Dietary sources of five nutrients in ethnic groups represented in the Multiethnic Cohort

Methods

The MEC was established in Hawaii and Los Angeles County, California to investigate the associations between diet and cancer among five ethnic/racial groups in the USA: African-Americans (AfAm); Native Hawaiians (NH); Japanese-Americans (JpAm); Latinos (born in Mexico and Central/ South America – Latino-Mexico; born in the USA – Latino-US); Caucasians. Study design, recruitment procedures and baseline characteristics have been reported elsewhere⁽¹⁴⁾. In brief, more than 215 000 men and women aged 45–75 years representing the five ethnic/racial groups were enrolled into the MEC study between 1993 and 1996. Ethnicity was self-defined.

Data were collected using a mailed self-administered questionnaire that was specifically developed for the MEC study⁽¹⁴⁾. The twenty-six-page questionnaire included a seventeen-page quantitative FFQ (QFFQ) that collected data on the consumption of more than 180 food items over the past year. The QFFQ was developed from 3 d measured dietary records from sixty men and sixty women of each of the five ethnic/racial groups^(14,15). Foods that contributed to more than 85 % of the intake of fat, dietary fibre, vitamin A, carotenoids and vitamin C were listed on the QFFQ. Ethnic-specific foods were also included irrespective of their contribution to the diet. The QFFQ measured the amount of food consumed based on a choice of three portion sizes (represented in photographs) specific to each food item listed on the QFFQ and the usual intake frequency based on the categories ranging from ‘never or hardly ever’ to ‘2 or more times a day’. A sub-study was conducted to validate and calibrate the QFFQ using three repeated 24 h dietary recalls collected in each ethnic–sex group⁽¹⁵⁾. Average correlations for nutrient densities ranged from 0·57 to 0·74 across ethnic–sex strata, and those for specific nutrient groups were from 0·32 to 0·67 for proteins, 0·46 to 0·77 for total fat, 0·56 to 0·7 for saturated fats and 0·68 to 0·78 for fibre.

Energy and nutrient intakes were calculated based on all QFFQ items⁽¹⁶⁾; similar foods were combined to determine the percentage contribution of each source to the total daily intake of energy, total fat, saturated fat, protein, dietary fibre and added sugar. Participants excluded in the analysis were those with extreme mean energy intake (beyond 3 sd) and/or macronutrient intake (beyond 3·5 sd). As the focus of the present study was to examine nutrient sources among persons of different ethnicities, those who identified themselves being from a mixed ethnic background were also excluded. Latinos born in the Caribbean were also excluded due to the very small number of participants. A total of 186 916 participants were included in the present analysis.

The present study was conducted according to the guidelines laid down in the Declaration of Helsinki, and the protocol was approved by the Institutional Review Boards of the University of Hawaii and the University of Southern California.

Results

Table 1 displays the demographic information of the participants enrolled in the MEC. NH men and AfAm women had the highest average BMI among all ethnic–sex groups, while NH men and NH women reported the highest mean daily energy intakes.

The top ten dietary sources of energy contributed from 36·2 % (Latino-Mexico women) to 49·6 % (JpAm men) to total energy intake (Table 2). Across all ethnic–sex groups, the dietary sources of energy intake were similar with rice and bread contributing most, ranging from 11·4 % (Latino-Mexico women) to 27·8 % (JpAm men). However, the percentages of contribution of each source varied among the ethnic–sex groups. Chicken/turkey dishes and pasta with tomato sauce or cheese were in the top ten list for all groups. Cereal was among the top contributors to energy for all ethnic–sex groups (2·3–4·7 %), except Latino-Mexico men. Regular soda contributed between 2·2 % (Latino-Mexico men) and 3·8 % (AfAm men) to total energy intake. Interestingly, beer was a major source of energy intake among men of all ethnic/racial groups with the exception of Latino-Mexico; conversely, beer did not appear among the ten major sources of energy for women of any of the ethnic/racial groups.

