ABSTRACT
Background
Whole-grain (WG) foods are defined by the Dietary Guidelines for Americans (DGA), FDA, AHA, American Association of Cereal Chemists International (AACCI), and Whole Grains Council (WGC) in different ways with diverse focuses on grain components only, whole foods, or nutrient contents.
Objectives
We aimed to compare estimated WG food intake among US adults using different definitions.
Methods
For each definition, we estimated the mean intake and trends of WG food consumption using survey-weighted 24-h dietary recalls from nationally representative samples of 39,755 US adults aged 20+ y from 8 cycles (2003–2018) of the NHANES. This is an observational study that used deidentified and publicly available datasets.
Results
The estimated mean consumption of WG foods (ounces equivalents/2000 kcal/d, oz. eq./d) varied by definition. In 2017–2018, the AHA (mean [SEM]: 1.05 [0.07] oz. eq./d) and WGC (0.95 [0.07]) definitions yielded the highest amounts, followed by the DGA (0.81 [0.06]), AACCI (0.73 [0.05]), and FDA (0.53 [0.04]). Using all definitions except for WGC, US adults increased WG food intake from 2003–2004 to 2017–2018 with the largest increase (61.5%) using the AHA (from 0.65 to 1.05 oz. eq./d), followed by DGA (0.50 to 0.81) and AACCI (0.51 to 0.73) definitions. For each definition, the main sources of WG foods consumed by US adults were ready-to-eat cereals, cooked grains and cereals, and breads (including rolls and tortillas). For all definitions except the AHA, non-Hispanic White adults and individuals with college degrees or above consumed higher levels of WG foods than non-Hispanic Blacks and those with lower levels of education.
Conclusions
Different definitions affect the determination of WG foods, estimated intakes, and associated trends in WG food consumption among US adults. These findings call for a standardized definition of WG foods to guide consumers, industry, and policymakers in promoting WG intake in the US.
Clinical Trial Registration: Not Applicable.
Keywords: whole-grain food definition, whole-grain food consumption, trends, food sources of whole grains, NHANES
See corresponding editorial on page 1470.
Introduction
A grain is considered to be a whole grain (WG) when it contains all 3 parts of the original kernel – the bran, germ, and endosperm (1). Despite a widespread agreement on the definition of WGs, a universally accepted standard definition of WG foods does not exist. In the US, government regulatory authorities, nonprofit organizations, and food industries define WG foods in different ways (2, 3). For example, the 2015–2020 Dietary Guidelines for Americans (DGA) defines WG foods as foods containing ≥50% (by weight) of the grain- or flour-containing component as WG ingredients (4, 5). The AHA uses a definition of WG foods as grain-rich foods having ≥1.1 g of fiber per 10 g of carbohydrates (6), and the industry-sponsored basic Whole Grain Stamp defines WG foods as foods containing ≥8 g of WG ingredients per labeled serving (7).
Despite strong evidence to support the health benefits of consuming WGs, consumer guidance, industry reformulations, and policy actions to promote WG intake may be limited by the lack of a standard definition of WG foods. We have previously shown that 40–50% of the respondents from a national survey panel could not correctly distinguish between foods that contained all or more than half of the grains as WGs and those that contained a little amount of WGs (8). For industry, different definitions of WG foods make it challenging to reformulate products and increase the availability of WG foods in the market (2, 9). Inconsistency in defining a WG product can also create barriers for governments to formulate policies to promote WG intake (2, 3). For instance, lobbying on the WG-specific standards of the National School Lunch Program substantially delayed the full implementation of this intervention and led to the rolling back of healthier criteria (10, 11).
To inform the formulation of a universally accepted definition of WG foods, it is essential to understand how different definitions may affect the estimated intake, trends, and food sources in the US population. In this study, we assessed WG food consumption and associated trends among US adults from 2003 to 2018 based on different definitions of WG foods. We further examined whether the WG food intake differed among population subgroups, as well as the types of foods that contributed the most to WG food intake in Americans.
Methods
Study design and population
We included adults aged 20 y and older who had a valid first-day 24-h diet recall across 8 2-y cycles (2003–2018) of NHANES (Supplemental Figure 1). NHANES is a nationally representative, cross-sectional survey conducted by the NCHS of the CDC (12). NHANES collects information on demographic, socioeconomic, dietary, and health-related data of the noninstitutionalized civilian subjects residing in the US. This survey program was approved by the research Ethics Review Board of NCHS, and all participants provided written informed consent (13).
Dietary assessment method
NHANES has documented the protocol and data collection methods (12). Briefly, trained interviewers conducted 24-h dietary recalls using the Automated Multiple Pass Method (AMPM) to record all foods and beverages consumed by the NHANES participants during the previous 24 h. One 24-h diet recall was conducted in person in the Mobile Examination Center (MEC). The NHANES interviewers and diet recall participants were monitored with established criteria to evaluate data acceptability. Food and nutrient intake were determined using the USDA Food and Nutrient Database for Dietary Studies (FNDDS) which contains information on What We Eat in America (WWEIA) food categories, food weights, ingredient lists, ingredient weights, and nutrient values (14). The Food Patterns Equivalents Database (FPED) and MyPyramid Equivalents Database (MPED) translated the foods and beverages in the FNDDS into 37 Food Pattern components (NHANES 2005–2018) and 32 MyPyramid food groups (NHANES 2003–2004), which includes the amounts (ounces equivalents) of total grains, WGs, and refined grains per 100 g of each food (15).
WG food definitions
Our analysis applied 5 different nationally used definitions for WG foods, specifically: 1) the DGA defines a WG food as food containing ≥50% of the total grain weight as WG ingredients (4); 2) the FDA defines a WG food as food containing 51% or more of the reference amount customarily consumed (RACC) food weight as WG ingredients (16); 3) the AHA defines WG foods as grain-rich foods having ≥1.1 g of fiber per 10 g of carbohydrates (6); 4) American Association of Cereal Chemists International (AACCI) specifies that a WG food must contain ≥8 g of WGs per 30 g of the product (17); and 5) the Whole Grain Council (WGC) permits a product to receive a Basic Whole Grain Stamp if it contains ≥8 g of WG ingredients per labeled serving (Table 1 and Supplemental Methods) (7).
TABLE 1.
