Abstract
Understanding the relationship between snacking and dietary intake in early life years is one key but understudied area. In this study, we examined snacking patterns in toddlers and preschool children and the associations between snacking frequency and daily energy intake. We analyzed data from children aged 12–72 months (N=1186) in the Newborn Epigenetic STudy (NEST). We used Bonferroni multiple comparison methods to examine the differences in snacking patterns across subgroups. Linear and quantile regression models were fit to investigate the association between dietary intake and snacking frequency. Our estimates suggest that Non-Hispanic blacks had the highest total daily energy intake from snacks (334 kcal/day) compared to non-Hispanic whites (270 kcal/day) and Hispanics (274 kcal/day) in 12-to-24-month-olds. In 2-to-6-year-olds, mean energy intake from snacks was 296 kcal/day without a significant racial/ethnic difference. Carbohydrate, fat and protein from snacks contributed about 17%, 9% and 4% respectively of the total energy intake in 12-to-24-month-olds while they contributed about 15%, 7% and 2% respectively of the total energy intake in the other age group. Snacking frequency was positively and significantly associated with total daily energy intake in both 12-to-24-month-olds and 2-to-6-year-olds as indicated by regression coefficient estimates of snacking frequency (β = 31.3 kcal/day with P = 0.027 and β = 175.4 kcal/day with P<0.0001, respectively, indicating a higher snacking frequency was associated with a greater total daily energy intake). In conclusion, snacking frequency was positively associated with daily energy intake. Carbohydrates and fats from snacks are significant energy contributors. Age differentiation was apparent regarding the relationship between snacking frequency and dietary intake. Differentiated interventions that are age-specific and focus on the dietary quality of snacks instead of quantity are needed.
Keywords: Dietary intake, Diet quality, Preschooler, Snacking, Toddler
INTRODUCTION
Over 1/3 of US children are now overweight/obese (Hales, Carroll, Fryar, & Ogden, 2017; Cynthia L Ogden, et al., 2016). Studies have shown that the incidence of obesity is more likely to occur at younger ages and the relative risk of adult obesity decreases with increased age of childhood obesity onset (Cunningham, Kramer, & Narayan, 2014; Ward, et al., 2017), suggesting that early childhood is a critical life stage to target for effective childhood and adult obesity prevention. Understanding the relationship between snacking which are eating occasions outside of main meals and dietary intake in early life years is one key area in such prevention efforts, given the significant increase of the prevalence of snacking and the portion of total energy intake from snacking in US children over the past four decades (E. Dunford & Popkin, 2017).
Nevertheless, most of the existing studies have focused on school-aged children and adolescents, for example, children aged 4–5 years (Rudy, et al., 2017), 6–19 years (Murakami & Livingstone, 2016), 9–14 years (Field, et al., 2004), fourth- and eighth-graders (Vader, Walters, Harris, & Hoelscher, 2009), and 12–18 years (Keast, Nicklas, & O’neil, 2010). There is one recent study that examined the association between snacking frequency and weight status among toddlers and preschoolers (aged 12–71 months) (Fisher, Davey, Kachurak, & Bailey, 2017) and two studies that used the Guelph Family Health Study data to examine the patterns of snacking (Hutchinson, et al., 2017) and effects of genetics of taste on snacking patterns of children aged 1.5–5 years (Chamoun, et al., 2018). However, in general, studies about young children, especially under 2 years of age, are lacking. Infants are usually introduced to solid foods after six months, and the American Academy of Pediatrics recommends to give 2–3 healthy snacks per day for children 12–36 months (American Academy of Pediatrics, 2017). US national data suggests that about 8% of children under the age of two have high weight-for-height (at or above 97.7th percentile of World Health Organization 2006 growth charts) (C. L. Ogden, Carroll, Kit, & Flegal, 2014), an indicator of excess weight and high risk for overweight and obesity in later life stages (Ward, et al., 2017). Few studies have examined snacking in young children and how this relates to their overall calorie intake and diet quality, and the lack of relevant research in this age group contributes to the dearth of evidence from which to mount interventions focused on snacking.
Racial/ethnic differences in the prevalence of overweight or obesity are evident in early childhood. Between birth to 2 years, recent National Health and Nutrition Examination Survey (NHANES) data suggest that 8.8% of Hispanic children are at or above the 97.7th percentile of WHO 2006 growth charts compared to 7.3% and 5.5% in Black children and White children respectively (C. L. Ogden, et al., 2014). Similarly, in 2-to-5-year-olds, 16.7% of Hispanic children have a body mass index (BMI) at or above the 95th percentile of CDC 2000 growth charts, followed by Black (11.3%) and White children (3.5%) (C. L. Ogden, et al., 2014). The differences in energy balance-related behaviors including dietary intake between racial/ethnic groups from birth to two years of age could be potential risk factors (Bernard F. Fuemmeler, et al., 2015). As snacking has become an increasingly important component of children’s diet, a good understanding of the snacking patterns and relevant racial/ethnic differences among young children is urgently needed to guide caregivers’ practice and inform interventions to reduce racial/ethnic disparities.
In addition, the association estimates from existing studies are primarily based on conditional mean regression. However, such methods are not sufficient because the effects of explanatory variables at different levels of dietary intake may have very different implications (Kan & Tsai, 2004). For example, positive association between snacking and fat intake at the upper quantiles of the fat intake distribution may suggest snacking as a risk factor for obesity, while such relationship at the lower quantiles may be considered as protective for underweight. Therefore, results reported in the prior studies may only provide a partial view of snacking.
