Associations between household food environment and daily intake of regular and diet soft drinks per BMI status of European children: Feel4Diabetes Study
Abstract
The objective of this study was to investigate how the availability of food in the household environment is associated with a daily intake of regular and diet soft drinks in European children, considering BMI status. This cross-sectional study utilised baseline data from 12 211 schoolchildren participating in the Feel4Diabetes European lifestyle modification intervention. Sociodemographics, soft drink intake and household food availability data were collected using parent-completed questionnaires. Anthropometry was recorded, and children were classified into BMI categories according to the International Obesity Task Force cut-offs. In the multivariate logistic regression analysis controlled for children's sex, mother's BMI, and educational level, frequent household availability of fruit juice (sugar added), regular soft drinks and salty snacks compared to less frequent were positively associated with daily regular soft drink intake in children, regardless of BMI group (ORs range 1.59–6.69). Conversely, frequent availability of fruit juice (no added sugar) was inversely related to regular soft drink intake in both BMI groups, as was the availability of fresh fruit in the overweight/obesity group, and the availability of diet soft drinks in the underweight/normal-weight (ORs range 0.31–0.54). In conclusion, habitual household availability of selected energy-dense foods/beverages was positively associated with a daily intake of regular soft drinks in European children, regardless of BMI status. Contrastingly, household availability of fresh fruit, fruit juice (no added sugar) and diet soft drinks were inversely associated with regular soft drink intake. Programmes focusing on reducing children's soft drink intake should consider reducing the availability of sugar-added beverages in the household food environment and encouraging water consumption, as a practical, healthier alternative suggestion.
INTRODUCTION
Obesity is defined as excessive accumulation of body fat which is detrimental to health (WHO, 2022). Most importantly, obesity has its roots in early childhood and is a strong risk factor for adult obesity and a precursor for a range of adult-onset chronic diseases increasingly observed in children such as hypercholesterolemia, high blood pressure, insulin resistance and fatty liver (Sahoo et al., 2015). Furthermore, obesity has a profound impact on quality of life, psychosocial co-morbidities and risk of early mortality along with individual and societal economic burdens attributed to increased medical costs, reduced work productivity, school absenteeism and poor academic performance (An et al., 2017; Sahoo et al., 2015; Tremmel et al., 2017). Of the 340 million children and adolescents (5–19 years) classified as overweight or obese globally (WHO, 2021), one-third dwell in the European region (WHO, 2022). Escalating rates of overweight and obesity in European schoolchildren necessitate urgent consideration of the household environment as a potential contributor. However, despite ongoing research and significant media attention, effective prevention of excess bodyweight in childhood remains elusive. Hence, elucidating risk factors associated with the development of obesity early in life is of utmost importance in public health.
Globalisation trends indicate a progressive transition towards a Westernised lifestyle, which includes increased dietary intake of energy and fat, reduction in physical activity levels and increased sedentariness (Costa-Font & Mas, 2016). An unhealthy dietary pattern, combined with increased portion size and snacking on high-energy foods, has been implicated as one of the many determinants of childhood obesity (Sahoo et al., 2015). Given the energy density of sugar-sweetened beverages (SSBs) including soft drinks and fruit juices, there has been considerable interest in the relationship between these beverages and excessive weight gain. Convincing evidence from systematic reviews supports an aetiological link between habitual intake of SSBs and obesity (Luger et al., 2017; Malik & Hu, 2022). Luger et al. (2017), in a systematic review of 26 prospective cohort studies and four randomised controlled trials that included data from 242 352 subjects (n = 56 340 children), revealed positive associations between SSB intake, weight gain and body mass index (BMI) in both adults and children (Luger et al., 2017). Conversely, in a meta-analysis of randomised controlled trials performed in children, Nissensohn et al. (2018) reported weak scientific evidence on the relationship between SSBs and childhood obesity due to limited data and high heterogeneity among study designs (Nissensohn et al., 2018).
Sugar-sweetened beverages are defined as beverages containing free sugars (including regular soft drinks, fruit juices, flavoured water, energy and sports drinks, ready-to-drink teas/or coffee and flavoured milks) and are the main source of added sugars in the diet (Malik & Hu, 2022). Contemporary data from round 4 of the WHO European Childhood Obesity Surveillance Initiative (COSI) conducted during 2015–2017 reported that 9.4% of 7-year-old children consumed soft drinks daily (Williams et al., 2020). Marked differences were observed across countries, with the lowest daily consumption rates noted in Ireland, Lithuania and Denmark (up to 2%) and the highest in Tajikistan and Turkmenistan (up to 32.8%) (Williams et al., 2020). In the same direction, according to EUROSTAT, in 2019, 9% of European adolescents and adults above 15 years of age consumed sugar-sweetened soft drinks daily, with the highest intake documented in Belgium at 20% and the lowest in Estonia at 2.5% (EUROSTAT, 2019). This is worrisome because a typical 355 mL serving of soft drink contains 35–37 g of sugar, providing 140–150 kcal, which equates to around 6.5% of total daily calorie intake for adults (Malik & Hu, 2022). One can of soft drink alone fulfils more than half of the WHO recommendation of consuming less than 10% of total energy intake from free sugars from food and beverages, including sugars that are naturally present in honey, syrups and fruit juices (WHO, 2015). This recommendation is based on the recognition of the link between SSB and weight gain (WHO, 2015). Worth consideration is that while previous studies have reported that the majority of children consume SSBs in the household environment (Hafekost et al., 2011; Wang et al., 2008), most public health strategies have focussed on external sources of SSBs (such as fast food outlets) (Hafekost et al., 2011). Therefore, to reduce intake, perhaps future interventions should target household consumption.
Parents are important agents of family dietary habits since they control aspects of the household food environment such as the availability and accessibility of healthy/unhealthy food and beverages, as well as rules surrounding food choices, meal habits and quality of family meals including parenting strategies employed regarding healthy food intake (Patrick & Nicklas, 2005). Grimm et al. reported that non-alcoholic carbonated beverage (regular soft drinks) availability at home was associated with three times increased odds of children consuming soft drinks ≥5 times per week (Grimm et al., 2004). High regular soft drink intake has been correlated with high BMI in children (Katzmarzyk et al., 2016; Papandreou et al., 2013). On the other hand, although artificially sweetened beverages (ASBs) defined as zero or low-calorie drinks, including diet soft drinks, have emerged as a healthy alternative to SSBs, the relationship between obesity and diet soft drink intake has not received the same amount of attention as SSBs. Nonetheless, contemporary data from meta-analyses support that the substitution of SSBs with ASBs results in modest reductions in energy intake and bodyweight in children and adults (Rios-Leyvraz & Montez, 2022; Rogers et al., 2016), although this is not recommended by the WHO (WHO, 2023). Bearing this in mind, more longitudinal studies are warranted to investigate the efficacy of ASBs in the context of a healthy diet as a preventive measure against childhood obesity and in the maintenance of a healthy weight.
