The prevalence of snacking behaviour and energy contribution from snacks have globally increased in recent decades(Reference Duffey, Rivera and Popkin1–Reference Fayet-Moore, Peters and McConnell4), especially among children(Reference Duffey, Rivera and Popkin1). For example, snacks provide around 25–30 % of total energy intake (EI) among children in the United States and Australia(Reference Wang, van der Horst and Jacquier5,Reference Deming, Reidy and Fox6) . Considering this situation, many countries and organisations provide recommendations about the quality and quantity of snacks(Reference Hess, Jonnalagadda and Slavin7,Reference Potter, Vlassopoulos and Lehmann8) . The most common rationale for these recommendations is that snacks contribute to meeting energy and nutrient requirements(Reference Potter, Vlassopoulos and Lehmann8). Indeed, snacks provide a sufficient amount of nutrients that should be encouraged, such as vitamin C and dietary fibre(Reference Deming, Reidy and Fox6,Reference Shriver, Marriage and Bloch9) , but also contribute to intakes of nutrients that should be limited, such as added sugars and SFA(Reference Wang, van der Horst and Jacquier5,Reference Deming, Reidy and Fox6,Reference Shriver, Marriage and Bloch9) . It has also been suggested that frequent snacking is associated with poor diet quality and obesity in children(Reference Murakami and Livingstone10,Reference Murakami and Livingstone11) . Thus, it is important to establish strategies to improve snack quality in children.
Snacking habits in Asia may differ from those in western countries. For example, the prevalence of snacking habits among children is low in China (66·1 %) compared to the United States (97·9 %)(Reference Wang, van der Horst and Jacquier5). However, studies regarding snacking habits in Asian children are limited to those conducted in China(Reference Wang, Zhai and Zhang2,Reference Wang, van der Horst and Jacquier5,Reference Taillie, Wang and Popkin12) . Moreover, these studies have not assessed how snacks contribute to the improvement of nutrient adequacy. In Asian children, excessive added sugar intake is less prevalent(Reference Afeiche, Koyratty and Wang13–Reference Fujiwara, Murakami and Asakura15), while insufficient calcium intake is more prevalent(Reference Murakami, Okubo and Livingstone16–Reference Lopez-Sobaler, Aparicio and Gonzalez-Rodriguez19) than in western children. The prevalence of childhood underweight/overweight is estimated to vary even among Asian countries(Reference Piernas, Wang and Du17,Reference Inokuchi, Matsuo and Takayama20) . Therefore, the kind and amount of nutrients supplied by snacks should be clarified in each country.
In this cross-sectional study among Japanese nursery school children, we evaluated the dietary characteristics of snacks (i.e. prevalence and food composition) and the percentage contribution of snacks to daily intakes of energy, nutrients and foods. Moreover, we also assessed the relationship of EI from snacks with the prevalence of nutritional inadequacy.
Methods
Participants
This study was based on a multi-regional dietary survey named Dietary Observation and Nutrient intake for Good Health Research in Japanese young children (DONGuRI), which was conducted from October to December 2015. Details of this study have been reported elsewhere(Reference Fujiwara, Murakami and Asakura15,Reference Murakami, Okubo and Livingstone16) . Briefly, with consideration to geographical diversity and feasibility, particularly for the recruitment of research dietitians, twenty-four of forty-seven prefectures were selected as the targeted area. The prefectures were selected from all over Japan.
Boys and girls aged 19–83 months were recruited at nursery schools by 323 dietitians who worked generally in nursery schools. In each prefecture, sixteen boys and sixteen girls were recruited: four aged 19–23 months, four aged 24–35 months, two aged 36–47 months, two aged 48–59 months, two aged 60–71 months and two aged 72–83 months. Children were considered eligible if they were apparently healthy and did not consume special diets, and their guardians were not dietitians or medical doctors and did not plan to move within a few months. Participation of only one child per household was permitted. As a result, 753 children from 315 nurseries participated. For analysis, we excluded 375 children aged <36 months (<3 years), because only 1-d dietary data were collected for these children. Thus, this analysis included 378 children (187 boys and 191 girls) aged 3–6 years. Written informed consent was obtained from a guardian of each child.
