Association of weight status with the types of foods consumed at snacking occasions among US adolescents

Abstract

Objective:

This study aimed to evaluate snack food-group composition by weight status among United States adolescents.

Methods:

Cross-sectional analysis of adolescent food-group-component intake from snacking occasions using two 24-hour dietary recalls from the 2007 through 2018 National Health and Nutrition Examination Survey (NHANES; n = 5264; 12–19 years) was conducted. ANCOVA models evaluated food intakes by BMI percentile (BMI%; normal weight [NW]: <85th BMI%; overweight [OW]: 85th-95th BMI%; and obesity [OB]: ≥95th BMI%), adjusting for energy misreporting and key covariates.

Results:

Adolescents with OB consumed greater total daily energy from snacks (mean [SE]: NW = 424 [10] kcal; OW = 527 [16] kcal; OB = 603 [22] kcal; p < 0.001) than adolescents with OW and NW. Adolescents with OW or OB consumed higher amounts of refined grains, dairy, protein, oil, solid fat, and added sugar from snacks than adolescents with NW (p < 0.05–0.001).

Conclusions:

Adolescents with OW or OB consume more calories and higher levels of overconsumed dietary components, i.e., added sugar, solid fats, and refined grains, from snacks than adolescents with NW. Age-specific snacking recommendations to inform dietary guidance are needed to prevent excess intake of overconsumed nutrients and calories.

INTRODUCTION

Adolescence is a transitional period for eating behaviors and nutritional health in which children have increased demands for energy and nutrients necessitated by rapid physical maturation as well as increased autonomy over food choices [1]. At the same time, diet quality declines markedly as children move into adolescence [2], when intakes of vegetables, fruits, dairy, and whole grains are low and intakes of refined grains, saturated fat, sodium, and added sugar are high [1, 3, 4]. Compared with other pediatric populations, United States (US) adolescents aged 12 to 19 years have the highest obesity (OB) prevalence (20.6%) [5], are more likely to continue to have OB as adults [6], share a disproportionate risk of low diet quality, and consistently have diets represented by energy-dense, nutrient-poor food [7]. These trends highlight the need to understand the eating habits and patterns that influence dietary intake and weight-related risk during this period.

Snacking is a prevalent dietary behavior during adolescence that has been suggested to contribute to poor diet quality and OB risk [8, 9]. The National Health and Nutrition Examination Survey (NHANES) 2017–2018 found that nearly 90% of US adolescents snacked daily, with intakes that contributed almost 25% of daily energy [10]. Increases in the number of daily snacking occasions as well as total energy intake from snacking have been seen among US adolescents in recent decades [11]. Whether snacks help adolescents meet nutrient needs or whether they are a source of extra and empty energy (kilocalories) that contributes to excess weight gain is debated. On one hand, adolescents obtain significant amounts of underconsumed nutrients at snacking occasions, such as 20% to 22% of dietary fiber and calcium [10]. Furthermore, some studies have suggested that snacking can promote satiety and suppress overconsumption at subsequent eating occasions, particularly when snacks include protein, fiber, and whole grains [12]. Reductions in blood glucose and insulin fluctuations may also result from increased eating frequency by lowering glycemic loads and delaying stomach emptying [9]. On the other hand, adolescents consistently have two or more snacks regardless of the number of meals consumed [13], and they preferentially consume snacks that contain higher levels of overconsumed food components (e.g., added sugar, sodium, saturated fat), particularly sugar-sweetened beverages, desserts, and salty snack foods [8]. NHANES 2017–2018, snack occasions contributed 33% to 34% of total sugar and 20% to 22% of saturated fat intake among US adolescents [10]. Other work has suggested that consumption of palatable snack foods outside of meals in the absence of hunger can contribute to overconsumption and OB [9, 11, 14, 15].

