate results shown in Table 1), dietary intakes of -carotene, carotene, lycopene, lutein, and -cryptoxanthin were not associated with altered colorectal cancer risk, either overall or for cancers of the colon or rectum when examined separately. Similarly, intake of carotenoids was not associated with altered risk of proximal or distal colon cancer (data not shown). Examining risk in association with decile levels of the carotenoids also revealed no association. Results were essentially unaltered when the carotenoids were mutually adjusted (data not shown). However, high positive correlations between several of the carotenoids, such as that observed between -carotene and -carotene (Pearson s R = 0.85), caution against mutual adjustment because of the possibility of collinearity, and removal of -carotene from the Table 1. Multivariate-Adjusted Incidence Rate Ratios for Daily Intake of Carotenoids a Colorectal Cancer (n = 295) Colon Cancer (n = 198) Rectal Cancer (n = 97) -Carotene, µg/day 2, (referent) 1.0 (referent) 1.0 (referent) 2,946 4, ( ) 0.7 ( ) 1.3 ( ) 4,168 5, ( ) 0.7 ( ) 0.8 ( ) 5,521 7, ( ) 0.6 ( ) 1.0 ( ) 7, ( ) 1.1 ( ) 1.7 ( ) P for trend b Carotene, µg/day (referent) 1.0 (referent) 1.0 (referent) ( ) 0.8 ( ) 0.8 ( ) 820 1, ( ) 1.0 ( ) 1.1 ( ) 1,166 1, ( ) 0.6 ( ) 0.8 ( ) 1, ( ) 0.9 ( ) 1.1 ( ) P for trend b Lycopene, µg/day 3, (referent) 1.0 (referent) 1.0 (referent) 3,925 6, ( ) 1.4 ( ) 0.8 ( ) 6,881 10, ( ) 0.8 ( ) 0.9 ( ) 10,438 16, ( ) 0.9 ( ) 0.5 ( ) 16, ( ) 1.3 ( ) 0.7 ( ) P for trend b Cryptoxanthin, µg/day (referent) 1.0 (referent) 1.0 (referent) ( ) 0.7 ( ) 0.2 ( ) ( ) 1.3 ( ) 0.6 ( ) ( ) 2.0 ( ) 0.7 ( ) ( ) 1.7 ( ) 0.7 ( ) P for trend b Lutein, µg/day 1, (referent) 1.0 (referent) 1.0 (referent) 1,607 2, ( ) 1.2 ( ) 1.1 ( ) 2,313 3, ( ) 0.8 ( ) 1.1 ( ) 3,182 4, ( ) 0.9 ( ) 0.9 ( ) 4, ( ) 1.0 ( ) 2.1 ( ) P for trend b Carotenoid index Lowest quartile 1.0 (referent) 1.0 (referent) 1.0 (referent) Second quartile 1.0 ( ) 1.0 ( ) 1.0 ( ) Third quartile 0.9 ( ) 1.0 ( ) 0.8 ( ) Highest quartile 1.0 ( ) 1.0 ( ) 1.1 ( ) P for trend b a: Multivariate models included age (in 5-yr age groups), smoking (never, past, current), body mass index (in quintiles), hours of vigorous physical activity (in quintiles), education (3 categories), multivitamin use (yes/no), and intakes of total energy (continuous), total fat, alcohol, dietary fiber, folate, and calcium (quintiles); n, number of cases. b: All P values are from 2-sided tests. Vol. 42, No

4 model did not alter the results for the remaining mutually adjusted carotenoids. Consistent with these findings, the index of total carotenoid intake was also unrelated to risk (Table 1). Additional adjustment for use of oral contraceptives, use of exogenous hormones, parity, age at first birth, height, treatment allocation, and study center did not alter the results, nor did excluding cases that occurred during the 1st yr of follow-up. Multivariate analyses in which dietary fiber intake (which was positively correlated with dietary carotenoids) was not included yielded results that were similar to those obtained in the full multivariate model. Also, excluding multivitamin users (6.2% of subjects) did not alter the results. The associations between intake of carotenoids and colorectal cancer risk did not vary according to strata of smoking status, Quetelet s index, menopausal status, or intakes of total fat, folate, total energy, or alcohol (Table 2). On formal testing, there was no evidence for interactions between dietary carotenoids and those factors in relation to colorectal cancer risk. Discussion In this large, prospective cohort study, we did not observe any clear association between intake of carotenoids and colorectal