Adherence to the USDA Food Guide, DASH Eating Plan, and Mediterranean Dietary Pattern Reduces Risk of Colorectal Adenoma 1 3

Transcription

1 The Journal of Nutrition Nutritional Epidemiology Adherence to the USDA Food Guide, DASH Eating Plan, and Mediterranean Dietary Pattern Reduces Risk of Colorectal Adenoma 1 3 L. Beth Dixon, 4 *AmyF.Subar, 5 Ulrike Peters, 6 Joel L. Weissfeld, 7 Robert S. Bresalier, 8 Adam Risch, 9 Arthur Schatzkin, 10 and Richard B. Hayes 10 4 Department of Nutrition, Food Studies, and Public Health, New York University, New York City, NY 10012; 5 Division of Cancer Prevention and Population Sciences, National Cancer Institute, Bethesda, MD 20892; 6 Division of Public Health Sciences, Cancer Prevention Program, Fred Hutchinson Cancer Research Center and Department of Epidemiology, University of Washington, Seattle, WA 98109; 7 University of Pittsburgh Medical Center, Pittsburgh, PA 15232; 8 MD Anderson Cancer Center, Houston, TX 77030; 9 Information Management Systems, Inc., Silver Spring, MD 20904; and 10 Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD Abstract The 2005 Dietary Guidelines for Americans include quantitative recommendations for 2 eating patterns, the USDA Food Guide and the Dietary Approaches to Stop Hypertension (DASH) Eating Plan, to promote optimal health and reduce disease risk. A Mediterranean dietary pattern has also been promoted for health benefits. Our objective was to determine whether adherence to the USDA Food Guide recommendations, the DASH Eating Plan, or a Mediterranean dietary pattern is associated with reduced risk of distal colorectal adenoma. In the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial, men and women aged y were screened for colorectal cancer by sigmoidoscopy at 10 centers in the U.S. After adjusting for potential confounders, men who most complied with the USDA Food Guide recommendations had a 26% reduced risk of colorectal adenoma compared with men who least complied with the recommendations (OR USDA score $ 5 vs. #2 ¼ 0.74, 95% CI ¼ ; P-trend, 0.001). Comparable results were found for men who had intakes most similar to the DASH Eating Plan or a Mediterranean dietary pattern. Women who most complied with the USDA Food Guide recommendations had an 18% reduced risk for colorectal adenoma, but subgroup analyses revealed protective associations only for current smokers (OR USDA score $ 5 vs. #2 ¼ 0.52, 95% CI ¼ ; P-trend, 0.01) or normal-weight women (OR USDA score $ 5 vs. #2 ¼ 0.74, 95% CI ¼ ; P-trend ¼ 0.08). Following the current U.S. dietary recommendations or a Mediterranean dietary pattern is associated with reduced risk of colorectal adenoma, especially in men. J. Nutr. 137: , Introduction Colorectal cancer is estimated to be the third leading cause of cancer mortality in men and women in the U.S. in 2007 (1). Between 24 and 47% of asymptomatic average-risk individuals over age 50 y are estimated to have adenomatous polyps, precursors to colorectal cancer (2). Several lifestyle factors, including diet, physical inactivity, excess body weight, and tobacco exposure early in life, are associated with developing colorectal cancer (3). To prevent or reduce risk of colorectal and other cancers, international and national organizations promote dietary recommendations that emphasize a plant-based diet of 1 Supported by the Intramural Research Program of the National Cancer Institute. The PLCO Cancer Screening Trial is funded by the National Cancer Institute, NIH, and the U.S. Department of Health and Human Services. 2 Author disclosures: L. B. Dixon, A. F. Subar, U. Peters, J. L. Weissfeld, R. S. Bresalier, A. Risch, A. Schatzkin, and R. B. Hayes, no conflicts of interest. 3 Supplemental Tables 1 5 are available with the online posting of this paper at jn.nutrition.org. * To whom correspondence should be addressed. beth.dixon@nyu.edu. fruits, vegetables, and whole grains, and dairy and meat products low in saturated and trans fatty acids, in addition to regular physical activity and avoidance of excess weight gain and alcohol intake (4 6). The 2005 Dietary Guidelines for Americans advise similar recommendations to promote health and reduce risk of major chronic diseases, including cancer (7). The 2005 Dietary Guidelines include quantitative recommendations for 2 eating patterns: the USDA Food Guide (formerly known as the U.S. Food Guide Pyramid) and the Dietary Approaches to Stop Hypertension (DASH) 11 Eating Plan. Both eating patterns suggest amounts of food to meet recommended nutrient intakes at different energy levels required by men and women throughout their lifetimes. Many nutritionists and epidemiologists also recommend a Mediterranean dietary pattern that emphasizes fruits, vegetables, legumes, whole grains, and foods high in monounsaturated fatty acids (MFA) like olive oil (8,9). A Mediterranean dietary pattern has also been 11 Abbreviations used: DASH, Dietary Approaches to Stop Hypertension; MFA, monounsaturated fatty acid; PLCO, Prostate, Lung, Colorectal, and Ovarian /07 $8.00 ª 2007 American Society for Nutrition Manuscript received 5 March Initial review completed 3 April Revision accepted 27 August 2007.