As with diet quality, the body of literature examining associations between snacking patterns and OB during adolescence is mixed [11, 16]. A review of 32 studies concluded that most studies either found no evidence of a relationship between snacking and weight status or an inverse association between snacking and weight, whereas only a smaller number of studies provided evidence of an increased risk for OB [16]. More recently, three studies [17–19], plus the present study, add to the growing body of evidence showing a positive correlation between snacking and OB. A key distinction between the literature that yielded mixed findings and recent work [16–19] that showed more consistent associations is the methodological approaches to examining the relationship. The studies that yielded inconclusive or inverse associations included smaller sample sizes and they did not account for energy misreporting. This is an important methodological consideration when examining associations between dietary intake and weight status, as weight status is a key predictor of misreporting [20]. Furthermore, research to date has provided limited insight on the parameters of snacking that might be most relevant (e.g., food-group components, snack portion size) and, instead, primarily focused on the frequency of snacking (e.g., number of snacks per day) [16]. A 2019 analysis of 2005 through 2016 NHANES data found that adolescents with overweight (OW) and OB snacked more frequently and consumed more calories per snack compared with adolescents with normal weight (NW) [19]. Whether the food-group components of foods and beverages consumed as snacks influence relationships with weight status remains unknown.

The primary aim of this study was to evaluate the association of food-group-component intakes from snacks with weight status (NW, OW, and OB) among US adolescents using NHANES data collected during the most recent decade (2007–2018). Given associations of snacking with overconsumed nutrients [8], we hypothesized that food-group-component intakes of added sugars, solid fats, and refined grains would be higher among adolescents with OW or OB than among adolescents with NW. Given significant contributions of snacking to energy intakes among US adolescents, population- representative data on the food and beverage composition of snacks and the association with OB risk will provide an evidentiary basis for the development of more specific and age-appropriate guidelines, policies, and intervention targets for snacking among adolescents [16].

METHODS

Study design and procedure

A cross-sectional secondary data analysis was conducted with data from adolescents, aged 12 to 19 years, who participated in the 2007 through 2018 NHANES. NHANES is an ongoing, nationally representative study of the nutritional and health status of the civilian, noninstitutionalized US population. NHANES design and data collection procedures have been described in detail elsewhere [21, 22]. In brief, NHANES uses a complex, multistage, probability-sampling design, with county as the primary sampling unit from which clusters of households and participants are randomly selected [23]. Data are collected by trained study staff, and participants complete an in-home interview during which sociodemographic and health-related information are collected about the household, followed by a medical examination, additional interviews, and a 24-hour dietary recall for collection of food-consumption data in the mobile examination center. A second 24-hour dietary recall is collected via phone 3 to 10 days after the initial recall. The National Center for Health Statistics (NCHS) Research Ethics Review Board approved the study protocols, and this project was determined to be exempt from review by the Temple University Institutional Review Board.

Measures

Sociodemographic characteristics

The following sociodemographic characteristics, collected by self-report using the NHANES computer-assisted personal interviewing system, were included in this analysis for descriptive purposes and as potential covariates: child age (years); child gender (male, female); child race and ethnicity (non-Hispanic White, Hispanic/Mexican-American, non-Hispanic Black, other/multiracial/non-Hispanic Asian); and head of household (HH) age, education, marital status, and income. Owing to small sample sizes, race and ethnicity categories were grouped; the “other” race category is defined as non-Hispanic races other than Asian, Black, or White, including non-Hispanic multiracial. HH education was coded as less than high school, high school/some college, and college graduate or above, whereas HH age (<40 years and ≥40 years), marital status (partnered vs. not partnered), and ratio of poverty to income (<125% vs. ≥125% of the poverty guidelines, representing above and below the poverty thresholds for determining eligibility to receive government assistance) [24] were coded as dichotomous variables.

Weight status

Standing height was measured in centimeters using a stadiometer, and weight was measured in kilograms using a digital scale by trained staff in the mobile examination center [25]. Height and weight were used to derive age- and gender-specific body mass index percentiles (BMI%) using the Centers for Disease Control and Prevention (CDC) growth charts [26]. Adolescent weight status was classified as NW (<85th percentile), OW (85th-95th percentile), and OB (≥95th percentile) [27].

