Snack food and beverage consumption and young child nutrition in low‐ and middle‐income countries: A systematic review

Abstract Although snacks can provide important nutrients for young children during the complementary feeding period, the increasing availability of snack foods and sugar‐sweetened beverages (SSB), often energy‐dense and nutrient‐poor, in low‐ and middle‐income countries (LMIC) is a concern. Such foods may displace consumption of nutritious foods in contexts where diets are often nutritionally inadequate and the burden of childhood malnutrition is high. This systematic review summarizes literature on the contribution of snack food/SSB consumption to total energy intakes (TEI) of children below 23 months of age in LMIC and associations between this consumption and nutritional outcomes. It also identifies areas where further research is needed. A systematic search of Embase, Global Health, and MEDLINE for literature published in January 1990–July 2018 was conducted. This search yielded 8,299 studies, 13 of which met inclusion criteria: Nine studies assessed % TEI from snack foods/SSB, and four studies assessed associations between snack food/SSB consumption and nutritional outcomes. Average % TEI from snack foods/SSB ranged from 13% to 38%. Findings regarding associations with growth were inconclusive, and no studies assessed associations with nutrient intakes. Variation in measurement of consumption and definitions of snack foods and SSB limited study comparisons. Further research is needed to understand how consumption of energy‐dense, nutrient‐poor snack foods and SSB influences undernutrition and overnutrition among young children during the complementary feeding period in settings that are experiencing dietary transitions and the double burden of malnutrition.

complementary foods, including snacks, at 6 months of age while continuing to breastfeed. The types of snacks fed to young children however are important to ensure diet quality. Snack food products and sugar-sweetened beverages (SSB) are often energy-dense, nutrient-poor, and high in salt or sugar (Lucan, Karpyn, & Sherman, 2010;Monteiro, Levy, Claro, de Castro, & Cannon, 2011;Sekiyama, Roosita, & Ohtsuka, 2012;Waseem et al., 2014), making them inappropriate for infant and young child feeding (Moodie et al., 2013).
The growing availability of unhealthy processed foods in many low-and middle-income countries (LMIC) is a concerning trend (Monteiro, Moubarac, Cannon, Ng, & Popkin, 2013). Overconsumption of snack food products and SSB has been shown to contribute to overweight and obesity among children in the United States (Nicklas, Yang, Baranowski, Zakeri, & Berenson, 2003;Phillips et al., 2004;Scientific Advisory Committee on Nutrition, 2014;Welsh, 2005) and Latin America (Asfaw, 2011;Rauber, Campagnolo, Hoffman, & Vitolo, 2015;Tavares, Fonseca, Garcia Rosa, & Yokoo, 2012). Additionally, energydense, nutrient-poor foods early in life can displace consumption of other nutritious foods, including breast milk (Kimmons et al., 2005), potentially increasing a child's risk of inadequate nutrient intakes and contributing to childhood undernutrition. The correlation between consumption of snack foods/SSB and lower consumption of nutrient-rich foods and/or reduced nutrient intakes has been shown in high-income settings (Bhargava & Amialchuk, 2007;Bowman, 1999;Kant, 2003;Marriott, Olsho, Hadden, & Connor, 2010;Murakami & Livingstone, 2015;Rennie & Livingstone, 2007;Webb et al., 2006). Prevalent consumption of snack foods and SSB among young children has been noted across Africa, Asia, and Latin America (Bentley et al., 2015;Faber & Benade, 2007;Pantoja-Mendoza, Melendez, Guevara-Cruz, & Serralde-Zuniga, 2014;Pries et al., 2017;Woo et al., 2013). However, the role these foods play in the overall diets of infants and young children in LMIC, as well as their impact on their nutrition, remains unclear. The influence of such foods on diet quality and nutritional outcomes is hypothesized to be different in LMIC, as compared with high-income settings, given the higher burden of undernutrition and limited accessibility of nutrient-dense foods in diets. The purpose of this systematic review was therefore to synthesize available literature on the contribution of snack food and SSB to total energy intakes (TEI) among children 0-23 months of age in LMIC and associations between consumption of such foods/beverages and nutritional status of young children in these settings, as well as to identify future research needs within this topic area. Because definitions for snack foods and SSB are wide ranging, the search strategy identified all studies related to children's diets, which were then screened to identify those that included specific consumption measurement of snack foods/SSB; this included measurement of types of foods (e.g., biscuits, candy, and soft drinks) or categories of foods (ultraprocessed foods, discretionary foods, snack foods, junk foods, noncore foods, etc.). Titles and abstracts were screened first for exclusion/inclusion, which was followed by full-text screening. All three researchers reviewed search strategy and terms (AP, SF, and EF), and screening was conducted by one researcher (AP). Screened studies were included if they met the following criteria: (a) They were conducted in an LMIC, (b) the study population included children below two years of age, and (c) they assessed the contribution of snack foods and/or SSB to children's TEI (based on kilocalories) or the association between children's consumption of snacks foods and/or SSB and nutrient intakes/micronutrient status/anthropometric status. Studies were excluded based on the following criteria: (a) studies published prior to January 1, 1990 (based on the assumption that availability and use of snack foods/SSB in LMIC have changed in the last two to three decades), (b) literature not published in English, (c) results published as conference/meeting abstracts only, (d) studies that assessed contribution to TEI from added sugars only, or (e) studies with a wider age range than children 0-23 months of age that did not present specific data within this age range. References of included studies were also hand-searched to identify relevant studies for inclusion; no additional studies were identified through this process.