We need a better understanding of the role that snacking plays in early childhood to inform effective interventions for healthy child growth and childhood obesity prevention. Previous studies that have looked at this relationship have had mixed findings with methodological limitations. To fill these gaps, using data from the Newborn Epigenetic STudy (NEST) cohort, this study 1) examined the snacking patterns of children 1–6 years of age and related gender, and racial/ethnic disparities, 2) assessed the association between snacking frequency and total energy intake 3) employed both conditional mean regression and quantile regression models to assess the differential impacts of snacking frequency on macronutrient and food intake.
METHODS
Data
Data collected from the participants in the Newborn Epigenetic STudy (NEST) was used. NEST is an ongoing prospective study of women and their children designed to identify early exposures associated with stable epigenetic alterations in infants that may alter chronic disease susceptibility later in life. The participants were recruited in three contiguous counties in central North Carolina (NC); Durham, Orange and Wake. The Institutional Review Board approved this study, and informed written consent was obtained from all participants. Participant identification and enrollment procedures are described elsewhere (Hoyo, et al., 2011). We used the data collected in the first wave of data collection in order to capture early life dietary patterns. Our sample included 1186 children aged between 12 months and 6 years with dietary assessment data. Non-Hispanic whites, non-Hispanic blacks, and Hispanics accounted for 35%, 42%, and 19% of our sample, respectively. 40% of the mothers had a high school or less education, while more than half had a college or some college education.
Technical information
Dietary Intake and Snacking Frequency:
The primary caregiver (in most cases the mother) reported on their child’s diet. The dietary intake of the children was assessed by telephone, using the Nutrition Data System for Research (NDSR, University of MN), a valid and established method for assessing energy intake(B. F. Fuemmeler, Lovelady, Zucker, & Ostbye, 2013). Snacking frequency was defined as eating occasions that were reported by parents as “snacks” based upon a defined list in the NDSR system. The list includes meal names in each eating occasion such as Breakfast, Brunch, Lunch, Dinner/supper, Snack, Beverage only (just a drink), and School lunch. The American Academy of Pediatrics recommends 2–3 healthy and nutritious snacks per day for infants (after 9 months) and toddlers (12–36 months), and 2 healthy and nutritious snacks per day for preschoolers (36 – 60 months) (American Academy of Pediatrics, 2017).
Demographic and socioeconomic data:
At the enrollment survey, parents reported on their level of educational attainment and age. Children’s race/ethnicity was based on maternal race/ethnicity.
Statistical analysis
We conducted descriptive analysis and used Bonferroni multiple comparison methods to examine the differences in snacking patterns across gender and racial/ethnic groups. Linear regression was used to generate conditional mean estimates of the mean effects of snacking frequency on dietary intake for both age groups with adjustment of age, gender, maternal education level. Previous studies have suggested that snacking behaviors may differ by age and gender (Evans, Jacques, Dallal, Sacheck, & Must, 2015; Wang, van der Horst, Jacquier, Afeiche, & Eldridge, 2018). For example, older children may make less healthy choices about snacking, consuming more energy from savory snacks than younger children (Barquera, et al., 2010) and consuming fewer fruits (Wang, et al., 2018). NHANES data suggested that female toddlers consumed slightly more macronutrients from snack occasions, though the differenced were not significant (U.S. Department of Agriculture, 2016). NHANES data also suggested that race/ethnicity was associated with sweet snack and salty snack consumption (Bleich & Wolfson, 2015). Maternal education is also associated with feeding style and children’s snacking patterns (Saxton, Carnell, Van Jaarsveld, & Wardle, 2009; Wijtzes, et al., 2013) Therefore, we included these variables as covariates in our analyses. Sensitivity analyses were conducted to examine the robustness of estimates by controlling the percentage energy of fat, carbohydrate, and protein from snacks. Quantile regression models with same-model specifications as linear regression models were fit to assess the differential impacts of snacking at different points in the dietary intake distributions (Koenker & Bassett Jr, 1978).
RESULTS
Snacking patterns
Table 1 through linear regression with maternal education, gender and age controlled. As shown in Table 1, on average, children 12 to 24 months old had snacks 2.6 times per day as compared to 2.2 times per day in children aged 2 to 6 years. There is no significant difference between boys and girls. In Bonferroni multiple comparisons across racial/ethnic groups, maternal education, gender and age were adjusted by producing conditional mean estimates from linear regression with these variables controlled. Non-Hispanic whites had the highest snacking frequency, followed by Hispanics, and non-Hispanic blacks in children aged 2–6 years. There is no significant racial/ethnic gradient in 12-to-24-month-olds.