Dietary intake

Adolescents’ self-reported dietary intake was collected by trained interviewers as part of the What We Eat in America survey and using the five-step Automated Multiple-Pass Method [23]. The mean of the 2 days of dietary intake was used to construct all dietary variables. Snacking was defined by self-identified eating occasions from predetermined options, including “snack” as well as “beverage” and “extended consumption,” and the Spanish equivalents (“merienda,” “bebida,” “bocadillo,” “tentempie,” and “entre comida”). Snack food and beverages consumed at the same time were aggregated into a single snack occasion, and, consistent with previous research, “trivial snacks” (i.e., snack occasions < 5 kcal) were excluded [19, 28]. Meal occasions were those identified by participants as “breakfast,” “lunch,” “dinner,” “supper,” “brunch,” or “other” and the Spanish equivalents (“desayuno,”“almuerzo,” “comida,” and “cena”). Meal occasions were used to derive mean meal size, which was defined as kilocalories per occasion and was included as a covariate in the main analyses to consider children’s non-snacking intake when evaluating associations of snacking with weight status.

The Food Patterns Equivalents Database (FPED) was used to convert the food and beverage intakes from the dietary recalls to the 37 US Department of Agriculture (USDA) food-group components [29]. Each participant’s snacking occasions from the two dietary recalls were broken down into the USDA food-pattern equivalents for each food-group component using the corresponding FPED files for each year and day of data collection. A total of 10 major food-group components were included in the present analysis: fruits, vegetables, legumes, whole grains, refined grains, dairy, protein, oils, solid fats, and added sugars. For the primary analyses, the mean food-pattern equivalents from each of the selected food-group components were calculated for snacking occasions and total daily intake. For descriptive purposes, the mean equivalents per day from each of the selected food-group components were also calculated as a percentage of total intake and as a percentage of daily recommendations for ages 12 to 19 years per the 2020 through 2025 Dietary Guidelines for Americans (DGA) healthy US-style dietary pattern [1].

Dietary reporting accuracy

To adjust for dietary reporting accuracy, the ratio of reported energy intake to estimated energy requirements (EI:EER) was calculated and included as a covariate in all models; this method has been used to adjust for misreporting without biasing sample selection by not removing improbable cases [18, 19]. EER was calculated using Dietary Reference Intake equations based on age, gender, height, weight, and physical activity [30]. Following methods of Murakami and Livingstone [18] to estimate the EI:EER ratio, a “low active” level of physical activity (≥1.4 to <1.6) was used for all adolescents [31]. Dietary 2-day sample weights were also used to adjust for dietary nonresponse and reporting day of the week [32].

Analyses and statistical methods

Individual data came from eight NHANES data set files: demographic files; body measures files (i.e., anthropometrics); Day 1 and Day 2 dietary interview individual foods files; Day 1 and Day 2 total nutrient intakes files; and diabetes and prescription drug files (used for exclusion criteria) [33]. A total of 7792 adolescents participated in the combined 2007 through 2018 cycles of NHANES. Cases without two dietary recalls (n = 1675), having a diabetes diagnosis (n = 44), using medications that impact weight status (n = 93), missing BMI data (n = 79), and missing covariate data (HH education, n = 200; mean meal size, n = 2) were excluded. Additionally, 435 cases from participants who did not report snacking occasions or who reported only snack occasions with <5 kcal were removed from the main analysis given the interest in examining intake from snacking occasions. The final analytical sample consisted of 5264 adolescents.

All analyses were conducted using R version 3.6.1. Sample sociodemographic characteristics were included for adolescents with (snackers) and without snacking occasions (non-snackers) and were expressed as percentages for categorical variables and means (standard errors [SE]) for continuous variables. SEs were estimated by Taylor Series linearization, including applicable dietary 2-day sample weights constructed using NHANES analytical guidelines for the six cycles (12 years) of NHANES data combined in this analysis, to account for the NHANES complex survey design [32]. For adolescents with snacking occasions, unadjusted mean intakes of the snack food-group components were calculated to describe snack intake. Generalized linear ANCOVA regression models examined predictive marginal means (±SE; i.e., probability-weighted averages) [34] for snack energy intake and food-group-component consumption by weight status (NW, OW, and OB). EI:EER and mean meal size were included in all ANCOVA models to adjust for reporting accuracy and relative size of meals, respectively. Backward stepwise elimination (using lowest Akaike information criterion value as the inclusion threshold for model selection) [35] and existing literature [8, 17, 19] were used to retain the following sociodemographic covariates for inclusion in all ANCOVA models: child race and ethnicity; child gender; child age; and HH education. Separate ANCOVA models were run for each of 10 main food-group components, followed by pairwise comparison to analyze between-group differences. Pairwise comparisons of mean equivalents per food-group components from snacks by weight status were performed using the Bonferroni test at the 5% level of significance to adjust for multiple comparisons; beta estimates (±SE) are presented in the text. For all tests, p < 0.05 was considered significant.