The ten major dietary sources of fat are presented in Table 3 for all ethnic–sex groups. The top ten sources of total fat contributed 35·1 % (Latino-Mexico women) to 43 % (JpAm men) of total fat intake (Table 3). A top-three food source of fat across all ethnic–sex groups, chicken/turkey dishes, contributed from 4·1 % (Caucasian women) to 7·4 % (AfAm men) to total fat intake. Butter was a top-five food source among all groups except Latino-Mexico men, contributing between 3·6 % (Latino-Mexico women) and 6·2 % (NH women). Peanuts/other nuts also appeared among the top five major sources for men and women in four of the ethnic/racial groups (AfAm, NH, JpAm and Latino-US); the percentage contribution to total fat intake ranged from 4 % for Latino-US men to 5·9 % for JpAm men.

The sources of saturated fats are shown in Table 4. Ice cream, ice milk or frozen yogurt was among the top five contributors to total saturated fat intake for all groups, except Latino-Mexico women. Butter, likewise, was a significant food source of saturated fat for all but Latino-Mexico men, contributing from 4·4 % among Latino-Mexico women to 7 and 7·2 % for NH men and women, respectively.

Chicken/turkey dishes were a major contributor to total protein intake across all ethnic–sex groups, topping the list for men and women in four of the ethnic/racial groups (AfAm, NH, Latino-Mexico and Latino-US; Table 5). The first dietary source of protein for JpAm men and women was rice, contributing 12·7 and 10·4 % to protein intake, respectively. Fish was also among the top five major sources in four ethnic/racial groups (AfAm, NH, JpAm and Latino-US); the percentage contribution varied from 4·8 % for AfAm men to 7·7 % for JpAm men. Bread contributed from 4·6 % (Latino-Mexico women) to 7·6 % (Latino-Mexico men) to total protein intake and also appeared among the top five sources across all ethnic–sex groups.

Bread and cereals were the top two food sources of dietary fibre among all groups, except Latino-Mexico men, for whom beans replaced cereals as the second highest contributor (Table 6). Together, bread and cereals accounted for 18·1 % (NH women) to 22 % (Caucasian men). Other important dietary sources of fibre were orange, grapefruit or pomelo across all ethnic–sex groups and bananas for men and women in all groups, except AfAm. Regular sodas were the top contributor to added sugar intake, ranging from 17·5 % for JpAm women to 35·2 % for Latino-Mexico men to total consumption (Table 7). The percentage contribution of regular sodas alone is greater than that of the other four major food sources combined for all groups except JpAm and Caucasian women. Other fruit juices/drinks were also a major source of added sugar across all ethnic–sex groups.

Discussion

The present study describes similarities and differences between ethnic/racial groups both in the types of foods that are consumed as well as the percentage contribution of these food sources to total energy and specific nutrient intake. In the present study population, we identified rice, bread, chicken/turkey dishes, sodas and pasta/tomato sauce dishes as the five major sources of energy. In contrast, Cotton et al.⁽¹⁷⁾ identified ‘yeast bread’, ‘beef’, ‘cakes/ cookies/quick breads/doughnuts’, ‘soft drinks/soda’ and ‘milk’ as the top five contributors to total energy intake among US adults from 1994 to 1996⁽¹⁸⁾. The differences may be attributed to several factors, including the use of different dietary assessment instruments and different sampling methodologies. Cotton et al. analysed 24 h dietary recalls collected as part of the Continuing Surveys of Food Intakes by Individuals (CSFII), whereas the present study utilised a validated QFFQ that was specifically developed to characterise the dietary consumption patterns of our multiethnic study population. Furthermore, the CSFII used personal weighting factors to compensate for demographic differences between the sampled population and the US population in order to obtain a nationally representative sample⁽¹⁸⁾. By contrast, the present study population comprises the diverse ethnic/racial groups of Hawaii and Los Angeles County, and requires no personal weighting factors to achieve the targeted population sample.

Consumption patterns in the USA have evolved as diverse population groups introduce new foods, novel products appear on the market, and diet trends emerge and fade. The US Department of Health and Human Services and the USDA Dietary Guidelines for Americans 2010 are designed to help Americans make better food choices to promote health and to reduce the risk for chronic diseases⁽⁹⁾. While the dietary guidelines contain technical information intended for policymakers, educators, dietitians and other health practitioners, the recommendations are ultimately targeted for the general public to embrace a healthier pattern of eating. The present findings, therefore, may serve to assist nutrition professionals in identifying foods for intervention and translating the dietary guidelines for practical application for specific ethnic/racial minority groups.