Table 1:
Snacking Patterns by Age, Gender, and Ethnicity
| Gender | Ethnicity | ||||||
|---|---|---|---|---|---|---|---|
| Boys (N=639) | Girls (N=547) | P-value | Non-Hispanic White (N=418) | Non-Hispanic Black (N=494) | Hispanic (N=220) | Significant group comparisons† | |
| Mean (SD) or % | Mean (SD) or % | Mean (SD) or % | Mean (SD) or % | Mean (SD) or % | |||
| 12-to-24 months (N=769) | |||||||
| Total daily energy intake (Kcal/day) | 1066.39(373.75) | 1003.49(377.34) | 0.021 | 1002.74(301.9) | 1178.44(402.42) | 917.83(377.14) | Black>White, Black>Hispanic* |
| Snacking frequency /day | 2.68(1.61) | 2.55(1.47) | 0.244 | 2.74(1.54) | 2.47(1.32) | 2.52(1.68) | NS‡ |
| Snack energy (Kcal/day) | 306.58(206.65) | 282.48(194.59) | 0.099 | 270.31(166.28) | 334.44(216.93) | 274.05(214.68) | Black>White, Black> Hispanic* |
| Snack Fat (g/day) | 10.49(9.03) | 10.23(8.36) | 0.681 | 10(7.73) | 11.91(9.42) | 8.57(8.53) | Black>Hispanic* |
| Snack Carbohydrate (g/day) | 44.92(30.3) | 39.79(28.89) | 0.017 | 37.45(24.19) | 48.26(31.85) | 41.64(32.58) | Black>White* |
| Snack Protein (g/day) | 9.58(8.17) | 9.04(7.16) | 0.342 | 8.93(6.68) | 9.77(7.8) | 9.27(8.96) | NS |
| Snack Sugar (g/day) | 29.51(21.12) | 26.21(19.89) | 0.027 | 24.5(17.82) | 31.27(20.76) | 28.3(23.44) | Black>White* |
| Snack Fruit (cup/day) | 0.54(0.89) | 0.42(0.7) | 0.053 | 0.43(0.72) | 0.39(0.69) | 0.68(1.07) | Hispanic>White, Hispanic>Black* |
| Snack Vegetable (cup/day) | 0.1(0.27) | 0.07(0.23) | 0.067 | 0.05(0.19) | 0.09(0.28) | 0.12(0.29) | Hispanic>White◦ |
| % total fat from snack | 28% | 28% | 0.64 | 28% | 28% | 27% | NS |
| % total carbohydrate from snack | 31% | 30% | 0.196 | 28% | 30% | 34% | Hispanic>White* |
| % total protein from snack | 24% | 24% | 0.883 | 23% | 23% | 25% | NS |
| % total sugar from snack | 35% | 35% | 0.496 | 33% | 35% | 37% | NS |
| % total fruit from snack | 35% | 33% | 0.517 | 26% | 31% | 50% | Hispanic>White, Hispanic>Black* |
| % total vegetable from snack | 10% | 7% | 0.159 | 5% | 9% | 12% | Hispanic>White◦ |
| % total energy from snacks | 29% | 28% | 0.525 | 27% | 28% | 30% | NS |
| % total energy intake from snack fat | 9% | 9% | 0.651 | 9% | 9% | 8% | NS |
| % total energy from snack carbohydrate | 17% | 16% | 0.126 | 15% | 16% | 18% | Hispanic>White* |
| % total energy from snack protein | 4% | 4% | 0.768 | 4% | 3% | 4% | NS |
| 2–6 years (N=417) | |||||||
| Total daily energy intake (Kcal/day) | 1610.18(467.58) | 1479.35(491.37) | 0.006 | 1517.48(484.07) | 1607.56(476.72) | 1312.01(509.82) | NS |
| Snacking frequency /day | 2.21(1.14) | 2.2(1.12) | 0.896 | 2.62(1.12) | 1.89(1.02) | 2.10(0.85) | White>Black, White>Hispanic* |
| Snack energy (g/day) | 395.32(240.6) | 379.17(277.76) | 0.525 | 402.1(222.66) | 385.93(288.25) | 298.24(184.1) | NS |
| Snack Fat (g/day) | 13.28(10.74) | 13.49(13.26) | 0.857 | 13.11(9.77) | 13.81(13.54) | 9.5(8.13) | NS |
| Snack Carbohydrate (g/day) | 61.61(38.67) | 57.31(35.82) | 0.24 | 64.11(36.11) | 57.84(39.27) | 48.46(27.98) | NS |
| Snack Protein (g/day) | 9.73(8.24) | 9.29(11.06) | 0.643 | 9.66(6.91) | 9.61(11.6) | 6.64(5.02) | NS |
| Snack Sugar (g/day) | 35.73(26.94) | 32.63(21.35) | 0.195 | 38.37(25.09) | 32.17(24.35) | 28.83(20.55) | NS |
| Snack Fruit (cup/day) | 0.57(1.19) | 0.44(0.73) | 0.202 | 0.61(1.3) | 0.39(0.72) | 0.68(0.67) | NS |
| Snack Vegetable (cup/day) | 0.11(0.34) | 0.09(0.29) | 0.553 | 0.09(0.29) | 0.11(0.35) | 0.08(0.21) | NS |
| % total fat from snack | 23% | 24% | 0.32 | 24% | 22% | 20% | NS |
| % total carbohydrate from snack | 28% | 28% | 0.936 | 30% | 26% | 29% | White>Black* |
| % total protein from snack | 16% | 16% | 0.666 | 17% | 15% | 14% | NS |
| % total sugar from snack | 32% | 32% | 0.962 | 35% | 29% | 37% | White>Black* |
| % total fruit from snack | 28% | 28% | 0.999 | 26% | 25% | 57% | Hispanic>White, Hispanic>Black* |
| % total vegetable from snack | 6% | 7% | 0.552 | 6% | 6% | 7% | NS |
| % total energy from snacks | 24% | 25% | 0.791 | 26% | 23% | 23% | White>Black* |
| % total energy intake from snack fat | 7% | 8% | 0.442 | 8% | 7% | 6% | NS |
| % total energy from snack carbohydrate | 15% | 15% | 0.978 | 17% | 14% | 16% | White>Black* |
| % total energy from snack protein | 2% | 2% | 0.686 | 3% | 2% | 2% | NS |
Note:
Bonferroni multiple comparisons.
