![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Cancer Etiology Program, Cancer Research Center of Hawaii, University of Hawaii, Honolulu, HI 96813 and * Department of Preventive Medicine, Keck School of Medicine, University of Southern California, Los Angeles, CA 90089
3To whom correspondence should be addressed. E-mail: spark{at}crch.hawaii.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT SUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
30 was strongly positively associated with the Fat and Meat pattern (OR = 2.14, 95% CI: 2.08–2.20, vs. BMI < 25). Current smokers showed a positive association with the Fat and Meat pattern (OR = 1.67, CI: 1.62–1.72, vs. nonsmokers) and inverse associations with the Vegetables (OR = 0.66, CI: 0.64–0.68) and Fruit and Milk patterns (OR = 0.53, CI: 0.52–0.55). Physical activity was positively associated with the Vegetables and Fruit and Milk patterns but not with the Fat and Meat pattern. These findings support the hypothesis that dietary patterns are influenced by interrelated sociocultural, demographic, and other lifestyle factors and may be useful in investigations of diet-disease relations.
KEY WORDS: • dietary patterns • factor analysis • food frequency questionnaire • ethnicity
Diet plays an important role in most of the chronic diseases that have been the largest cause of morbidity and mortality in the world (1,2). Although the role of individual dietary components has been a focus of considerable research, foods are consumed in many combinations that are likely to be complex. As a result, nutrient intakes are often highly correlated with each other and may have interactive and synergistic effects (3,4). Dietary pattern analysis, which reflects the complexity of dietary intake, has recently received greater attention from nutritional epidemiologists (5–7). Researchers have examined health outcomes in relation to dietary patterns created by a variety of approaches, including simple or complex scores to assess desirable dietary characteristics and statistical techniques (8). Among these approaches, factor analysis has been used to identify the number and nature of the underlying factors that are responsible for covariation in the data; an advantage of factor analysis is that the resulting factors are uncorrelated variables that often can account for most of the variability in the original data (9,10).
In dietary pattern analysis, foods are usually aggregated into food groups. Various schemes for food grouping have been used, which makes a comparison of the results from different studies difficult. There is inconsistency not only in the grouping method, but also in allocating foods that are part of mixed dishes (11). Food group servings developed by the USDA for the Food Guide Pyramid might be helpful in standardizing food grouping methods (12).
Ethnic groups in the United States have different dietary cultures and distinct rates of chronic disease (13,14). However, ethnic differences have rarely been the focus of the diet-disease studies using dietary pattern analysis. We defined dietary patterns among subjects in the Hawaii-Los Angeles Multiethnic Cohort Study, a large study of 5 ethnic groups (15). We also examined the relation of the patterns with nutrient intake, sociodemographic, and health-related characteristics.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Data collection.
The baseline questionnaire included various demographic, lifestyle, and medical history items as well as a quantitative food frequency questionnaire (QFFQ). Three-day measured food records were used to develop a single, self-administered QFFQ appropriate for all ethnic groups (15,16). The minimal set of foods contributing at least 85% of the intakes of fat, dietary fiber, vitamin A, carotenoids, and vitamin C were selected for each ethnic group. In addition, unique ethnic foods were included regardless of nutrient contribution. The questionnaire included 8 frequency categories for foods and 9 for beverages, with 3 choices of portion size to permit adequate specificity in defining daily intakes. For food items, the highest frequency category was 2 times/d, whereas for beverages the highest category was 4 times/d. As an aid to quantification, photographs showing selected foods in representative portion sizes were provided at the top of several pages of the questionnaire. A calibration substudy was conducted and showed satisfactory correlations between the QFFQ and three 24-h recalls for all ethnic and gender groups being studied (16).