Whole-grain food definitions used by industry, advocacy, and regulatory groups in the US
| Type of grain- or flour-containing foods | |||
|---|---|---|---|
| Noncombination foods | Combination foods (mixed dishes) | ||
| Agency/organization | Total weight | Grain or flour component weight | |
| 1) DGA1 | ≥50% of the grain component are whole grains | ≥50% of the grain component are whole grains | |
| 2) FDA2 | ≥51% whole grains by weight per RACC3 | ≥51% whole grains by weight per RACC3 | |
| 3) AHA | CHO: fiber <10:1 | CHO: fiber <10:1 | |
| 4) AACCI2 | ≥8 g whole grains per 30 g food | ≥8 g whole grains per 30 g food | |
| 5) WGC4 | ≥8 g whole grains per RACC | ≥8 g whole grains per RACC | |
AACCI, American Association of Cereal Chemists International; DGA, Dietary Guidelines for Americans; FNDDS, Food and Nutrient Database for Dietary Studies; FPED, Food Patterns Equivalents Database; RACC, reference amount customarily consumed; WGC, Whole Grain Council.
If the only grains the foods contain are whole grains, they were defined as 100% whole-grain foods.
Single-ingredient whole-grain foods (e.g. cooked brown rice, cooked oatmeal etc.) may not meet the criteria by definition but they are whole-grain foods by nature. Therefore, we determined whether a single-ingredient grain- or flour-based food is a whole-grain food based on its FNDDS food description and the corresponding FPED amounts of whole grains and total grains.
RACCs for specified product categories were established by the FDA for manufactures to determine the required labeled serving size for foods and beverages commonly consumed in the US. The categories in the RACC tables are generally considered to be broad descriptions of the types of products that are assigned to the applicable RACC.
If a product bears the 100% Stamp, all grain ingredients in the food are whole grains, and there is a minimum requirement of 16 g of whole grains per labeled serving. If a product bears the 50%+ Stamp, at least half of its grain ingredients are whole grain, and there is a minimum requirement of 8 g of whole grain per labeled serving. If a product bears a basic Stamp, it contains ≥8 g of whole grain but may contain more refined grain than whole. Labeled servings for foods and beverages are not provided in FNDDS or FPED databases. We used RACCs to approximate the labeled servings given RACCs for specified products are established by FDA for manufactures to determine the required labeled serving size.
Based on these definitions, we first identified WG foods in each cycle-specific FNDDS database and then estimated the consumption of WG foods for each NHANES participant (Supplemental Figure 2). We reviewed all foods and corresponding ingredients in FNDDS, and categorized foods that contain any flour or grain ingredients into 1 of the 4 groups according to the food and ingredient descriptions and the associated WWEIA codes: 1) Flour-based noncombination foods (e.g. breads, cookies, tortillas, etc.), 2) grain-based noncombination foods (e.g. cereals, nutrition bars, cooked rice, etc.), 3) flour-based mixed dishes (e.g. pizza, burgers, sandwiches, etc.), and 4) grain-based mixed dishes (e.g. dumplings, sushi, stir-fries, etc.) (Supplemental Table 1) (18). During this process, 2 investigators initially performed the classification by independently reviewing the foods and ingredients and assigning each food/beverage into 1 of the 4 groups, with discrepancies resolved through multiple rounds of discussions with the senior investigators. We then linked the FNDDS database to the FPED/MPED databases to quantify the amounts of total grains, refined grains, WGs, fiber, and carbohydrates per 100 g of the food and per 100 g of each ingredient of the food. By referring to the FDA's guidance on food serving sizes and the USDA FoodData Central system that provides FNDDS-linked food serving sizes (19, 20), we determined the RACC value for each grain- or flour-containing food item. Following the 5 definitions, foods were classified as WG foods, refined-grain foods which are grain- or flour-containing foods but do not meet the criteria of being WG foods, and foods that do not contain grains or flours. We further categorized WG foods and refined-grain foods into 6 subgroups: 1) Breads (including rolls and tortillas), 2) snack/meal bars (including quick breads), 3) ready-to-eat cereals, 4) cooked grains and cereals, 5) savory snacks/crackers, and 6) mixed dishes (including sweet snacks and sweets, among others) (Supplemental Table 2). We subsequently estimated the mean consumption of WG foods, refined-grain foods, and subcategories.
Statistical analysis
The primary outcome was the mean consumption of WG foods assessed using each definition among total and socioeconomic population subgroups in NHANES from 2003–2004 to 2017–2018. Population subgroups were defined by age (20–34 y, 35–49 y, 50–64 y, 65 y and older), sex, race/ethnicity (non-Hispanic White, non-Hispanic Black, and Hispanic), education levels [<high school graduate, high school graduate or general equivalency diploma (GED), some college or Associate of Arts (AA) degree, and college graduate or above], and the ratio of family income to the federal poverty levels (<1.30, 1.30–1.84, 1.85–2.99, and ≥3.00). Secondary outcomes included the percentage of WG foods among total grain- or flour-containing foods consumed and the mean intake of WG food subgroups.
The mean consumption of WG foods was estimated in ounces equivalents per 2000 kcal per d (oz. eq./d), with energy adjustment using the residual method to reduce measurement error and day-to-day variation in dietary intake. To assess the percentage of WG foods among total grain- or flour-containing foods consumed, we computed the mean ratio of WG food intake to total grain- or flour-containing food intake among all participants. WG foods contributing the most to the least to WG intake among Americans by different definitions were ranked based on the mean intake of WG food subgroups. Polynomial regression was performed to test the nonlinearity of the trends in WG food consumption by including a continuous, a quadratic, and/or a cubic term of the survey cycle. If the trend was determined as linear, the significance of trends was assessed by treating the survey cycle as a continuous variable in a linear regression model. If nonlinearity was indicated, the NCI's Joinpoint software was used to locate the joinpoint(s) and piecewise regression was performed to examine trends in the linear segments. Overall changes in intake were calculated as the difference in mean intake between the first (2003–2004) and last (2017–2018) cycles. Segmental changes were computed as the difference in mean intake between the first and last cycle in each segment. To assess statistical heterogeneity of changes by subgroups, the Wald test was used to evaluate an interaction term between the survey cycle and categorical variables (age, sex, race/ethnicity) or ordinal variables (education and income levels). Sensitivity analyses were conducted after adjusting for age, sex, race/ethnicity, education, and income level within each cycle to assess whether the observed changes are affected by demographic factors. To ensure nationally representative estimates, all analyses incorporated the appropriate survey weights and account for complex survey design. All analyses were conducted in SAS (version 9.4) and a P value <0.05 was considered statistically significant.