The difference is still significant (significance level = 0.05) after controlling for maternal education level, gender and age.
The difference became insignificant (significance level = 0.05) after controlling for maternal education level, gender and age.
Non-significant.
In 12-to-24-month-olds non-Hispanic Blacks had the highest energy intake from snacks (334 kcal/day) compared to Whites (270 kcal/day) and Hispanics (274 kcal/day). The difference was statistically significant (significance level = 0.05) after controlling for age, gender, and socioeconomic status indicated by maternal education. There was no significant difference between Hispanics and Whites in energy intake from snacks. In children aged 2 to 6 years, mean energy intake from snacks was about 387 kcal/day, with no statistically significant differences across gender and racial/ethnic groups. Expressed as a percentage of total daily energy intake, approximately 29% and 25% of the total daily energy intake was from snacks in children aged 12 to 24 months and 2 to 6 years, respectively.
Carbohydrate from snacks was the major source of snack energy. The 2-to-6-year-olds consumed an average of 59.6 g/day of carbohydrate from snacks, contributing about 15% of their total daily energy intake. In the 12-to-24-month-olds, there is significant gender difference, with boys having a higher mean snack carbohydrate intake compared to girls (44.9 g/day vs. 39.8 g/day, P =0.017). Mean daily fat intake from snacks was about 10–13 g/day in both age groups, accounting for 7%−9% of total daily energy intake. Protein from snacks only contributed about 2–4% of the total daily energy. The non-Hispanic Black group had the highest fat intake from snacks, compared to whites and Hispanics.
Associations between snacking frequency and dietary intake
Table 2 depicts the linear regression estimates of the relationships between snacking frequency and dietary intake controlled for age, gender, ethnicity, and maternal education level. In the group of 12-to-24-month-olds, snacking frequency was positively associated with total daily energy intake. One more eating occasion defined as snacking was associated with 31.3 kcal/day total energy increase (P =0.027). There was also a positive association between snacking frequency and total carbohydrate intake in boys (β = 7.1 g/day, P = 0.011). On the other hand, a higher snacking frequency was also associated with a higher fruit intake (β = 0.18 cup/day, P = 0.013) in 12-to-24-month boys only. In addition, maternal education level was not significantly associated with snacking outcomes based on the regression.
Table 2:
Conditional Mean Regression Estimates of the Association between Dietary Intake and Snacking Frequency
| Full Sample | Boys | Girls | ||||
|---|---|---|---|---|---|---|
| Coefficient of snacking frequency | P-value | Coefficient of snacking frequency | P-value | Coefficient of snacking frequency | P-value | |
| 12 -to-24 months | ||||||
| Total daily energy (kcal/day) | 31.29 | 0.027 | 41.19 | 0.029 | 21.4 | 0.329 |
| Daily carbohydrate intake (g/day) | 6.04 | 0.004 | 7.13 | 0.011 | 4.83 | 0.129 |
| Daily fat intake (g/day) | 0.64 | 0.304 | 0.86 | 0.29 | 0.46 | 0.636 |
| Daily protein intake (g/day) | 0.41 | 0.535 | 1.26 | 0.153 | −0.4 | 0.698 |
| Daily Sugar intake (g/day) | 3.36 | 0.008 | 4.04 | 0.02 | 2.43 | 0.204 |
| Daily Fruit intake (cup/day) | 0.14 | 0.003 | 0.18 | 0.013 | 0.08 | 0.193 |
| Daily Vegetable intake (cup/day) | −0.02 | 0.658 | −0.03 | 0.574 | 0 | 0.937 |
| 2–6 years | ||||||
| Total daily energy (kcal/day) | 175.41 | <0.0001 | 165.22 | 0.001 | 181.75 | 0.001 |
| Daily carbohydrate intake (g/day) | 23.17 | <0.0001 | 24.31 | <0.0001 | 21.76 | 0.003 |
| Daily fat intake (g/day) | 7.24 | <0.0001 | 6.29 | 0.008 | 7.79 | 0.002 |
| Daily protein intake (g/day) | 5.01 | 0.002 | 3.61 | 0.078 | 6.43 | 0.017 |
| Daily Sugar intake (g/day) | 12.73 | <0.0001 | 16.14 | <0.0001 | 8.64 | 0.027 |
| Daily Fruit intake (cup/day) | 0.24 | 0.026 | 0.28 | 0.087 | 0.21 | 0.147 |
| Daily Vegetable intake (cup/day) | 0.09 | 0.285 | 0.06 | 0.665 | 0.11 | 0.363 |
Note: Separated regression models were fit with total daily energy intake, daily carbohydrate intake, daily fat intake, daily protein intake, daily sugar intake, daily fruit intake, and daily vegetable intake as dependent variables, and snacking frequency as independent variables, controlled for age, gender, race/ethnicity, and maternal educational level.
Similarly, in the age group of 2 to 6 years, snacking frequency was also positively associated with total daily energy intake. One more eating occasion defined as snacking was associated with about 175 kcal more total energy intake per day (P < 0.0001). The association was significant in both boys (β = 165.2 kcal/day, P = 0.001) and girls (β = 181.8 kcal/day, P = 0.001). Daily carbohydrate and fat intakes were significantly positively associated with snacking frequency with one more snacking occasion contributing about 23.2 g/day and 7.2 g/day of carbohydrate and fat respectively. These associations were significant in both gender groups. Due to a small portion of the sample that had a snacking frequency greater than 5 times a day (<5%), we were not able to examine if such negative association existed in snackers with a snacking frequency much higher than the recommended levels.