Food groupings. Food Guide Pyramid servings were computed for each cohort member as follows. The Pyramid Servings Database file identified the number of servings from 30 food groups provided per 100 g for the wide variety of foods reported during the 1994–1996 Continuing Survey of Food Intake by Individuals (12). Foods that were mixtures were first disaggregated into their components and each ingredient was assigned to the appropriate Pyramid food group. The Cancer Research Center of Hawaii food composition table was linked to this database so Pyramid servings could be calculated for the basic foods in the table. Food Guide Pyramid servings of many traditional mixtures consumed in Hawaii were determined using local recipes. The daily number of Food Guide Pyramid servings was computed for each individual by summing the servings across the reported food items (11,17,18). The 30 Pyramid food groups include the 5 main food groups in the Food Guide Pyramid (grains, vegetables, fruits, dairy, and meat), 22 subgroups, and a further 3 groups that make up the Pyramid tip (added sugars, discretionary fat, and alcohol) (see supplemental Appendix A). Added sugars are defined as all sugars in the diet that have been added to foods during processing, preparation, or at the table. Discretionary fat is defined as all fat in the diet that could have been avoided by making lower fat choices; it includes all fat from dairy products, fat in all but the leanest meats, and all fat added to foods. In our data, a few food groups, notably soybean products, yogurt, and organ meats, had a high proportion of non- or very rare consumers (46, 40, and 22%, respectively). Because such high proportions of null values will violate the assumptions of the factor analysis, the servings of these 3 groups were combined with other food groups having a similar nutrient profile. Soybean products were combined with cooked dry beans and peas and nuts and seeds; yogurt with milk; and organ meats with meat. Thus, 21 food groups were available in the pattern analyses; however, 1 additional group, alcohol, had to be eliminated (see below), leaving a total of 20.
Statistical analysis. All analyses were performed with SAS software, version 8.2 (SAS Institute). Because the gender and age group distribution differed across ethnic groups, we adjusted sociodemographic and health-related characteristics by gender and age for descriptive analyses. This adjustment was completed by the poststratification method described by Rossi et al. (19), weighted to the gender and 10-y age group distribution of the entire cohort of 195,298. Because the distribution of the dietary variables was not normal, data were transformed using the Box-Cox transformation (see supplemental Appendix A for a list of the transformations). Two types of analyses were performed: 1) exploratory factor analysis [essential fatty acids (EFA)] to determine dietary patterns using principal components factor analysis and 2) confirmatory factor analysis (CFA) to validate the factor model. For the EFA, we used principal component factor analysis using the PROC FACTOR procedure to define dietary patterns among the study population, while PROC CALIS was used for the CFA (10).
First, the entire sample was divided randomly into 2 groups. EFA was performed with 1 group, and then the other group was used to confirm the model by CFA. We excluded 1 variable, alcohol, from the analysis because the normality assumption for factor analysis could not be met with transformation. Therefore, the remaining 20 food groups were used in EFA. Factors were retained based on the following criteria: eigenvalue >1.25, a scree plot, and the interpretability of the factors. An orthogonal rotation was then used to achieve a structure with independent factors. To increase the interpretability of the factor structure, the food groups not contributing to a pattern significantly were excluded based on the criterion that the factor loadings be >0.6. EFA was performed again using the remaining food groups to obtain the final factor model.
Second, CFA was performed on the second half of the sample to verify the final factor model found by EFA. By splitting the sample in half, we can determine whether the 3-factor model with the factor loadings found in the EFA described that data set specifically or whether it broadly describes patterns from the underlying population. The measurement model consisted of the identified patterns and the corresponding indicator variables. The process for determining whether a model fit the data involved review of overall goodness of fit statistics [the comparative fit index (CFI), the non-normed fit index (NNFI), and the root mean squared error approximation (RMSEA)] and a more detailed fit assessment (significance test for factor loadings). The chi-square test was not used, since with large sample sizes it can indicate that the model is not a good fit even if the model fits the data well. CFI and NNFI values >0.9 indicate that the model provides an acceptable fit. A RMSEA value <0.1 suggests an acceptable fit of the model. Verification of the importance of independent variables contributing to the factors was judged by the significance of the parameters, based on the t test, as well as visual inspection to ensure that factor loadings were relatively large (10). This process was repeated for the overall cohort and separately by ethnic-gender group.