RESULTS

Sociodemographic characteristics

Sociodemographic characteristics are presented in Table 1. The sample consisted of 5264 adolescents (mean age: 15.5 [0.05] years), of whom approximately 50% were female and 54% identified as non-Hispanic White, 23% as Mexican-American or Hispanic, 14% as non-Hispanic Black, and 8% as other/multiracial or non-Hispanic Asian. In addition, 20% of adolescents had a HH with less than a high school education, about three quarters (72%) lived in a household with partnered caregivers, and close to a third (32%) had a low poverty to income ratio. The mean adolescent BMI was 24.0 (0.15), and the mean BMI z score was 0.66 (0.03); 62% of the sample were classified as NW, and the remainder were classified as having OW (16%) or OB (22%).

TABLE 1.

Sociodemographic and anthropometric characteristics of 5264 US adolescents aged 12 to 19 years reporting snacking and 435 US adolescents aged 12 to 19 years not reporting snacking and participating in NHANES 2007 through 2018

	Snackers		Non-snackers
Adolescent
Age (y), mean (SE)	15.46 (0.05)		15.54 (0.15)
BMI, mean (SE)	24.01 (0.15)		25.31 (0.47)
BMI z score, mean (SE)	0.66 (0.03)		0.91 (0.08)
Gender
Male	2644	50%	232	52%
Female	2620	50%	203	48%
Race and ethnicity
Non-Hispanic White	1482	54%	111	51%
Mexican-American/Hispanic	1774	23%	128	21%
Non-Hispanic Black	1297	14%	132	19%
Other/multiracial/non-Hispanic Asian	711	8%	64	9%
Weight status
NW	3121	62%	225	54%
OW	924	16%	75	20%
OB	1219	22%	135	26%
Head of household
Age
<40 years	1785	30%	137	28%
≥40 years	3479	70%	298	72%
Education
Less than high school	1425	20%	108	20%
High school/some college	2761	53%	249	59%
College graduate or above	1078	27%	61	21%
Marital statusa
Partnered	3377	72%	279	73%
Not partnered	1605	28%	132	27%
Poverty to income ratiob
<125%	1941	32%	158	28%
≥125%	2923	68%	232	72%

Note: Snackers reported snacking, and non-snackers did not report snacking; snacking = eating occasions self-identified as snacks or other eating between meals (beverages and extended consumption) consumed at the same time and >5 kcal; BMI-for-age percentiles: NW (<85th percentile); OW (85th-95th percentile); OB (≥95th percentile) [26]; poverty to income ratio calculated by dividing family income by the poverty guidelines specific to the survey year.

Abbreviations: NHANES, National Health and Nutrition Examination Survey; NW, normal weight; OB, obesity; OW, overweight.

^aSnackers: n = 4982; non-snackers: n = 411.

^bSnackers: n = 4864; non-snackers: n = 390.

Source: Data were derived from demographic and body measures files, NHANES 2007 through 2018, weighted.

Snack energy and composition

Mean daily snacking energy intake among adolescents was 480 (9) kcal/d, and snacking accounted for 22% (0.003%) of total daily energy intake. Mean equivalents of each food-group component from snacks and total daily intake are shown in Table 2. Contributions from snacking intakes of each food-group component are shown in Figure 1. Food-group-component intakes showed that snacking occasions contributed 37% (0.01%), 23% (0.01%), and 17% (0.004%) of total daily intakes from added sugar, solid fats, and refined grains food-group components, respectively; these intakes represented 62% (0.02%), 37% (0.01%), and 31% (0.01%) of daily DGA recommendations for added sugar, solid fats, and refined grains, respectively. Snacks also made significant contributions to adolescent intakes of fruits (31% [0.01%] of total daily intake; 17% [0.01%] of daily recommendations) and oil (23% [0.01%] of total daily intake; 17% [0.01%] of daily recommendations) while making negligible contributions to intakes of legumes, vegetables, and protein foods.

TABLE 2.