In the descriptive analyses, among the 2-to-6-year-olds, Non-Hispanic Whites had the highest snacking frequency, while energy intake was not significantly different across racial/ethnic groups. These estimates suggest evident differences across race/ethnic groups. Our regression estimates further indicate the within-race/ethnic-group relationship between snacking frequency and dietary intake. As the estimates suggest, a higher snacking frequency was associated with a higher energy intake in each racial/ethnic group.
Quantile estimates of the association between snacking frequency and dietary intake
In addition to the conditional mean regression to evaluate the relationships between snacking frequency and dietary intake, we also conducted quantile regression to analyze the associations across the whole intake distributions (Table 3).
Table 3:
Quantile Regression Estimates of the Association between Dietary Intake and Snacking Frequency
| Full Sample | Boys | Girls | |||||
|---|---|---|---|---|---|---|---|
| Coefficient of Snacking Frequency | P value | Coefficient of Snacking Frequency | P value | Coefficient of Snacking Frequency | P value | ||
| 12 -to-24 months | |||||||
| Total daily energy (kcal/day) | β at 25th | −1.63 | 0.927 | 4.51 | 0.844 | −10.24 | 0.748 |
| β at 50th | 36.71 | 0.04 | 18.41 | 0.456 | 38.57 | 0.217 | |
| β at 75th | 30.42 | 0.134 | 54.01 | 0.065 | −13.9 | 0.651 | |
| Daily carbohydrate intake (g/day) | β at 25th | 1.81 | 0.438 | 3.08 | 0.343 | −2.88 | 0.482 |
| β at 50th | 3.24 | 0.203 | 2.85 | 0.445 | 2.82 | 0.512 | |
| β at 75th | 7.41 | 0.031 | 9.86 | 0.034 | 5.6 | 0.263 | |
| Daily fat intake (g/day) | β at 25th | 0.16 | 0.865 | 0.08 | 0.948 | −0.25 | 0.86 |
| β at 50th | 0.84 | 0.3 | 0.23 | 0.834 | 0.97 | 0.507 | |
| β at 75th | 0.92 | 0.245 | 0.77 | 0.472 | 1.34 | 0.318 | |
| Daily protein intake (g/day) | β at 25th | 0.39 | 0.663 | 0.16 | 0.901 | −0.27 | 0.86 |
| β at 50th | 0.58 | 0.536 | 1.65 | 0.224 | −0.64 | 0.663 | |
| β at 75th | 1.36 | 0.125 | 1.8 | 0.155 | 0 | >0.999 | |
| Daily Sugar intake (g/day) | β at 25th | 1.39 | 0.299 | 4.34 | 0.02 | 0.77 | 0.72 |
| β at 50th | 2.39 | 0.141 | 2.51 | 0.251 | 0 | 0.81 | |
| β at 75th | 4.03 | 0.038 | 4.46 | 0.09 | 3.04 | 0.346 | |
| Daily Fruit intake (cup/day) | β at 25th | 0 | >0.999 | 0.09 | 0.171 | 0 | >0.999 |
| β at 50th | 0.06 | 0.255 | 0.15 | 0.034 | −0.05 | 0.498 | |
| β at 75th | 0.13 | 0.11 | 0.2 | 0.099 | 0.14 | 0.218 | |
| Daily Vegetable intake (cup/day) | β at 25th | 0.01 | 0.73 | 0.01 | 0.899 | 0 | 0.934 |
| β at 50th | −0.02 | 0.65 | −0.06 | 0.382 | −0.01 | 0.929 | |
| β at 75th | −0.05 | 0.397 | −0.05 | 0.503 | 0.03 | 0.671 | |
| 2–6 years | |||||||
| Total daily energy (kcal/day) |
β at 25th | 103.11 | 0.019 | 61.53 | 0.38 | 111.85 | 0.089 |
| β at 50th | 136.17 | 0.004 | 142.18 | 0.043 | 133.97 | 0.081 | |
| β at 75th | 186.02 | <0.0001 | 205.54 | 0.003 | 210.33 | 0.019 | |
| Daily carbohydrate intake (g/day) | β at 25th | 19.92 | 0.002 | 21.66 | 0.028 | 23.6 | 0.011 |
| β at 50th | 16.5 | 0.016 | 23.88 | 0.025 | 15.37 | 0.141 | |
| β at 75th | 14.13 | 0.042 | 18.73 | 0.112 | 22.48 | 0.044 | |
| Daily fat intake (g/day) | β at 25th | 4.87 | 0.037 | 6.06 | 0.088 | 3.38 | 0.313 |
| β at 50th | 8.06 | <0.0001 | 7.16 | 0.033 | 6.98 | 0.068 | |
| β at 75th | 8.22 | <0.0001 | 8.25 | 0.007 | 10.77 | 0.018 | |
| Daily protein intake (g/day) | β at 25th | 3.83 | 0.053 | 0.46 | 0.886 | 3.74 | 0.26 |
| β at 50th | 3.52 | 0.107 | 4.38 | 0.177 | 1.7 | 0.626 | |
| β at 75th | 5.38 | 0.035 | 5.05 | 0.128 | 4.62 | 0.202 | |
| Daily Sugar intake (g/day) | β at 25th | 10.38 | 0.003 | 12.54 | 0.018 | 7.47 | 0.162 |
| β at 50th | 13.36 | 0.002 | 15.56 | 0.011 | 9.74 | 0.122 | |
| β at 75th | 12.03 | 0.014 | 16.15 | 0.034 | 8.06 | 0.214 | |
| Daily Fruit intake (cup/day) | β at 25th | 0.03 | 0.78 | 0 | >0.999 | 0 | >0.999 |
| β at 50th | 0.21 | 0.087 | 0.41 | 0.029 | 0.12 | 0.523 | |
| β at 75th | 0.49 | 0.001 | 0.49 | 0.014 | 0.42 | 0.068 | |
| Daily Vegetable intake (cup/day) | β at 25th | 0.12 | 0.175 | 0.18 | 0.206 | 0.06 | 0.68 |
| β at 50th | 0.1 | 0.374 | 0.17 | 0.284 | 0.05 | 0.777 | |
| β at 75th | 0.06 | 0.674 | −0.03 | 0.866 | 0.18 | 0.499 | |
Note: Separated quantile regression models were fit with total daily energy intake, daily carbohydrate intake, daily fat intake, daily protein intake, daily sugar intake, daily fruit intake, and daily vegetable intake as dependent variables, and snacking frequency as independent variables, controlled for age, gender, race/ethnicity, and maternal educational level.