cancer risk when the carotenoids were examined individually or when they were examined in combination. There was also no clear indication of any important differences in the association over strata of various risk factors for colorectal cancer in our study population. Our results do not confirm the inverse association between dietary lutein intake and colorectal cancer risk observed in a previous case-control study (13). Among the strengths of our study was the large sample size of our cohort of women and the relatively long-term follow-up. The completeness of follow-up of the cohort (23,24) reduces the likelihood that our results reflect bias due to differential follow-up of exposed compared with unexposed women. It is also unlikely that undiagnosed early stages of colorectal cancer altered carotenoid intake, since we also observed no association after excluding cases that occurred during the 1st yr of follow-up. Moreover, the prospective assessment of diet in our study eliminates the potential for differential recall bias, which is of particular concern because preconceptions about the healthy effects of fruit and vegetables are common among the public. Our data were limited, however, by the possibility of measurement error with respect to carotenoid intake (27). Nondifferential misclassification would tend to attenuate true associations and can stem from measurement error in the assessment of diet due Table 2. Incidence Rate Ratios a for Colorectal Cancer According to an Increase in One Standard Deviation b of Daily Carotenoid Intake in Subgroups Defined by Selected Variables Variable n -Carotene -Carotene -Cryptoxanthin Lycopene Lutein Smoking status Never ( ) 1.18 ( ) 1.06 ( ) 1.01 ( ) 1.24 ( ) Past ( ) 0.94 ( ) 1.40 ( ) 1.01 ( ) 1.05 ( ) Current ( ) 1.12 ( ) 1.11 ( ) 1.28 ( ) 1.31 ( ) Body mass index, kg/m ( ) 1.04 ( ) 1.29 ( ) 0.96 ( ) 1.03 ( ) ( ) 0.93 ( ) 1.14 ( ) 0.93 ( ) 1.03 ( ) ( ) 1.05 ( ) 1.13 ( ) 1.29 ( ) 1.25 ( ) Total fat, g/day ( ) 0.97 ( ) 1.09 ( ) 1.17 ( ) 1.18 ( ) ( ) 1.10 ( ) 1.11 ( ) 0.98 ( ) 1.06 ( ) ( ) 1.00 ( ) 1.26 ( ) 1.06 ( ) 1.07 ( ) Folate intake, µg/day ( ) 0.99 ( ) 0.99 ( ) 0.94 ( ) 0.95 ( ) ( ) 0.83 ( ) 1.36 ( ) 1.15 ( ) 0.79 ( ) ( ) 1.14 ( ) 1.02 ( ) 1.07 ( ) 1.13 ( ) Energy intake, kcal/day 1, ( ) 0.96 ( ) 1.16 ( ) 1.22 ( ) 1.10 ( ) 1,753 2, ( ) 0.92 ( ) 1.11 ( ) 0.87 ( ) 1.08 ( ) 2, ( ) 1.15 ( ) 1.16 ( ) 1.08 ( ) 1.20 ( ) Alcohol intake, g/day ( ) 0.85 ( ) 0.96 ( ) 0.84 ( ) 1.04 ( ) ( ) 1.15 ( ) 1.24 ( ) 1.17 ( ) 1.16 ( ) ( ) 0.94 ( ) 1.17 ( ) 1.07 ( ) 1.05 ( ) Menopausal status Premenopausal ( ) 1.00 ( ) 1.33 ( ) 1.02 ( ) 0.99 ( ) Postmenopausal ( ) 1.07 ( ) 1.06 ( ) 1.08 ( ) 1.17 ( ) a: Multivariate models included variables listed in Table 1 footnote. b: Standard deviations for daily carotenoid intake used in these analyses are those observed in entire cohort and are as follows: -carotene, 3,556 µg; -carotene, 1,027 µg; -cryptoxanthin, µg; lycopene, 12,085 µg; lutein, 3,583 µg. 170 Nutrition and Cancer 2002

5 to inaccurate recall of past diet (27), changes in diet over time (27), or degradation of nutrients during cooking and storage of fruit and vegetables (27,28). Moreover, given the limited number of cancer cases in analyses for the various cancer subsites, such as with those of proximal and distal colon cancers, the interpretation of subgroup analyses should be made cautiously. Finally, although we adjusted our estimates for a wide range of potentially confounding variables, uncontrolled confounding from dietary (or other) factors cannot be excluded. A recent case-control study found an inverse association between lutein intake and colon cancer risk (13), especially for proximal cancers and among subjects whose cancer occurred before 