Dietary assessment
Dietary intake was assessed using semi-weighed 3-d dietary records. Recording days consisted of 2 weekdays (nursery days) and 1 weekend day, and were not consecutive. It is common practice in Japanese nursery schools to serve lunch and at least one snack. Foods and beverages the children consumed, as well as leftovers, were weighed and recorded by the research dietitians in the nursery schools, and by the guardians outside the nursery schools. The research dietitians instructed guardians on how to keep dietary records and gave them a digital kitchen scale (KD-812WH; Tanita), measuring spoon, measuring cup, recording sheets and an instruction sheet. The starting and ending time of each eating occasion was also recorded. When weighing was difficult (e.g. when eating out), guardians recorded the estimated amount of foods and drinks consumed. On the following nursery day of each recording day, recording sheets were checked by the research dietitians. In accordance with the research manual, research dietitians converted the estimated portion sizes into weights and coded all food items on the basis of values in the Standard Tables of Food Composition in Japan(21). Food items not listed on the tables were assigned the nutrient values of similar foods. The dietary records were then submitted to the central office and checked by two dietitians.
The recording sheets consisted of the following sections of eating occasions: breakfast, lunch, dinner and snacks. Eating occasions other than snacks were considered as ‘meals’ regardless of the time of day or foods and beverages consumed. Thus, the definition of meals and snacks in this study was based on what guardians and dietitians considered as meals and snacks. In accordance with previous studies(Reference Piernas and Popkin3,Reference Murakami and Livingstone10,Reference Murakami and Livingstone11,Reference Leech, Worsley and Timperio22) , guardians and dietitians were instructed to consider two eating events as distinct eating occasions when these were separated by a minimum of 15 min. Calculations of eating frequency were based on all eating occasions, or eating occasions providing a minimum of 209·2 kJ, considering comparability with previous studies(Reference Murakami and Livingstone10,Reference Murakami and Livingstone11,Reference Leech, Worsley and Timperio22) . Intakes of energy, nutrients and foods were calculated as the mean value of 3 d. Supplements were not considered in nutrient intake calculation, not only because the dietary supplement use in the previous month was rare in our population (3·5 %) but also because of the lack of a dietary supplement database in Japan. We estimated dietary free sugar, which was defined as all mono- and disaccharides added to foods and beverages by the manufacturer, cook or consumer, and naturally occurring sugars in honey, syrups, fruit juices and fruit juice concentrate(23). Each mono- and disaccharide content was listed for only 880 of the 2222 food items in the Standard Tables of Food Composition in Japan(21), with no information on free sugar. Thus, we calculated free sugar consumption with reference to a recently developed sugar database(Reference Fujiwara, Murakami and Asakura15,Reference Fujiwara, Murakami and Asakura24) . Briefly, the remaining 1342 items without saccharide content were complemented with assignment of available carbohydrate content of foods, results of literature search, or calculated values based on recipes(Reference Rand, Pennington and Murphy25). Then, based on the saccharide content, free sugar content of 2222 food items was estimated in each of the food groups using a published step-wise method(Reference Louie, Moshtaghian and Boylan26,Reference Mok, Ahmad and Rangan27) . Food grouping was mainly based on the Standard Tables of Food Composition in Japan(21).
Assessment of other factors
Questionnaires designed for this study were answered by the guardians before dietary assessment was conducted. The questionnaires asked about parental educational level, parental occupation and household income. Information on time spent on certain activities (vigorous, moderate or sedentary activity) both on weekdays and weekend days was also obtained. For weekdays, activities outside of the nursery were reported because we assumed that guardians could not assess their children’s activity in nurseries. Vigorous activity (running, ball play or swimming) was used as a surrogate marker of the habitual physical activity of children. Anthropometric measurements were conducted by research dietitians as part of this study or at routine health check-ups (within 1 month before the study period). Body height (to 0·1 cm) and weight (to 0·1 kg) of children were measured while children were wearing light clothing and no shoes. BMI was calculated by dividing weight (kg) by height-squared (m2).