We estimated the associations at 25th, 50th, and 75th percentiles in the distribution of daily intake of energy, carbohydrate, fat, protein, sugar, fruit, and vegetable with adjustment of age, gender, ethnicity, and maternal education level. Generally, the quantile regression provided similar but more detailed information about the relationship between snacking frequency and dietary intake.
In 12-to-24-month-olds, the effect of the snacking frequency on total energy intake was significant at the middle of the distribution (50th percentile, β =36.7 kcal/day, P = 0.004). The association between the snacking frequency and the daily carbohydrate intake was only significant at the higher end of the distribution (75th percentile, β =7.41, P = 0.031).
For children aged 2–6 years, quantile estimates suggested significant associations between the dietary intake and snacking frequency from 25th to 75th percentiles with the highest impact at 75th percentile (β =186.0 kcal/day, P < 0.0001). Similarly, for fat intake and protein intake, the association between snacking occasion and the intake was more pronounced at the higher end of the intake distributions. In contrast, snacking may contribute more to carbohydrate intake for those at the lower end of the carbohydrate intake distribution with the lowest percentile having the highest snack carbohydrate intake ( β at 25th percentile =19.9 g/day, P = 0.002).
Sensitivity analyses were also conducted by controlling for snack quality measured by daily energy intake of carbohydrate, fat, and protein from snacks as percentages of total daily energy intake. The results are similar (not reported).
DISCUSSION
The current study examined snacking patterns and their associations with dietary intake and weight status among preschool children. Our results suggest that snacking frequency was positively associated with total daily energy intake in children and snacks were a significant source of energy in toddlers and preschoolers.
The mean snacking frequency was higher in the 12-to-24-month age group compared to the 2-to-6-year group. This may be attributable to feeding practice as children age. This finding is in line with a study comparing snacking patterns between children aged 2–5 years with children aged 6–12 years, which found that parents of younger children gave more snacks than parents of older children (Blaine, et al., 2015). Demographic characteristics were also associated with snacking frequency. Non-Hispanic Whites had the highest snacking frequency, followed by Hispanics and Non-Hispanic Blacks, in children aged 2-to-6 years. The reasons for these differences deserve further study, but could reflect the availability of snacks in the home or cultural differences in feeding practice. Despite the variations across age and racial/ethnic groups, however, the mean snacking frequency was in line with the stated recommendation from the American Academy of Pediatrics.
We found that snacks accounted for about 25–30% of total daily energy. This is in line with the finding from a study using the Guelph Family Health Study data (Hutchinson, et al., 2017). Carbohydrate and fat from snacking were major contributors of energy in toddlers and preschoolers. Our estimates suggested that carbohydrate and fat from snacks combined accounted for approximately one-quarter of total daily energy intake in this population. In addition to confirm the findings from previous studies which found similar patterns in the US children aged 2-to-5 years (Shriver, et al., 2018) or 2-to-18 years (E. K. Dunford & Popkin, 2018), our results provide insight of the quality of snacks consumed by children aged 1-to-2 years. This is of particular concern because this period is critical for children to develop healthy dietary habits. While children in this age group are introduced to a wide variety of new foods for the first time, the over-consumption of high energy snacks may lead to the onset of unhealthy dietary habits and subsequent risk of obesity and chronic diseases later in life (Grummer-Strawn, Li, Perrine, Scanlon, & Fein, 2014).
Our results suggest that snacks play an important role in children’s daily food and nutrient intakes. Healthy snacking is thus essential in influencing diet quality. Consuming snacks with smaller portion size and lower energy density is effective to reduce energy intake from snacks. Consumption of snacks containing high solid fat and added sugars should be limited to reduce empty energy intake. Additionally, providing fruits and vegetables as snacks could also reduce energy intake, promote overall healthier diet quality, and improve children’s dietary behavior. Moreover, despite the lack of a clear consensus regarding the health effects of snacking, healthy snacking should focus on balanced dietary intake instead of focusing on a single food or nutrient.
