These findings reveal sex and ethnic differences in grain foods consumption. In particular, JpAm men had the highest contribution of refined grain food, while AfAm women had the highest contribution of whole grain food to total grain intake. Ethnic differences in the consumption of grain foods have been previously reported [38, 39]. For example, Champagne et al.  found that cereal consumption was lower in European Americans when compared with AfAm. Differences in grain consumption by sex have also been previously described by Maras et al. ; women consumed more whole grain and less refined grain products than men. In contrast, AfAm women have the highest rates of overweight/obesity in the US . Thus, it is important to evaluate ethnic-sex differences in food group consumption in light of the current obesity epidemic in the United States.
Major sources of whole grains in this study were whole wheat/rye bread, popcorn, cooked cereals, and wild rice. In the Baltimore Longitudinal Study of Aging, including 1,516 men and women (aged 62.1 ± 16.0 years), Maras et al.  found that top contributors of whole grains were ready-to-eat breakfast cereals, hot breakfast cereals, multi-grain bread, and whole wheat bread. A recent report of the 2001–2002 NHANES indicated that ready-to-eat cereals and hot cereals contributed more than 40% to whole grain consumption, followed by yeast breads, popcorn, and crackers . In our study, major contributors of refined grains were white rice and white bread, except for Latinos who primarily consumed Latino corn/tortilla/bread and rolls/buns/biscuits. Data from the 2001–2002 NHANES indicated that yeast breads contributed 26% to non-whole grains followed by grain-based desserts, pizza, pasta and pasta dishes, and Mexican dishes . However, most previous studies included mainly Caucasian populations and discrepancy between results could also be due to different methods, and geographical and cultural influences [42, 43].
Results of the present study also indicate that the contribution of refined grain foods exceeded that of whole grain products among all ethnic/racial groups, except for AfAm women who primarily consumed whole grain foods. Such trends were previously observed in the US [7, 44], the UK , and Mediterranean countries . Factors responsible for these trends include confusion about which foods contain whole grains as well as the assumption that whole grains are darker in color than refined grains even though color additives are often used to change the color of foods . Furthermore, many consumers associate whole grains with the coarse textures and flavors found in oat bran, wheat germ, and whole grain flours . According to Baker et al. , consumers may not purchase whole grain foods due to higher cost. Finally, it is important to note that although many refined grain products are enriched with niacin, riboflavin, thiamin, and iron, enrichment doesn’t restore insoluble fiber and other nutrients that are lost during the milling process . Our results underscore the need to encourage the consumption of whole grains instead of refined grains among ethnic/racial groups to reduce risk of chronic diseases.
Grain foods made an important contribution to the mean daily intakes of thiamin, riboflavin, niacin, pyridoxine, and folic acid. Previous studies have also demonstrated a positive impact of grain foods on intake of B vitamins [48, 49]. Breakfast cereal consumption, for example, was associated with an adequate level of micronutrient intake . Furthermore, some studies have shown that breakfast cereal consumers had higher intakes of thiamin, riboflavin, and folic acid compared with non-consumers [50, 51]. The high level of vitamin B fortification in many breakfast cereals may explain the high contribution of these foods to vitamin B intake observed in this study . Conversely, intake of refined grains, white rice and breads in particular, was relatively high among all ethnic-sex groups in this study, but the contribution of these foods to B vitamins was comparatively low. These findings are not surprising, given that the level of vitamin B fortification for breakfast cereals tends to be higher than that for rice [52, 53]. Additional fortification of refined grains which make a considerable contribution to this food group intake may present an opportunity to address nutritional deficiencies in some populations.
Intakes of thiamin, riboflavin, and niacin have been reported to be low among older men and women (60–80 years) in the lipid research clinics program prevalence study . In 2003–2006, 13% of non-Hispanic white women as well as Mexican American women had inadequate total folic acid intakes . Another previous study found that 13% of AfAm, 6% of Caucasian, and 3% of Hispanic women reported inadequate intakes of vitamin B6 . Results from this paper may potentially be used to help ease some of these concerns through the development of food-based ethnic-specific dietary guidelines. For example, recommendations based on increased consumption of whole wheat/rye bread could help increase thiamin, riboflavin, niacin, and vitamin B6 intakes, and would result in a diet with better nutrient quality. Results of this study underscore the need for both individualized nutrition interventions as well as environmental interventions to influence consumption patterns of these five ethnic groups.
Limitations of this study include possible recall bias. Also, measurement error is known to be higher with FFQs compared to other methods . Unfortunately, the methodology used to validate the accuracy of the QFFQ used in this study to capture nutrient intake did not assess B vitamins . While this limitation is a concern, the results from that study indicate that the tool is a valid instrument for assessment of several nutrients, and thus the findings from that study do support the application to a wider range of nutrients. The major strength of this paper lies in its large sample size and the use of a QFFQ developed and validated specifically for the multiethnic group to assure standardized data collection among the five ethnic/racial groups. A standardized food grouping methodology for grains and their subgroups was used and based on the USDA food groupings. Further, a comparison of the cohort distributions by educational levels and marital status with corresponding census data for the two geographic areas confirmed that the cohort was representative of the general population. However, the data available for this analyses were collected over 15 year ago, thus it is possible that changes in dietary sources or consumption patterns have occurred over time, which may have impacted the generalisability to current populations. Analyses of more recent data would be useful to clarify and confirm the finding from this study.