Energy misreporting
Energy misreporting, especially underreporting, cannot be completely excluded from any dietary assessment(Reference Livingstone and Black28). Thus, we examined if reported daily EI was comparable to the estimated energy requirement in Japanese Dietary Reference Intakes (DRIs), 2015(29) at the ‘good’ level. The plausibility of reported EI from snacks was roughly assessed by comparing body size (weight, height and BMI) and habitual physical activity according to EI from snacks.
Assessment of nutritional inadequacy
Nutrient intakes were compared to the age- and sex-specific reference values in Japanese DRIs, 2015(29). In Japanese DRIs, estimated average requirement (EAR) is defined as ‘the amount that would meet the nutrient requirements of 50 % of the population’. The percentage of children with intake less than the EAR suggests an estimated prevalence of inadequate intake within a group. Of the seventeen nutrients with EAR, three (iodine, selenium and molybdenum) were excluded from the evaluation because of insufficient data for these nutrients in the Standard Tables of Food Composition in Japan. In Japanese DRIs, tentative dietary goals (DG) for the prevention of lifestyle-related diseases were developed for the percentage of energy from protein, total fat, SFA and carbohydrate, and intakes of Na, K and dietary fibre. To estimate the prevalence of inadequate intake, we calculated the percentage of children who consumed a nutrient outside (below or above) the range of the DG.
For children aged 3–5 years, a DG is not available for dietary fibre, K and SFA. For dietary fibre and K, we determined a tentative DG mainly based on the median intake in children of that age in the National Health and Nutrition Survey 2010 and 2011(Reference Murakami, Okubo and Livingstone16). For SFA, we tentatively applied the DG of adults to children of that age(Reference Murakami, Okubo and Livingstone16,Reference Asakura and Sasaki30) . DG and EAR values for children aged 3–6 years are shown in Supplemental Table 1. Due to the lack of a recommendation for free sugar intake for Japanese, excessive free sugar intake was defined as ≥10 % of total EI, in accordance with the recommendation of the WHO(23).
Statistical analysis
Because the results did not differ by sex, analyses were conducted for boys and girls combined. We compared intakes of energy, nutrients and foods derived from snacks with those from meals using a paired t test. To evaluate the characteristics of snacks, the percentage contribution of snacks to the daily intakes of energy, nutrients and foods was also calculated. In accordance with previous studies(Reference Wang, van der Horst and Jacquier5,Reference Deming, Reidy and Fox6) , the percentage contribution to EI was used as a standard from which the degree of contribution of each nutrient was evaluated.
Then, children were classified into tertile categories of EI from snacks. The associations of demographic characteristics of children and their parents with the categories of EI from snacks were tested by the Mantel–Haenszel χ 2 test for categorical variables and by regression analysis for continuous variables. Relationships of the categories of EI from snacks to the daily intakes of energy and nutrients were tested using regression analysis, with median values of EI from snacks for each category as independent variables. The associations between the categories of EI from snacks and the prevalence of children with inadequate nutrient intakes were tested by the Mantel–Haenszel χ 2 test.
All statistical analyses were performed using SAS (version 9.4; SAS Institute Inc.). P-values <0·05 were considered statistically significant.