Finally, we applied the final EFA model to the entire study population of 195,298 to obtain factor scores. Factor scores were created for each subject as the linear combination of the dietary variables, weighted by an equivalent of the factor loadings. The scores were then used to study the association between dietary patterns and the other variables.
Partial Pearson’s correlation coefficients adjusted for energy were calculated between factor scores and nutrient intakes. A pattern’s association with age, gender, ethnicity, and health characteristics was assessed using multivariate logistic regression analysis to model the odds of being in the upper half of the factor scores. The logistic regression was performed with 182,998 subjects with complete covariate information. The results were similar when a linear regression approach was also used to model the continuous factor scores. These results are not shown because the coefficients are difficult to interpret because the scale of the factor scores is not meaningful.
|
RESULTS |
|---|
|
TOPABSTRACTSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
|
When the factor analysis was stratified by ethnicity and gender, similar results were found. Three factors were found for each group, with loadings to that described from the overall model. The first factor that explained the most variance was Fat and Meat for all groups. The variability explained by the first factor varied from 27% for white women to 33% for Hawaiian women. Vegetables and Fruit and Milk were the second and the third factors for all groups. The total variability explained by the 3 factors was between 59% for white women and 69% for Hawaiian women.
Factor scores were computed for all cohort members using the overall model in Table 2, and associations with other factors were examined. All factors were positively correlated with total energy intake (Table 3). Therefore, correlations of other nutrients were adjusted for energy intake. The Fat and Meat pattern had the strongest correlation with energy, as might be expected, and showed a positive relation with fats and a negative relation with carbohydrates, vitamins, and dietary fiber in men and women. The correlation coefficients between pattern scores and the energy-adjusted nutrient intakes differed across 3 dietary patterns. The Vegetables pattern showed negative correlations with fat and positive correlations with vitamins and dietary fiber. The Fruit and Milk pattern had a positive correlation with calcium and iron intakes as well as with vitamin C and dietary fiber.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Several researchers have reported dietary patterns derived from FFQ in cohort studies using factor analysis. Although several distinct patterns were defined, 2 patterns were relatively dominant. One is a healthful or prudent or vegetable-fruit dietary pattern characterized by low fat and high fruit, vegetable, and whole grain intake. The other is a Western or traditional or red meat-starch pattern characterized by high fat, meat, and refined grain consumption. Additionally, "sweet," "drinker," and "Southern" patterns have been identified (1,9,24–27). Food intakes were represented by different methods in past factor analyses. Some studies used individual food items in the questionnaire for factor analysis. Others aggregated food items for the FFQ into food groups to reduce the number of variables considered and improve their distribution. Some authors did food groupings based on the similarity of nutrient profiles or culinary usage, while others used predefined food groups. The resulting number of food group variables input into factor analysis differed widely, from 26 to 49. We used a different scheme of aggregating foods, where groups were based on the Food Guide Pyramid groups and servings defined by the USDA. Although some mixed dishes may contribute to distinct dietary patterns (5), the USDA system groups basic foods and ingredients to define dietary patterns. The use of USDA food groupings may provide a common system for comparison of results from many different dietary pattern studies.
The first pattern found in our study, Fat and Meat, was characterized in other studies as the Western pattern. The second and third patterns, Vegetables and Fruit and Milk, respectively, appear consistent with a set of food items that has been labeled the "prudent" pattern. This pattern did not necessarily include dairy products, even low-fat dairy products, as ours did. Interestingly, in our study fruits and milk were extracted as 1 pattern but vegetables did not load with any dairy products. The Vegetables pattern showed relatively high loadings for fruits but small negative loadings for cheese and milk and yogurt. This result suggests that vegetables and fruits should not be aggregated into 1 food group. In this study, milk and yogurt were associated with fruits, resulting in separate patterns for fruits and vegetables. When we tried to extract only 2 factors instead of 3, fruits and vegetables were defined in the same pattern. But the loading of milk and yogurt weakened significantly and did not contribute to either factor. Additionally this 2-factor model did not show an acceptable goodness of fit of the data when modeled by CFA.