2 associated with reduced mortality and morbidity, including lower risk of colorectal and other cancers (10,11). Although intervention studies like the Polyp Prevention Trial did not reduce recurrence of colorectal adenoma, actual adherence to the multiple dietary goals (low fat, high dietary fiber, and high fruit and vegetable intake) was limited to a small sample of subjects (12,13). Results from observational studies of dietary patterns and colorectal cancer have been more consistent, but most of these studies use data-driven methods (e.g. factor analysis and cluster analysis) that are not intended for evaluation of formal dietary guidelines (14). The creation of dietary scores is a common, intuitively simple way to assess adherence to recommended eating patterns in relation to health outcomes (15). The objective of this study was to determine whether dietary scores that assess adherence to the USDA Food Guide recommendations, the DASH Eating Plan, and a Mediterranean dietary pattern are associated with reduced risk of distal colorectal adenoma, a precursor of colorectal cancer, in a large sample of U.S. men and women. Methods The Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial. This investigation of dietary patterns and colorectal adenoma was carried out in the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial (16), a 14-y, 2-armed, randomized trial designed to determine the effectiveness of early detection and identify etiologic determinants for these cancers. Between 1993 and 2000, men and women aged y were enrolled and screened at 1 of 10 centers (Birmingham, AL; Denver, CO; Detroit, MI; Honolulu, HI; Marshfield, WI; Minneapolis, MN; Pittsburgh, PA; Salt Lake City, UT; St. Louis, MO; and Washington DC) in the US. An equal number of controls have also been enrolled. For colorectal cancer, screening arm participants had a 60-cm flexible sigmoidoscopy at study entry. Participants with polypoid lesions or masses indicative of colorectal neoplasia were referred for endoscopic follow-up, which usually results in a histopathologic exam. Trained medical abstractors obtained and coded all available medical-pathologic reports on all lesions removed during the diagnostic endoscopy and related surgical procedures. Questionnaire data regarding sociodemographic factors, current and past smoking behavior, history of cancer and other diseases, use of selected drugs, recent history of screening examinations, reproductive history, and diet were collected from participants in the screening arm at baseline (17). Approval from the institutional review boards of the U.S. National Cancer Institute and the 10 screening centers and informed consent from all participants were obtained. Study participants. Successful sigmoidoscopy examinations (defined as insertion to at least 50 cm with.90% of mucosa visible or suspect lesion found) were carried out for 56,176 men and women aged y in the screening arm of the trial. Of these participants, 51,028 (91%) completed a FFQ and other risk factor questionnaires at study entry. We excluded participants for 1 or more of the following reasons: missing data for more than 7 items on the FFQ (n ¼ 451); extreme values for energy intake defined as lowest and highest 1% of gender-specific energy intake (n ¼ 947); previous history of cancer, except basal-cell skin cancer (n ¼ 2329); and self-reported history of ulcerative colitis, Crohn s disease, familial polyposis, colorectal polyps, or Gardner s syndrome (n ¼ 4614). Adenoma was defined as advanced if the polyps were large ($1 cm), had high-grade dysplasia (including cancer in situ), or had villous elements (including tubulo-villous adenomas). We also excluded participants if they had hyperplastic polyps only (n ¼ 1501), benign lesions not further specified (n ¼ 108), colorectal lesions (polyps or cancer) of unknown location (n ¼ 279), polyps of uncertain histology or cancer (n ¼ 1,513), histology data pending at the time of analysis (n ¼ 83), or were under medical follow-up but did not have additional endoscopy after a positive screening (n ¼ 2565). After all exclusions, this study included 3592 cases with pathologically verified left-sided adenomatous polyps of the distal colon (descending colon, sigmoid colon, or rectum) and 33,971 controls with no suspicion of neoplasia at this site on either sigmoidoscopic screening exam (n ¼ 32,415) or from a follow-up endoscopy after a suspect screen (n ¼ 1556). Data for this analysis were updated through September Creation of dietary scores. We assessed usual dietary intake over the 12 mo prior to enrollment using a 137-item FFQ that included 77 items with usual portion size, 10 questions about meat cooking practices, and 14 questions about vitamin/mineral supplement use. This FFQ was modeled after 3 well-established and validated questionnaires: the National Cancer Institute s Diet History Questionnaire (18), the Block FFQ (19), and the Willett FFQ (20). Values for daily intake of energy, nutrients (e.g. macronutrients, vitamins, minerals, and types of fatty acids), added sugars, drinks of alcohol, and servings from 27 food groups (e.g. grains, vegetables, fruits, dairy, meat and meat equivalents, and subgroups within the 5 main food groups) were determined from national dietary data and the Pyramid food group servings database from the Continuing Survey of Food Intakes by Individuals using the method developed by Subar et al. (21). This method disaggregates food mixtures (e.g. pizza) into their component ingredients or foods (e.g. pizza dough, tomato sauce, and cheese) before assigning items to their respective discrete food groups (e.g. grains, vegetables, and dairy). We created an 8-point USDA score to reflect adherence to the USDA Food Guide recommendations included in the appendix of the 2005 Dietary Guidelines for Americans (7). Suggested amounts of foods for this eating pattern vary across a range of