Dairy, Meat, Seafood, and Plant Sources of Saturated Fat: United States, Ages Two Years and Over, 2017–2020

Abstract

Background

Research suggests that the effects of saturated fat (SF) on health differ depending on its food source. Dairy SF has been associated with lower cardiovascular disease (CVD) risk, whereas meat SF is linked to a higher CVD risk.

Objectives

To estimate the contribution to the total intake of SF of 1) 5 food groups - dairy, meats, seafood, plants, and “other,” and 2) the top 10 specific food category sources in the United States population overall and by sociodemographic subgroup.

Methods

The analysis included data from 11,798 participants in the 2017–March 2020 National Health and Nutrition Examination Survey aged 2+ y. Grams of SF from the food sources expressed as a percentage of the total grams of SF consumed, were estimated using the population ratio method.

Results

Mean daily intake of SF was 28.1 g [95% confidence interval (CI): 27.6–28.6 g], comprising 11.9% (95% CI: 11.7–12.1%) of total energy intake. Dairy contributed 28.4% of SF, followed by meats (22.1%), plant sources (7.5%), fish and seafood (1.2%), and the remaining foods (41.6%). Youth had higher SF intake from dairy than adults (P < 0.001), whereas non-Hispanic Whites had higher intake than non-Hispanic Blacks (P < 0.001) and Hispanics (P = 0.016). Adults had higher SF intake from meats than youth (P = 0.002), males more than females (P < 0.001), and non-Hispanic Blacks more than non-Hispanic Asians (P = 0.016) and Hispanics (P < 0.001). The top 10 specific sources of SF were unprocessed red meats, sweet bakery products, cured meats, milk, cheese, pizza, unprocessed poultry, Mexican mixed dishes, eggs, and combined fruits and vegetables.

Conclusions

Although dairy contributed ∼30% of SF compared to ∼20% for total meat, the top specific food category source of SF was unprocessed red meats, which ranked in the top 2 food category sources of SF for most subgroups. These findings may be useful for further research to examine the relationship between the different sources of SF and health outcomes.

Introduction

The Dietary Guidelines for Americans recommend no >10% of calories from saturated fat (SF) because increased intake of SF has been associated with CVD risk factors [1]. The AHA recommends replacing SF with unsaturated fat, especially polyunsaturated fat, reducing blood concentrations of total cholesterol and LDL cholesterol [2]. Replacing SF with omega (ω)–3 polyunsaturated fat can decrease risk of CVD by ≤30% [2] and lower risk of CVD events [3].

SF occurs naturally in animal foods such as meat and dairy and some plant foods like coconut and palm oil. SF can also be added to foods while processing or preparing fried foods and mixed dishes [1]. Associations between SF and health outcomes, such as CVD risk, may vary by the food source of SF, highlighting the importance of considering the food source [4,5]. For example, butter and cheese have been reported to have similar effects on HDL cholesterol among adults with abdominal obesity, but LDL cholesterol concentrations were lower after cheese consumption than butter [6]. Red meat tends to be higher in SF than white meat and has been associated with CVD in observational studies [7]. However, consumption of SF from either white or red meat has been equally positively associated with blood LDL cholesterol, suggesting that risk factors associated with red meat go beyond SF [7]. Intake of processed meat such as sausage, which often contains a high amount of minced fatty tissue, has consistently been associated with increased chronic disease risk, including cardiometabolic diseases [8,9] and some types of cancer [10,11].

Since sources of SF vary and dietary patterns vary by sociodemographic characteristics, this study describes sociodemographic differences in percentage contribution to the total intake of SF for 1) 5 food groups - dairy, meats, seafood, plants, and “other,” and 2) the top 10 more specific food category sources. This information may help inform public health messaging and interventions aimed at changing dietary patterns and for further research to examine the relationship between the different food category sources of SF and health outcomes.