| Data extraction and synthesis
For studies presenting data on the proportion of TEI derived from snack foods and/or SSB, the following information was extracted:

Key messages
• Limited evidence suggests that snack foods and sugarsweetened beverages (SSB) are providing a substantial proportion of energy intakes among children below 2 years of age in Latin American and South-east Asian LMIC.
• There is limited and inconclusive evidence on relationships between snack food/SSB consumption patterns and young children's overall dietary adequacy, micronutrient status, and growth. Given that diets in LMIC are often nutrient-poor and young children's nutrient requirements are high, the high consumption of such foods may contribute to undernutrition, as well as increased risk of overnutrition, and more research is needed.
• Definitions of snack foods/SSB and measurement of their consumption among children are wide ranging.
There is a need for standardized approaches in order to allow comparison between regions and over time, as well as a need to expand research on contribution of snack foods/SSB to young children's diets and nutritional outcomes in regions where malnutrition is highest. reference, study population, sample size, location, study design, dietary assessment methods, diet findings, and their definition of snack food/SSB. For studies that tested associations between snack food/SSB consumption and child nutritional status, the following additional information was extracted: the nutritional outcomes tested and results of the associations tested.

| RESULTS
After deletion of 3,019 duplicates, this search resulted in 8,299 studies. After title/abstract screening, 205 studies were identified as relevant for full-text review. The majority of studies were excluded because the study population did not include children below two years of age. During full-text review, 30 relevant studies were identified with age ranges that included children 0-23 months of age in their samples. Just under half of these studies however presented data on children specifically within the 0-23 months of age range. A total of 13 studies met the selection criteria: Nine studies detailed the proportion of energy intake derived from snack foods or SSB, and four examined associations between snack food/SSB consumption and nutritional status (Figure 2).

| Contribution to energy intake
Nine studies identified in this review assessed the contribution of snack foods/SSB to dietary energy intakes among children within the complementary feeding period. Details of these studies are presented in Table 1. Five of these studies were from Latin America, three studies were from East/South-east Asia, and one study was from Egypt.
Four studies presented differences in % TEI from snacks/SSB across age groups, with most showing an increase in % TEI from such foods among older children. Across the entire complementary feeding period, Lander et al. (2010) and Denney et al. (2017) found that % TEI from snack foods and SSB increased with age (27.0% among 6-to 8month-olds vs. 35.0% among 9-to 11-month-olds vs. 40.0% among 12-to 23-month-olds; and 8.0% among 6-to 11-month-olds vs. 19.6% among 12-to 23-month-olds, respectively). However, a decrease in % TEI from processed snack foods/SBB with age was shown by Kavle et al. (2015) in rural and periurban Egypt.

| Relationships between consumption and nutritional outcomes
Four studies assessed relationships between consumption of snack foods/SSB and nutritional outcomes among children during the complementary feeding period in LMIC (Table 2). Three studies looked at associations with anthropometry, one study looked at the association with anaemia, and no studies reported association testing between snack food/SSB consumption and dietary nutrient intakes. These four studies were conducted in countries that spanned three separate regions, with three conducted in urban/periurban locations and one in a rural location (Faber, 2007). Three studies included consumption of both snack foods and SSB in their analyses, whereas one study included consumption of snack foods only (not SSB; Vakili, Kiani, Saeidi, Hoseini, & Anbarani, 2015).
For associations between snack food/SSB consumption and anthropometric outcomes, one study assessed differences in mean zscores (Budree et al., 2017), one study assessed associations with overweight/obesity, and one study assessed associations with growth delay (Vakili et al., 2015). Budree et al. (2017) found no relationship between snack food/SSB consumption and body mass index z-score (BMIZ), height-for-age z-score (HAZ), or weight-for-age z-score (WAZ), comparing mean z-scores among 12-month-olds who had consumed snack foods/SSB daily with those who consumed these foods among 5-to 24-month-olds who consumed high-fat snack foods and SSB at least once a week; this relationship was also noted for con-  Vakili et al. (2015) noted a positive association between regular feeding of junk food among 6-to 24-month-olds and growth delays. The definition and measurement of growth delay however were not presented in the paper. One study (Faber, 2007) assessed the association between consumption of various types of snack foods/SSB-biscuits, sweets, savoury snacks, or soft drinks-and anaemia among 6-12 months of age in rural South Africa, with no statistical differences in proportions of anaemic versus nonanaemic children noted.