Given the gap in the current literature addressing the relationship between household food availability and soft drink intake in children with high BMI specifically in the European region (Gallagher et al., 2021; Papandreou et al., 2013; Zheng et al., 2014), the scope of the present study was to investigate the availability of food in the household environment associated with a daily intake of regular (sugar-containing) and diet soft drinks in European children of normal weight compared to those of overweight and obesity. Further insight into key aspects of the household food environment concerning childhood overweight will help in the formulation of interventions and obesity prevention programmes. We hypothesised that the availability of less healthy foods/beverages in the household (exposure variable) is positively associated with regular and diet soft drink intake (outcome of interest) in children with overweight/obesity.
METHODS
Subjects and procedures
This current cross-sectional study utilised baseline data from children participating in the EU-funded Feel4Diabetes Project conducted from April 2016 to 2018 (https://feel4diabetes-study.eu/). The primary aim of the Feel4Diabetes study was to develop, implement and evaluate a school- and community-based intervention to prevent type 2 diabetes among families from vulnerable groups across Europe. In essence, the parent study was a lifestyle intervention study of cluster-randomised design that included families recruited from the overall population located in selected provinces and municipalities of low socio-economic areas in high-income Northern European countries (Belgium, Finland), low-middle income Central European countries (Bulgaria, Hungary) and from Southern European countries under austerity measures (Greece, Spain). Children attending the first 3 years of primary school were eligible to participate. Additional details regarding study design, recruitment and procedures have been published elsewhere (Manios et al., 2018). The Feel4Diabetes study has been registered at clinicaltrials.gov (NCT02393872).
Written consent/ethical approval
Before the enrolment of subjects into the Feel4Diabetes study, protocol approval was granted from the relevant human ethical committee of all institutions and local authorities in the six participating countries (Medical Ethics Committee of the Ghent University Hospital, Belgium [Ethical Approval {EA} No.: B670201524237; 21/04/15]; Ethics Committee of the Medical University of Varna, Bulgaria [EA No.: 52/10–3-201r; 10/03/16] and the Municipalities of Sofia and Varna, as well as the Ministry of Education and Science local representatives; Hospital district of Southwest Finland ethical committee [EA: No.: 174/1801/2015; 13/03/15]; Bioethics Committee of Harokopio University, Greece and the Greek Ministry of Education; [EA No.: 46/3-4-2015; 03/04/15]; National Committee for Scientific Research in Medicine, Hungary [EA No.: 20095/2016/EKU; 29/03/16]; Clinical Research Ethics Committee and the Department of Consumers´ Health of the Government of Aragón, Spain [EA No.: CP03/2016; 08/04/15]). Additional information is available in a previous publication (Manios et al., 2018). Guardians of children received details of the aims and study procedures and signed written informed consent. The study protocol was performed in accordance with the standards set out in the Declaration of Helsinki and the conventions of the Council of Europe on Human Rights and Biomedicine.
Measurements
All assessments and data for these analyses were obtained from families during the baseline measurements in the school setting by trained research personnel. Questionnaires designed with detailed instructions for guardians/parents were self-completed. Sociodemographic information related to parents and children – age, sex, maternal/paternal educational level and parents' anthropometry weight and height, was collected via questionnaires.
Anthropometry
Mothers' anthropometry details were collected from self-reported questionnaires. Children's height and weight were measured by study staff using standardised WHO procedures (WHO, 2008) and tools in all countries. Weight was measured in children with shoes removed and in light clothing using calibrated scales to the nearest 1/10 of a kilogram and standing height was recorded to the nearest 1/10 of a cm using calibrated digital scales (SECA, 813) and an upright stadiometer (SECA 217). Both measurements were triplicated to ensure accuracy, while average values of bodyweight and height were calculated. Then BMI was calculated using the Quételet equation (kg/m2). Children were classified into one of four categories underweight (UW), normal-weight (NW), overweight (OW) and obese (OB) based on the International Obesity Task Force (IOTF) cut-off points which are based on age- and sex-specific growth charts for children from 2 to 18 years (Cole & Lobstein, 2012). For analytical purposes, underweight and normal weight categories were combined and overweight with obese to form a dichotomous variable underweight/normal weight (UW/NW) versus overweight/obese (OW/OB). In addition, BMI z-scores were computed by (measured BMI value – average BMI value of the sample population divided by the standard deviation [SD]), where the z-score represents the number of SD units above or below the average/or mean. Mothers' BMI ≥25 kg/m2 signified overweight and obesity based on WHO thresholds (WHO, 2021).
Children's daily intake of soft drinks
Details on children's soft drink intake were collected by a questionnaire self-reported by guardians that assessed the daily and weekly intake of selected foods and drinks (Data S1). This validated short questionnaire (Anastasiou et al., 2020) consisted of nine questions that assessed the frequency of food (fresh fruit, canned fruit, vegetables, sweets, salty snacks and fast food) and beverage intake (water, freshly squeezed fruit juice or commercial fruit juice with no added sugar, regular sugar-containing soft drinks and diet soft drinks) ranging from <1 time/week to more than 6 times/day. Portion sizes were defined by conventional household units using kitchen tools (for example ½ cup of cut or canned fruit, 1 cup of fruit juice, 1 cup of water, or soft drink). For this analysis, data on diet and regular sugar-containing soft drinks were utilised. Children's water and canned fruit intake were not assessed since they were beyond the scope of this analysis. The frequency of daily intake was estimated by re-categorising the eight options into two categories (<0 times/day vs. ≥1 time/day), based on previous studies (Millar et al., 2014; Williams et al., 2020). The cut-off of 1 time/day was selected to identify the habitual daily intake of soft drink consumers (Park et al., 2015).
Household food environment
The household food environment was assessed via a validated 8-item questionnaire (Anastasiou et al., 2020) completed by parents using the following question: ‘On a weekly basis, how often are the following foods available at your home?’ (Data S1). Specific foods investigated were fresh fruit, freshly squeezed fruit juice, commercial fruit juice with no added sugar and sugar-added, regular soft drinks containing sugar, diet soft drinks, vegetables, sweets and salty snacks. Possible responses were ‘always’, ‘often’, ‘sometimes’, ‘rarely’ and ‘never’ which were recoded into three categories ‘always/often’, ‘sometimes’ and ‘rarely/never’.