Initially, EFA based on 20 food groups was performed separately for the 5 ethnic groups and for women and men, resulting in 10 ethnic-gender groups (see supplemental Appendix A). Other starchy vegetables loaded in Factor 2 for 7 ethnic-gender groups, but did not make it into the overall model (loading = 0.59). Whole grains loaded for 6 groups and added sugars for 2 groups in Factor 3, although they did not load in the overall model (loading of 0.59 and 0.46, respectively). Therefore, the pattern for Factor 3, Fruit and Milk in the final factor model, may be a breakfast pattern. EFA, with 13 food groups, showed almost identical patterns across 10 ethnic-gender groups, although factor loadings for the food groups varied slightly (see supplemental Appendix B). Additionally, CFA proved that the indices for the goodness of fit of the data to the model were acceptable for all groups. Therefore, we applied the 3-factor solution to the whole study population rather than using separate factors for each gender and ethnic group. This may result in a decrease in the accuracy of the measurement of ethnic- or gender-specific dietary patterns. However, using the same weightings allowed for comparison of the differences in mean scores between men and women and among ethnicities.
Previous studies have reported that women tended to have generally higher scores on healthy dietary patterns (28–32). Similarly, our results showed that men had much higher scores for the Fat ant Meat pattern and slightly lower scores on the Vegetables and Fruit and Milk patterns than women. We also found differences in pattern scores across ethnicities. Hawaiians and Latinos had high scores for the Fat and Meat pattern, but the differences in the Vegetables and Fruit and Milk patterns across ethnicities were much larger. Japanese Americans showed the highest score on the Vegetables pattern, whereas they showed the lowest score on the Fruit and Milk pattern. A few studies have compared dietary patterns among different ethnicities. Kerver et al. (32) reported that African Americans and Mexican Americans tended to have lower scores for an "American-healthy" pattern than did whites, whereas North and Emmett (33) reported that more non-white children showed healthy patterns than white children in the United Kingdom. In contrast, Bell et al. (34) found no significant differences among whites, African Americans, and Native Americans.
Our observation that the 3 dietary patterns are differently related to age and health characteristics is consistent with earlier studies. Age showed a negative association with a Western or red meat-starch pattern and a positive association with vegetable-fruit patterns in other studies as well as ours (1,9,21). More educated people showed higher scores for the Vegetables and Fruit and Milk patterns and lower scores for the Fat and Meat pattern, which was also observed in other studies (9,21).
In our study, the Fat and Meat pattern was positively associated with BMI, smoking, and alcohol consumption but negatively associated with dietary supplement use, whereas Vegetables and Fruit and Milk patterns were negatively associated with smoking but positively associated with physical activity and supplement use. Similar findings have been reported in previous studies (1,9,21,22). Although our 2 healthy patterns showed similar relations with covariates, only the Fruit and Milk pattern was negatively associated with BMI and alcohol consumption. Thus the Fruit and Milk pattern was more strongly associated with healthy lifestyles than the Vegetables pattern.
We did not find any effect of baseline personal or family history of cancer on dietary patterns. However, other studies have found an association between dietary patterns and later risk of cancer, although the findings are inconsistent (20,25,35–40). Further investigations of the association of baseline dietary patterns with subsequent cancer incidence are planned for the Multiethnic Cohort Study.