Results
Population characteristics
After excluding individuals aged younger than 20 y (n = 35,522) and individuals with invalid diet recall data (n = 5,035), a total of 39,755 US adults aged 20 y and older (weighted mean age ± SEM, 47.3 ± 0.22 y; 51.9% females; 67.7% non-Hispanic white individuals) were included in this analysis (Table 2). From 2003 to 2018, the proportion of older adults (age ≥65 y) increased from 17.7% to 20.6%, the adults who are non-Hispanic white decreased from 73.0% to 62.0%, and the proportion of respondents with college degrees or above increased from 23.3% to 30.2%.
TABLE 2.
Characteristics of US adults aged 20+ y, NHANES 2003–20181
| No. of participants (weighted %) | ||||||||
|---|---|---|---|---|---|---|---|---|
| Characteristics | 2003–2004 (n = 4448) | 2005–2006 (n = 4520) | 2007–2008 (n = 5419) | 2009–2010 (n = 5762) | 2011–2012 (n = 4801) | 2013–2014 (n = 5047) | 2015–2016 (n = 5017) | 2017–2018 (n = 4741) |
| Age group, y | ||||||||
| 20–34 | 1207 (29.5) | 1378 (27.8) | 1260 (27.7) | 1442 (27.9) | 1317 (27.7) | 1321 (27.8) | 1273 (27.8) | 1079 (27.7) |
| 35–49 | 1044 (28.8) | 1141 (30.1) | 1368 (30.1) | 1522 (28.9) | 1178 (26.6) | 1322 (26.1) | 1244 (24.8) | 1057 (24.2) |
| 50–64 | 901 (24.0) | 953 (24.1) | 1395 (25.6) | 1419 (25.8) | 1274 (28.1) | 1298 (27.4) | 1290 (26.3) | 1384 (27.4) |
| ≥65 | 1296 (17.7) | 1048 (18.0) | 1396 (16.6) | 1379 (17.4) | 1032 (17.6) | 1106 (18.6) | 1210 (21.1) | 1221 (20.6) |
| Sex | ||||||||
| Males | 2135 (48.1) | 2163 (48.0) | 2661 (47.1) | 2789 (48.2) | 2394 (48.7) | 2414 (48.5) | 2415 (48.1) | 2307 (48.0) |
| Females | 2313 (51.9) | 2357 (52.0) | 2758 (52.9) | 2973 (51.8) | 2407 (51.3) | 2633 (51.5) | 2602 (51.9) | 2434 (52.0) |
| Race/ethnicity | ||||||||
| Non-Hispanic White | 2391 (73.0) | 2276 (72.7) | 2547 (70.2) | 2786 (68.7) | 1842 (66.6) | 2233 (65.7) | 1711 (64.4) | 1694 (62.0) |
| Non-Hispanic Black | 867 (11.2) | 1012 (11.5) | 1136 (11.3) | 1025 (11.4) | 1274 (11.5) | 1009 (11.4) | 1060 (11.0) | 1128 (11.6) |
| Hispanic | 1016 (10.8) | 1049 (10.6) | 1525 (13.2) | 1647 (13.6) | 932 (14.3) | 1125 (14.7) | 1543 (15.0) | 1057 (15.8) |
| Other | 174 (5.0) | 183 (5.2) | 211 (5.3) | 304 (6.3) | 753 (7.7) | 680 (8.2) | 703 (9.6) | 862 (10.6) |
| Education level | ||||||||
| Less than high school graduate | 1288 (18.1) | 1234 (17.1) | 1665 (20.1) | 1634 (18.8) | 1102 (16.3) | 1028 (14.7) | 1150 (13.8) | 891 (10.7) |
| High school graduate or GED | 1111 (26.4) | 1093 (25.4) | 1340 (25.8) | 1316 (22.4) | 1008 (19.9) | 1141 (22.3) | 1108 (21.0) | 1146 (28.1) |
| Some college | 1216 (32.1) | 1290 (31.2) | 1398 (29.4) | 1624 (31.1) | 1463 (32.7) | 1578 (32.9) | 1492 (33.3) | 1554 (30.9) |
| College graduate or above | 827 (23.3) | 901 (26.3) | 1012 (24.6) | 1175 (27.5) | 1225 (31.0) | 1297 (30.0) | 1265 (32.0) | 1142 (30.2) |
| Ratio of family income to poverty level | ||||||||
| <1.30 | 1209 (20.3) | 1115 (16.8) | 1498 (19.8) | 1746 (20.6) | 1564 (23.7) | 1592 (23.5) | 1429 (19.4) | 1175 (18.7) |
| 1.30–1.84 | 570 (9.5) | 540 (9.1) | 699 (10.6) | 702 (9.9) | 594 (9.9) | 519 (9.6) | 658 (10.3) | 635 (9.5) |
| 1.85–2.99 | 830 (18.5) | 838 (19.1) | 985 (16.8) | 923 (15.6) | 702 (16.5) | 749 (15.3) | 916 (18.1) | 820 (15.8) |
| ≥3 | 1602 (46.4) | 1832 (51.6) | 1752 (45.3) | 1857 (46.5) | 1574 (44.0) | 1826 (45.4) | 1534 (44.5) | 1553 (46.4) |
GED, general educational development.
Values are presented as numbers of observations (n) and percentages (weighted %) with adjustment for complex survey design.
WG foods identified by different definitions
The numbers and types of WG foods identified differed substantially according to the different definitions. In 2017–2018, the WGC definition identified the largest number of foods defined as WG foods (n = 644), followed by the AHA (n = 621) and DGA (n = 581), and then AACCI (n = 329) and FDA (n = 230) (Table 3). Almost all of the WG foods identified by the DGA, FDA, and AACCI definitions were also captured using the WGC definition except for a few; however, the overlaps between WG foods determined by the DGA, FDA, and AACCI definitions and those identified by the AHA definition were small. The WGC and AHA definitions had the largest number of nonoverlapping WG foods (Supplemental Figure 3). The WG foods that were uniquely identified by the AHA, WGC, DGA, and AACCI were 230, 44, 7, and 2, respectively.