Covariates Children's sex, maternal educational level and BMI status were based on parents'/guardians' self-report. Maternal educational level was assessed by six options: < 6 years, 7–9 years, 10–12 years, 13–14 years, 15–16 years and > 16 years which were re-grouped into two categories ≤ 12 years and > 12 years, where 12 years is the cut-off for completion of primary and secondary schooling in most European countries (Baïdak & Sicurella, 2021).
Statistical analysis
All continuous variables were checked if they conformed to the normal distribution using the Kolmogorov–Smirnov test and visually by the histogram plot. In tables, normally distributed continuous variables are presented as means and standard deviations (SD) and categorical variables as counts (n) and percentages (%). Group differences were assessed by applying Pearson's X2 test.
After stratification by BMI group (underweight/normal vs. overweight/obese), cross-sectional associations between household food availability (independent variable) and children's daily soft drink intake (regular and diet) (dependent variables) were estimated by applying logistic regression models. Adjustments were made for confounding factors and specifically for children's sex, maternal overweight/obesity and maternal educational level based on evidence from the literature (Ratajczak & Petriczko, 2020; Ruiz et al., 2016; Shah et al., 2020). Two regression analyses were executed. The first analysis used the daily intake of regular soft drinks as the dependent variable, and the second, the daily intake of diet soft drinks. The magnitude of associations is expressed as odds ratios (OR) and 95% confidence intervals (CI). The model fit was estimated by the Nagelkerke coefficient (R2) which ranges from 0 (poor fit of the model) to 1 (good fit). An R2 value of ≤0.2 indicates a weak relationship between the predictors and the outcome (Baguley, 2012). Significant associations at p < 0.05 in the univariate analysis were entered simultaneously as independents in the multivariate analysis and adjusted for the same confounding factors. Computations were executed using SPSS version 27 (IBM), and for all statistical tests conducted significance was defined at p < 0.05.
RESULTS
A total of 12 211 children completed baseline assessments, half were boys (n = 6031), with a mean age of 8.2 SD (1.0) years, and one-quarter of European children were classified as overweight/obese according to IOTF thresholds. The mean BMI z-score was 0.56 (1.09) kg/m2 and ranged from −6.87 to 4.70. Mothers were the main caregivers; the majority were under 45 years of age (90.4%) and had over 12 years of schooling (71.3%). Concerning maternal weight status, approximately one-third (34.4%) presented with overweight/obesity. Sociodemographic data for the total sample is displayed in Table 1.
| Frequency % (n) | |
|---|---|
| Country of residence | |
| Belgium | 14.7% (1800/12280) |
| Bulgaria | 24.9% (3059/12280) |
| Finland | 12.3% (1506/12280) |
| Greece | 18.6% (2287/12280) |
| Hungary | 15.6% (1921/12280) |
| Spain | 13.9% (1707/12280) |
| Children's data | |
| Sex (Boys) | 49.4% (6031/12211) |
| Age (years) (Mean ± S.D) | 8.2 ± 1.00 |
| BMI (z-scores) (Mean ± S.D) | 0.56 ± 1.09 |
| BMI group | |
| Underweight/normal | 74.5%(8962/12030) |
| Overweight/obese | 25.5% (3068/12030) |
| Overweight/Obese | |
| Boys | 23.7% (1408/5933) |
| Girls | 27.2% (1660/6097) |
| Mothers' data | |
| Age < 45 years | 90.4% (10 651/11787) |
| Education > 12 years | 71.3% (8076/11328) |
| BMI group Overweight/obese | 34.4% (3726/10828) |
- Abbreviation: BMI, body mass index.
Regarding children's soft drink intake, about 15% (n = 1657/10878) of European children consumed regular soft drinks ≥1 time/day, and 5% (501/9989) diet soft drinks. Sex differences were observed for regular soft drink intake with 54.2% (896/1654) of boys and 45.8% of girls (758/1654) consuming regular soft drinks ≥1 time/day (p < 0.001 [X2 test]), but not for daily intake of diet soft drinks (Boys: 50.3% [252/501]; Girls: 49.7% [249/501; p = 0.76]).
Univariate associations between the household food environment and daily soft drink intake (regular and diet) by BMI category after controlling for children's sex, maternal BMI group and educational level are presented in Table 2. Data for the crude analysis is presented in the Online Table S2.
| Regular soft drinks ≥1 time/day | Diet soft drinks intake ≥1 time/day | |||
|---|---|---|---|---|
| Underweight/normal | Overweight/obese | Underweight/normal | Overweight/obese | |
| Adjusted | Adjusted | Adjusted | Adjusted | |
| Household food availability | ORadj (95%CI), padj* | ORadj (95%CI), padj* | ORadj (95%CI), padj* | ORadj (95CI), padj* |
| Fresh fruit | ||||
| Rarely/never | Ref | |||
| Sometimes | 1.96 (0.91–4.20), padj = 0.08 | 0.40 (0.18–0.86), padj = 0.019 | padj = 1.00 | 0.54 (0.09–3.10), padj = 0.49 |
| Always/often | 1.12 (0.55–2.31), padj = 0.75 | 0.29 (0.15–0.56), padj <0.001 | padj = 1.00 | 0.98 (0.23–4.21), padj = 0.98 |