Unadjusted estimated mean daily intakes of selected food-group components from snacks for US adolescents aged 12 to 19 years reporting snacking and participating in NHANES 2007 through 2018 (N = 5264)

	Snack intake	SE	Total intake	SE
Fruits (cups/d)	0.35	0.02	0.90	0.03
Vegetables (cups/d)	0.13	0.01	1.06	0.02
Legumes (cups/d)	0.01	0.01	0.08	0.01
Whole grains (oz-eq/d)	0.17	0.01	0.79	0.02
Refined grains (oz-eq/d)	1.18	0.04	6.44	0.08
Dairy (cups/d)	0.39	0.01	1.99	0.04
Protein foods (oz-eq/d)	0.49	0.03	5.07	0.09
Oils (g/d)	5.47	0.17	23.22	0.39
Solid fats (g/d)	8.80	0.25	36.87	0.51
Added sugars (teaspoons/d)	7.73	0.22	19.46	0.29

Note: Snacking = eating occasions self-identified as snacks or other eating between meals (beverages and extended consumption) consumed at the same time and >5 kcal.

Abbreviations: NHANES, National Health and Nutrition Examination Survey; oz-eq, ounce equivalent.

Source: Data were derived from demographic, body measures, and What We Eat in America, Day 1 and Day 2 dietary intake and Food Pattern Equivalent Database (FPED) files, NHANES 2007 through 2018, weighted.

FIGURE 1.

Contributions from snacks to daily intake and recommendations by food-group component among US adolescents aged 12 to 19 years reporting snacking and participating in NHANES 2007 through 2018 (N = 5264)

Snack energy and composition by weight status

Adolescents with NW consumed 424 (10) kcal, on average, or 22% (0.004%) of total daily energy from snacks; adolescents with OW consumed 527 (16) kcal or 23% (0.01%) of total daily energy from snacks; and adolescents with OB consumed 603 (22) kcal or 24% (0.01%) of total daily energy from snacks. Pairwise analyses showed significant differences in mean daily snack energy intake by weight status (OW-NW: b = 103.2 [16.2], p < 0.001; OB-NW: b = 178.8 [22.2], p < 0.001; OB-OW: b = 75.5 [19.3], p < 0.001), but not in mean percent of daily energy from snacks (OW-NW: b = 0.01 [0.01], p = 0.21; OB-NW: b = 0.02 [0.01], p = 0.08; OB-OW: b = 0.01 [0. 01], p = 1.00).

Predictive marginal means and pairwise comparisons of mean equivalents per food-group component from snacks by weight group are shown in Table 3. Snacking intake for several food-group components differed by weight status. Compared with adolescents with NW, adolescents with OW or OB consumed significantly greater amounts of refined grains (OW: b = 0.3 [0.1] ounce equivalents [ozeq]/d, p < 0.001; OB: b = 0.5 [0.1] oz-eq/d, p < 0.001); dairy (OW: b = 0.1 [0.04] cups/d, p = 0.01; OB: b = 0.2 [0.04] cups/d, p < 0.001); protein (OW: b = 0.1 [0.1] oz-eq/d, p = 0.01; OB: b = 0.2 [0.1] oz-eq/d, p = 0.04); oil (OW: b = 1.5 [0.4] g/d, p < 0.001; OB: b = 2.2 [0.4] g/d, p < 0.001); solid fats (OW: b = 2.2 [0.5] g/d, p < 0.001; OB: b = 3.5 [0.7] g/d, p < 0.001); and added sugar (OW: b = 1.8 [0.5] teaspoons/d, p < 0.001; OB: b = 2.9 [0.5] teaspoons/d, p < 0.001) from snacks. Apart from solid fats (OW-NW: b = 2.2 [0.5] g/d, p < 0.001; OB-NW: b = 3.5 [0.7] g/d, p < 0.001; OB-OW: b = 1.4 [0.5] g/d, p = 0.01), food-group-component intakes from snacks showed few differences between adolescents with OW and OB.

TABLE 3.