There are several strengths to this study. A very large multiethnic population was used to determine the patterns, and we were able to perform both EFA and CFA. Several methodological concerns have been pointed out in dietary pattern analysis such as the food grouping scheme, quantifying and transforming input variables, and deriving patterns separately for gender and ethnicity (6,7). We were able to overcome many of these problems in our study.
However, there are still some limitations. The Food Guide Pyramid was developed to focus on Americans’ diets; therefore this food grouping scheme might not be appropriate in non-U.S. populations. Also, the study sample for the Multiethnic Cohort is from Hawaii and California, which may not be nationally representative. Furthermore, the statistics for the goodness of fit from CFA were acceptable but not excellent on the basis of the recommendation that CFI and NNFI should both exceed 0.9 and, ideally, be equal to 1.0 (10). Schulze et al. (41) reported a good fit of the 2-factor model using only 8 food groups. In factor analysis, a good fit is represented by variables having a relatively high factor loading on only 1 factor and near zero loadings on the other factors. However, in pattern analysis of foods, most researchers have over 20 food group variables, many of which have relatively high loading on more than 1 factor. Thus it may be difficult to get an excellent fit of the model from CFA in dietary pattern analysis.
In conclusion, 3 factors explained 63.5% of the variability in diet in the Multiethnic Cohort, and these same factors were found in each ethnic-gender group. Our data also show that dietary patterns were associated with age, gender, ethnicity, education, and health behavior in this large multiethnic population. These findings support the hypothesis that dietary patterns are influenced by interrelated sociocultural, demographic, and other lifestyle factors, which may be important for the implementation of programs to improve public health through dietary modification. In future analyses of dietary patterns and disease associations in the Multiethnic Cohort and other studies, these variables must be considered potential confounding factors.
|
FOOTNOTES |
|---|
2 The stratified analyses by ethnic-gender group are included in the Appendix tables available with the online posting of this paper at www.nutrition.org.
4 Abbreviations used: CFA, confirmatory factor analysis; CFI, comparative fit index; CRCH, Cancer Research Center of Hawaii; CSFII, 1994–1996 Continuing Survey of Food Intake by Individuals; EFA, exploratory factor analysis; NNFI, nonnormed fit index; QFFQ, quantitative FFQ; RMSEA; root mean squared error approximation; RSD, robust standard deviation.
Manuscript received 4 November 2004. Initial review completed 30 November 2004. Revision accepted 31 December 2004.
|
LITERATURE CITED |
|---|
|
TOPABSTRACTSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
1. Sánchez-Villegas, A., Delgado-Rodríguez, M., Martínez-González, M. Á. & de Irala-Estévez, J. (2003) Gender, age, socio-demographic and lifestyle factors associated with major dietary patterns in the Spanish Project SUN (Seguimiento Universidad de Navarra). Eur. J. Clin. Nutr. 57:285-292.[Medline]
2. Yach, D., Hawkes, C., Gould, C. L. & Hofman, K. J. (2004) The global burden of chronic diseases: overcoming impediments to prevention and control. J. Am. Med. Assoc. 291:2616-2622.
3. Mishra, G., Ball, K., Arbuckle, J. & Crawford, D. (2002) Dietary patterns of Australian adults and their association with socioeconomic status: results from the 1995 National Nutrition Survey. Eur. J. Clin. Nutr. 56:687-693.[Medline]
4. Jacobs, D. R., Jr & Steffen, L. M. (2003) Nutrients, foods, and dietary patterns as exposures in research: a framework for food synergy. Am. J. Clin. Nutr. 78:508S-513S.
5. Hu, F. B. (2002) Dietary pattern analysis: a new direction in nutritional epidemiology. Curr. Opin. Lipidol. 13:3-9.[Medline]
6. Kant, A. K. (2004) Dietary patterns and health outcomes. J. Am. Diet. Assoc. 104:615-635.[Medline]
7. Newby, P. K. & Tucker, K. L. (2004) Empirically derived eating patterns using factor or cluster analysis: a review. Nutr. Rev. 62:177-203.[Medline]
8. Kant, A. K., Graubard, B. I. & Schatzkin, A. (2004) Dietary patterns predict mortality in a national cohort: the National Health Interview Surveys, 1987 and 1992. J. Nutr. 134:1793-1799.