| Fruit juice fresh/ commercial fruit juice (no sugar added) | ||||
| Rarely/never | Ref | |||
| Sometimes | 1.08 (0.90–1.28), padj = 0.41 | 0.83 (0.61–1.12), padj = 0.23 | 1.38 (0.99–1.91), padj = 0.05 | 0.87 (0.48–1.59), padj = 0.66 |
| Always/often | 0.70 (0.60–0.83), padj <0.001 | 0.60 (0.46–0.79), padj <0.001 | 1.18 (0.88–1.60), padj = 0.27 | 1.41 (0.88–2.27), padj = 0.15 |
| Fruit juice commercial (sugar added) | ||||
| Rarely/never | Ref | |||
| Sometimes | 2.31 (1.85–2.89), padj <0.001 | 2.83 (1.98–4.03), padj <0.001 | 1.96 (1.40–2.74) padj <0.001 | 1.16 (0.66–2.03), padj = 0.59 |
| Always/often | 8.01 (6.64–9.68), padj <0.001 | 8.74 (6.36–12.01), padj <0.001 | 2.78 (2.07–3.74), padj <0.001 | 3.03 (1.92–4.78), padj <0.001 |
| Regular soft drinks | ||||
| Rarely/never | Ref | |||
| Sometimes | 1.91 (1.56–2.34), padj <0.001 | 2.08 (1.50–2.87), padj <0.001 | 1.50 (1.06–2.12), padj = 0.023 | 1.65 (0.98–2.78), padj = 0.06 |
| Always/often | 6.33 (5.39–7.43), padj <0.001 | 8.30 (6.28–10.97), padj <0.001 | 3.11 (2.37–4.08), padj <0.001 | 3.49 (2.24–5.44), padj <0.001 |
| Diet soft drinks | ||||
| Rarely/never | Ref | |||
| Sometimes | 1.21 (0.99–1.47), padj = 0.06 | 1.08 (0.78–1.49), padj = 0.66 | 2.45 (1.68–3.56), padj <0.001 | 1.93 (1.00–3.90), padj = 0.048 |
| Always/often | 1.43 (1.21–1.69), padj <0.001 | 1.54 (1.16–2.04), padj = 0.003 | 6.91 (5.23–9.13), padj <0.001 | 8.73 (5.53–13.79), padj <0.001 |
| Vegetables | ||||
| Rarely/never | Ref | |||
| Sometimes | 0.85 (0.45–1.60), padj = 0.60 | 0.66 (0.29–1.50), padj = 0.32 | 0.56 (0.22–1.40), padj = 0.21 | 1.05 (0.21–5.13), padj = 0.95 |
| Always/often | 0.66 (0.37–1.18), padj = 0.17 | 0.53 (0.26–1.09), padj = 0.08 | 0.36 (0.16–0.81), padj = 0.014 | 0.91 (0.21–3.92), padj = 0.90 |
| Sweets | ||||
| Rarely/never | Ref | |||
| Sometimes | 1.11 (0.79–1.56), padj = 0.53 | 1.10 (0.66–1.82), padj = 0.72 | 0.71 (0.43–1.18), padj = 0.19 | 0.76 (0.35–1.66), padj = 0.49 |
| Always/often | 2.28 (1.67–3.10), padj <0.001 | 2.45 (1.53–3.93), Padj <0.001 | 1.16 (0.74–1.83), padj = 0.52 | 1.57 (0.77–3.18), padj = 0.21 |
| Salty snacks | ||||
| Rarely/never | Ref | |||
| Sometimes | 1.80 (1.47–2.20), padj <0.001 | 1.49 (1.09–2.04), padj = 0.013 | 1.24 (0.86–1.80), padj = 0.25 | 1.21 (0.71–2.06), padj = 0.48 |
| Always/often | 3.63 (3.00–4.39), padj <0.001 | 3.94 (2.92–5.31), padj <0.001 | 3.03 (2.18–4.20), padj <0.001 | 2.86 (1.75–4.67), padj <0.001 |
- Note: In bold text statistically significant p-values at the 5% level.
- Ref: Reference group.
- Dependent 1: Regular soft drink intake (0 times/day vs. ≥1 time/day).
- Dependent 2: Diet soft drink intake (0 times/day vs. ≥1 time/day).
- Independent 1: Household food availability (1 = always/often, 2 = sometimes, 3 = rarely/never).
- *p-value estimated applying the logistic regression model adjusted for children's sex, maternal BMI group and maternal education.
Univariate regression model
Daily intake of regular soft drinks (≥1 time/day)
Table 2 indicates that in the adjusted univariate regression analysis, a more frequent household food availability of fresh fruit ‘always/often’ or ‘sometimes’ was associated with lower odds of daily regular soft drink intake in children with OW/OB ([always/often] OW/OB: ORadj: 0.29, 95%CI: 0.15–0.56; [sometimes] ORadj: 0.40, 95%CI: 0.18–0.86) as compared to less frequent household availability ‘rarely/never’, whereas frequent home availability of fresh fruit juice/or commercial juice (without added sugar) ‘always/often’ was associated with lower odds of daily intake of regular soft drinks for both the UW/NW and OW/OB groups ([UW/NW] ORadj: 0.70, 95%CI: 0.60–0.83; [OW/OB] ORadj: 0.60, 95%CI: 0.46–0.79). In contrast, independent of the BMI group, frequent availability of diet soft drinks and sweets ‘always/often’ were associated with approximately 1.5- to 2-fold higher odds of daily intake of regular soft drinks (‘Always/often’ [Diet soft drinks UW/NW: {ORadj:1.43, 95%CI: 1.21–1.69}; OW/OB {ORadj: 1.54, 95%CI: 1.16–2.04}]; [Sweets UW/NW: {ORadj: 2.28, 95%CI: 1.67–3.10}; OW/OB: {ORadj: 2.45, 95%CI: 1.53–3.93}]). Frequent availability of sugar-sweetened fruit juice and regular soft drinks and salty snacks ‘always/often’ was associated with about four to almost nine times higher odds of daily intake of regular soft drinks in both group (‘Always/often’ [Fruit juice added sugar: UW/NW {ORadj: 8.01, 95%CI: 6.64–9.68}; OW/OB {ORadj: 8.74, 95%CI: 6.36–12.01}]; [regular soft drinks UW/NW {ORadj: 6.33, 95%CI: 5.39–7.43}; OW/OB: {ORadj: 8.30, 95%CI: 6.28–10.97}]; [salty snacks UW/NW: {ORadj: 3.63, 95%CI: 3.00–4.39}; OW/OB: {ORadj: 3.94, 95%CI: 2.92–5.31}]). The same positive trend was observed for household availability of these food items at a rate of ‘sometimes’ (‘Sometimes’ [Fruit juice added sugar: UW/NW {ORadj: 2.31, 95%CI: 1.85–2.89}; OW/OB {ORadj: 2.83, 95%CI: 1.98–4.03}]; [regular soft drinks UW/NW {ORadj: 1.91, 95%CI: 1.56–2.34}; OW/OB {ORadj: 2.08, 9%CI: 1.50–2.87}]; [salty snacks UW/NW {ORadj: 1.80, 95%CI: 1.47–2.20}; OW/OB {ORadj: 1.49, 95%CI: 1.09–2.04}]).