Our linear regression estimates suggested that snacking frequency was positively associated with total energy intake. Compared to the previous studies that identified such relationship in adults (Bertéus Forslund, Torgerson, Sjöström, & Lindroos, 2005), adolescents (Sebastian, Cleveland, & Goldman, 2008) and school-age children (Vader, et al., 2009), the present study focused on toddlers and preschoolers adding additional data about snacking during an earlier phase in childhood. Our analyses on the components of energy intake indicated that the increased consumption of snacks rich in carbohydrate and fat due to increased frequency of snacking was the major cause of increased energy intake, especially in the 2-to-6-year-olds. Interestingly, another study on adolescent snacking behavior suggested that fat intake was negatively associated with snacking frequency (Sebastian, et al., 2008). The comparison between our study and the previous study reveal potential age differentiation with respect to the quality of dietary intake from snacks that deserve further study. Moreover, one recent study identified the association between the genetics of taste and the snacking patterns of children (Chamoun, et al., 2018). For example, genetic variation in taste receptors may influence the total energy density of snacks and the intake of sugar. Therefore, the interactions between the genetics of taste and age differentiation should be examined in future research.
Our quantile regression method provided additional details regarding the relationship between snacking frequency and dietary intake. In the 12-to-24-month-olds, snacking frequency was positively associated with carbohydrate intake in those with relatively higher carbohydrate intake. In contrast, snacking frequency was positively associated with carbohydrate intake for those at the lower end of the carbohydrate intake distribution in the 2-to-6-years-olds. This finding indicates that encouraging more snacking may help increase carbohydrate intake in those with low carbohydrate intake in 2-to-6-years-olds, while leading to an undesired increase of carbohydrate intake for those who have high carbohydrate already in 12-to-24-month olds. Therefore, differentiated age-specific snacking recommendations should be implemented to avoid unintended consequences.
Our study has a number of strengths. First, the dietary intake data analyzed in the study was from 24-hour dietary recall administered using the NDSR system, a rigorous method for assessing dietary intake. The multiple-pass interview methodology and detailed search algorithms implemented in the NDSR help to ensure the accuracy of food intake recall, which is lacking in the dietary studies using other methods, such as food frequency questionnaires. Second, our study included a large sample of children between 12 and 24 months of age – a period of development where snacking is quite common but has been overlooked in the larger research literature. Moreover, the use of quantile regression techniques allowed us to examine further the relationships between snacking and dietary intake, in addition to conditional mean estimates, which provided a bigger picture and more relevant evidence for interventions. Nevertheless, the findings here should be interpreted within the context of the limitations. For instance, although our sample included children from a diverse range of demographic and socioeconomic environments and allows us to evaluated correlations between snacking intake, caloric intake and dietary quality, a larger nationally representative sample would be needed to inform national nutrition policies. The study also did not evaluate long-term consequences which would provide greater insight into what impact poor dietary intake from snacks during early development might relate to later unhealthy eating and obesity. As we follow this cohort, we hope to provide additional data on these outcomes.
CONCLUSIONS
Snacking frequency was positively associated with daily energy intake. Carbohydrates and fats from snacks are significant energy contributors. Age differentiation was apparent regarding the relationship between snacking frequency and dietary intake. Differentiated interventions that are age-specific and focus on the dietary quality of snacks instead of quantity are needed.
Acknowledgments
The research was supported by National Institute of Environmental Health Sciences of the National Institutes of Health (P01ES022831, R21ES014947, R01ES016772, 1R21AG041048, R01HD084487, P30ES025128) and by the U.S. Environmental Protection Agency (RD-83543701). Additional support was provided by the National Center for Advancing Translational Sciences (UL1TR001117).
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
CONFLICT OF INTEREST
Dr. Kollins reports personal fees from Aevi, grants and personal fees from Akili, personal fees from Arbor, grants from Bose, grants and personal fees from Ironshore, personal fees from Jazz, other from Kashiv, grants and personal fees from KemPharm, personal fees and non-financial support from News, personal fees from Otsuka, grants and personal fees from Rhodes, personal fees from Shire, personal fees from SK Life Sciences, grants and personal fees from Sunovion, personal fees from Tris, other from Behavioral Innovations Group, outside the submitted work.
Contributor Information
Hong Xue, Department of Health Behavior and Policy, School of Medicine, Virginia Commonwealth University, Richmond, VA.
Rachel L. Maguire, Department of Biological Sciences, College of Sciences, North Carolina State University, Raleigh, NC.
Jin Liu, Department of Biostatistics, School of Medicine, Virginia Commonwealth University, Richmond, VA.
Scott H. Kollins, Department of Psychiatry and Behavioral Sciences, School of Medicine, Duke University, Durham NC.
Susan K. Murphy, Department of Obstetrics and Gynecology, School of Medicine, Duke University, Durham, NC.
Cathrine Hoyo, Department of Biological Sciences, College of Sciences, North Carolina State University, Raleigh, NC.
Bernard F. Fuemmeler, School of Medicine, Virginia Commonwealth University, Richmond, VA.