67 yr of age. That study found no clear association with other carotenoids. Our results suggest that lutein intake is not associated with risk of cancer at any site in the colorectum, even though the range of intake in our data (the daily median intakes in the lowest and highest quintile levels were 1,231 and 6,833 µg, respectively) was higher than that observed in the previous study (300 and 1,395 µg, respectively). However, some studies suggest that individuals with very low intake of fruit and vegetables (29) and antioxidants (30) are at the highest risk of cancer, including colorectal cancer (31). Indeed, Slattery and colleagues (13) observed the greatest risk among individuals in the lowest lutein intake category (median intake 300 µg/day). Lutein is primarily found in broccoli and green leafy vegetables (15,28). A recent case-control study of serum carotenoids and adenomatous polyps of the distal colon and rectum (32) found no association with serum lutein, zeaxanthin, -cryptoxanthin, -carotene, -carotene, lycopene, or total carotenoids after multivariate adjustment. Although the null results of that study are consistent with our own results, epidemiological studies of adenomas may have little relevance to the evolution of adenomas to colorectal cancer (33). In conclusion, our data do not support the hypothesis that relatively high dietary intake of carotenoids is associated with reduced risk of colorectal cancer, although our data are consistent with weak associations in either direction. Our null results are generally consistent with those of a previous case-control study in which only very low lutein intake was associated with increased risk (13). Given the scarcity of studies of carotenoid intake in relation to colorectal cancer risk, further investigations are warranted. Acknowledgments and Notes This study was funded in part by the National Cancer Institute of Canada. Address correspondence to Dr. Paul Terry, Dept. of Epidemiology and Social Medicine, Albert Einstein College of Medicine, 1300 Morris Park Ave., Rm A, Bronx, NY Phone: (718) FAX: (718) pterry@aecom.yu.edu. Submitted 4 October 2001; accepted in final form 20 December References 1. Parkin DM, Pisani P, and Ferlay J: Estimates of the worldwide incidence of 25 major cancers in Int J Cancer 80, , Zaridze D, Filipchenko V, Kustov V, Serdyuk V, and Duffy S: Diet and colorectal cancer: results of two case-control studies in Russia. Eur J Cancer 29A, , Whittemore AS, Wu-Williams AH, Lee M, Zheng S, Gallagher RP, et al.: Diet, physical activity, and colorectal cancer among Chinese in North America and China. JNCI 82, , Bostick RM, Potter JD, McKenzie DR, Sellers TA, Kushi LH, et al.: Reduced risk of colon cancer with high intake of vitamin E: The Iowa Women s Health Study. Cancer Res 53, , Ferraroni M, La Vecchia C, D Avanzo B, Negri E, Franceschi S, et al.: Selected micronutrient intake and the risk of colorectal cancer. Br J Cancer 70, , Breuer-Katschinski B, Nemes K, Marr A, Rump B, Leiendecker B, et al.: Colorectal adenomas and diet: a case-control study. Colorectal Adenoma Study Group. Dig Dis Sci 46, 86 95, Enger SM, Longnecker MP, Chen MJ, Harper JM, Lee ER, et al.: Dietary intake of specific carotenoids and vitamins A, C, and E, and prevalence of colorectal adenomas. Cancer Epidemiol Biomarkers Prev 5, , Omenn GS: Chemoprevention of lung cancer: the rise and demise of carotene. Annu Rev Public Health 19, 73 99, Omenn GS, Goodman GE, Thornquist MD, Balmes J, Cullen MR, et al.: Effects of a combination of -carotene and vitamin A on lung cancer and cardiovascular disease. N Engl J Med 334, , Varis K, Taylor PR, Sipponen P, Samloff IM, Heinonen OP, et al.: Gastric cancer and premalignant lesions in