Methods

Study design

These analyses utilized data from participants aged 2 y and older from the NHANES between 2017 and March 2020. NHANES is a complex, stratified, multistage probability sample of the United States civilian, noninstitutionalized population administered by the NCHS. The survey includes an in-home interview and a standardized health examination at a mobile examination center (MEC). Additional details about the NHANES study design, implementation, data sets, and other relevant information can be found online [12]. The NCHS ethics review board approved the NHANES protocol [13]. For participants under the age of 18, written parental consent was obtained, and for participants between the ages of 7 and 17, assent was also obtained. Written informed consent was obtained from adults.

Because of the COVID-19 pandemic, data collection for the NHANES 2019–2020 cycle was not completed, and the collected data were not nationally representative. Therefore, data from 2019 to March 2020 were combined with data from the NHANES 2017–2018 cycle to create a nationally representative NHANES 2017-March 2020 prepandemic data file [14,15]. This sample included oversampling to improve the reliability of estimates between various subgroups, including Hispanic, non-Hispanic Black, and non-Hispanic and non-Black Asian persons [15]. The unweighted NHANES MEC response rate from 2017 to March 2020 was 48.0% for all ages combined [16].

Dietary intake

NHANES included 2 24-h dietary recalls. The first interview was conducted in person at the MEC, whereas the second interview was conducted by telephone 3–10 d later [13,[17], [18], [19]]. Trained interviewers used a computer-assisted dietary interview system that included an automated multiple-pass method with standardized probes [20] to gather details about the type and quantity of all foods and beverages consumed during the day before the recall interview. Participants aged 12 y and older completed their dietary interviews independently, whereas children aged 6–11 were assisted by an adult, and proxies were reported for children aged 5 y or younger.

For this analysis, only records considered reliable from the first 24-h dietary recall were used to represent the mean population intake on a given day [12]. The dietary data quality control criteria and methods are described in detail elsewhere [19,21,22].

Defining specific food category sources and food groups

Each food and beverage reported by NHANES participants were coded using the Food and Nutrient Database for Dietary Studies (FNDDS) food codes and nutrient contents for each food were publicly released on the NHANES website [23]. FNDDS; food codes include food descriptions, ingredient lists, gram weights, and nutrient and energy values for each reported food and beverage. For the current analysis, these FNDDS food codes were linked to the USDA’s What We Eat in America (WWEIA) food classification scheme, which categorizes food items as they are consumed. This classification system includes over 150 food categories, such as milk with various fat contents and different types of fruits and vegetables [24].

Food groups

Reported foods were categorized into 5 broad groups based on the specific 2017–2020 WWEIA food categories. These food groups were: 1) dairy, which included milk, yogurt, cheese, dairy desserts, pizza, macaroni and cheese, coffee, and tea; 2) meats, which included unprocessed red meat, poultry, and cured meats; 3) fish and seafood; 4) plant sources; and 5) “other,” which included the remaining WWEIA food categories such as savory snacks, bread, breakfast cereals, 100% fruit juice, sweetened beverages, sugars, fats and oils, protein and nutritional powders, and foods not included in a food category (see Supplementary Table 1). Mixed dishes that didn’t fit into any of the first 4 groups were also classified under the “other” group. Deli and cured meat sandwiches, and meat and barbecue sandwiches, were included in either the cured meats or unprocessed red meats and poultry category based on the ingredients listed in the food descriptions. For example, a barbecue chicken sandwich on a wheat bun would be classified as unprocessed poultry. Infant formula and infant food were not included in these analyses.

Specific food category sources

The 150 WWEIA food categories were also organized into 32 specific food categories within each of the 5 major food groups. Supplementary Table 1 contains both the 32 specific food categories within each of the 5 major food groups along with a description of the foods within the 32 specific food categories.