energy intakes 12 ( kcal/d) to meet the needs of different age and gender groups in the U.S. population. We used recommendations from the 2000 kcal/d eating plan for men and from the 1600 kcal/d eating plan for women that are appropriate for men and women who are $51 y and have lifestyles that include mostly light physical activity associated with typical daily life (characteristics of the PLCO study population). For their USDA scores, men and women received 1 point if the minimum recommended number of daily servings from each of the 5 main USDA food groups was met according to the sex-specific eating plans (i.e. for men at 2000 kcal/d: $6 servings 13 /d of grains, $3 servings/d of dairy products, $4 servings/d of fruits, $5 servings/d of vegetables, $5.5 oz 14 /d of meat and meat equivalents; for women at 1600 kcal/d: $5 servings/d of grains, $3 servings/d of dairy products, $ 3 servings/d of fruits, $ 4 servings/d of vegetables, $5 oz/d of meat and meat equivalents). Men and women also received 1 point if their average daily consumption of saturated fat was less than the recommended 10% of total energy intake, if their added sugar intake was #7% of total energy intake, and if their alcohol intake was #2 drinks 15 /d for men and 1 drink/d for women. We included these 3 components because the 2005 Dietary Guidelines for Americans state that the discretionary calorie allowance covers all calories from added sugars, alcohol, and the additional fat found in even moderate fat choices from the milk and meat group, which is primarily saturated fat. The USDA Food Guide also includes a separate recommendation for intake of oils in grams, which are primarily polyunsaturated and MFA. We did not include this recommendation as a separate component in the USDA score because the USDA servings database used with the PLCO FFQ did not separate gram amounts from oils and gram amounts from solid fat. We used similar methods to create a 9-point DASH score to reflect adherence to the DASH Eating Plan included in the appendix of the 2005 Dietary Guidelines for Americans (7). The DASH Eating Plan states that whole grains are recommended for most servings to meet fiber recommendations. We considered intake of 67% of grains as whole grains as most servings. Accordingly, men and women received 1 point if they consumed the minimum recommended number of daily servings from each of the 6 main food groups according to their sex-specific eating plan [i.e. for men at 2000 kcal/d: $4.7 (67% 3 7) servings/d of 12 1 kcal ¼ kj. 13 See (7). 14 1oz.¼ g drink ¼ 12 oz beer, 5 oz wine, or 1.5 oz 80-proof distilled spirits Dixon et al.

3 whole grains, $4 servings/d of vegetables, $4 servings/d of fruits, $2 servings/d of dairy products,,6 oz/d of meat and meat equivalents, and $4 servings/d of nuts, seeds, and legumes; for women at 1600 kcal/d: $4 (67% 3 6) servings/d of grains, $3 servings/d of vegetables, $4 servings/d of fruits, $2 servings/d of dairy products,,6 oz/d of meat and meat equivalents, $ 3 servings/d of nuts, seeds, and legumes]. Men and women also received 1 point if their average daily consumption of saturated fat was less than the recommended 5% of total energy intake, if their added sugar intake was #3% of total energy intake, and if their alcohol intake was at or below the recommended 2 drinks/d for men and 1 drink/d for women. We used a cut-point of 5% for saturated fat intake based on estimates generated from a 1-wk menu of the DASH Eating Plan provided in the 2005 Dietary Guidelines for Americans. To be consistent with the USDA score, we used the average recommended contribution of added sugars to total energy intake (i.e. 3%) for the different energy levels of DASH Eating Plans as the cut-point for both men and women. For comparison, we created a dietary score representing a Mediterranean dietary pattern based on the methodology of Trichopoulou et al. (11). The MED score has a maximum of 9 points, counting 1 point if: daily servings of vegetables, legumes, fruits and nuts, cereals, and fish, and the ratio of grams of MFA to grams of SFA were equal to or greater than the sex-specific median intake of the study population; daily servings of meat (including red meat, poultry, organ meats, and processed meats) and dairy foods were less than the sex-specific median intake of the study population; and alcohol intake was g/d for men and 5 25 g/d for women. In our study population, the respective cut-points for men were: $4.86 servings/d of vegetables, $0.17 servings/d of legumes, $3.15 servings/d of fruits and nuts, $6.37 servings/d of cereals, $0.51 oz/d of fish, $1.14 MFA:SFA ratio,,3.23 oz/d of meat, and,1.58 servings/d of dairy. The respective cut-points for women were: $4.39 servings/d of vegetables, $0.11 servings/d of legumes, $3.33 servings/d of fruits and nuts, $4.98 servings/d of cereals, $0.45 oz/d of fish, $1.10 MFA:SFA ratio,,2.09 oz/d of meat, and,1.42 servings/d of dairy. Statistical analysis. To determine prevalence OR of colorectal adenoma, we created logistic regression models with the USDA, DASH, and MED dietary scores entered as continuous or categorical variables. USDA and DASH score categories of #2, 3, 4, and $5 and MED score categories of #2, 3, 4, 5, and $6 were selected due to lower numbers of cases in the extreme categories. Outcome variables included all colorectal adenoma, colon adenoma, rectal adenoma, nonadvanced adenoma, and advanced adenoma. Because our findings differed by sex, we conducted separate models for men and for women. All models were adjusted for screening center, age at randomization into the trial, and daily energy intake. For men, all models were also adjusted for ethnic origin (American Indian/Alaskan Native, Asian, Hispanic, non-hispanic black, non-hispanic white, or Pacific Islander), educational attainment (,12 y of school, 12 y of school/high school equivalent, some college, college graduate, or