Daily intake of diet soft drinks (≥1 time/day)
Concerning factors affecting diet soft drink intake, in the adjusted univariate analysis, a more frequent household availability of sugar-sweetened fruit juice, regular soft drinks and salty snacks ‘Always/often’, were associated with about 3-fold higher odds of daily intake of diet soft drinks for both the UW/NW and OW/OB groups as compared to less frequent ‘Rarely/never’ (‘Always/often’ [Fruit juice added sugar UW/NW: ORadj: 2.78, {95%CI: 2.07-3.74}; OW/OB {ORadj: 3.03, 95%CI: 1.92-4.78}]; [regular soft drinks UW/NW: {ORadj: 3.11, 95%CI: 2.37-4.08}; OW/OB: {ORadj: 3.49, 95%CI: 2.24-5.44}]; [salty snacks UW/NW: {ORadj: 3.03, 95%CI: 2.18-4.20}; OW/OB: {ORadj: 2.86, 95%CI:1.75-4.67}]). However, household availability of sugar-sweetened fruit juice and regular soft drinks ‘Sometimes’ was associated with almost 2-fold higher odds of daily intake of diet soft drinks in the UW/NW group only (‘Sometimes’ [fruit juice sugar added UW/NW: {ORadj: 1.96, 95%CI: 1.40–2.74}]; [regular soft drinks {ORadj:1.50, 95%CI: 1.06–2.12}]). Regarding availability of diet soft drinks, household availability ‘Sometimes’ and ‘Always/often’ were associated with two and almost nine times higher odds of daily intake of diet soft drinks in both groups (‘Sometimes’: UW/NW: [ORadj: 2.45, 95%CI: 1.68–3.56]; OW/OB: [ORadj:1.93, 95%CI: 1.00–3.90]; ‘Always/often’ UW/NW: [ORadj: 6.91, 95%CI: 5.23–9.13]; OW/OB: [ORadj: 8.73, 95%CI: 5.53–13.79]). On the other hand, frequent household availability of vegetables ‘Always/often’ was associated with lower odds of daily intake of diet soft drinks in children of UW/NW, but not for the OW/OB group (‘Always/often’ [UW/NW: ORadj: 0.36, 95%CI: 0.16–0.81]).
Multivariate regression model
Significant associations at p < 0.05 in the univariate analysis were added simultaneously as independent variables in the multivariate analysis. Multivariate associations between household food availability and daily intake of regular and diet soft drinks per children's BMI category after adjusting for children's sex, maternal BMI group and educational level are presented in Table 3. Most of the associations found to be significant in the univariate analysis were retained in the multivariate analysis except for associations between household availability of sweets and daily intake of regular soft drinks, which became non-significant in both the UW/NW and OW/OB groups. Likewise, the association between household availability of diet soft drinks and daily intake of regular soft drinks became non-significant in the OW/OB group and inversely associated with daily regular soft drink intake in the UW/NW (‘Always/often’ UW/NW: ORadj: 0.53, 95%CI: 0.43–0.65). The same pattern was observed for the daily intake of diet soft drinks. The association between household availability of regular soft drinks and daily intake of diet soft drinks was no longer significant for both BMI groups, whereas the availability of vegetables and salty snacks was insignificant in the UW/NW and OW/OB groups, respectively. In contrast, household availability of sweets became inversely associated with the daily intake of diet soft drinks in the UW/NW group (Sweets UW/NW ‘Sometimes’: [ORadj: 0.51, 95%CI: 0.29–0.90]; ‘Always/often’ [ORadj: 0.41, 95%CI: 0.23–0.73]). The outcome of the crude regression analysis is displayed in the Online Table S3.
| Regular soft drink intake ≥1 time/day | Diet soft drinks intake ≥1 time/day | |||
|---|---|---|---|---|
| Underweight/normal weight | Overweight/obese | Underweight/normal weight | Overweight/obese | |
| Adjusted | Adjusted | Adjusted | Adjusted | |
| Household food availability | ORadj (95%CI), padj* | ORadj (95%CI), padj* | ORadj (95%CI), padj* | ORadj (95%CI), padj* |
| Fresh fruit | ||||
| Rarely/never | Ref | |||
| Sometimes | 2.65 (1.07–6.51), padj = 0.034 | 0.40 (0.15–1.04), padj = 0.06 | padj = 1.00 | 1.07 (0.11–10.79), padj = 0.95 |
| Always/often | 1.43 (0.60–3.39), padj = 0.42 | 0.31 (0.13–0.74), padj = 0.009 | padj = 1.00 | 1.42 (0.17–11.75), padj = 0.74 |
| Fruit juice fresh/commercial fruit juice (no sugar added) | ||||
| Rarely/never | Ref | Ref | ||
| Sometimes | 0.91 (0.74–1.11), padj = 0.35 | 0.75 (0.52–1.07), padj = 0.12 | 1.22 (0.86–1.73), padj = 0.27 | 0.80 (0.41–1.58), padj = 0.53 |
| Always/often | 0.54 (0.45–0.66), padj <0.001 | 0.46 (0.33–0.65), padj <0.001 | 0.93 (0.67–1.29), padj = 0.66 | 0.99 (0.56–1.73), padj = 0.97 |
| Fruit juice commercial (sugar added) | ||||
| Rarely/never | Ref | Ref | ||
| Sometimes | 1.94 (1.52–2.49), padj <0.001 | 2.73 (1.82–4.09), padj <0.001 | 1.56 (1.07–2.27), padj = 0.021 | 1.22 (0.65–2.30), padj = 0.54 |
| Always/often | 5.59 (4.50–6.94), padj <0.001 | 6.69 (4.56–9.83), padj <0.001 | 1.57 (1.10–2.25), padj = 0.013 | 2.09 (1.17–3.74), padj = 0.012 |
| Regular soft drink | ||||
| Rarely/never | Ref | |||
| Sometimes | 1.44 (1.14–1.81), padj = 0.002 | 1.33 (0.91–1.94), padj = 0.14 | 0.79 (0.52–1.19), padj = 0.26 | 1.10 (0.59–2.07), padj = 0.76 |
| Always/often | 3.74 (3.04–4.59), padj <0.001 | 3.81 (2.68–5.40), padj <0.001 | 1.16 (0.81–1.65), padj = 0.41 | 1.25 (0.70–2.25), padj = 0.46 |
| Diet soft drinks | ||||
| Rarely/never | Ref | |||
| Sometimes | 0.81 (0.65–1.03), padj = 0.08 | 0.76 (0.51–1.13), padj = 0.18 | 2.38 (1.58–3.60), padj <0.001 | 1.85 (0.90–3.80), padj = 0.09 |
| Always/often | 0.53 (0.43–0.65), padj <0.001 | 0.73 (0.51–1.03), padj = 0.08 | 6.13 (4.44–8.58), padj <0.001 | 6.87 (4.04–11.67), padj <0.001 |
| Vegetables | ||||
| Rarely/never | Ref | |||
| Sometimes | 0.75 (0.35–1.62), padj = 0.46 | 0.96 (0.33–2.77), padj = 0.94 | 0.65 (0.20–2.11), padj = 0.47 | 1.10 (0.19–6.20), padj = 0.91 |
| Always/often | 0.65 (0.32–1.34), padj = 0.24 | 0.86 (0.32–2.33), padj = 0.77 | 0.41 (0.14–1.21), padj = 0.11 | 0.83 (0.17–4.04), padj = 0.82 |
| Sweets | ||||
| Rarely/never | Ref | |||
| Sometimes | 0.73 (0.50–1.06), padj = 0.10 | 0.97 (0.53–1.77), padj = 0.91 | 0.51 (0.29–0.90), padj = 0.020 | 0.66 (0.27–1.61), padj = 0.36 |
| Always/often | 0.72 (0.50–1.05), padj = 0.09 | 0.97 (0.54–1.76), padj = 0.93 | 0.41 (0.23–0.73), padj = 0.002 | 0.67 (0.28–1.60), padj = 0.37 |
| Salty snacks | ||||
| Rarely/never | Ref | |||
| Sometimes | 1.24 (0.98–1.57), padj = 0.07 | 1.10 (0.76–1.59), padj = 0.63 | 1.11 (0.72–1.70), padj = 0.64 | 0.85 (0.46–1.59), padj = 0.62 |
| Always/often | 1.59 (1.24–2.04), padj <0.001 | 1.75 (1.18–2.61), padj = 0.006 | 2.02 (1.29–3.16), padj = 0.002 | 1.39 (0.78–2.65), padj = 0.32 |
- Note: Significant associations at p < 0.05 in the univariate analysis were entered simultaneously as independents in the multivariate analysis.