Demographic variables

In the analysis, age was categorized into 6 groups: 2–5, 6–11, 12–19, 20–39, 40–59, and 60 y and older, and included an overall estimate for children and adolescents (youth) age 2–19 y, and for adults aged 20 y and older. The analyses were restricted to race and Hispanic origin groups for which data were statistically reliable, including non-Hispanic White (NHW), non-Hispanic Black (NHB), non-Hispanic Asian (NHA), and Hispanic persons. The “other race” group included persons who reported multiple races and were included in the overall estimates but were not shown separately.

Family income was defined based on the federal poverty level (FPL) and was categorized as FPL < 130%, FPL 130 to < 350%, and FPL ≥ 350%. These levels were determined based on the income-to-poverty ratio, which is calculated by dividing the annual total family income by the United States Department of Health and Human Services poverty threshold, taking into account inflation and family size [25]. The recommended threshold for eligibility for the Supplemental Nutrition Assistance Program [26] and the free and reduced-price school lunch program [27] is 130% of the poverty threshold.

Statistical analysis

Mean SF intake (grams) on a given day was calculated overall and by sex, age, race and Hispanic origin, and family income. The mean percent contribution of the 5 food groups and the 32 specific food categories to the total SF intake on a given day was estimated using the population ratio method [28,29]. This method for determining the contribution of each food group and food category involves summing the grams of SF provided by each food group or food category for all persons within a specific sociodemographic group and dividing by the total grams of SF consumed by all persons in that group.

$$\frac{\sum _{i=1}^n\left(F_i\cdot W_i\right)}{\sum _{i=1}^n\left(T_i\cdot W_i\right)}\times 100$$

Where n = sample size, $F_i$= SF contributed by the food group or WWEIA food categories for the ith individual, $T_i$ = total SF from all foods for the ith individual, and $W_i$ = sample weight for the ith individual.

To obtain nationally representative estimates, the NHANES survey design variables and day 1 dietary sample weights were used. These sample weights account for differential probabilities of selection, nonresponse, noncoverage, and day of the week of the 24-h recall in NHANES. SEs of the means were estimated using Taylor series linearization [30,31], a method that incorporates the NHANES sampling design. All reported mean estimates were deemed reliable using the criterion of a relative SE < 30%, calculated as [(SE of the mean / mean) • 100].

To compare the relative contribution of the top 10 food category contributors of SF from the 32 WWEIA food categories, pairwise differences between males and females, youth and adults, and race and Hispanic origin groups were evaluated using a t statistic. Tests for linear trends by age for youth and adults and by family income were evaluated using orthogonal polynomials. All statistical significance levels were set at P < 0.05, and Bonferroni’s method of correction was used to adjust for multiple comparisons, and the adjusted P values are presented [32].

Given the differences in nutritional requirements and total food and caloric intake by age, overall estimates were age-adjusted using the direct method to 2000 projected United States Census population and age groups 2–5, 6–11, 12–19, 20–39, 40–59, 60 and over. Statistical analyses were performed using SAS version 9.4 (SAS Institute Inc.) [33] and SUDAAN version 11.0 (RTI International) [34].

After excluding participants who did not provide a dietary recall (n = 1415), those reported to have consumed breast milk (n = 18), and those who had unreliable recalls (n = 184), the final sample size consisted of 11,798 out of the 13,415 participants aged 2 y and above who visited the MEC during NHANES 2017–March 2020 (Supplementary Figure 1).

Results

On a given day, during 2017–March 2020, the mean consumption of SF in the United States population was 28.1 g (95% CI: 27.6–28.6 g), or 11.9% (95% CI: 11.7–12.1%) of total daily caloric intake (Table 1).

TABLE 1.