postgraduate), BMI in kg/m 2, physical activity (none,,1, 1, 2, 3, 41 h/wk of vigorous activity), smoking (never, quit smoking $20 y ago and smoked #1 pack/d, quit smoking $20 y ago and smoked.1 pack/d, quit smoking,20 y ago and smoked #1 pack/d, quit smoking,20 y ago and smoked.1 pack/d, ever smoked cigar or pipe, currently smokes cigarettes), use of aspirin and ibuprofen separately (no regular use,,2, 2 3, 4, 8, 12 16, 30, 60 times per month), and calcium intake from dietary supplements (in milligrams per day). For women, all models were adjusted for BMI, smoking, calcium intake from dietary supplements, and use of hormone replacement therapy (never, former, or current). These covariates were selected based on the literature and because they were significant (P, 0.05) in most models for all colorectal adenoma for each sex. The test for trend was determined using the median value of each category of each score included in each model. We conducted stratified analyses for smoking status (never, former, or current) and BMI (,25 kg/m 2 for normal weight vs kg/m 2 for overweight vs. $30 kg/m 2 for obese), because the interaction terms with smoking status or BMI category and each dietary score in the fully adjusted models for all colorectal adenoma had P-values,0.1. We also created multivariate models to determine whether individual components of the USDA, DASH, and MED dietary scores were independently associated with the colorectal adenoma outcomes. All analyses were conducted using SAS version 9.1 (SAS Institute). Significance was determined by a 2-sided P-value,0.05. Results The PLCO study population included mostly well-educated non- Hispanic white men and women (Table 1). Participants tended toward overweight; most were never or former smokers and reported healthful behaviors such as weekly exercise or taking a multivitamin regularly. Within the respective sociodemographic and behavioral categories, adherence to the USDA Food Guide recommendations (USDA score $5) was greater for women, participants from ethnicities other than non-hispanic white or black, and participants who had years of education beyond high school, exercised more frequently, never smoked, and took aspirin, ibuprofen, or a multivitamin regularly (Supplemental Table 1). Adherence to the DASH Eating Plan (DASH score $5) was difficult to achieve for all participants, whereas participant characteristics associated with adherence to a Mediterranean dietary pattern (MED score $6) were similar to the USDA Food Guide. In general, participants age and BMI did not vary but TABLE 1 Characteristics of the PLCO study population 1 Characteristics Study population n 33,971 Age, y Men, % 51.3 Ethnicity, % Non-Hispanic White 90.4 Non-Hispanic Black 3.2 Hispanic 1.5 Asian 4.2 Pacific Islander 0.5 American Indian or Alaskan Native 0.2 Education, %,12 y y y 71.0 BMI, % ,18.5 kg/m kg/m kg/m $30 kg/m Vigorous physical activity $1 h/wk, % 68.9 Smoking status, % Never 46.4 Former 45.7 Current 8.0 Aspirin use.13/wk, % 39.6 Ibuprofen use.13/wk, % 18.9 Supplemental calcium use, mg/d Hormone replacement therapy use, % Never 31.3 Former 15.0 Current 53.7 Energy intake, 2 kcal/d Values are means 6 SEM or percentages. 2 To convert kcal to kj, multiply by Dietary patterns and risk of colorectal adenoma 2445

4 energy intake and calcium intake from dietary supplements increased with adherence to all 3 dietary scores. After adjusting for potential confounders, we observed an 8% reduced risk in men and 5% reduced risk in women for colorectal adenoma per unit increase in the USDA score (Table 2). Men who most complied with the USDA Food Guide recommendations (USDA score $5) had 26% reduced risk of colorectal adenoma, 23% reduced risk of colon adenoma (descending colon and sigmoid), and 40% reduced risk of rectal adenoma compared with men who least complied with the recommendations (USDA score #2). Men who most complied with the USDA Food Guide recommendations also had 29% reduced risk for nonadvanced adenoma and 22% reduced risk for advanced adenoma. Women who most complied with the 2446 Dixon et al. TABLE 2 Risk of distal colorectal adenoma according to USDA score 1 Continuous #2 3 4 $5 P-trend Men Cases, 2 n Multivariate-adjusted OR , % CI 0.88, , , , 0.85 Cases, n Multivariate-adjusted OR % CI 0.88, , , , 0.91 Cases, n Multivariate-adjusted OR % CI 0.79, , , , 0.84 Cases, n Multivariate-adjusted OR , % CI 0.87, , , , 0.85 Cases, n Multivariate-adjusted OR % CI 0.88, , , , 0.97 Women Cases, 4 n Multivariate-adjusted OR % CI 0.91, , , , 0.99 Cases, n Multivariate-adjusted OR % CI 0.90, , , , 1.02 Cases, n Multivariate-adjusted OR % CI 0.84, , , , 1.18 Cases, n Multivariate-adjusted OR % CI 0.91, , , , 1.06 Cases, n Multivariate-adjusted OR % CI 0.86, , , , For men, the USDA score ¼ 8 points if $6 servings/d of grains, $ 3 servings/d of dairy products, $ 4 servings/d of fruits, $ 5 servings/d of vegetables, $5.5 oz/d of meat and meat equivalents,,10% of daily energy intake from saturated fat, #7% of daily energy intake from added sugars, and #2 drinks/d of alcohol. For women, the USDA score ¼ 8 points if $5 servings/d of grains, $ 3 servings/d of dairy products, $ 3 servings/d of fruits, $ 4 servings/d of vegetables, $5 oz/d of meat and meat equivalents,,10% of energy intake from saturated fat, #7% of energy intake from added sugars, and #1 drink/d of alcohol. 2 Number of controls for men ¼ 17,435 (5256, 3928, 3821, and 4430 within the respective score categories). 3 All models for men were adjusted for center, age, energy intake, race, education, current BMI, physical activity, smoking, aspirin use, ibuprofen use, and calcium supplement use. 4 Number of controls for women ¼ 16,536 (3799, 3602, 3943, and 5192 within the respective score categories). 5 All models for women were adjusted for center, age, energy intake, current BMI, smoking, calcium supplement use, and hormone replacement therapy use.