TABLE 3.
Numbers of whole-grain foods captured in the 2017–2018 Food and Nutrient Database for Dietary Studies (FNDDS) by different definitions and whole-grain food subgroups1
| DGA | FDA | AHA | AACCI | WGC | |
|---|---|---|---|---|---|
| Total, N | 581 | 230 | 621 | 329 | 644 |
| Whole-grain foods that uniquely identified, N | 7 | — | 230 | 2 | 44 |
| Whole-grain food subgroups, N (%) | |||||
| Snack/meal bars | 56 (9.64) | 13 (5.65) | 29 (4.67) | 28 (8.51) | 59 (9.16) |
| Savory snacks/crackers | 55 (9.47) | 47 (20.43) | 51 (8.21) | 56 (17.02) | 57 (8.85) |
| Breads | 38 (6.54) | 27 (11.74) | 61 (9.82) | 41 (12.46) | 69 (10.71) |
| Ready-to-eat cereals | 86 (14.8) | 52 (22.61) | 56 (9.02) | 92 (27.96) | 99 (15.37) |
| Cooked grains and cereals | 75 (12.91) | 72 (31.3) | 59 (9.5) | 72 (21.88) | 75 (11.65) |
| Pizza, burgers, sandwiches | 118 (20.3) | 0 (0) | 128 (20.6) | 6 (1.82) | 138 (21.4) |
| Mexican mixed dishes (e.g. burritos, tacos, nachos) | 2 (0.34) | 0 (0) | 39 (6.28) | 0 (0) | 2 (0.31) |
| Other mixed dishes | 115 (19.8) | 1 (0.43) | 102 (16.4) | 7 (2.13) | 116 (18.0) |
| Soups | 0 (0) | 0 (0) | 36 (5.8) | 0 (0) | 0 (0) |
| Infant foods | 28 (4.82) | 11 (4.78) | 41 (6.6) | 15 (4.56) | 16 (2.48) |
| Other | 8 (1.38) | 7 (3.04) | 19 (3.06) | 12 (3.65) | 13 (2.02) |
AACCI, American Association of Cereal Chemists International; DGA, Dietary Guidelines for Americans; FNDDS, Food and Nutrient Database for Dietary Studies; WGC, Whole Grains Council.
Values are numbers of whole-grain food items identified by each definition with or without percentages (%).
More snack/meal bars were defined as WG foods by the DGA (n = 56) and WGC (n = 59) than other definitions (n = 13–29). In contrast, twice as many breads were defined as WG foods by the AHA (n = 61) and WGC (n = 69) as other definitions (n = 27–41). The DGA, AACCI, and WGC definitions included more ready-to-eat cereals as WG cereals (n = 86, 92, and 99) than the AHA (n = 56) and FDA (n = 52) definitions. Differences were greatest for mixed dishes, including pizza, burgers, sandwiches, Mexican mixed dishes, other mixed dishes, and soups. The DGA, AHA, and WGC definitions captured 235, 305, and 256 mixed dishes as WG foods, respectively, whereas the AACCI definition detected 13 WG mixed dishes and the FDA only detected 1 WG mixed dishes. The AHA definition identified the most Mexican mixed dishes (n = 39) and soups (n = 36) and the least cooked grains and cereals (n = 59) as WG foods when compared with other definitions.
WG food consumption by different definitions
The mean consumption of WG foods among US adults varied considerably by each definition. For example, in 2017–2018, the highest population mean consumption was identified by the AHA (mean [SEM]: 1.05 [0.07] oz. eq./d) and WGC (0.95 [0.07]) definitions followed by the DGA (0.81 [0.06]) and AACCI (0.73 [0.05]), and then the FDA (0.53 [0.04]) (Table 4). Using all definitions except for WGC, we observed increasing WG food intake between 2003–2004 and 2017–2018 with the largest increase based on the AHA definition (from 0.65 to 1.05 oz. eq./d; mean difference [95% CI]: 0.40 [0.22, 0.59], equal a 61.5% increase), followed by the DGA (0.50 to 0.81; 0.30 [0.18, 0.43]) and AACCI (0.51 to 0.73; 0.21 [0.09, 0.33]) definitions (Table 4 and Supplemental Figure 4). Overall, the trend from 2003 to 2018 was nonlinear, with a significantly increasing trend from 2003 to 2012 but no significant trends (for the DGA, AHA, and WGC definitions) or a significantly decreasing trend from 2011 to 2018 (for the FDA and AACCI definitions) (Supplemental Table 3). Only the use of the AHA definition detected a significantly increased linear trend in WG food consumption.
TABLE 4.