Age-adjusted¹ percentage contribution of dairy, meats, plant sources, fish and seafood, and other sources to total saturated fat intake (g) on a given day among United States persons aged 2 y and over, NHANES, 2017– March 2020

	Sample size	Mean, (g)	Percent, (kcal) from saturated fat	Percent of total saturated fats2 (CI)
				Dairy	Meats			Plant sources	Fish and seafood	Other
					Total	Cured meats	Unprocessed
Overall	11,798	28.1 (27.6–28.6)	11.9 (11.7–12.1)	28.4 (27.1–29.6)	22.1 (20.8–23.5)	7.2 (6.7–7.7)	14.9 (13.7–16.2)	7.5 (6.9–8.2)	1.2 (1.0–1.3)	41.6 (40.4–42.8)
Overall3	11,798	28.1 (27.6–28.6)	11.9 (11.7–12.1)	27.8 (26.5–29.1)	22.3 (20.9–23.7)	7.2 (6.7–7.8)	15.1 (13.8–16.4)	7.8 (7.1–8.5)	1.2 (1.1–1.4)	41.7 (40.5–42.9)
Age (y)
2–19	4091	26.3 (25.7–27.0)4	12.2 (12.0–12.4)	35.1 (33.6–36.6)4	18.8 (17.1–20.6)4	6.7 (6.0–7.5)	12.1 (10.5–13.7)4	4.4 (4.1–4.8)4	0.6 (0.4–0.9)4	41.2 (39.9–42.6)
2–53	940	20.4 (19.6–21.2)5	11.7 (11.5–11.9)	42.3 (40.1–44.5)5	16.0 (14.3–17.8)5	7.0 (5.4–8.7)	9.0 (7.5–10.5)5	4.6 (4.0–5.2)	0.5 (0.3–0.7)	36.2 (33.9–38.5)6
6–113	1411	27.2 (26.1–28.3)	12.4 (12.1–12.7)	34.1 (31.8–36.3)	16.7 (14.9–16.4)	6.8 (5.7–7.9)	9.9 (8.3–11.5)	3.7 (3.2–4.2)	0.6 (0.3–1.0)	44.4 (43.1–45.9)
12–193	1740	28.6 (27.7–29.6)	12.2 (12.0–12.5)	32.5 (29.7–35.2)	21.8 (19.1–24.6)	6.6 (5.2–7.9)	15.3 (12.9–17.7)	4.9 (4.2–5.6)	0.7 (0.5–1.0)	40.9 (38.7–43.0)
20 and over	7707	28.7 (28.1–29.4)	11.8 (11.6–12.0)	25.9 (24.4–27.5)	23.3 (21.8–24.8)	7.3 (6.7–8.0)	16.0 (14.7–17.3)	8.7 (7.8–9.5)	1.3 (1.2–1.5)	41.7 (40.2–43.2)
20–393	2358	29.6 (28.6–30.6)	11.7 (11.5–12.0)	25.1 (23.3–26.9)	25.2 (23.4–26.9)5	7.7 (6.4–9.0)	17.5 (15.7–19.2)5	7.6 (7.0–8.3)5	0.9 (0.7–1.1)5	42.0 (40.3–43.9)
40–593	2548	28.7 (27.7–29.8)	11.7 (11.4–12.0)	27.4 (25.7–29.1)	22.6 (20.4–24.8)	6.7 (5.8–7.7)	15.9 (14.0–17.7)	9.2 (7.8–10.6)	1.4 (1.1–1.7)	40.0 (37.8–42.1)