5 USDA Food Guide recommendations had an 18% reduced risk of colorectal adenoma compared with women who least complied with the recommendations, but significance was not achieved when the outcome was either colon or rectal adenoma or if colorectal adenoma was advanced. In men, results for both the DASH score and MED score were similar to results for the USDA Food Guide recommendations. After adjusting for potential confounders, we observed a 10% reduced risk of colorectal adenoma per unit increase in the DASH score (Table 3). Men who most complied with the DASH Eating Plan (DASH score $5) had a 25% reduced risk of colorectal adenoma and a 33% reduced risk for advanced adenoma compared with men who least complied with the DASH Eating Plan (DASH score #2). We also observed 5% reduced risk of TABLE 3 Risk of distal colorectal adenoma according to DASH score 1 Continuous #2 3 4 $5 P-trend Men Cases, 2 n Multivariate-adjusted OR , % CI 0.86, , , , 0.91 Cases, n Multivariate-adjusted OR , % CI 0.85, , , , 0.95 Cases, n Multivariate-adjusted OR % CI 0.77, , , , 1.03 Cases, n Multivariate-adjusted OR % CI 0.86, , , , 1.01 Cases, n Multivariate-adjusted OR % CI 0.82, , , , 0.91 Women Cases, 4 n Multivariate-adjusted OR % CI 0.91, , , , 1.30 Cases, n Multivariate-adjusted OR % CI 0.95, , , , 1.62 Cases, n Multivariate-adjusted OR % CI 0.75, , , , 1.27 Cases, n Multivariate-adjusted OR % CI 0.93, , , , 1.46 Cases, n Multivariate-adjusted OR % CI 0.83, , , , For men, the DASH score ¼ 9 points if $67% of grains recommended ($7 servings/d) were whole grains ($4.7 servings/d), $4 servings/d of vegetables, $4 servings/d of nuts and legumes, $4 servings/d of fruits, $2 servings/d of dairy products,,6 oz/d of meat and meat equivalents,,5% of daily energy intake from saturated fat, #3% of daily energy intake from added sugars, and #2 drinks/d of alcohol. For women, the DASH score ¼ 9 points if $67% of grains recommended ($ 6 servings/d) were whole grains ($ 4 servings/d), $3 servings/d of vegetables, $3 servings/d of nuts and legumes, $4 servings/d of fruits, $2 servings/d of dairy products,,6 oz/d of meat and meat equivalents,,5% of daily energy intake from saturated fat, #3% of daily energy intake from added sugars, and #1 drink/d of alcohol. 2 Number of controls for men ¼ 17,435 (5848, 6026, 4038, and 1528 within the respective score categories). 3 All models for men were adjusted for center, age, energy intake, race, education, current BMI, physical activity, smoking, aspirin use, ibuprofen use, and calcium supplement use. 4 Number of controls for women ¼ 16,536 (3558, 6269, 4698, and 2011 within the respective score categories). 5 All models for women were adjusted for center, age, energy intake, current BMI, smoking, calcium supplement use, and hormone replacement therapy use. Dietary patterns and risk of colorectal adenoma 2447

6 TABLE 4 Risk of distal colorectal adenoma according to MED score 1 Continuous # $6 P-trend Men Cases, 2 n Multivariate-adjusted OR % CI 0.92, , , , , 0.92 Cases, n Multivariate-adjusted OR % CI 0.93, , , , , 1.00 Cases, n Multivariate-adjusted OR % CI 0.89, , , , , 0.92 Cases, n Multivariate-adjusted OR % CI 0.93, , , , , 1.03 Cases, n Multivariate-adjusted OR % CI 0.89, , , , , 0.89 Women Cases, 4 n Multivariate-adjusted OR % CI 0.94, , , , , 1.23 Cases, n Multivariate-adjusted OR % CI 0.94, , , , , 1.23 Cases, n Multivariate-adjusted OR % CI 0.89, , , , , 1.73 Cases, n Multivariate-adjusted OR % CI 0.94, , , , , 1.40 Cases, n Multivariate-adjusted OR % CI 0.91, , , , , For men and women, the MED score ¼ 9 points (1 point if servings/d of vegetables, legumes, fruits and nuts, cereals, and fish intake is greater than or equal to the sex-specific median intake; 1 point if ratio of MFA:SFA is greater than or equal to the sex-specific median intake; 1 point if servings/d of meat and dairy intake is less than the sex-specific median intake; and 1 point if alcohol is g/d for men or 5 25 g/d for women). 2 Number of controls for men ¼ 17,435 (2753, 3005, 3689, 3602, and 4386 within the respective score categories). 3 All models for men are adjusted for center, age, energy intake, race, education, current BMI, physical activity, smoking, aspirin use, ibuprofen use, and calcium supplement use. 4 Number of controls for women ¼ 16,536 (2664, 3114, 3611, 3422, and 3725 within the respective score categories). 5 All models for women are adjusted for center, age, energy intake, current BMI, smoking, calcium supplement use, and hormone replacement therapy use. colorectal adenoma per unit increase in the MED score (Table 4). Men whose diets most represented a Mediterranean dietary pattern (MED score $6) had a 21% reduced risk of colorectal adenoma and 29% reduced risk for advanced adenoma compared with men whose diets were least similar to a Mediterranean dietary pattern (MED score #2). In women, following the DASH Eating Plan or a Mediterranean dietary pattern was not associated with risk of colorectal adenoma (Tables 3 and 4). Results from stratified analyses for smoking status and body weight showed differences between men and women for the 3 eating patterns (Supplemental Tables 2 4). In men who were former smokers, compliance with all 3 eating patterns was associated with reduced risk of colorectal adenoma, but in men who currently smoked, only adherence to the USDA Food Guide recommendations was associated. In men who never smoked, risk of colorectal adenoma was not associated with any dietary 2448 Dixon et al.