Results
Snacks were consumed by all children on the weekdays and 92·9 % of children in the weekend day. Mean snack frequency was 2·8 times per day (sd 1·2), but was 1·6 times per day (sd 0·65) for snacks providing a minimum of 209·2 kJ (Table 1). The mean frequency for meals was 3·0 times per day (sd 0·06). Table 2 shows that the mean reported daily EI was comparable to the estimated energy requirement for children aged 3–5 years (5439·2 kJ/d for boys and 5230 kJ/d for girls(29)). On average, EI from snacks was 1118·4 kJ/d (sd 445·9) and accounted for 19·5 % (sd 6·9) of total EI. As expected, snacks provided a lower amount of all nutrients than meals, except for no difference in free sugar. Relative to their energy contribution, snacks accounted for a small proportion of daily intake of all nutrients except for free sugar (46·2 %), calcium (34·8 %), SFA (28·1 %) and riboflavin (25·4 %). Tea/coffee drinks (34·5 %) and milk (23·7 %) accounted for the greatest part of the food weight of snacks (Fig. 1). Confectionaries accounted for the largest part of EI from snacks (35·3 %), followed by milk (19·5 %). Most foods were mainly consumed as part of meals, but 73·3 % of confectionaries, 59·7 % of milk and 57·4 % of sweetened drinks were consumed in snacks (Supplemental Table 2).
Characteristics of study participants (n 378)
* Missing (n 1).
† Breakfast, lunch and dinner were considered as ‘meals’ regardless of the time of the day or foods and beverages consumed. Eating occasions other than meals were considered as ‘snacks’.
Daily intake of nutrients with dietary goals or estimated average requirement among Japanese nursery school children (n 378)
NE, niacin equivalent; RAE, retinol activity equivalent.
* Breakfast, lunch and dinner were considered as ‘meals’ regardless of the time of day or foods and beverages consumed. Eating occasions other than meals were considered as ‘snacks’.
† Paired t test.
‡ 1 g NaCl equivalent = 58·5/23 × Na (g).
§ 1 μg RAE = sum of retinol (μg) + β-carotene (μg) × 1/12 + α-carotene (μg) × 1/12 + β-cryptoxanthin (μg) × 1/24.
|| 1 mg NE = niacin (mg) + protein (mg)/6000.
Weight-based and energy-based contribution of food types to daily snack intake among Japanese nursery school children (n 378). Breakfast, lunch and dinner were considered as ‘meals’ regardless of the time of day or foods and beverages consumed. Eating occasions other than meals were considered as ‘snacks’. Population proportions were calculated by summing the weights (or energies) of a given food group for all persons and then dividing it by the sum of weights (or energy intakes) for all food groups consumed by all persons. 
, White rice, noodle, breads; 
, fruit; 
, dairy products; 
, tea and coffee drinks; 
, potato; 
, fruit/vegetable juice; 
, fats and oils; 
, sugar; 
, fish, meat, eggs; 
, confectionaries; 
, pulses, nuts, total vegetables; 
, milk; 
, sweetened drinks
Mean age and BMI as well as the sex prevalence of children did not differ among tertile categories of EI from snacks (Table 3). However, children with a higher intake of EI from snacks were taller, heavier and spent more time in vigorous outdoor activity on weekend days. Mean EI from snacks in the lowest, middle and highest tertile categories was 664·6 kJ/d (sd 140·6), 1068·4 kJ/d (sd 125·0) and 1622·2 kJ/d (sd 313·8), respectively (Fig. 2). The tertile categories of EI from snacks were not associated with EI from meals (P trend = 0·63) but were positively associated with total EI (P trend < 0·0001). EI from snacks was positively associated with daily intakes of all nutrients with EAR or DG, except for vitamin A, vitamin B6, vitamin B12 and vitamin C (Table 4). The prevalence of children consuming K below the DG, and thiamine, riboflavin, calcium and iron below the EAR, was low in children with high EI from snacks (Table 5). However, 50 % of children consumed inadequate calcium even in the highest category of EI from snacks. The prevalence of children exceeding the DG for Na was >70 % in all categories, and was higher among children with high EI from snacks. Over 90 % of children in each category had an SFA intake exceeding the DG, with no significant differences among categories. The prevalence of excessive free sugar intake (≥10 % of EI) was higher among children with a higher EI from snacks. EI from snacks was positively associated with the daily intakes of sugars as seasonings, fruit, milk, confectionaries and sweetened drinks (Supplemental Table 3).