9. Schulze, M. B., Hoffmann, K., Kroke, A. & Boeing, H. (2001) Dietary patterns and their association with food and nutrient intake in the European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam study. Br. J. Nutr. 85:363-373.[Medline]
10. Hatcher, L. (1994) A step-by-step approach to using SAS for factor analysis and structural equation modeling 1994 SAS Institute, Inc Cary, NC.
11. Sharma, S., Murphy, S. P., Wilkens, L. R., Au, D., Shen, L. & Kolonel, L. N. (2003) Extending a multiethnic food composition table to include standardized food group servings. J. Food Compost. Anal. 16:485-495.
12. Cook, A. & Friday, J. E. (2002) Documentation: Pyramid Servings database for USDA survey food codes [Online] 2002 U.S. Department of Agriculture http://www.ba.ars.usda.gov/cnrg/services/toc.html [accessed October 4, 2004].
13. Brown, L. K. & Mussell, K. (1984) Ethnic and regional foodways in the United States 1984 University of Tennessee Press Knoxville, TN.
14. National Center for Heath Statistics (2004) Summary Health Statistics for U.S. Adults: National Health Interview Survey, 2002. Vital and Health Statistics, Series 10, No. 222 [Online] 2004 Department of Health and Human Services Publication no. (PHS) 2004–1550. http://www.cdc.gov/nchs/data/series/sr_10/sr10_222acc.pdf [accessed October 4, 2004].
15. Kolonel, L. N., Henderson, B. E., Hankin, J. H., Nomura, A. M., Wilkens, L. R., Pike, M. C., Stram, D. O., Monroe, K. R., Earle, M. E. & Nagamine, F. S. (2000) A multiethnic cohort in Hawaii and Los Angeles: baseline characteristics. Am. J. Epidemiol. 151:346-357.
16. Stram, D. O., Hankin, J. H., Wilkens, L. R., Pike, M. C., Monroe, K. R., Park, S., Henderson, B. E., Nomura, A. M., Earle, M. E., Nagamine, F. S. & Kolonel, L. N. (2000) Calibration of the dietary questionnaire for a multiethnic cohort in Hawaii and Los Angeles. Am. J. Epidemiol. 151:358-370.
17. Sharma, S., Murphy, S. P., Wilkens, L. R., Shen, L., Hankin, J. H., Henderson, B. & Kolonel, L. N. (2003) Adherence to the Food Guide Pyramid recommendations among Japanese Americans, Native Hawaiians, and whites: results from the Multiethnic Cohort Study. J. Am. Diet. Assoc. 103:1195-1198.[Medline]
18. Sharma, S., Murphy, S. P., Wilkens, L. R., Shen, L., Hankin, J. H., Monroe, K. R., Henderson, B. & Kolonel, L. N. (2004) Adherence to the Food Guide Pyramid Recommendations among African Americans and Latinos: results from the Multiethnic Cohort Study. J. Am. Diet. Assoc. 104:1873-1877.[Medline]
19. Rossi, P. H. Wright, J. D. Anderson, A. B. eds. Handbook of survey research (quantitative studies in social relations) 1983 Academic Press San Diego, CA. .
20. Kim, M. Y., Sasaki, S., Sasazuki, S. & Tsugane, S. (2004) Prospective study of three major dietary patterns and risk of gastric cancer in Japan. Int. J. Cancer. 110:435-442.[Medline]
21. Tseng, M. & DeVellis, R. F. (2001) Fundamental dietary patterns and their correlates among US whites. J. Am. Diet. Assoc. 101:929-932.[Medline]
22. Slattery, M. L., Boucher, K. M., Caan, B. J., Potter, J. D. & Ma, K. N. (1998) Eating patterns and risk of colon cancer. Am. J. Epidemiol. 148:4-16.