60+3	2801	27.4 (25.9–28.8)	12.1 (11.6–12.7)	25.0 (22.8–27.1)	21.3 (19.0–23.6)	7.7 (6.2–9.2)	13.6 (11.3–15.9)	9.6 (8.2–11.0)	2.0 (1.6–2.4)	42.7 (44.0–44.5)
Sex
Male	5813	31.7 (31.0–32.4)	11.8 (11.6–12.0) 7	27.9 (26.2–29.5)	24.6 (23.2–26.1)7	8.0 (7.3–8.8)	16.6 (15.2–17.9)7	7.0 (6.4–7.6)	1.2 (1.0–1.51)	40.3 (38.7–42.0)
Female	5985	24.7 (24.2–25.2)	12.0 (11.8–12.3)	29.0 (27.9–30.1)	19.1 (17.5–20.6)	6.1 (5.3–6.9)	12.9 (11.5–14.4)	8.2 (7.5–9.0)	1.1 (0.9–1.2)	43.1 (41.9–44.4)
Race and Hispanic origin8
Non-Hispanic White	4106	29.0 (28.4–29.7)	12.3 (12.1–12.6)	30.0 (28.8–31.2)9,10	22.2 (20.2–24.1)	8.0 (7.2–8.7)11	14.2 (12.5–15.9)	7.8 (6.9–8.8)	1.0 (0.7–1.2)9	39.9 (38.3–41.5)10
Non-Hispanic Black	3155	26.6 (25.7–27.6)	11.4 (11.1–11.6)	23.3 (22.1–24.4)	28.5 (26.7–30.3)10,11	8.8 (7.5–10.0)11	19.7 (18.3–21.2)10,11	6.9 (6.3–7.5)11	1.9 (1.6–2.2)	40.2 (38.7–41.7)10
Non-Hispanic Asian	1168	22.2 (21.3–23.2)	10.4 (9.9–10.9)	26.5 (24.9–28.1)	16.9 (13.2–20.6)	3.3 (2.4–4.3)	13.6 (10.4–16.8)	10.5 (9.2–11.8)10	2.2 (1.5–2.9)	44.1 (41.1–47.1)
Hispanic	2616	26.8 (25.7–27.8)	11.2 (10.9–11.5)	25.5 (24.4–26.6)	19.5 (18.1–20.8)	5.3 (4.2–6.4)	14.1 (13.2–15.1)	6.6 (6.0–7.3)	1.2 (0.9–1.5)	48.0 (46.3–49.7)
FPL12
<130% FPL	3391	26.7 (25.6–27.9)13	11.5 (11.1–11.9)13	28.4 (26.3–30.5)	22.8 (21.3–24.2)	6.8 (5.7–7.8)	16.0 (14.6–17.4)5	6.3 (5.5–7.2)5	1.3 (0.9–1.8)	41.8 (39.4–44.3)
130 to < 350% FPL	3961	28.4 (27.7–29.1)	11.9 (11.8–12.1)	28.2 (26.6–29.8)	22.5 (20.6–24.5)	7.9 (6.8–9.0)	14.6 (12.9–16.3)	7.7 (7.0–8.3)	1.2 (0.9–1.4)	41.2 (39.5–42.9)
≥350% FPL	3062	28.7 (27.8–29.7)	12.1 (11.9–12.4)	28.5 (27.0–29.9)	21.3 (19.3–23.3)	7.0 (5.9–8.2)	14.3 (12.6–15.9)	8.1 (7.3–8.9)	1.1 (0.8–1.4)	41.8 (40.1–43.5)