| DISCUSSION
This review indicates that the % TEI contributed by snack foods and SSB among children in the 0-to 23-month age range in LMIC ranged from 13% in rural Peru to 38% in urban Cambodia. Evidence regarding the influence of snack food and SSB consumption on children's dietary adequacy and nutritional status in these contexts is however limited.
Results from the three studies that explored associations between snack food/SSB consumption and child growth outcomes show mixed findings: one found no significant relationship with z-scores (Budree et al., 2017), one found a positive relationship with child overweight/obesity (Jimenez-Cruz et al., 2010), and one identified a positive relationship with child growth delays (Vakili et al., 2015). No studies were identified that assessed associations between snack food/SSB consumption and dietary nutrient intakes, and only one assessed child micronutrient status, specifically, anaemia status.
Although this review indicates that snack foods/SSB are potentially providing a substantial proportion of dietary energy among young children in LMIC, the low number of studies and their limited   (Anderson et al., 2008;Karnopp et al., 2017;Lander et al., 2010;Valmórbida & Vitolo, 2014), and one was conducted in a conflict area (Jeharsae et al., 2011). Two studies used national datasets (Denney et al., 2017;Rodríguez-Ramírez et al., 2016); however, neither disaggregated data among 0-to 23-month-olds by rural/urban area of residence. Kavle et al. (2015) and Roche et al. (2011)    Note. NS: not significant; BMIZ: body mass index z-score; HAZ: height-for-age z-score; WAZ: weight-for-age z-score; MUAC: mid upper arm circumference. and among Filipino 15-year-olds, 21% TEI came from snack foods (Adair & Popkin, 2005). In high-income settings, 31% and 27% TEIs came from snack foods/SSB among 2-to 6-year-olds in Russia and the United States, respectively (Adair & Popkin, 2005), and 31% of TEI among American children and adolescents 8-18 years of age came from low-nutrient density foods, such as processed snack foods and SSB (Kant, 2003). Although high consumption of snack foods and SSB has often been thought to be a problem specific to school-age children and children in higher socio-economic settings, the findings from this systematic review suggest that these foods are now making up a significant portion of total dietary intake among infants and young children in low-income settings in some regions of the world.
This systematic review also identified a need for further research to examine the relationship between snack food/SSB and dietary nutrient intake adequacy during the crucial complementary feeding period in LMIC. Although no studies in this review explored the relationship between snack food/SSB consumption and micronutrient intakes, based on the average % TEI noted, it is plausible that such consumption patterns are contributing to reduced dietary nutrient intakes among young children in LMIC settings. There is increasing evidence that high intakes of energy-dense, nutrient-poor snack foods/SSB contribute to micronutrient dilution and reduced nutrient intakes among adolescents and adults (Louzada et al., 2015). In a systematic review of evidence evaluating the nutritional significance of added sugar consumption, Gibson (2007) concluded that very high intakes of added sugars (over 20% of energy intake)-particularly when consumed in the form of soft drinks, sugar, and sweets-are correlated with lower intakes of some micronutrients among school-age children in high-income settings. Among U.S. children 8-18 years of age, Kant (2003) found that mean intakes of vitamins A and B6 and folate, as well as calcium, magnesium, iron, and zinc, all declined with increased consumption of low-nutrient-dense foods (candy, baked and dairy desserts, salty snacks, and SSB), with these foods contributing 30% TEI on average in the study sample. Among Australian 16-to 24-month-olds, Webb et al. (2006) noted reduced intakes of many nutrients, including calcium, zinc, and vitamin A, among the highest consumers of snack foods/SSB, with these foods contributing 27% TEI on average. Among South African 1-to 3-year-olds, those in the highest quartile of added sugar consumption (based on % TEI) had lower intakes of calcium, iron, and zinc, as compared with toddlers with lower % TEI from added sugar (Maunder, Nel, Steyn, Kruger, & Labadarios, 2015). Five studies in this review (Anderson et al., 2008;Jeharsae et al., 2011;Karnopp et al., 2017;Lander et al., 2010;Rodríguez-Ramírez et al., 2016) (Rennie & Livingstone, 2007), as it has not yet been established if such displacement translates into micronutrient deficiencies or growth faltering (Gibson, 2007 (Kuriyan et al., 2012). A positive association between SSB consumption and overweight/obesity among preschoolers in the United States has also been noted (Welsh, 2005), whereas another U.S. study found an association between SSB consumption and growth faltering among preschoolers (Dennison, Rockwell, & Baker, 1997). The influence of such consumption patterns on growth outcomes would likely be different in LMIC contexts, where constrained diets are often nutrient-poor, and among young children, whose nutrient requirements are high. Given these differing circumstances, micronutrient dilution from high consumption of energy-dense/nutrientpoor foods could plausibly contribute to micronutrient deficiencies and poor growth outcomes and requires further research.