atrophic gastritis: a controlled trial on the effect of supplementation with -tocopherol and carotene. The Helsinki Gastritis Study Group. Scand J Gastroenterol 33, , Hennekens CH, Buring JE, Manson JE, Stampfer M, Rosner B, et al.: Lack of effect of long-term supplementation with -carotene on the incidence of malignant neoplasms and cardiovascular disease. N Engl J Med 334, , Albanes D, Malila N, Taylor PR, Huttunen JK, Virtamo J, et al.: Effects of supplemental -tocopherol and -carotene on colorectal cancer: results from a controlled trial (Finland). Cancer Causes Control 11, , Slattery ML, Benson J, Curtin K, Ma KN, Schaeffer D, et al.: Carotenoids and colon cancer. Am J Clin Nutr 71, , Nishino H, Tokuda H, Murakoshi M, Satomi Y, Masuda M, et al.: Cancer prevention by natural carotenoids. Biofactors 13, 89 94, 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 selected multicomponent foods. J Am Diet Assoc 93, , Rothman KJ and Greenland S: Modern Epidemiology, 2nd ed. Philadelphia, PA: Lippincott-Raven, 1998, p Miller AB, Baines CJ, To T, and Wall C: Canadian National Breast Screening Study. Breast cancer detection and death rates among women aged 4 49 years and breast cancer detection and death rates among women aged 5 59 years. Can Med Assoc J 147, , Miller AB, Howe GR, and Wall C: The National Study of Breast Cancer Screening Protocol for a Canadian randomized controlled trial of screening for breast cancer in women. Clin Invest Med 4, , Jain M, Howe GR, and Rohan T: Dietary assessment in epidemiology: comparison on food frequency and a diet history questionnaire with a 7-day food record. Am J Epidemiol 143, , Jain M, Howe GR, Johnson KC, and Miller AB: Evaluation of a diet history questionnaire for epidemiologic studies. Am J Epidemiol 111, , Morgan RW, Jain M, Miller AB, Choi NW, Matthews V, et al.: A comparison of dietary methods in epidemiologic studies. Am J Epidemiol 107, , Jain MG, Harrison L, Howe GR, and Miller AB: Evaluation of a selfadministered dietary questionnaire for use in a cohort study. Am J Clin Nutr 36, , Vol. 42, No

6 23. Robles SC, Marrett LD, Clarke EA, and Risch HA: An application of capture-recapture methods to the estimation of completeness of cancer registration. J Clin Epidemiol 41, , Shannon HS, Jamieson E, Walsh C, Julian JA, Fair ME, et al.: Comparison of individual follow-up and computerized record linkage using the Canadian Mortality Data Base. Can J Public Health 80, 54 57, Huber PJ: The Behaviour of Maximum Likelihood Estimates Under Non-Standard Conditions. Berkeley, CA: University of California Press, Willett W and Stampfer MJ: Total energy intake: implications for epidemiologic analyses. Am J Epidemiol 124, 17 27, Willett WC: Nutritional Epidemiology, 2nd ed. New York: Oxford University Press, Micozzi MS, Beecher GR, Taylor PR, and Khachik F: Carotenoid analyses of selected raw and cooked foods associated with a lower risk for cancer. JNCI 82, , Terry P, Terry JB, and Wolk A: Fruit and vegetable consumption in the prevention of cancer: an update. J Int Med 250, , Blot WJ, Li JY, Taylor PR, Guo W, Dawsey S, et al.: Nutrition intervention trials in Linxian, China: supplementation with specific vitamin/mineral combinations, cancer incidence, and disease-specific mortality in the general population. JNCI 85, , Terry P, Giovannucci E, Michels KB, Bergkvist L, Hansen H, et al.: Fruit, vegetables, dietary fiber, and the risk of colorectal cancer. JNCI 93, , Shikany JM, Witte JS, Henning SM, Swendseid ME, Bird CL, et al.: Plasma carotenoids and the prevalence of adenomatous polyps of the distal colon and rectum. Am J Epidemiol 145, , Byers T: Diet, colorectal adenomas, and colorectal cancer. N Engl J Med 342, , Nutrition and Cancer 2002

7