Mean consumption of whole-grain foods (ounces equivalents per 2000 kcal per d) among US adults aged 20+ y, NHANES 2003–2018, by different whole-grain food definitions1
| Ounces equivalents per 2000 kcal per d,2 Survey weighted mean (SEM) | Difference, 2003–2004 vs. 2017–2018 Mean (95% CI) | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Whole-grain foods by different definitions | 2003–2004 (n = 4448) | 2005–2006 (n = 4520) | 2007–2008 (n = 5419) | 2009–2010 (n = 5762) | 2011–2012 (n = 4801) | 2013–2014 (n = 5047) | 2015–2016 (n = 5017) | 2017–2018 (n = 4741) | |
| DGA | |||||||||
| Whole-grain foods | 0.50 (0.03) | 0.61 (0.03) | 0.56 (0.04) | 0.68 (0.03) | 0.93 (0.05) | 0.82 (0.03) | 0.89 (0.04) | 0.81 (0.06) | 0.30 (0.18, 0.43) |
| % Whole-grain foods | 8.44 (0.41) | 11.0 (0.57) | 9.90 (0.70) | 11.6 (0.45) | 15.3 (0.92) | 14.5 (0.63) | 15.7 (0.87) | 13.7 (0.63) | 5.27 (3.74, 6.80) |
| FDA | |||||||||
| Whole-grain foods | 0.38 (0.02) | 0.41 (0.02) | 0.39 (0.03) | 0.50 (0.03) | 0.64 (0.04) | 0.61 (0.03) | 0.57 (0.03) | 0.53 (0.04) | 0.15 (0.06, 0.24) |
| % Whole-grain foods | 6.26 (0.36) | 7.40 (0.47) | 6.78 (0.46) | 8.35 (0.54) | 10.3 (0.58) | 10.7 (0.49) | 10.0 (0.61) | 9.18 (0.41) | 2.92 (1.81, 4.03) |
| AHA | |||||||||
| Whole-grain foods | 0.65 (0.05) | 0.71 (0.03) | 0.74 (0.06) | 0.91 (0.05) | 1.04 (0.05) | 1.02 (0.06) | 1.10 (0.05) | 1.05 (0.07) | 0.40 (0.22, 0.59) |
| % Whole-grain foods | 10.2 (0.70) | 12.2 (0.46) | 12.6 (1.03) | 14.2 (0.57) | 15.8 (0.77) | 16.0 (0.82) | 18.1 (0.75) | 16.7 (0.84) | 6.45 (4.23, 8.67) |
| AACCI | |||||||||
| Whole-grain foods | 0.51 (0.03) | 0.62 (0.03) | 0.73 (0.05) | 0.89 (0.04) | 0.94 (0.05) | 0.85 (0.03) | 0.78 (0.04) | 0.73 (0.05) | 0.21 (0.09, 0.33) |
| % Whole-grain foods | 8.56 (0.44) | 11.2 (0.59) | 13.0 (0.90) | 15.3 (0.56) | 15.4 (0.89) | 14.8 (0.63) | 13.8 (0.75) | 12.4 (0.60) | 3.85 (2.33, 5.37) |
| WGC | |||||||||
| Whole-grain foods | 0.91 (0.04) | 0.96 (0.04) | 0.88 (0.06) | 1.09 (0.04) | 1.11 (0.06) | 1.02 (0.03) | 1.04 (0.05) | 0.95 (0.07) | 0.04 (–0.12, 0.21) |
| % Whole-grain foods | 15.0 (0.62) | 16.8 (0.57) | 15.4 (0.93) | 18.4 (0.62) | 18.2 (1.04) | 17.5 (0.64) | 18.2 (0.91) | 16.0 (0.74) | 1.02 (–0.95, 2.99) |
AACCI, American Association of Cereal Chemists International; DGA, Dietary Guidelines for Americans; WGC, Whole Grains Council.
Values are means ± SEMs for each NHANES cycle with adjustment for complex survey design. Means and 95% CIs were presented for differences between the first and the last NHANES cycles. Both the mean intakes of whole-grain foods and percentages of whole-grain foods among all grain- or flour-containing foods consumed by US adults aged 20+ y were presented.
The mean consumption of whole-grain foods was estimated in ounces equivalents per 2000 kcal per d (oz. eq./d). For grain products such as breads, bagels, biscuits, muffins, cakes, cookies, pancakes, and waffles made with flour, each 16 g of flour present in a food was used as the basis for defining 1 ounce grain equivalent; for intact grains or grain products such as cream of wheat, barley, bulgur, millets, oats, pasta, rice, and ready-to-eat cereals, 28.35 g of grains was defined as equal to 1 ounce equivalent (23).
The percentage of grain foods consumed as WG foods by US adults also varied by definition. In 2017–2018, the highest percentage was estimated using the AHA (mean [SEM]: 16.7% [0.84%]) and WGC (16.0% [0.74%]) definitions, each nearly 2-fold higher than using the FDA definition (9.18% [0.41%]). Under all definitions, these percentages increased from 2003 to 2018, with the largest increase using the AHA (from 10.2% to 16.7%; mean difference [95% CI]: 6.45% [4.23%, 8.67%]) and DGA (8.44% to 13.7%; 5.27% [3.74%, 6.80%]) definitions, followed by AACCI (8.56% to 12.4%; 3.85% [2.33%, 5.37%]) and FDA (6.26% to 9.18%; 2.92% [1.81%, 4.03%]). Patterns were similar after adjusting for changes in sociodemographic characteristics during this period (Supplemental Tables 4 and 5).
Consumption of WG food subgroups by different definitions
Across all definitions, the food category that contributed to the largest amount of WG foods consumed by US adults in 2017–2018 was WG breads (Figure 1 and Supplemental Table 6). WG ready-to-eat cereals were the second largest contributor in all definitions, except for the AHA, under which mixed dishes were the second largest contributor of WG foods. From 2003 to 2018, US adults had the largest increase (mean difference [95% CI]: DGA: 0.26 [0.18, 0.34] oz. eq./d; AACCI: 0.25 [0.17, 0.33]; AHA: 0.24 [0.10, 0.37]; and FDA: 0.20 [0.12, 0.27]) in WG food from WG breads for each definition except WGC, for which the largest increase was from WG mixed dishes (0.08 [0.04, 0.12]). A significant increase in WG mixed dishes was also observed using the AHA, but not other definitions.
FIGURE 1.
Changes in mean consumption of whole-grain food subgroups among US adults aged 20+ y from 2003–2004 to 2017–2018, based on different definitions of whole-grain foods. Values are survey-weighted means. The mean consumption of whole-grain foods was estimated in ounces equivalents per 2000 kcal per d (oz. eq./d). For grain products such as breads, bagels, biscuits, muffins, cakes, cookies, pancakes, and waffles made with flour, each 16 g of flour present in a food was used as the basis for defining 1 ounce grain equivalent; for intact grains or grain products such as cream of wheat, barley, bulgur, millets, oats, pasta, rice, and ready-to-eat cereals, 28.35 g of grains was defined as equal to 1 ounce equivalent (24). AACCI, American Association of Cereal Chemists International; DGA, Dietary Guidelines for Americans; WGC, Whole Grains Council.
In 2017–2018, as a proportion of all products, ready-to-eat cereals had the highest percentage of being classified as WG foods, with proportions ranging from 43.3% to 77.2% using different definitions. The next highest proportions were for cooked grains and cereals (18.3–29.4%), savory snacks/crackers (17.0–22.1%), breads (13.9–32.2%), and snack/meal bars (0.86–12.9%), with the relative ranking of these subgroups being similar across definitions (Figure 2). The main exceptions were for mixed dishes. The FDA definition did not identify any mixed dishes as WG foods, whereas the AHA definition identified a higher percentage (8.5%) of WG mixed dishes than other definitions (DGA: 2.8%, AACCI: 0.32%, and WGC: 3.5%).