7 score. All men, regardless of BMI category, had reduced risk of colorectal adenoma with compliance to the USDA Food Guide recommendations and DASH Eating Plan. Both overweight and obese men whose diets were most similar to a Mediterranean dietary pattern also showed reduced risk of colorectal adenoma. In women, adherence to the USDA Food Guide recommendations was associated with reduced risk of colorectal adenoma only in current smokers and women who were normal weight. Following the DASH Eating Plan or a Mediterranean dietary pattern was not associated with colorectal adenoma in any subgroup of women based on smoking status or body weight. Results from analyses of individual components of the 3 scores showed that men who met the USDA score criteria for grains, fruits, saturated fat, and alcohol intake had reduced risk of colorectal adenoma compared with men who did not meet the respective criteria (Supplemental Table 5). Meeting the DASH score criteria for fruits and alcohol intake and the MED score criteria for cereals, fruits and nuts, and meat intake were also associated with reduced risk of colorectal adenoma in men. Men who met the Mediterranean diet criterion for alcohol intake had increased risk of colorectal adenoma. In women, only meeting the USDA or MED score criterion for intake of grains or cereals was associated with reduced risk of colorectal adenoma. Discussion In this large study of.3500 colorectal adenoma cases, consuming a variety of foods from different food groups at the minimum daily amounts recommended by the U.S. Food Guide Pyramid while maintaining saturated fat, added sugars, and alcohol intakes below the recommended daily intake limits was associated with significant reduction in risk of colorectal adenoma in men and women. Similar reduction in risk of colorectal adenoma was observed in men who followed a DASH or Mediterranean dietary pattern. These findings were true for all men regardless of weight status, but results varied by smoking status. In men, compliance with all 3 eating patterns was associated with reduced risk of colorectal adenoma among former smokers, but only compliance with the USDA Food Guide recommendations was associated in current smokers and no associations were found in men who never smoked. In women, following the USDA Food Guide recommendations was associated with reduced risk of colorectal adenoma. These results remained significant only in women who were current smokers or normal weight. Other studies have evaluated cancer risk with dietary scores that included aspects of the USDA Food Guide recommendations and reported mixed results. In general, higher variety or diversity scores have been associated with reduced risk of colorectal cancer and other cancers (22 26), although increased dietary diversity was associated with increased risk of colorectal cancer in men from western New York (27). The USDA Healthy Eating Index score was not associated with overall cancer risk in U.S. nurses or health professionals (28,29), whereas an index score that included servings from food groups, body weight, and physical activity was associated with reduced risk of colon cancer in postmenopausal women (30). Scores created to assess a Mediterranean dietary pattern have also shown reduced risk of different cancers (11,31). To date, no studies have evaluated a DASH Eating Pattern with cancer. The DASH diet was developed to reduce hypertension and risk of cardiovascular disease, although key dietary recommendations (e.g. high fruit and vegetable intake) are important for cancer prevention. Our study has the advantage of large numbers of men and women from diverse U.S. geographic areas examined for screening purposes, largely avoiding the selection biases associated with clinical studies. It is difficult, however, to explain the risk differences observed between men and women for the 3 dietary scores. Like men, women showed reduced risk of colorectal adenoma with increasing adherence to the USDA Food Guide recommendations, but significance was achieved only in women who currently smoked or were normal weight. Among overweight and obese women, it is possible that the adiposity overwhelms the protective effect of healthful eating. It is also possible that