23. Sichieri, R. (2002) Dietary patterns and their associations with obesity in the Brazilian city of Rio de Janeiro. Obes. Res. 10:42-48.[Medline]
24. Hu, F. B., Rimm, E., Smith-Warner, S. A., Feskanich, D., Stampfer, M. J., Ascherio, A., Sampson, L. & Willett, W. C. (1999) Reproducibility and validity of dietary patterns assessed with a food-frequency questionnaire. Am. J. Clin. Nutr. 69:243-249.
25. Tseng, M., Breslow, R. A., DeVellis, R. F. & Ziegler, R. G. (2004) Dietary patterns and prostate cancer risk in the National Health and Nutrition Examination Survey Epidemiological Follow-up Study cohort. Cancer Epidemiol. Biomarkers Prev. 13:71-77.
26. Khani, B. R., Ye, W., Terry, P. & Wolk, A. (2004) Reproducibility and validity of major dietary patterns among Swedish women assessed with a food-frequency questionnaire. J. Nutr. 134:1541-1545.
27. Togo, P., Heitmann, B. L., Sørensen, T. I. & Osler, M. (2003) Consistency of food intake factors by different dietary assessment methods and population groups. Br. J. Nutr. 90:667-678.[Medline]
28. Williams, D. E., Prevost, A. T., Whichelow, M. J., Cox, B. D., Day, N. E. & Wareham, N. J. (2000) A cross-sectional study of dietary patterns with glucose intolerance and other features of the metabolic syndrome. Br. J. Nutr. 83:257-266.[Medline]
29. Gex-Fabry, M., Raymond, L. & Jeanneret, O. (1988) Multivariate analysis of dietary patterns in 939 Swiss adults: sociodemographic parameters and alcohol consumption profiles. Int. J. Epidemiol. 17:548-555.
30. Barker, M. E., McClean, S. I., Thompson, K. A. & Reid, N. G. (1990) Dietary behaviors and sociocultural demographics in Northern Ireland. Br. J. Nutr. 64:319-329.[Medline]
31. Osler, M., Helms Andreasen, A., Heitmann, B., Høidrup, S., Gerdes, U., Mørch Jørgensen, L. & Schroll, M. (2002) Food intake patterns and risk of coronary heart disease: a prospective cohort study examining the use of traditional scoring techniques. Eur. J. Clin. Nutr. 56:568-574.[Medline]
32. Kerver, J. M., Yang, E. J., Bianchi, L. & Song, W. O. (2003) Dietary patterns associated with risk factors for cardiovascular disease in healthy US adults. Am. J. Clin. Nutr. 78:1103-1110.
33. North, K. & Emmett, P. (2000) Mutivariate analysis of diet among three-year-old children and associations with socio-demographic charateristics. Eur. J. Clin. Nutr. 54:73-80.[Medline]
34. Bell, R. A., Quandt, S. A., Vitolins, M. Z. & Arcury, T. A. (2003) Dietary patterns of older adults in a rural, tri-ethnic community: a factor analysis approach. Nutr. Res. 23:1379-1390.
35. Fung, T., Hu, F. B., Fuchs, C., Giovannucci, E., Hunter, D. J., Stampfer, M. J., Colditz, G. A. & Willett, W. C. (2003) Major dietary patterns and the risk of colorectal cancer in women. Arch. Intern. Med. 163:309-314.
36. Sieri, S., Krogh, V., Pala, V., Muti, P., Micheli, A., Evangelista, A., Tagliabue, G. & Berrino, F. (2004) Dietary patterns and risk of breast cancer in the ORDET Cohort. Cancer Epidemiol. Biomarkers Prev. 13:567-572.