FIGURE 2.
Percentage of grain- or flour-containing foods that are whole-grain or refined-grain foods consumed by US adults aged 20+ y, by definition and whole-grain food subgroups, NHANES 2017–2018. Values are percentages – the mean ratios of whole-grain food intake or the refined-grain food intake to total grain- or flour-containing food intake among all participants, with adjustment for complex survey design. Refined-grain foods were defined as foods that are grain- or flour-containing foods but do not meet the definitions of being whole-grain foods. AACCI, American Association of Cereal Chemists International; DGA, Dietary Guidelines for Americans; WGC, Whole Grains Council.
WG food consumption in population subgroups
In each cycle, older adults, males, individuals who are non-Hispanic white, college graduates or above, and individuals with higher income (ratio of family income to poverty ≥3.00) consumed higher levels of WG foods than younger adults, females, non-Hispanic Black or Hispanic individuals, individuals with less than high school education, and those with lower levels of income, respectively, under almost all definitions (Figure 3 and Supplemental Table 7). Exceptions were observed for the AHA definition, under which the Hispanic subgroup had the highest WG food consumption and individuals with less than high school education had similar WG food intakes as those of the college graduates or above in 2017–2018.
FIGURE 3.
Changes in estimated mean consumption of whole-grain foods (ounces equivalents per 2000 kcal per d) among US adults aged 20+ y from 2003–2004 to 2017–2018, by race/ethnicity and education. Values are means with adjustment for complex survey design. P values for significance of the mean difference were obtained by performing t-tests. P values for interactions were obtained by conducting Wald tests to evaluate interaction terms between the survey cycle and categorical variables (age, sex, race/ethnicity) or ordinal variables (education and income levels). For grain products such as breads, bagels, biscuits, muffins, cakes, cookies, pancakes, and waffles made with flour, each 16 g of flour present in a food was used as the basis for defining 1 ounce grain equivalent; for intact grains or grain products such as cream of wheat, barley, bulgur, millets, oats, pasta, rice, and ready-to-eat cereals, 28.35 g of grains was defined as equal to 1 ounce equivalent (24). AACCI, American Association of Cereal Chemists International; DGA, Dietary Guidelines for Americans; GED, general equivalency diploma; WGC, Whole Grains Council.
From 2003 to 2018, greater increases in WG food intake were detected among younger adults aged 20–34 y and older adults aged 65+ y, males, individuals who are non-Hispanic White, individuals who are college graduates or above, individuals with lower income (ratio of family income to poverty = 1.30–1.84) compared with the corresponding sociodemographic subgroups, using all definitions. Exceptions were observed for the WGC definition, where no significant changes were detected for age, sex, race/ethnicity, and education subgroups, and for the AHA definition that greater increases in WG food intake were detected for adults aged 20–34 y and 35–49 y. Significant interactions between education subgroups and survey cycles were observed for the DGA (P value for interaction: 0.02), AHA (0.01), and AACCI (0.03) definition.
Discussion
In this nationally representative investigation, different definitions used to define WG products affected estimated intakes and trends of WG foods consumed by US adults from 2003 to 2018. The highest mean consumption in 2017–2018 and the greatest increase over time were observed using the AHA definition, and the lowest using the FDA definition, with ∼2-fold differences between the 2. Four of the 5 definitions identified significant increases in WG food intake. Under all definitions, the largest increase in consumption was seen for breads; and the greatest proportion of foods that are defined as WG foods was observed for ready-to-eat cereals. Under all definitions except the AHA, non-Hispanic white adults and individuals with college degrees or above consumed higher WG foods than non-Hispanic Black adults and people with lower levels of education.
These findings may partly relate to different origins and goals of the organizations when defining WG foods. The FDA definition, for example, originated from a health claim application on WG foods submitted by General Mills in 1999, which specified that the percentage of WG ingredients by weight per RACC serving should be ≥51% (16). Such a percentage, however, could be affected by the density, moisture, or other ingredients in the product. Consequently, products with WGs but also have nuts, seeds, dried fruits, oils, or others may not qualify as WG foods, resulting in the misclassification of many foods that are actually high in WGs. According to AACCI, the 8 g of WGs was selected to meet the DGA's recommendation on WGs, which was interpreted as Americans should eat ≥3 servings (16 g) of WGs or 6 servings of WG foods that have ≥8 g of WGs (17). Compared with the FDA definition, the AACCI definition is also weight based (≥8 g of WGs per 30 g food, or ≥26.7% by weight) but less restrictive, allowing a higher proportion of either refined-grain or nongrain ingredients (17). The WGC definition, developed to label industrial WG products using the Whole Grain Stamp, is in some ways the simplest (≥8 g per serving), but also most permissive in allowing anything else to be added, in any quantity, per serving (7). This is likely why other reports have documented that grain and cereal foods meeting the WGC definition also contain more calories and added sugars than those not meeting the definition (21). This could be problematic for consumers aiming to eat WG foods for health, and policies aiming to encourage this. As one component of The 2020 Impact Goals, the AHA recommended increasing the consumption of fiber-rich WGs (6). To provide a practical guide for consumers that captures the relative balance of WGs versus sugars, refined grains, and other fiber-rich ingredients like nuts or seeds, the AHA's definition utilizes the total carbohydrate to dietary fiber ratio (6, 21). Like the DGA and WGC definitions, the AHA definition captures the second largest number of WG foods among all grain- or flour-containing foods.
The heterogeneities in these definitions also affect the characterization of WG foods in different subgroups of foods. The WGC definition tends to identify more breads and ready-to-eat cereals as WG foods. According to the WGC, in 2021, ∼13,000 products were labeled with the Whole Grain Stamp, with 22% being breads and 25% breakfast cereals (7). In comparison, our findings suggest that the DGA, AHA, and WGC are more sensitive to identifying WG mixed dishes, such as pizza, burgers, and sandwiches; the AHA definition to identifying traditional Hispanic mixed dishes such as burritos, tacos, and nachos, whereas the DGA and WGC definitions are not. Based on nutrition facts, the carbohydrate to fiber ratios of tortilla wraps, taco shells, and tortilla chips are 7.7:1, 7:1, and 9.5:1, respectively, illustrating why higher means of mixed dishes were estimated using the AHA definition. Related to this, under the AHA definition, Hispanic individuals had higher WG food intake than other population subgroups, which could be due to WG cultural foods being more frequently consumed by Hispanic populations. There is a critical need to understand the strengths and gaps of differing WG food definitions in assessing and tracking dietary disparities by race/ethnicity.