References
- American Academy of Pediatrics. (2017). Healthy Active Living for Families (HALF) [Google Scholar]
- Forslund H. Bertéus, Torgerson JS, Sjöström L, & Lindroos AK (2005). Snacking frequency in relation to energy intake and food choices in obese men and women compared to a reference population. International journal of obesity, 29, 711. [DOI] [PubMed] [Google Scholar]
- Blaine R, Fisher J, Taveras E, Geller A, Rimm E, Land T, Perkins M, & Davison K (2015). Reasons Low-Income Parents Offer Snacks to Children: How Feeding Rationale Influences Snack Frequency and Adherence to Dietary Recommendations. Nutrients, 7, 5265. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chamoun E, Hutchinson J, Krystia O, Mirotta J, Mutch D, Buchholz A, Duncan A, Darlington G, Haines J, & Ma D (2018). Single nucleotide polymorphisms in taste receptor genes are associated with snacking patterns of preschool-aged children in the Guelph family health study: A pilot study. Nutrients, 10, 153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cunningham SA, Kramer MR, & Narayan KV (2014). Incidence of childhood obesity in the United States. New England Journal of Medicine, 370, 403–411. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dunford E, & Popkin B (2017). 37 year snacking trends for US children 1977–2014. Pediatric Obesity [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dunford EK, & Popkin BM (2018). 37 year snacking trends for US children 1977–2014. Pediatric Obesity, 13, 247–255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Field AE, Austin SB, Gillman MW, Rosner B, Rockett HR, & Colditz GA (2004). Snack food intake does not predict weight change among children and adolescents. International journal of obesity, 28, 1210–1216. [DOI] [PubMed] [Google Scholar]
- Fisher JO, Davey A, Kachurak A, & Bailey RK (2017). Association of Snacking Frequency with Patterning of Daily Energy Intake and Weight Among US Toddlers and Preschoolers. The FASEB Journal, 31, 29.24–29.24.27682203 [Google Scholar]
- Fuemmeler BF, Lovelady CA, Zucker NL, & Ostbye T (2013). Parental obesity moderates the relationship between childhood appetitive traits and weight. Obesity (Silver Spring), 21, 815–823. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fuemmeler BF, Stroo M, Lee C-T, Bazemore S, Blocker B, & Østbye T (2015). Racial Differences in Obesity-Related Risk Factors Between 2-Year-Old Children Born of Overweight Mothers. Journal of Pediatric Psychology, 40, 649–656. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Grummer-Strawn LM, Li R, Perrine CG, Scanlon KS, & Fein SB (2014). Infant Feeding and Long-Term Outcomes: Results From the Year 6 Follow-Up of Children in the Infant Feeding Practices Study II. Pediatrics, 134, S1–S3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hales CM, Carroll MD, Fryar CD, & Ogden CL (2017). Prevalence of Obesity Among Adults and Youth: United States, 2015–2016. NCHS Data Brief, 288. [PubMed] [Google Scholar]
- Hoyo C, Murtha AP, Schildkraut JM, Jirtle RL, Demark-Wahnefried W, Forman MR, Iversen ES, Kurtzberg J, Overcash F, & Huang Z (2011). Methylation variation at IGF2 differentially methylated regions and maternal folic acid use before and during pregnancy. Epigenetics, 6, 928–936. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hutchinson JM, Watterworth JC, Haines J, Duncan AM, Mirotta JA, Ma DW, Buchholz AC, & Study GFH (2017). Snacking patterns of preschool-aged children: Opportunity for improvement. Canadian Journal of Dietetic Practice and Research, 79, 2–6. [DOI] [PubMed] [Google Scholar]
- Kan K, & Tsai W-D (2004). Obesity and risk knowledge. Journal of Health Economics, 23, 907–934. [DOI] [PubMed] [Google Scholar]
- Keast DR, Nicklas TA, & O’neil CE (2010). Snacking is associated with reduced risk of overweight and reduced abdominal obesity in adolescents: National Health and Nutrition Examination Survey (NHANES) 1999–2004. The American journal of clinical nutrition, 92, 428–435. [DOI] [PubMed] [Google Scholar]
- Koenker R, & Bassett G Jr (1978). Regression quantiles. Econometrica: journal of the Econometric Society, 33–50. [Google Scholar]
- Murakami K, & Livingstone MBE (2016). Associations between meal and snack frequency and overweight and abdominal obesity in US children and adolescents from National Health and Nutrition Examination Survey (NHANES) 2003–2012. British Journal of Nutrition, 115, 1819–1829. [DOI] [PubMed] [Google Scholar]
- Ogden CL, Carroll MD, Kit BK, & Flegal KM (2014). Prevalence of childhood and adult obesity in the united states, 2011–2012. Jama, 311, 806–814. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ogden CL, Carroll MD, Lawman HG, Fryar CD, Kruszon-Moran D, Kit BK, & Flegal KM (2016). Trends in obesity prevalence among children and adolescents in the United States, 1988–1994 through 2013–2014. Jama, 315, 2292–2299. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rudy E, Bauer K, Hughes S, O’connor T, Vollrath K, Davey A, Correa N, Chen TA, & Fisher J (2017). Interrelationships of child appetite, weight and snacking among Hispanic preschoolers. Pediatric Obesity [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sebastian RS, Cleveland LE, & Goldman JD (2008). Effect of Snacking Frequency on Adolescents’ Dietary Intakes and Meeting National Recommendations. Journal of Adolescent Health, 42, 503–511. [DOI] [PubMed] [Google Scholar]
- Shriver LH, Marriage BJ, Bloch TD, Spees CK, Ramsay SA, Watowicz RP, & Taylor CA (2018). Contribution of snacks to dietary intakes of young children in the United States. Maternal & Child Nutrition, 14, e12454. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Vader AM, Walters ST, Harris TR, & Hoelscher DM (2009). Television viewing and snacking behaviors of fourth- and eighth-grade schoolchildren in Texas. Prev Chronic Dis, 6, A89. [PMC free article] [PubMed] [Google Scholar]
- Ward ZJ, Long MW, Resch SC, Giles CM, Cradock AL, & Gortmaker SL (2017). Simulation of Growth Trajectories of Childhood Obesity into Adulthood. New England Journal of Medicine, 377, 2145–2153. [DOI] [PMC free article] [PubMed] [Google Scholar]