- In bold text statistically significant p-values at the 5% level.
- Dependent 1: Regular soft drink intake (0 times/day vs. ≥1 time/day).
- Dependent 2: Diet soft drink intake (0 times/day vs. ≥1 time/day).
- Independent 1: Household food availability (1 = always/often, 2 = sometimes, 3 = rarely/never).
- Ref: Reference group.
- *p-values estimated applying the logistic regression model adjusted for children's sex, mothers BMI group and educational level.
- Daily intake of regular soft drinks: (Adjusted analysis) Underweight/normal weight group Nagelkerke R2 = 26.4%, overweight/obese group R2 = 28.8%.
- Daily intake of diet soft drinks: (Adjusted analysis) Underweight/normal weight group Nagelkerke R2 = 18.2%, overweight/obese group R2 = 17.6%.
Concerning sex differences and the influence of mothers' weight status and educational level, in both BMI groups lower odds of daily intake of regular soft drinks were observed in girls than in boys (UW/NW group: [ORadj:0.79, 95%CI: 0.68–0.92; Padj = 0.003]; OW/OB: [ORadj: 0.76 95%CI: 0.58–0.99; Padj = 0.042]). There was no influence of mothers' BMI group on children's consumption of regular soft drinks (UW/NW group: [ORadj: 1.12, 95%CI: 0.95–1.33; Padj = 0.16]; OW/OB: [ORadj: 1.27, 95%CI: 0.98–1.66; Padj = 0.07]). However, in children with UW/NW, those having mothers with <12 years of schooling had higher odds of daily intake of regular soft drinks (ORadj: 1.77, 95%CI: 1.50–2.08; Padj <0.001), but not in the OW/OB group (ORadj: 1.22, 95%CI: 0.92–1.61; Padj = 0.17). Regarding the model fit, the Nagelkerke coefficient R2 revealed that the factors explaining the consumption of regular soft drinks at least once daily in the UW/NW group were 26.4%, and in the OW/OB group, 28.8%.
For daily intake of diet soft drinks, children with UW/NW having mothers with OW/OB (ORadj: 1.33, 95%CI: 1.02–1.73; Padj = 0.038) and ≤ 12 years of education (ORadj: 2.64; 95%CI: 2.02–3.44; Padj <0.001) had higher odds of daily consumption. Contrastingly, only mothers with low education levels appear to strengthen the daily intake of diet soft drinks in children of OW/OB (ORadj: 1.57, 95%CI: 1.01–2.45; Padj = 0.044). In both BMI groups, children's sex did not influence daily intake of diet soft drinks (UW/NW group: Padj = 0.68, OW/OB: Padj = 0.77). The Nagelkerke coefficient R2 indicated that the model explained only 18.2% of daily diet soft drink intake in the UW/NW group and 17.6% in the OW/OB group.
DISCUSSION
Childhood is a window of opportunity to influence children's eating habits and weight status, factors impacting future health in adult life. The present study builds upon previous research by investigating multiple aspects of the household food environment associated with soft drink intake in children of normal weight and overweight/obesity residing in different countries in the EU. We sought to identify potential targets for future public health intervention.
A major finding of the multivariate analysis was that more frequent household availability (‘Always/often’) of energy-dense, processed foods (including commercial sugar-sweetened fruit juice, regular soft drinks and salty snacks) compared to less frequent household availability (‘Rarely/never’) was positively associated with daily intake of regular soft drinks in European schoolchildren, regardless of BMI group. This refutes our original hypothesis that the household availability of unhealthy foods or beverages is associated with regular and diet soft drink intake in children with overweight/obesity only. Our observations corroborate with the findings of previous studies reporting that easy access to SSBs including carbonated drinks mediated by increased availability in the household enhanced children's daily intake (Hafekost et al., 2011; Wang et al., 2008). In line with our findings, Couch et al., in a cohort study of 699 child–parent pairs, reported that household availability of energy-dense/nutrient-poor foods (such as sweets, candy, pastries, carbonated/non-carbonated sodas, sports drinks, sweetened breakfast cereals and fruit juice drinks) was positively associated with high-energy beverage intake and inversely with fruit and vegetable (FV) intake (Couch et al., 2014). This is important because daily intake of SSB including sugar-sweetened soft drinks and high-fat foods is associated with elevated BMI (Millar et al., 2014; Zheng et al., 2014) and visceral fat (Gallagher et al., 2021) in children. Gallagher et al., in a cross-sectional analysis of data from the Healthy Growth Study that included data from 2665 Greek children, reported that high SSB consumers had 1.3 units higher visceral fat than low consumers (Gallagher et al., 2021). In another cross-sectional study of 607 Greek children 7–15 years, Papandreou et al. reported that out of the 79.4% of children consuming SSBs daily, consumption was higher in children with OW/OB than in their NW counterparts. In addition, children consuming SSBs were 2.57 times more likely to have obesity (Papandreou et al., 2013). Comparably, Zheng et al., in a longitudinal study of 283 children participating in the European Youth Heart Study, showed that subjects who increased SSB intake from 9 to 15 years had higher increases in BMI and waist circumference from 15 to 21 years than peers reporting no change in consumption patterns (Zheng et al., 2014). Millar et al., in an Australian longitudinal study of 4164 children (4–10 years), reported that daily consumption of SSBs and high-fat foods (defined as fast food, salty snacks and sweets) were related to elevated BMI z-scores (Millar et al., 2014).