Estimated mean daily intakes of selected food-group components from snacks, by weight group, for US adolescents aged 12 to 19 years reporting snacking and participating in NHANES 2007 through 2018 (N = 5264)

	Snack intake			Total intake
	NW	OW	OB	NW	OW	OB
Fruits (cups/d)	0.34a	0.34a	0.39a	0.89a	0.88a	0.97a
Vegetables (cups/d)	0.12a	0.13a	0.16a	0.89b	0.95ab	1.04a
Legumes (cups/d)	0.01a	0.02a	0.03a	0.07a	0.08a	0.10a
Whole grains (oz-eq/d)	0.16a	0.18a	0.20a	0.75b	0.78ab	0.91a
Refined grains (oz-eq/d)	1.03b	1.31a	1.49a	6.16b	6.75a	7.02a
Dairy (cups/d)	0.34b	0.45a	0.50a	1.86b	2.17a	2.24a
Protein foods (oz-eq/d)	0.42b	0.56a	0.60a	4.85b	5.29ab	5.52a
Oils (g/d)	4.74b	6.23a	6.97a	21.51c	23.91b	27.50a
Solid fats (g/d)	7.67c	9.84b	11.21a	34.37b	39.83a	41.74a
Added sugars (teaspoons/d)	6.79b	8.62a	9.71a	17.77b	21.27a	22.90a

Note: ANCOVA models were adjusted for survey cycle year, child gender, child race and ethnicity, child age, energy intake to estimated energy requirements, head of household education, and mean meal size; means followed by a common letter are not significantly different by the Bonferroni test at the 5% level of significance; BMI-for-age percentiles: NW (<85th percentile); OW (85th-95th percentile); OB (≥95th percentile) [26]; snacking = eating occasions self-identified as snacks or other eating between meals (beverages and extended consumption) consumed at the same time and >5 kcal.

Abbreviations: NHANES, National Health and Nutrition Examination Survey; NW, normal weight; OB, obesity; OW, overweight; oz-eq, ounce equivalent.

Source: Data were derived from demographic, body measures, What We Eat in America, Day 1 and Day 2 dietary intake and Food Pattern Equivalent Database (FPED) files, NHANES 2007 through 2018, weighted.

DISCUSSION

The objective of this study was to examine the differences in food-group-component intakes from snacks by weight status among US adolescents aged 12 to 19 years using the most recent decade of NHANES data. For all adolescents, top food-group-component sources of snacking came from fruits, refined grains, oils, solid fats, and added sugars. Furthermore, adolescents with OW or OB consumed greater amounts of refined grains, dairy, protein, oils, solid fats, and added sugar food-group components from snack occasions compared with adolescents with NW. Results also showed that adolescents with OB consumed greater daily energy from snacks compared with adolescents with NW. Taken together, these findings suggest that snacking is potentially associated with excessive dietary intakes from snacks that might be associated with weight status in this age group.

Findings from this study advance previous work that has shown associations of snacking frequency and size with weight status [19] by providing new evidence that snacking intakes of overconsumed food-group components, specifically added sugar, solid fats, and refined grains, also vary by weight status. Results suggest that, in addition to increased frequency and larger size of daily snacks seen among adolescents with OW or OB [19], the types of foods consumed as snacks may have implications for weight status. Food-group components consumed in excess may contribute to excess energy intake and weight gain as well as displace recommended foods and nutrients that are important to overall health [8, 12, 17].

Additionally, this study adds to a growing body of evidence showing positive associations between snacking and OB risk. Compared with work that has shown a protective effect of snacking on weight status or no association [11, 16], this study leverages a large, nationally representative sample and addresses key methodological considerations, including accounting for energy misreporting and defining snacking by total energy and food-group components. Characterizing snacks by food-group components as opposed to by general types (e.g., sweets, savory snacks) allows for a more accurate and comprehensive investigation of what snacks are contributing to intake and how that is associated with weight status. Although future work is warranted to further elucidate the mechanisms underlying snacking and weight status, the present findings add to the body of literature identifying snacking as a behavioral target for addressing excessive consumption of overconsumed nutrients in this population.