The percentage of the US population claiming minority racial or ethnic heritage is growing^(19,20). The US population is becoming increasingly diverse with the nation's minority population reaching 102·5 million in 2007, or 34 % of the total⁽²¹⁾. It has been shown that ethnic/racial groups are adopting US dietary practices, which are typically high in fat and low in fruits and vegetables⁽²²⁾. This trend is particularly alarming when observed in combination with other lifestyle changes, such as decreased physical activity, which may also increase the risk for chronic disease⁽⁴⁾. In the present study, we found that the Latino-US group identified different dietary sources of energy and other nutrients compared with the foods consumed by the Latino-Mexico group. For example, while peanuts/other nuts were a top contributor to total fat intake for Latino-US, peanuts/other nuts did not appear among the major sources for Latino-Mexico. Instead, taco salad was a top-ten contributor to total fat among Latino-Mexico compared with Latino-US, for whom taco salad was not an important source. Ethnic differences in dietary sources, as described in the present study, demonstrate a need to consider cultural identities and unique customs when formulating and conveying public health messages and developing and implementing nutrition education and intervention programmes aimed at improving food choices among diverse populations.

While numerous ethnic differences in major dietary sources of energy, total fat, saturated fat, protein, dietary fibre and added sugar have been noted between ethnic/racial groups, an appreciable similarity exists in dietary consumption patterns, consistent among all groups represented in the MEC. Some similarities would be expected, given that the MEC participants were recruited primarily from residents living in two specific geographic areas of the USA. Of note, regular sodas were the primary source of added sugar across all ethnic–sex groups. Combined, regular sodas and other fruit juices/drinks contribute up to almost 50 % to total added sugar consumption. These findings present a tremendous opportunity for public health advocates to help reduce the intake of added sugar, specifically the consumption of sugar-sweetened beverages, which has been linked with an increased risk for a number of diet-related chronic diseases⁽²³⁾.

The strengths of the present study include the large sample size and the population-specific dietary assessment instrument used to collect data. The MEC QFFQ was specifically developed and validated to describe dietary consumption patterns for the five ethnic/racial groups represented in the present study population. The non-Caucasian sample sizes generated for the MEC are substantially larger than those that are available from other national surveys. The CSFII 1994–6 collected dietary intake data from 16 103 adults⁽¹⁸⁾, which included AfAm and Latinos, but the samples were comparatively small. Furthermore, no national survey has sampled populations in Hawaii, a state that comprises a relatively large number of JpAm and NH. In addition, with the exception of a slightly higher education among cohort participants, baseline characteristics were comparable with census data, supporting the generalisability of these results to the larger US population⁽¹⁴⁾.

Some limitations also warrant mention. Unfortunately, there were a relatively large number of exclusions due to missing data, and the proportion of excluded participants did vary somewhat for the different ethnic–sex groups, ranging from 15 % in Caucasian and JpAm women, to only 2·5 % for NH of both sexes. Although these differences do raise concern regarding introduction of selection bias, with the considerable sample sizes that were still maintained in these analyses, considerable dietary variation would have to have occurred in order to have an impact on these results. Recall bias is also a possible limitation; however, the QFFQ utilised in the MEC characterises total consumption relatively well as demonstrated in calibration and validation assessments^(14,15). Further, previous studies have demonstrated that FFQ are suitable tools for quantitative assessment of food and nutrient intake⁽²⁴⁾, although some have suggested that the validity may be higher among women⁽²⁵⁾, and could vary by food group⁽²⁶⁾. Another limitation is the collection of data over 15 years ago, thus more recent data would be useful to determine whether changes in the dietary patterns of specific ethnic groups over time may have had an impact on the generalisability of these results to the current population. Nevertheless, no other study on the dietary sources exists for a large sample of the five main ethnic/racial groups in the USA.