Characteristics of Japanese nursery school children and their parents according to tertile categories of energy intakes from snacks*
* Breakfast, lunch and dinner were considered as ‘meals’ regardless of the time of day or foods and beverages consumed. Eating occasions other than meals were considered as ‘snacks’. Median (range) energy intake from snacks in each category was 703·0 kJ/d (208·9–855·0), 1068·4 kJ/d (867·9–1272·5) and 1531·4 kJ/d (1273·4–3009·1), respectively.
† Regression analysis using median values of energy intakes from snacks for each category as independent variables. Mantel–Haenszel χ 2 test or Fisher’s exact test for categorical variables.
‡ Data were missing for height (n 1), weight (n 1), BMI (n 1), vigorous outdoor activity on weekday and weekend (n 1), paternal education (n 17), maternal education (n 1), paternal occupation (n 16), maternal occupation (n 1) and household income (n 7).
Mean daily energy intake from meals and snacks according to tertile categories of energy intakes from snacks among Japanese nursery school children (n 378). Breakfast, lunch and dinner were considered as ‘meals’ regardless of the time of day or foods and beverages consumed. Eating occasions other than meals were considered as ‘snacks’. P
trend was calculated based on a regression analysis using median values of energy intakes from snacks for each category (703·0, 1068·4 and 1531·4 kJ/d, respectively) as independent variables. P
trend < 0·0001 for energy intake from snacks, P
trend = 0·63 for energy intake from meals, and P
trend < 0·0001 for total energy intake. 
, Energy intake from meals; 
, energy intake from snacks
Daily intake of nutrients with dietary goals or estimated average requirement according to tertile categories of energy intakes from snacks* among Japanese nursery school children
NE, niacin equivalent; RAE, retinol activity equivalent.
* Breakfast, lunch and dinner were considered as ‘meals’ regardless of the time of day or foods and beverages consumed. Eating occasions other than meals were considered as ‘snacks’. Median (range) energy intake from snacks in each category was 703·0 kJ/d (208·9–855·0), 1068·4 kJ/d (867·9–1272·5) and 1531·4 kJ/d (1273·4–3009·1) respectively.
† Regression analysis using median values of energy intakes from snacks for each category as independent variables.
‡ 1 g NaCl equivalent = 58·5/23 × Na (g).
§ 1 μg RAE = sum of retinol (μg) + β-carotene (μg) × 1/12 + α-carotene (μg) × 1/12 + ββ-cryptoxanthin (μg) × 1/24.
|| 1 mg NE = niacin (mg) + protein (mg)/6000.
Prevalence of inadequate intake of selected nutrients according to tertile categories of energy intakes from snacks among Japanese nursery school children*
DG, dietary goal; EAR, estimated average requirement.
* Breakfast, lunch and dinner were considered as ‘meals’ regardless of the time of day or foods and beverages consumed. Eating occasions other than meals were considered as ‘snacks’. Median (range) energy intake from snacks in each category was 703·0 kJ/d (208·9–855·0), 1068·4 kJ/d (867·9–1272·5) and 1531·4 kJ/d (1273·4–3009·1), respectively.
† Mantel–Haenszel test was conducted for nutrients for which at least 5 % of participants were classified as having inadequate intake.
Discussion
To our knowledge, this is the first study to show the characteristics of snacks and the prevalence of nutritional inadequacy according to EI from snacks among Japanese nursery school children. Snacking habits are very prevalent, and snacks provide one-fifth of daily EI. However, the largest part of EI from snacks is supplied by confectionaries, and snacks are generally not nutrient-dense but rather energy-dense. Children with higher EI from snacks were less likely to have inadequate intake of thiamine, riboflavin, calcium and iron, but more likely to have excess intake of free sugar and Na. There is a room for improvement in food choices and diet quality at snack time among nursery school children.