The different definitions could impact the evaluation of the health implications of WG food intake. A prior study found that the AHA definition tends to identify grain foods with healthier nutritional profiles and other definitions tended to perform less well, whereas the WGC definition identifies products with higher sugars and calories (21). This highlights the need for investigations on the nutritional profile of WG foods identified by various definitions and their correlation with health outcomes.
The lack of a clear definition of WG foods has important policy implications for food labeling. Public health experts recommend consuming WGs over refined grains but research indicates that consumers have difficulty differentiating between the two (8). Food packaging displays various WG or WG food claims, such as touting the “first ingredient” as WGs, using “whole grain” in the product names, displaying the Basic Whole Grain Stamp, or disclosing the amounts of WGs present. Yet, none of these statements convey the ratio of WGs to refined grains. Thus, unless a label states the product is “100% whole grain,” consumers have no method to identify products that are primarily WG foods. A policy that requires product labels displaying WG or WG food claims to disclose the percent or grams of both whole and refined grains, would provide such information to consumers (22). An additional policy avenue to consider is whether certain WG claims, such as the use of “whole grain” in the product names, should be restricted to foods that are 100% WGs because of the deceptive nature of such claims (8).
Strengths
The most recent nationally representative data were used in this examination, depicting the up-to-date WG food intake under 5 different definitions over 16 y and ensuring the generalizability to US adults. The findings are among the first to facilitate the understanding of how different definitions can affect the estimated intake, trends, and sources of WG foods consumed by Americans. The WG food intake among the total population and subgroups and the intake of WG food subgroups were examined for each definition, providing comprehensive results and allowing the characterization of potential disparities. Sensitivity analyses were performed to adjust for shifting demographics when estimating trends, improving the robustness of study findings. When identifying WG foods, 2 researchers independently assign the groups and have multiple rounds of discussions with the senior investigators to resolve the discrepancies, which helped reduce the risk of misclassification and ensure the accuracy of study results.
Limitations
Self-reported dietary recalls are subject to measurement errors. To address this, the interviewer-administered 24-h diet recalls were used to collect intake data and energy adjustment was performed in analyses. The classification and RACC serving were manually assigned to food items, and subject to errors. Yet, we relied on the best available evidence (i.e. FNDDS food and ingredient descriptions, FDA's guidance on food serving sizes, and USDA FoodData Central) and applied the same criteria to all definitions to reduce random errors. Updates in NHANES foods from cycle to cycle could affect the estimated means of WG food intake (15). These updates were made to reflect the current foods available on the market, which may allow for more precise estimations.
Conclusions
The different definitions of WG foods affected the estimated mean intakes, trends, and main sources of WG foods consumed by US adults. These findings call for a standard definition of WG foods across the nation to guide consumers, industry, and policymakers to promote WG intake among Americans.
Supplementary Material
Acknowledgements
The authors’ contributions were as follows—MD, DM, JBW, and FFZ: designed research; MD: conducted research, analyzed the data, and wrote the manuscript; DM, JBW, JLP, PW, and FFZ: reviewed and edited the manuscript; FFZ: had primary responsibility for final content; and all authors: read and approved the final manuscript. The authors report no conflicts of interest for the submitted work. JBW reports leadership or fiduciary role in the US Preventive Services Task Force. DM reports research funding from the NIH, Gates Foundation, Rockefeller Foundation, and Vail Institute for Global Research; consulting fees from Acasti Pharma, Barilla, Danone, and Motif FoodWorks; participating on scientific advisory boards of start-up companies focused on innovations for health including Beren Therapeutics Brightseed, Calibrate, DayTwo, Elysium Health, Filtricine, Foodome, HumanCo, January Inc., Perfect Day, Season, and Tiny Organics; and chapter royalties from UpToDate. All of the above is outside the submitted work.
Notes
Supported by National Institute on Minority Health and Health Disparities (NIH/NIMHD) R01 MD011501.
Supplemental Figures 1–4, Supplemental Methods, and Supplemental Tables 1–7 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/ajcn/.
Abbreviations used: AACCI, American Association of Cereal Chemists International; DGA, Dietary Guidelines for Americans; FNDDS, Food and Nutrient Database for Dietary Studies; FPED, Food Patterns Equivalents Database; MPED, MyPyramid Equivalents Food Database; RACC, reference amount customarily consumed; SEM, standard error of the mean; WG, whole grain; WGC, The Whole Grains Council; WWEIA, What We Eat In America.
Contributor Information
Mengxi Du, Friedman School of Nutrition Science & Policy, Tufts University, Boston, MA, USA.
Dariush Mozaffarian, Friedman School of Nutrition Science & Policy, Tufts University, Boston, MA, USA; Tufts School of Medicine and Division of Cardiology, Tufts Medical Center, Boston, MA, USA.
John B Wong, Division of Clinical Decision Making, Tufts Medical Center, Boston, MA, USA.
Jennifer L Pomeranz, School of Global Public Health, New York University, New York, NY, USA.
Parke Wilde, Friedman School of Nutrition Science & Policy, Tufts University, Boston, MA, USA.
Fang Fang Zhang, Friedman School of Nutrition Science & Policy, Tufts University, Boston, MA, USA.
Data Availability
Data described in the manuscript, code book, and analytic code will be made available upon request pending application and approval.