One more interesting point in the present study was that, in the adjusted multivariate analysis, compared to infrequent household availability (‘Rarely/never’), frequent availability of fresh fruit ('Always/often') was associated with 69% lower odds of daily intake of regular soft drinks in children of OW/OB, while frequent availability of fresh fruit juice/or commercial fruit juice without added sugar (‘Always/often’) with approximately 50% lower odds in both the UW/NW and OW/OB groups. In the same line, frequent household availability of diet soft drinks (‘Always/often’) was inversely associated with a daily intake of regular soft drinks in children of UW/NW by 47% lower odds. Prior research has demonstrated that increased accessibility and household availability of FV not only promoted FV intake in children (Bassul et al., 2020; Ong et al., 2017) but was associated with reduced intake of SSBs (Bassul et al., 2020). Studies by Birch and Fisher established that the mechanism by which household availability of FV is associated with intake is via increased food exposure to these healthy foods which eventually results in acceptance and preference for such foods by children (Birch & Fisher, 1998). One might argue that children's preference for fresh fruit and fruit juice, both freshly squeezed and commercial with no added sugar, would be healthier substitutes (USDA, 2015) for regular soft drink intake. Alternatively, increased household availability of FV might signify an overall healthier dietary pattern of parents and children favouring increased intake of FV including fruit juice, and lower intake of SSBs.
Concerning children's daily intake of diet soft drinks, in the adjusted multivariate analysis, in both BMI groups, a more frequent household availability of sugar-sweetened commercial fruit juice and diet soft drinks (‘Always/often’) was positively associated with daily diet soft drink intake compared to less frequent (‘Rarely/never’). Similarly, the frequent availability of salty snacks (‘Always/often’) and availability of sugar-sweetened commercial fruit juice and diet soft drinks ‘Sometimes’ were only positively associated with children of UW/NW. Paradoxically, the availability of sweets ‘Sometimes’ or ‘Always/often’ was inversely associated with diet soft drink intake in children of UW/NW. A possible explanation for the association between household availability of fruit juice (with added sugar) and increased daily intake of diet soft drinks could be that parents might offer and encourage intake of diet or artificially-sweetened beverages with zero calories or low-calories to balance the high-calorie intake of these beverages and as prophylaxis against weight gain (Rios-Leyvraz & Montez, 2022; Rogers et al., 2016).
It is noteworthy that in our study, associations for the frequency of regular soft drink intake were stronger in boys than in girls in both BMI groups and in children with UW/NW having mothers of low educational level, which is consistent with the existing literature (Schneider et al., 2021). Interestingly, there was no influence of mothers' BMI group on children's intake of regular soft drinks. Regarding the frequency of daily diet soft drink intake, in both BMI groups, associations were stronger in children having mothers of low education attainment, while in children with UW/NW, having mothers with OW/OB. Unlike regular soft drink intake, children's sex did not influence diet soft drink intake. Consumption patterns of SSBs have been documented to be higher in boys and households of low educational levels (Petrauskienė et al., 2015; Schneider et al., 2021). In fact, Petrauskiene et al. reported that daily consumption of regular soft drinks was 2.6 times lower among children of mothers with high education attainment and 1.7 times lower among children of fathers with high education as compared to children of low-education parents (Petrauskienė et al., 2015). This suggests that maternal educational level influences children's soft drink consumption patterns to a greater extent than fathers. A plausible explanation is that having only a few years of schooling could be related to poor health literacy and limited knowledge of healthy nutritional habits, thereby leading to unhealthy dietary choices for both mothers and their children (Zarnowiecki et al., 2014). One might postulate that the skills obtained via formal education enable parents to process and interpret basic health/nutrition information and make appropriate health decisions (Zarnowiecki et al., 2014). In reference to associations for daily intake of diet soft drinks in children having mothers of OW/OB, as discussed above, mothers with OW/OB might encourage children's consumption of diet soft drinks as a healthier alternative to regular soft drinks, to balance the overall daily total energy intake and as a weight management strategy (Rios-Leyvraz & Montez, 2022; Rogers et al., 2016).
Strengths/limitations
The findings of the current study should be interpreted considering its strengths and limitations. To our knowledge, most studies investigating the home food environment have focused on children's intake of fruit and vegetables (Ong et al., 2017), and more work is needed to determine factors that reduce the intake of SSBs, specifically regular soft drinks. Given that humans consume a variety of food groups when eating, examining the influence of household food availability on beverage intake and weight status in children has great relevance. We examined soft drink intake in both children of UW/NW and OW/OB, therefore our findings apply to all BMI groups. Although preliminary, our observations provide important directions for future research in EU obesity-prevention strategies. Nonetheless, longitudinal studies are required to confirm and elaborate upon these findings. Additional strong points of the present study are the strict standard methods and procedures used by all participating sites to measure BMI, record soft drink intake and household food availability along with the large homogeneous sample of children across a diverse group of countries.
Limitations of the current study include the cross-sectional nature of the design which is useful in identifying associations but cannot determine directionality. Bi-directionality in parent–child interactions cannot be precluded since child weight status could influence regular soft drink and diet soft drink consumption patterns along with parent responses in questionnaires. Moreover, it is unknown whether a child's weight status is driving the household food environment, or the household environment is driving child obesity. Further research should consider the role of environments outside the household on children's soft drink intake such as the school food environment and weekend habits with peers. A further limitation is the variation between the household food environment categories from one household to another, which could lead to misclassification and statistical error. In addition, the household availability of only eight food items was investigated and therefore, it does not reflect the habitual dietary pattern of European children. We did not have data on the availability of flavoured water, milk-based beverages, cordials, energy or sports drinks and fast foods that could contribute to children's soft drink intake. Furthermore, given that participating families were selected from low, low-middle income groups and those under austerity measures, our findings cannot be extrapolated to all populations. From a statistical point of view, low values of the Nagelkerke coefficient R2 (<20%) indicate that there are other factors influencing regular or diet soft drink intake independent of children's BMI that have not been accounted for (e.g. total energy intake) and that the matter is more complex than we currently understand. Another source of bias is that data were collected by guardian/parent self-reported questionnaires which may have introduced recall bias and social desirability bias. In addition, parents surveyed in the Feel4Diabetes study were highly educated which limits the generalisability of our findings. In terms of implications of the study, the findings indicate that the presence of healthy foods in the household is inversely associated with a daily intake of regular soft drinks, which can have a positive consequence on children's health.