dietary patterns high in fruits and vegetables may be more beneficial in smokers, a finding previously shown for lung cancer in male smokers compared with nonsmokers (32). Our results in women may also reflect the inability to completely adjust for other colorectal tumor risk factors or lower statistical power from smaller sample sizes across categories of dietary scores. Underreporting energy intake may be another explanation and has been shown to reduce the likelihood of meeting the criteria for certain food groups (33), which differ across the 3 eating patterns. However, we have no reason to believe that women would have underreported more so than men in our study population. Dietary scores have general limitations, described in detail elsewhere (14), that may also contribute to our findings. Measurement error in the dietary data is especially important to mention because it may be much larger than previously appreciated (34,35), although how this error applies to dietary scores is unknown. Interpretation of adherence to dietary guidelines or what constitutes a Mediterranean dietary pattern varies among researchers, as shown by the many dietary scores that have been created and evaluated for different health outcomes in different populations (36 38). The scores in our study had similar criteria for a few components (e.g. grains, fruits, dairy), but further differentiation of certain food groups and evaluation of their respective cut-points may add insight into risk differences observed between men and women. For example, in the MED score, meat and fish are separate components with cut-points of 2 3 oz/d for meat and ;0.5 oz/ d for fish. In contrast, the meat component in the USDA and DASH scores includes fish and meat equivalents (e.g. eggs and soy products) and 5 6 oz/d is recommended. Recommendations for daily alcohol intake also differ between the U.S. recommendations and a Mediterranean dietary pattern, yielding different results for that component with risk for colorectal adenoma. However, the exclusion of alcohol intake from the 3 scores and subsequent adjustment for alcohol intake as a potential confounder in the regression models did not change our findings. In conclusion, following the recommendations of the USDA Food Guide, the DASH Eating Pattern, or a Mediterranean dietary pattern was associated with reduced risk of colorectal adenoma of the distal large bowel in men. Women who followed the USDA Food Guide recommendations also had reduced risk of colorectal adenoma, but this was true only for women who currently smoked or were normal weight. Findings from our study suggest that following current U.S. dietary recommendations or a Mediterranean dietary pattern could improve colorectal cancer prevention and control, especially in men. Literature Cited 1. Jemal A, Siegel R, Ward E, Murray T, Xu J, Thun MJ. Cancer statistics, CA Cancer J Clin. 2007;57: Dietary patterns and risk of colorectal adenoma 2449

8 2. Anderson WF, Guyton KZ, Hiatt RA, Vernon SW, Levin B, Hawk E. Colorectal cancer screening for persons at average risk. J Natl Cancer Inst. 2002;94: Martinez ME. Primary prevention of colorectal cancer: lifestyle, nutrition, exercise. Recent Results Cancer Res. 2005;166: WHO. Joint WHO FAO expert report on diet, nutrition, and the prevention of chronic disease. Geneva: WHO; World Cancer Research Fund and American Institute for Cancer Research. Food, nutrition, and the prevention of cancer: a global perspective. Washington DC: American Institute for Cancer Research; Kushi LH, Byers T, Doyle C, Bandera EV, McCullough M, Gansler T, Andrews KS, Thun MJ. The American Cancer Society 2006 Nutrition and Physical Activity Guidelines Advisory Committee. American Cancer Society Guidelines on Nutrition and Physical Activity for cancer prevention: reducing the risk of cancer with healthy food choices and physical activity. CA Cancer J Clin. 2006;56: U.S. Department of Health and Human Services and USDA. Dietary guidelines for Americans th ed. Washington DC: U.S. Government Printing Office; January Willett WC. The Mediterranean diet: science and practice. Public Health Nutr. 2006;9: Nestle M. Mediterranean diets: historical and research overview. Am J Clin Nutr. 1995;61:S Serra-Majem L, Roman B, Estruch R. Scientific evidence of interventions using the Mediterranean diet: a systematic review. Nutr Rev. 2006;64:S Trichopoulou A, Costacou T, Bamia C, Trichopoulos D. Adherence to a Mediterranean diet and survival in a Greek population. N Engl J Med. 2003;348: Lanza E, Schatzkin A, Daston C, Corle D, Freedman L, Ballard-Barbash R, Caan B, Lance P, Marshall J, et al. Implementation of a 4-y, highfiber, high-fruit-and-vegetable, low-fat dietary intervention: results of dietary changes in the Polyp Prevention Trial. Am J Clin Nutr. 2001; 74: Wanke KL, Daston C, Slonim A, Albert PS, Snyder K, Schatzkin A, Lanza E. Adherence to the polyp prevention trial dietary intervention is associated with a behavioral pattern of adherence to nondietary trial requirements and general health recommendations. J Nutr. 2007; 137: Moeller SM, Reedy J, Millen AE, Dixon LB, Newby PK, Tucker KL, Krebs-Smith SM, Guenther PM. Dietary patterns: challenges and opportunities in dietary patterns research. An Experimental Biology Workshop, April 1, J Am Diet Assoc. 