Consistent with previous work, this research shows that snacking makes significant contributions to daily intakes of energy and can impact diet quality [36, 37]. Studies have suggested that, overall, snacking has the potential to increase diet quality and assist children in meeting national recommendations but is dependent on the quality of the snacks consumed [38–40]. In previous analyses of NHANES data (2001–2004 and 2005–2006, children aged 12–19 years), snacks provided 11% to 39% of daily intakes for the main MyPyramid food groups (grains, fruits, vegetables, milk, meat/beans, and oils) while providing 34% to 43% of added sugar and 20% to 24% of solid fat intake [36, 37]. Our findings reveal similar intake patterns, with snacks providing low intakes (3%–11%) of legumes, protein, and vegetables versus substantial intakes (17%–31%) of fruit, grain, dairy, and oil food-group components, and extend previously reported contributions by showing that these food-group-component intakes from snacks provide 1% to 17% of recommendations. Despite potentially valuable contributions to some aspects of diet quality, particularly fruit intake, our findings also show that snacks provide notable contributions of overconsumed food-group components. Added sugars, solid fats, and refined grains contribute 37%, 23%, and 17%, respectively, of daily intakes and meet one- to two-thirds of national recommendations, showcasing the disproportionate consumption from snacks alone among US adolescents.

Recommendations in the 2020–2025 DGA reflect a shift away from a focus on nutrients to a focus on dietary patterns and behavioral and lifestyle changes [1]. Eating patterns that follow dietary guidelines have been associated with a lower risk of mortality, which has led to studies evaluating associations between various eating patterns and diet-related health outcomes such as weight status [41]. Current DGA guidelines recommend limits on added sugar, solid fats, and refined grains but do not include specific guidance on the role of snacking, particularly for different life stages. The findings of this study indicate that snacking represents a key source of overconsumed nutrients in the diets of US adolescents. Moreover, this work highlights that snacking should be considered as part of the age-specific recommendations for healthy dietary patterns in adolescents.

Study findings should be interpreted in the context of a few notable limitations. First, causality cannot be inferred from this cross-sectional analysis; the current analysis did not consider important behavioral or environmental determinants of snack food and beverage choices. Second, despite being the current gold-standard approach for large epidemiological studies, dietary recalls are subject to well-characterized sources of random and systematic measurement error, including reporting/recall bias [23]. Foods consumed during snacking occasions compared with main meals are more likely to be forgotten or excluded, leading to underreported energy intake, which is common among adolescents, especially among those in higher BMI categories [20]. Adolescents with OW and OB tend to underreport dietary intake as well as restrict eating patterns in the form of dieting or irregular meal consumption, whereas those with NW tend to overreport energy intake [20]. This analysis included estimated reporting bias in models to take these biases into consideration [18]. Last, the use of 2 days of dietary recall reduces sample size but follows recommended approaches for examining snacking-pattern associations by weight status [42]; 2-day dietary weights were included in models to adjust for dietary nonresponse on the second day.

CONCLUSION

In conclusion, these findings provide population-representative evidence that US adolescents, aged 12 to 19 years, with OW or OB consumed greater daily energy from snacks and higher amounts of overconsumed food-group components from snacks compared with those with NW. Specifically, snacking contributed significant amounts of added sugars, solid fats, and refined grains to the diets of US adolescents. Moreover, the dietary components of snacks consumed varied by weight status; adolescents with OW or OB consumed more total calories and overconsumed nutrients from snacks than NW adolescents. To our knowledge, outside of the context of broader weight management interventions, there are few studies that target the key dimensions of snacking (e.g., frequency, snack size, food-group components), and more comprehensive intervention efforts are warranted. Efforts to improve snacking by providing clear and age-specific recommendations for the types and amounts of foods and beverages that are served as snacks can encourage consumption patterns that are more consistent with national recommendations. These findings warrant further inquiry, particularly using longitudinal and experimental designs, to understand temporal and causal influences of snacks on diet quality and weight status among US adolescents.

Study Importance.

What is already known?

• Current literature examining associations of snacking and obesity during adolescence is mixed.

• Recent analysis of population-representative data has shown that more frequent daily snacks and larger snacks (more calories per snack) were associated with overweight and obesity in adolescents.

What does this study add?

• Snacking contributes significantly to adolescent intake of overconsumed food-group components, specifically added sugar, solid fats, and refined grains, and these associations vary by weight status.

How might these results change the direction of research or the focus of clinical practice?

• These results might change the focus of research or clinical practice to include age-specific recommendations on the types of foods consumed as snacks and their implications for obesity risk and excessive consumption of overconsumed nutrients in adolescents.

• These results highlight the need for population-representative research on composition of snacks to provide an evidentiary basis for the development of guidelines, policies, and intervention targets for snacking among adolescents.