CONCLUSION
Habitual household availability of energy-dense foods/beverages was positively associated with the daily intake of sugars-added soft drinks in European children, regardless of BMI status. In contrast, household availability of fresh fruit, freshly squeezed fruit juice and commercial fruit juice with no added sugar and diet soft drinks were inversely associated with sugar-added soft drink intake. This study suggests that programmes focusing on reducing children's regular soft drinks intake should consider reducing the availability of sugars-added beverages in the household food environment and encouraging water consumption as a practical, healthier alternative suggestion.
AUTHOR CONTRIBUTIONS
Conceptualization: YM, MP. Methodology: YM, KR, MP. Formal Analysis: YM, MP conceived the concept for the analysis. KR conducted the statistical analysis and is the principal author of the first and final draft of the manuscript. Investigation and data collection: KR, MP. Data curation: KR, MP, GM, and YM. Writing-original and final draft preparation, KR, MP, GM. Review and editing, YM, MP, GM, GC, VI, NC, PV, NU, SL, KM, RI, EA, LM. Supervision and Project Administration: YM, GM. All authors have read and agreed to the published version of the manuscript.
ACKNOWLEDGEMENTS
The authors would like to thank the members of the Feel4Diabetes-study group: Coordinator: Yannis Manios, Steering Committee: Yannis Manios, Greet Cardon, Jaana Lindström, Peter Schwarz, Konstantinos Makrilakis, Lieven Annemans, Winne Ko, Harokopio University (Greece): Yannis Manios, Kalliopi Karatzi, Odysseas Androutsos, George Moschonis, Spyridon Kanellakis, Christina Mavrogianni, Konstantina Tsoutsoulopoulou, Christina Katsarou, Eva Karaglani, Irini Qira, Efstathios Skoufas, Konstantina Maragkopoulou, Antigone Tsiafitsa, Irini Sotiropoulou, Michalis Tsolakos, Effie Argyri, Mary Nikolaou, Eleni-Anna Vampouli, Christina Filippou, Kyriaki Apergi, Amalia Filippou, Gatsiou Katerina, Efstratios Dimitriadis, Finnish Institute for Health and Welfare (Finland): Jaana Lindström, Tiina Laatikainen, Katja Wikström, Jemina Kivelä, Päivi Valve, Esko Levälahti, Eeva Virtanen, Tiina Pennanen, Seija Olli, Karoliina Nelimarkka, Ghent University (Belgium), Department of Movement and Sports Sciences: Greet Cardon, Vicky Van Stappen, Nele Huys, Department of Public Health: Lieven Annemans, Ruben Willems, Department of Endocrinology and Metabolic Diseases: Samyah Shadid, Technische Universität Dresden (Germany): Peter Schwarz, Patrick Timpel, University of Athens (Greece), Konstantinos Makrilakis, Stavros Liatis, George Dafoulas, Christina-Paulina Lambrinou, Angeliki Giannopoulou, International Diabetes Federation European Region (Belgium): Winne Ko, Ernest Karuranga, Universidad De Zaragoza (Spain): Luis Moreno, Fernando Civeira, Gloria Bueno, Pilar De Miguel-Etayo, Esther Mª Gonzalez-Gil, María L. Miguel-Berges, Natalia Giménez-Legarre; Paloma Flores-Barrantes, Alelí M. Ayala-Marín, Miguel Seral-Cortés, Lucia Baila-Rueda, Ana Cenarro, Estíbaliz Jarauta, Rocío Mateo-Gallego, Medical University of Varna (Bulgaria): Violeta Iotova, Tsvetalina Tankova, Natalia Usheva, Kaloyan Tsochev, Nevena Chakarova, Sonya Galcheva, Rumyana Dimova, Yana Bocheva, Zhaneta Radkova, Vanya Marinova, Yuliya Bazdarska, Tanya Stefanova, University of Debrecen (Hungary): Imre Rurik, Timea Ungvari, Zoltán Jancsó, Anna Nánási, László Kolozsvári, Csilla Semánova, Éva Bíró, Emese Antal, Sándorné Radó: Extensive Life Oy (Finland): Remberto Martinez, Marcos Tong.
FUNDING INFORMATION
The Feel4Diabetes study has received funding from the European Union's Horizon 2020 research and innovation programme [Grant Agreement: n° 643708]. The content of this article reflects only the authors' views and the European Community is not liable for any use that may be made of the information contained therein. The funding body had no role in the design of this study and collection, analysis and interpretation of the data, and in writing this manuscript.
CONFLICT OF INTEREST STATEMENT
The authors have no conflict of interest to disclose.
ETHICS STATEMENT
This study was conducted conforming to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were granted approval by the human ethics committees of all of the participating tertiary institutions (Medical Ethics Committee of the Ghent University Hospital, Belgium [Ethical Approval {EA} No.: B670201524237; 21/04/15]; Ethics Committee of the Medical University of Varna, Bulgaria [EA No.: 52/10–3-201r; 10/03/16] and the Municipalities of Sofia and Varna, as well as the Ministry of Education and Science local representatives; Hospital district of Southwest Finland ethical committee [EA: No.: 174/1801/2015; 13/03/15]; Bioethics Committee of Harokopio University, Greece and the Greek Ministry of Education; [EA No.: 46/3-4-2015; 03/04/15]; National Committee for Scientific Research in Medicine, Hungary [EA No.: 20095/2016/EKU; 29/03/16]; Clinical Research Ethics Committee and the Department of Consumers´ Health of the Government of Aragón, Spain [EA No.: CP03/2016; 08/04/15]). Parents/guardians of children signed written informed consent.
TRANSPARENCY DECLARATION
The lead author affirms that this manuscript is an honest, accurate and transparent account of the study being reported. The reporting of this work is compliant with STROBE guidelines. The lead author affirms that no important aspects of the study have been omitted and that any discrepancies from the study (clinicaltrials.gov [NCT02393872]) have been explained.
Open Research
DATA AVAILABILITY STATEMENT
The authors confirm that all data generated or analysed during this study are included in this published article and its supplementary online resource information files.