References
- 1. Academy of Nutrition and Dietetics . What is a Whole Grain?2020. [Internet]. [Accessed 2020 Jan 26]. Available from: https://www.eatright.org/food/nutrition/healthy-eating/what-is-a-whole-grain. [DOI] [PubMed]
- 2. Ferruzzi MG, Jonnalagadda SS, Liu S, Marquart L, McKeown N, Reicks Met al. Developing a standard definition of whole-grain foods for dietary recommendations: summary report of a multidisciplinary expert roundtable discussion. Adv Nutr. 2014;5(2):164–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Korczak R, Marquart L, Slavin JL, Ringling K, Chu Y, O'Shea Met al. Thinking critically about whole-grain definitions: summary report of an interdisciplinary roundtable discussion at the 2015 Whole Grains Summit. Am J Clin Nutr. 2016;104(6):1508–14. [DOI] [PubMed] [Google Scholar]
- 4. U.S. Department of Agriculture, Department of Health and Human Services . Dietary Guidelines for Americans, 2020–2025. 2020. [Internet]. [Accessed 2020 Dec 20]. Available from: https://www.dietaryguidelines.gov/sites/default/files/2020-12/Dietary_Guidelines_for_Americans_2020-2025.pdf. [Google Scholar]
- 5. U.S. Department of Health and Human Services, U.S. Department of Agriculture . Dietary Guidelines for Americans 2015–2020. 2015. [Internet]. [Accessed 2019 Jun 20]. Available from: https://health.gov/dietaryguidelines/2015/. [Google Scholar]
- 6. Lloyd-Jones DM, Hong Y, Labarthe D, Mozaffarian D, Appel LJ, Van Horn Let al. Defining and setting national goals for cardiovascular health promotion and disease reduction: the American Heart Association's strategic impact goal through 2020 and beyond. Circulation. 2010;121(4):586–613. [DOI] [PubMed] [Google Scholar]
- 7. Whole Grains Council . Whole Grain Stamp. 2021. [Internet]. [Accessed 2021 Aug 3]. Available from:https://wholegrainscouncil.org/whole-grain-stamp.
- 8. Wilde P, Pomeranz JL, Lizewski LJ, Zhang FF. Consumer confusion about wholegrain content and healthfulness in product labels: a discrete choice experiment and comprehension assessment. Public Health Nutr. 2020;23(18):3324–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. McKeown NM, Jacques PF, Seal CJ, de Vries J, Jonnalagadda SS, Clemens Ret al. Whole grains and health: from theory to practice—highlights of the Grains for Health Foundation's Whole Grains Summit 2012. J Nutr. 2013;143(5):744S–58S. [DOI] [PubMed] [Google Scholar]
- 10. Fox News . USDA delays whole grains rule for school pastas. Fox News. 2021. [Internet]. [Accessed 2021 Dec 20]. Available from:https://www.foxnews.com/health/usda-delays-whole-grains-rule-for-school-pastas.
- 11. Wheeler L. USDA delays healthy school lunch requirements. TheHill. 2021. [Internet]. [Accessed 2021 Dec 20]. Available from:https://thehill.com/business-a-lobbying/362434-usda-delays-healthy-school-lunch-requirements.
- 12. Centers for Disease Control and Prevention (CDC). National Center for Health Statistics (NCHS) . About the National Health and Nutrition Examination Survey. Hyattsville (MD): U.S. Department of Health and Human Services, Centers for Disease Control and Prevention; 2020. [Internet]. [Accessed 2020 Oct 21]. Available from: [Google Scholar]
- 13. Centers for Disease Control and Prevention . NCHS Research Ethics Review Board (ERB) Approval. 2020. [Internet]. [Accessed 2020 Oct 21]. Available from:https://www.cdc.gov/nchs/nhanes/irba98.htm. [Google Scholar]
- 14. U.S. Department of Agriculture, Agricultural Research Service . USDA Food and Nutrient Database for Dietary Studies 2017–2018. 2021. [Internet]. [Accessed 2021 Aug 3]. Available from:http://www.ars.usda.gov/nea/bhnrc/fsrg. [Google Scholar]
- 15. U.S. Department of Agriculture, Agricultural Research Service . Beltsville (MD): Food Surveys Research Group; 2020. [Internet]. [Accessed 2020 Oct 21]. Available from:https://www.ars.usda.gov/northeast-area/beltsville-md-bhnrc/beltsville-human-nutrition-research-center/food-surveys-research-group/. [Google Scholar]
- 16. U.S. Food and Drug Administration . Health Claim Notification for Whole Grain Foods. Accessed 03 August, 2021. [Internet]. [Accessed 2021 Aug 3]. Available from:https://www.fda.gov/food/food-labeling-nutrition/health-claim-notification-whole-grain-foods. [Google Scholar]
- 17. Cereals and Grains Association . AACCI's Whole Grains Working Group Unveils New Whole Grain Products Characterization. 2021. [Internet]. [Accessed 2021 Aug 3]. Available from:https://www.cerealsgrains.org/about/newsreleases/Pages/WholeGrainProductCharacterization.aspx. [Google Scholar]
- 18. U.S. Department of Agriculture, Agricultural Research Service . What We Eat in America Food Categories 2017–2018. Available from: www.ars.usda.gov/nea/bhnrc/fsrg. [Google Scholar]
- 19. U.S. Department of Agriculture , Agricultural Research Service. FoodData Central. 2020. [Internet]. [Accessed 2020 Oct 21]. Available from:https://fdc.nal.usda.gov/. [Google Scholar]
- 20. U.S. Department of Health and Human Services, Food and Drug Administration, Center for Food Safety and Applied Nutrition . Reference Amounts Customarily Consumed: List of Products for Each Product Category: Guidance for Industry. 2020. [Internet]. [Accessed 2020 Oct 21]. Available from:https://www.fda.gov/media/102587/download. [Google Scholar]
- 21. Mozaffarian RS, Lee RM, Kennedy MA, Ludwig DS, Mozaffarian D, Gortmaker SL. Identifying whole grain foods: a comparison of different approaches for selecting more healthful whole grain products. Public Health Nutr. 2013;16(12):2255–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Pomeranz JL, Lurie PG. Harnessing the power of food labels for public health. Am J Prev Med. 2019;56(4):622–5. [DOI] [PubMed] [Google Scholar]
- 23. Bowman SA, Clemens JC, Friday JE, Lynch KL, LaComb RP, and Moshfegh AJ. Food Patterns Equivalents Intakes by Americans: What We Eat in America, NHANES 2003–04 and 2013–14. Food Surveys Research Group. Dietary Data Brief No. 17. May 2017. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data described in the manuscript, code book, and analytic code will be made available upon request pending application and approval.