The prevalence of snacking was much higher in our participants than Chinese children(Reference Wang, van der Horst and Jacquier5). The percentage of EI from snacks was lower than that reported in some countries (26·6 % in the United States and 32·9 % in Australia)(Reference Wang, van der Horst and Jacquier5), but higher than that in other countries (11·9 % in China and 15·4 % in Mexico)(Reference Wang, van der Horst and Jacquier5). A high prevalence of snacks in this population might be explained by the fact that snacks are regularly offered at nursery schools in Japan, even though snacks are also prevalent on the weekend. Snacks might have a greater impact on the diet quality of children in Japan than in China. EI from snacks was independent of that from meals, suggesting that EI from snacks is not compensated for by EI from meals. Although overweight is not prevalent (<5 %) in Japanese nursery school children(Reference Inokuchi, Matsuo and Takayama20), excess EI and overweight due to frequent snacking is of concern(Reference Murakami and Livingstone11). However, this does not mean that snacking should be avoided. In developed countries, children commonly have four or more eating occasions per day(Reference Murakami and Livingstone10,Reference Murakami and Livingstone11,Reference Taillie, Afeiche and Eldridge31,Reference Taylor, Iosua and Heath32) . The EI needed for a healthy growth of children is considered to be not fully supplied by breakfast, lunch and dinner alone. Strategies aimed at promoting a healthy growth of children should not attempt to stop snacking but rather must improve the quality of snacks.
A large part of EI from snacks is derived from confectionaries, and a higher EI from snacks is associated with excess free sugar intake. Confectionaries as well as sweetened drinks are the main contributors to free sugar consumption in Japanese nursery school children(Reference Fujiwara, Murakami and Asakura15). Among children in Scotland, confectionaries are consumed not only in snacks but also with meals, and soft drinks are more frequently consumed as part of a meal than a snack(Reference Macdiarmid, Loe and Craig33). Considering that the consumption of these foods as meals is not common in our participants, snacking habits appear to be the main target in controlling free sugar intake among Japanese nursery school children.
Compared to sweetened drinks, tea/coffee drinks and milk account for a larger part of the food weight of snacks. Considering free sugar consumption, this is a favourable aspect of beverage intake during snacking in Japanese nursery school children. Moreover, the relatively adequate calcium and riboflavin intakes among children with higher EI from snacks could be explained by the higher intake of milk and dairy products during snacking. Milk and dairy products are also popular snack items in other countries(Reference Wang, van der Horst and Jacquier5,Reference Deming, Reidy and Fox6,Reference Taillie, Afeiche and Eldridge31) , and thus snacking on such foods is an effective way to improve calcium and riboflavin intake in children. In contrast, intakes of SFA should be limited both in meals and snacks because excessive intake of SFA prevails regardless of EI from snacks. Replacing regular-fat dairy with low-fat dairy could reduce EI from SFA without changing calcium intake(Reference Rehm, Drewnowski and Monsivais34,Reference Hendrie and Golley35) . Although low-fat dairy is recommended for children in some countries(Reference Gidding, Dennison and Birch36,37) , there is no consensus on this topic in Japan. Considering all the foods consumed as snacks by Japanese children (Supplemental Table 2), a low EI from milk fat might be compensated for by a higher free sugar intake from soft drinks or confectionaries. Further discussion is needed based on scientific evidence in children regarding the effects of SFA restriction on blood lipids, growth and development(Reference Simell, Niinikoski and Ronnemaa38–Reference Niinikoski, Pahkala and Ala-Korpela40), as well as the energy compensation caused by a shift to low-fat milk(Reference Hendrie and Golley35,Reference Lee, Gerrior and Smith41) . On the other hand, although EI from snacks was positively associated with inadequate Na intake, interventions for Na reduction should mainly target meals, because the contribution to Na intake from snacks was relatively small (8·8 %).
Although snacks did not substantially contribute to intakes of iron and thiamine, the prevalence of inadequate thiamine and iron intake was low in children with a higher EI from snacks. This was because many children had intakes near the EAR, and a small increase in intake changed the proportion of children with inadequate intake. Considering that anaemia is common after first menstruation(Reference Asakura, Sasaki and Murakami42), iron-rich foods should be incorporated into snacks for girls for both this purpose and for growth.