FPL, federal poverty level.

¹Percentages are age-adjusted by the direct method to the 2000 projected United States Census population using age groups 2–5, 5–11, 12–19, 20–39, 40–59, and 60 and over.

²Percentages are based on What We Eat in America Food Categories 2017– March 2020; unweighted sample size includes all reliable and complete recalls. Percentages may not add ≤100% because of the rounding. All pairwise tests were performed using t statistic, tests for trends by age were evaluated using orthogonal polynomials, significance levels for statistical testing were at the P < 0.05 significance level, and P values for pairwise comparisons adjusted using Bonferroni’s method of correction for multiple comparisons.

³Unadjusted estimates.

⁴Significantly different from the 20 y and over group.

⁵Linear trends

⁶Significantly different from 6–11.

⁷Significantly different from females.

⁸Other race category not shown (n = 753).

⁹Significantly different from non-Hispanic Black.

¹⁰Significantly different from Hispanic.

¹¹Significantly different from non-Hispanic Asian.

¹²Participants with missing FPL information (n = 1384) were included in analyses that did not involve family income.

¹³Significantly different from 130 to < 350% FPL and ≥350% FPL.

Food group sources of SF

Dairy sources, including milk, yogurt, cheese, pizza, and ice cream, accounted for 28.4% of SF intake. Meats, including cured and unprocessed red meat and poultry, accounted for 22.1% of SF intake. Plant sources accounted for 7.5%, whereas fish and seafood accounted for only 1.2%. The largest contributor to SF intake was the “other” food group, which accounted for 41.6% of SF intake and included the remaining food categories such as fats and oils, processed foods, and Asian and Mexican mixed dishes. The unadjusted and age-adjusted estimates were nearly identical.

Adjusted for age, dairy contributed more to SF intake for youth (35.1%) compared to adults (25.9%), P < 0.001 (Table 1). On the contrary, the contribution of total meats (23.3% compared with 18.8%, P = 0.002), plant sources (8.7% compared with 4.4%, P < 0.001), and fish and seafood (1.3% compared with 0.6%, P < 0.001) were higher for adults than youth. There were no differences in the contribution of dairy, fish, and seafood or “other” by sex, whereas total meats (P < 0.001) and unprocessed meats (P = 0.016) contributed more SF for males than females.

Between the different race and Hispanic origin groups, the contribution of dairy was less for NHB (P < 0.001) and Hispanic persons (P = 016) compared to NHW persons. Conversely, the contribution of total meats was higher among NHB persons compared to NHA (P = 0.016) and Hispanic persons (P < 0.001). The contribution of plant sources was higher for NHA persons compared to NHB (P = 0.03) and Hispanic persons (P = 0.03), whereas the contribution from fish and seafood sources was higher for NHB than for NHW persons (P = 0.016). The contribution from the “other” category was higher among Hispanic persons than NHW (P < 0.001) and NHB persons (P < 0.001).

The top 10 specific WWEIA food category sources of SF

Table 2 shows the age-adjusted percent contribution to SF of the top 10 specific WWEIA food categories overall and by sex and age. The colors reflect the 5 broader food groups. Overall, 3 of the top 10 specific food sources were meats (in red), 3 were dairy (in orange), 1 was a plant source (in green), and 3 food sources were categorized as other (in yellow). Pizza was among the top 10 contributors for males but not for females, whereas fats and oils were among the top 10 for females but not males.

TABLE 2.

Age-adjusted percent contribution of top 10 food category sources of total saturated fats (g) on a given day, by sex and age, among United States persons aged 2 y and over, NHANES, 2017– March 2020

Among youth, 4 or 5 of the top 10 specific WWEIA food category sources of SF belonged to the dairy food group compared to 3 of the top 10 for all age groups of adults. Plant sources were among the top 10 specific sources in adults but not in youth. The contribution of milk (P < 0.001) and cheese (P = 0.014) decreased with age, whereas the contribution of unprocessed red meats (P < 0.001), pizza (P = 0.001) and Mexican mixed dishes (P = 0.008) increased with age among youth. Among adults, the contribution of Mexican mixed dishes (P < 0.0001) and pizza (P < 0.0001) decreased with age, and the contribution of cheese (P = 0013) increased with age.

Table 3 displays the age-adjusted contributions of the top 10 specific WWEIA food sources of SF by race and Hispanic origin. Plant-based protein foods were 2 of the top 10 specific sources, and meats were 2 of the specific sources among NHA persons, whereas meat was 3 of the sources, and plants were 1 of the top 10 for the other race/Hispanic origin subgroups. Moreover, differences between the race/Hispanic origin subgroups included cheese and cured meats being among the top sources for all groups except for among NHA persons. Similarities included unprocessed red meats as the top 2 sources for all adult groups and sweet bakery products as 1 of the top 4 sources for both youth and adults.

TABLE 3.