Limitations in study design, particularly related to sampling and measurement, challenge the ability to draw conclusions across papers identified in this review. Among the nine studies assessing % TEI from snack foods/SSB, only two provided nationally representative estimates, both from Mexico (Denney et al., 2017;Rodríguez-Ramírez et al., 2016). Three of the remaining studies assessed nonrandom samples, including convenience samples of stunted children and mothers attending health services (Anderson et al., 2008;Roche et al., 2011;Valmórbida & Vitolo, 2014), limiting conclusions from these papers to wider populations. The four studies testing associations with nutritional outcomes also did not utilize representative samples-three studies systematically sampled mothers attending health centres for child health/vaccination services (Budree et al., 2017;Jimenez-Cruz et al., 2010;Vakili et al., 2015), and one included all children within catchment areas of selected health facilities (Faber, 2007). Additionally, limitations in measurement of exposure and outcomes are noted among the four studies testing associations between consumption and nutritional outcomes. First, none of these studies utilized a comparable measurement of consumption of snack foods/SSB; the four separate measurements of consumption included any consumption in the last week, consumption on at least 4 days in the previous week, daily consumption based on weekly recall, and "use" or "non-use" of foods for child feeding. In addition to restricting comparability across the four studies, such measurements crudely capture consumption of snack foods and SSB and would not accurately estimate the magnitude of consumption that would theoretically influence nutritional outcomes. A comparable measure of consumption, specifically one that quantifies the intakes of snack foods/SSB such as % TEI, would aid investigations into the relationship between consumption and diet/nutritional outcomes. Finally, although two studies used standardized measurements and definitions for anthropometrics, the lack of definition of "growth delay" in the paper by Vakili et al. (2015) prohibits understanding of the study findings and comparison with other papers.
The wide range of definitions used for snack foods and SSB, in both studies testing associations and studies describing % TEI, is clear from this review. Varying definitions included: the NOVA classification (with further variation between studies on inclusion of processed and/or ultraprocessed foods; Moubarac, Parra, Cannon, & Monteiro, 2014), specific food types such as "desserts" and "candy," and food categorization such as "sugary foods." Additionally, three studies did not provide a specific definition of snack foods, or indicate if both foods and SSB were included in their definition. As diets continue to evolve in LMIC, there is a need for a standardized definition to allow for comparisons between geographic areas, rural and urban populations, and across time. Such a definition could include food/beverage types that are typically common across geographies (i.e., candies, biscuits, and soft drinks) but also context-specific foods (e.g., instant noodles and aguas frescas) and could also differentiate snack foods versus SSB given differential trends in use for young child feeding. The underlying hypothesis for influence of these foods on diet/nutritional status is centred upon these foods being both energy-dense and nutrient-poor; therefore, it is also recommended that the nutrient profile of these foods be assessed when possible.
This review indicates that snack foods and SSB contribute a substantial proportion of dietary energy intakes among young children in LMIC. However, there is a need to standardize definitions of such foods and coordinate measurement in order to better understand the influence of these consumption patterns on nutritional outcomes.
Though diet displacement and micronutrient dilution from processed foods and added sugars have been noted among children in highincome settings, the impact of such dilution in contexts struggling with undernutrition may be significantly higher and may be contributing to childhood undernutrition. Additional studies exploring the relationship between % TEI from processed foods and SBB and nutritional outcomes-including nutrient intakes, micronutrient status, and growthamong representative samples of young children in LMIC, particularly in Asia and Africa, are needed to better understand this issue. As economies develop and food systems change, there is a timely need for further investigation into the role of these foods in child nutritional outcomes in order to protect and promote nutritious and appropriate young child feeding.