2007;107: Dixon LB, Cronin FJ, Krebs-Smith SM. Let the Pyramid guide your food choices: capturing the total diet concept. J Nutr. 2001;131:S Gohagan JK, Prorok PC, Hayes RB, Kramer BS, for the PLCO Project Team. The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial of the National Cancer Institute: history, organization, and status. Control Clin Trials. 2000;21:S Hayes RB, Reding D, Kopp W, Subar AF, Bhat N, Rothman N, Caporaso N, Ziegler RG, Johnson CC, et al. Etiologic and early marker studies in the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial. Control Clin Trials. 2000;21:S Subar AF, Thompson FE, Kipnis V, Midthune D, Hurwitz P, McNutt S, McIntosh A, Rosenfeld S. Comparative validation of the Block, Willett, and National Cancer Institute food frequency questionnaires: the Eating at America s Table Study. Am J Epidemiol. 2001;154: Block G, Woods M, Potosky A, Clifford C. Validation of a selfadministered diet history questionnaire using multiple diet records. J Clin Epidemiol. 1990;43: Willett WC, Sampson L, Stampfer MJ, Rosner B, Bain C, Witschi J, Hennekens CH, Speizer FE. Reproducibility and validity of a semiquantitative food frequency questionnaire. Am J Epidemiol. 1985;122: Subar AF, Midthune D, Kulldorff M, Brown CC, Thompson FE, Kipnis V, Schatzkin A. Evaluation of alternative approaches to assign nutrient values to food groups in food frequency questionnaires. Am J Epidemiol. 2000;152: Kant AK, Schatzkin A, Ziegler RG. Dietary diversity and subsequent cause-specific mortality in the NHANES I Epidemiologic Follow-up Study. J Am Coll Nutr. 1995;14: Fernandez E, D Avanzo B, Negri E, Franceschi S, La Vecchia C. Diet diversity and the risk of colorectal cancer in northern Italy. Cancer Epidemiol Biomark Prev. 1996;5: Fernandez E, Negri E, La Vecchia C, Franceschi S. Diet diversity and colorectal cancer. Prev Med. 2000;31: Franceschi S, Favero A, La Vecchia C, Negri E, Dal Maso L, Salvini S, Decarli A, Giacosa A. Influence of food groups and food diversity on breast cancer risk in Italy. Int J Cancer. 1995;63: La Vecchia C, Munoz SE, Braga C, Fernandez E, Decarli A. Diet diversity and gastric cancer. Int J Cancer. 1997;72: McCann SE, Randall E, Marshall JR, Graham S, Zielezny M, Freudenheim JL. Diet diversity and risk of colon cancer in Western New York. Nutr Cancer. 1994;21: McCullough ML, Feskanich D, Rimm EB, Giovannucci EL, Ascherio A, Variyam JN, Spiegelman D, Stampfer MJ, Willett WC. Adherence to the Dietary Guidelines for Americans and risk of major chronic disease in men. Am J Clin Nutr. 2000;72: McCullough ML, Feskanich D, Stampfer MJ, Rosner BA, Hu FB, Hunter DJ, Variyam JN, Colditz GA, Willett WC. Adherence to the Dietary Guidelines for Americans and risk of major chronic disease in women. Am J Clin Nutr. 2000;72: Harnack L, Nicodemus K, Jacobs DR, Folsom AR. An evaluation of the Dietary Guidelines for Americans in relation to cancer occurrence. Am J Clin Nutr. 2002;76: Bosetti C, Gallus S, Trichopoulou A, Talamini R, Franceschi S, Negri E, La Vecchia C. Influence of a Mediterranean diet on the risk of cancers of the upper aerodigestive tract. Cancer Epidemiol Biomark Prev. 2003;12: Balder HF, Goldbohm RA, van den Brandt PA. Dietary patterns associated with male lung cancer risk in the Netherlands Cohort Study. Cancer Epidemiol Biomarkers Prev. 2005;14: Krebs-Smith SM, Graubard BI, Kahle LL, Subar AF, Cleveland LE, Ballard-Barbash R. Low energy reporters vs others: a comparison of reported food intakes. Eur J Clin Nutr. 2000;54: Subar AF, Kipnis V, Troiano RP, Midthune D, Schoeller DA, Bingham S, Sharbaugh CO, Trabulsi J, Runswick S, et al. Using intake biomarkers to evaluate the extent of dietary misreporting in a large sample of adults: the OPEN Study. Am J Epidemiol. 2003;158: Kipnis V, Midthune D, Freedman LS, Bingham S, Schatzkin A, Subar A, Carroll RJ. Empirical evidence of correlated biases in dietary assessment instruments and its implications. Am J Epidemiol. 2001;153: Kant AK. Dietary patterns and health outcomes. J Am Diet Assoc. 2004;104: Bach A, Serra-Majem L, Carrasco JL, Roman B, Ngo J, Bertomeu I, Obrador B. The use of indexes evaluating the adherence to the Mediterranean diet in epidemiological studies: a review. Public Health Nutr. 2006;9: Fogli-Cawley JJ, Dwyer JT, Saltzman E, McCullough ML, Troy LM, Jacques PF. The 2005 Dietary Guidelines for Americans Adherence Index: development and application. J Nutr. 2006;136: Dixon et al.