37. Dixon, L. B., Balder, H. F., Virtanen, M. J., Rashidkhani, B., Männistö, S., Krogh, V., van Den Brandt, P. A., Hartman, A. M., Pietinen, P., Tan, F., Virtamo, J., Wolk, A. & Goldbohm, R. A. (2004) Dietary patterns associated with colon and rectal cancer: results from the Dietary Patterns and Cancer (DIETSCAN) Project. Am. J. Clin. Nutr. 80:1003-1011.
38. Terry, P., Hu, F. B., Hansen, H. & Wolk, A. (2001) Prospective study of major dietary patterns and colorectal cancer risk in women. Am. J. Epidemiol. 154:1143-1149.
39. Terry, P., Suzuki, R., Hu, F. B. & Wolk, A. (2001) A prospective study of major dietary patterns and the risk of breast cancer. Cancer Epidemiol. Biomarkers Prev. 10:1281-1285.
40. Masaki, M., Sugimori, H., Nakamura, K. & Tadera, M. (2003) Dietary patterns and stomach cancer among middle-aged male workers in Tokyo. Asian Pac. J. Cancer Prev. 4:61 (abs.).[Medline]
41. Schulze, M. B., Hoffmann, K., Kroke, A. & Boeing, H. (2003) Risk of hypertension among women in the EPIC-Potsdam Study: comparison of relative risk estimates for exploratory and hypothesis-oriented dietary patterns. Am. J. Epidemiol. 158:365-373.
This article has been cited by other articles:
|
|
 |
|
  C. D. Williams, J. A. Satia, L. S. Adair, J. Stevens, J. Galanko, T. O. Keku, and R. S. Sandler Dietary Patterns, Food Groups, and Rectal Cancer Risk in Whites and African-Americans Cancer Epidemiol. Biomarkers Prev., May 1, 2009; 18(5): 1552 - 1561. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. K. Gao, A. L. Fitzpatrick, B. Psaty, R. Jiang, W. Post, J. Cutler, and M. L. Maciejewski Suboptimal Nutritional Intake for Hypertension Control in 4 Ethnic Groups Arch Intern Med, April 13, 2009; 169(7): 702 - 707. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Darmon and A. Drewnowski Does social class predict diet quality? Am. J. Clinical Nutrition, May 1, 2008; 87(5): 1107 - 1117. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Esmaillzadeh, M. Kimiagar, Y. Mehrabi, L. Azadbakht, F. B Hu, and W. C Willett Dietary patterns, insulin resistance, and prevalence of the metabolic syndrome in women Am. J. Clinical Nutrition, March 1, 2007; 85(3): 910 - 918. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. A. McNaughton, G. D. Mishra, A. M. Stephen, and M. E. J. Wadsworth Dietary Patterns Throughout Adult Life Are Associated with Body Mass Index, Waist Circumference, Blood Pressure, and Red Cell Folate J. Nutr., January 1, 2007; 137(1): 99 - 105. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L Hutton, L. Martin, C. J Field, W. V Wismer, E. D Bruera, S. M Watanabe, and V. E Baracos Dietary patterns in patients with advanced cancer: implications for anorexia-cachexia therapy. Am. J. Clinical Nutrition, November 1, 2006; 84(5): 1163 - 1170. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. C. Howarth, S. P. Murphy, L. R. Wilkens, J. H. Hankin, and L. N. Kolonel Dietary Energy Density Is Associated with Overweight Status among 5 Ethnic Groups in the Multiethnic Cohort Study J. Nutr., August 1, 2006; 136(8): 2243 - 2248. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. J. Loos, T. Rankinen, T. Rice, D. Rao, A. S Leon, J. S Skinner, C. Bouchard, and G. Argyropoulos Two ethnic-specific polymorphisms in the human Agouti-related protein gene are associated with macronutrient intake Am. J. Clinical Nutrition, November 1, 2005; 82(5): 1097 - 1101. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