Age-adjusted percent contribution of top 10 food category sources of total saturated fats (g) on a given day, by race/Hispanic origin, among United States persons aged 2 y and over, NHANES, 2017– March 2020

Table 4 displays the age-adjusted contributions of the top 10 specific WWEIA food sources of SF intake by family income. Overall, the number and the food categories representing dairy, meats, and plant sources were similar across income levels. Unprocessed red meats and sweet bakery products were the top 2 sources for all income levels. Overall, the contribution of fruits and vegetables (P = 0.003) increased with increasing family income, and the contribution of unprocessed red meats (P = 0.044) and Mexican mixed dishes (P = 0.012) decreased with increasing family income. Among adults, the contribution to SF of sweet bakery products (P = 0.045) and milk (P = 0.012) decreased as income increased.

TABLE 4.

Age-adjusted percent contribution of top 10 food category sources of total saturated fats (g) on a given day, by family income, among United States persons aged 2 y and over, NHANES, 2017– March 2020

Discussion

Between 2017 and March 2020, dairy products and meats were the primary sources of SF among United States individuals aged 2 y and older. Dairy products accounted for 28.4% of SF intake, whereas meats contributed 22.1%. On the contrary, plant sources, including nuts, seeds, legumes, and fish and seafood, made relatively smaller contributions, accounting for 7.5% and 1.2%, respectively. The largest contributor to SF intake was the “other” food group, which includes SF added during processing, such as those found in potato chips accounting for ∼42% of SF intake. Notably, sweet bakery products emerged as the second-largest contributor to SF intake among the top 10 food categories, and they ranked either first or second for most demographic subgroups studied. In addition, unprocessed red meat ranked 1 or 2 among adults in all subgroups. Only among NHA persons was >1 of the top 10 sources of SF from plants.

Our findings related to SF from meat sources were similar to previous publications. Analyses from NHANES 2003–2006 using a different food categorization scheme [35] also ranked meat higher than most other sources of SF. In 2017–March 2020, unprocessed red meat was the only food in the top 10 food sources with differences between males and females, aligning with previously reported higher consumption of red meat among males [36]. Cured meats were found to contribute slightly more to SF intake in males than females, though the contribution was not significantly different. Moreover, meats contributed more to SF intake among NHB persons than NHA and Hispanic persons, a finding that also aligns with a previous analysis of NHANES 2003–2006 data [37]. Additionally, both NHW and NHB persons had a higher SF intake from cured meats than NHA persons. Considering the potential CVD risks associated with the intake of high sodium and preservative foods [4], it is particularly important to assess the contribution of cured meats to overall dietary intake. Excess sodium consumption increases risk for hypertension, which is a leading risk factor for CVD.

Dairy products provide essential nutrients such as high-quality protein, vitamins, minerals, and essential FAs [4,38], but some dairy products can also be high in fat. There have been inconsistent findings regarding the association between dairy consumption and CVD risk [4,[39], [40], [41], [42]], despite the established association between overall SF intake and CVD risk. In general, it has been reported that consuming ≤200 g/d of low-fat or full-fat dairy products is not associated with all-cause mortality or CVD risk in healthy people [5]. Furthermore, dairy food consumption has been associated with reduced risk for heart disease mortality [43]. A study of adults aged 45–84 y from the Multi-Ethnic Study of Atherosclerosis also found that intake of dairy SF was inversely associated with CVD risk, whereas meat SF was positively associated with risk of CVD [4]. When considering specific dairy food categories, fermented dairy products such as yogurt and cheese intake have been inversely associated with all-cause mortality and CVD risk. However, the association between milk intake and CVD risk has been inconsistent, with the relationship varying based on the fat content of the milk. A positive association has been reported between fatal and non-fatal CHD events and full-fat milk, whereas no association has been observed for low-fat milk [5]. In the current study, milk was among the top 10 food category contributors to SF consumption for all subgroups analyzed, ranking higher for youth than adults, which may reflect a preference for full-fat milk among youth. Cheese generally ranked lower and contributed less to SF intake than milk, unlike in 2003–2006 NHANES [35], suggesting some changes in the United States dietary pattern. It is worth noting that different racial and ethnic groups may have varying concentrations of lactose intolerance, which can influence the contribution of dairy sources to SF intake [44,45].