Snacking is a common way to incorporate fruits into the diet in the United States and China(Reference Taillie, Wang and Popkin12,Reference Moore, Hamner and Kim43) . Contrary to studies in these countries(Reference Wang, van der Horst and Jacquier5,Reference Deming, Reidy and Fox6) , however, the contribution to daily vitamin C and dietary fibre intake from snacks, relative to its energy contribution in our participants, was small, which might be due to a relatively low consumption of fruits during snacking. Substituting confectionaries with fruits might be an effective way to reduce free sugar and SFA intakes from snacks, as well as to increase intakes of dietary fibre, vitamins C and K.
Several limitations of this study should be mentioned. First, systematic and random errors in dietary assessment could not be completely eliminated. Dietary assessment was based on only three recording days in November (autumn), and thus day-to-day and seasonal variations might have biased our findings. The positive association between EI from snacks and daily nutrient intake might be partly explained by an underreporting of eating frequency(Reference Lioret, Touvier and Balin44). However, children with a higher EI from snacks were physically more active on weekend days than those with a lower EI from snacks. Moreover, the mean reported daily EI was comparable to the estimated energy requirement for children aged 3–5 years(29). Thus, we considered that the reported EI from snacks was plausible. Second, eating occasions were classified based on the self-reports of guardians. There are various definitions of snacks(Reference Johnson and Anderson45), and the definition may affect the observed association between snacking frequency and diet quality(Reference Murakami and Livingstone46,Reference Llaurado, Albar and Giralt47) . Thus, the results of this study should be interpreted based on the definition used in this study. Third, generalisability is likely limited because all of our study participants were nursery school children, and the selection of participants as well as of study areas was not randomised. Around 50 % of children aged 3–6 years attend nursery schools in Japan(48). Snacks offered at nursery schools are controlled by dietitians and would be healthier than those consumed in other locations. Thus, the positive (or negative) impact of snacking on nutrient adequacy might be overestimated (or underestimated). Although geographic diversity was considered, study areas were limited to twenty-four prefectures because collaborators (research dietitians) were recruited from the personal connections of dietitians. Further studies based on a more representative data are needed to clarify the differences between snacks consumed at nursery schools, at home and elsewhere(Reference Ziegler, Briefel and Ponza49).
Conclusions
Although snacks contribute to meeting the requirements for some nutrients among Japanese nursery school children, snacks are generally not nutrient-dense and are associated with an excessive intake of free sugar. Food choice at snack time has room for improvement, such as substituting confectionaries with fruits. Further studies in Asia should be conducted to establish the effective ways to improve nutrient adequacy among children, with consideration to specific eating occasions.
Acknowledgements
Acknowledgements: We thank Nana Kimoto and Nanako Koe and other colleagues for the management of dietary data. Financial support: This work was supported in part by a Health and Labour Sciences Research Grant (no. H26-jyunkankitou (seisaku)-shitei-001) from the Ministry of Health, Labour and Welfare, Japan. The Ministry of Health, Labour and Welfare had no role in the design, analysis or writing of this article. Conflict of interest: None. Authorship: R.T. formulated the research, analysed and interpreted the data, prepared the first draft of the manuscript and had primary responsibility for the final content. K.M. formulated the research, interpreted the data and provided critical input into the final draft of the manuscript. K.A., A.F., K.U., M.S. and H.-C.W. contributed to the design of the survey, data collection and data management. S.M. managed the study field establishment and recruitment and contributed to data collection. S.S. directed the survey and contributed to the research plan. All authors read and approved the final manuscript. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving human subjects were approved by the Ethics Committee of the University of Tokyo Faculty of Medicine. Written informed consent was obtained from each participant’s parent or guardian.
Supplementary material
For supplementary material accompanying this article visit https://doi.org/10.1017/S1368980019005007