Plant sources of SF, such as nuts and seeds, are rich in fiber, protective PUFAs, and phytochemicals [4]. Higher amounts of PUFAs are found in nuts and seeds, in some fish and seafood, and in lower amounts in beef, chicken, eggs, and milk [46]. Flaxseed, chia seeds, and walnuts contain higher amounts of ALA, an ω–3 PUFA [46]. This study found that plant sources contributed <10% of SF, although adults consumed a higher amount of SF from plant sources than youth. Furthermore, the contribution of plant sources to SF intake was slightly higher among all females than all males, although this difference was not statistically significant. Research has shown that women tend to consume more vegetable servings, and a higher percentage report consuming vegetables than men [47]. Additionally, higher family income was associated with a greater intake of SF from plant-based sources. These findings are consistent with previous research demonstrating a relationship between the female [48,49] and income with overall diet quality, particularly among adults [48]. Finally, among NHA individuals, 2 of the top 10 sources of SF were plant-based (fruits and vegetables; plant-based protein). Taken together, 10.5% of SF was obtained from plant sources among NHA individuals.

One limitation of our study is the use of self-reported dietary data, which may be prone to misreporting, particularly underestimating total EI [[50], [51], [52]]. However, since our analysis focused on the percentage of SF intake from different sources, independent of EI, this may have reduced some of the potential for error. Another limitation is that our food categorization approach may have over or underestimated the contribution of certain food groups, such as dairy. For example, our dairy category included non-typical dairy foods such as pizza and macaroni and cheese, which could contain SF from meat toppings or animal-based oils used in dough or baking. To avoid double-counting, all SF in these categories were attributed solely to dairy. Our groupings may also differ from those used in previous studies, which could limit comparability across food groups. Finally, we acknowledge that the NHANES surveys have seen a gradual decline in response rates, which may increase risk of nonresponse bias [53].

The study has several strengths. First, the results are nationally representative, providing a comprehensive overview of the population’s dietary habits. Second, using the automated multiple-pass method to collect 24-h dietary recalls in NHANES is widely regarded as the gold standard for both self-report and proxy reporting of 24-h recalls. Third, the WWEIA food categorization scheme, similar to the 2020–2025 Dietary Guidelines for Americans [1], was utilized to categorize foods, allowing for comparisons across different food categories. Finally, the study used ratios of means to summarize the dietary data, which is a more suitable technique for population-level research questions, and more closely reflects usual intakes than the mean of ratios method. Therefore, the results can be accurately interpreted at the population level [29].

In conclusion, between 2017 and March 2020, dairy contributed just over 28% of SF intake, meat contributed to just over one-fifth of intake, whereas plant sources contributed <10% of SF in the United States diet. SF added during mainly food processing from the “other” food group represented the largest proportion of SF intake at around 42%. Unprocessed red meats and sweet bakery products were identified as the top 2 food category contributors to SF for most demographic subgroups. These findings may be useful for further research to examine the relationship between the different food category sources of SF and health outcomes.

Funding

This work was performed under the employment of the United States Federal Government, and the authors did not receive any outside funding.

The opinions expressed in this manuscript are the author's own and do not reflect the view of the NIH, the CDC, the Department of Health and Human Services, or the United States government.

Author contributions

The authors’ responsibilities were as follows– EAW, NA, AT, CF, and CO: designed the research and are responsible for the critical revision of the manuscript and the final content; EAW and NA: contributed to the analysis of data; EAW: drafted the manuscript and all authors: read and approved the final manuscript.

Conflict of Interest

The authors report no conflicts of interest.

Data availability

Data described in the manuscript and codebook is publicly and freely available without restriction at the NCHS’ NHANES website (https://wwwn.cdc.gov/nchs/nhanes/Default.aspx) and USDA data at DMR - Food Categories: USDA ARS. Analytic code will be made available upon request from the corresponding author.

Appendix A. Supplementary data

The following are the Supplementary data to this article: