Childhood obesity is a major public health problem not only owing to the global increase in the prevalence during the last four decades [1
], but also to the adverse health consequences in child and adult life [3
]. The estimated worldwide prevalence of childhood overweight and obesity in 2010 was 6.7% and is expected to reach up to 9.1% in 2020 [1
]. In Europe, approximately 7% of children under the age of 10 years were estimated to have obesity according to the International Obesity Task Force (IOTF) criteria [5
], similar to figures observed in Spain, where the prevalence of obesity in children from 2 to 10 years was 5.6% in boys and 6.8% in girls [6
It has been pointed out that obesity is a complex disorder that is affected by many interacting genetic and non-genetic factors that should be further investigated [7
]. In a systematic review on the causes of obesity in children that included 61 articles, it was concluded that there was sufficient evidence to support the association between low physical activity and genetic factors and the development of excessive fatness in children and adolescents, although there was less evidence for other factors such as sedentary behavior or dietary factors among which were included sugar-containing beverages (SCB) intake [8
The SCB are non-alcoholic beverages whose main nutritional component is sugar, naturally present (e.g., fruit juices) or industrially added (e.g., fruit drinks or sugar-sweetened soft drinks). The consumption of SCB has increased in parallel with the obesity epidemic in the last few decades in the World [10
] and therefore, the World Health Organization considers that it may be implicated in the obesity epidemic [11
], as a probable result of the high sugar content and the low capacity of satiation of this type of drinks compared to other foods, thus providing extra calories above energy requirements and leading to weight gain [12
The association between SCB and obesity has been extensively investigated, and despite some controversy [15
], the present evidence supports a positive association between SCB consumption and weight gain, particularly among adults [17
] and, to a lesser extent, among children [20
]. In a recently published review based on 37 cohort studies carried out in children under 12 years of age, a positive association was shown between the total consumption of SCB and excess weight gain, and a positive association for fruit juices among children under 5 years of age [23
]. This positive association between fruit juices and obesity in children under 5 years should be highlighted because fruit juices are the most commonly consumed SCB at this age [24
]. The positive association between SCB and obesity has also been reported in children between 5 and 18 years of age from Mediterranean countries [25
], although the association has not been reported by others [28
]. Thus, it seems there is increasing evidence that the consumption of fruit juices is related to obesity in young children, although the available information is still insufficient and not fully consistent, particularly in Mediterranean areas [24
The aim of this study was to assess the association between the consumption of SCB, including packaged juices and sugar-sweetened soft drinks, and obesity in children aged 4–5 years in Spain.
The overall prevalence of obesity was 5.9%, ranging from 3.5% in Gipuzkoa to 9.3% in Asturias. The main sociodemographic and lifestyle characteristics of the parents and children according to the non-obesity and obesity status of children are shown in Table 1
. A lower maternal educational level and social class, smoking during pregnancy, and higher parental BMI and maternal weight gain were more common in children with obesity (p
< 0.05). Moreover, children with obesity showed a higher use of formula feeding during infancy, higher level of television viewing, and intake of trans-fatty acids, and lower overall physical activity and milk intake (p
The mean daily consumption of SCB was significantly higher (118.4 vs 76.6 mL) in children with obesity than children without (Table 2
). Children with obesity also showed a higher consumption of packaged juice (p
= 0.03) than children without obesity (Table 2
In the most fully adjusted multivariate models, the consumption of SCB and packaged juices were positively associated with obesity in children at the age of 4–5 years (Table 3
). Compared to the consumption of less than one drink/week of SCB, the consumption of more than one drink/day was associated with higher prevalence of obesity, OR = 3.23 (IC 95% 1.48-6.98). The consumption of one drink of SCB a day was associated with a significantly higher prevalence of obesity, OR = 1.55 (IC 95% 1.14–2.09). We also observed positive significant associations for the consumption of one drink of packaged juice, OR = 1.55 (IC 95% 1.09–2.15), and for one drink of sugar-sweetened soft-drinks, OR = 1.59 (IC 95% 0.76–3.39), although this association was not statistically significant. When total energy intake was not included in the multivariate analysis, the association between the consumption of SCB, packaged juice and sugar-sweetened soft drinks, and obesity remained practically unchanged.
shows the results of sensitivity analyses for the association between the consumption of one drink of SCB, packaged juice and sugar-sweetened soft drinks, and obesity. The associations were practically unchanged for all the sensitivity analyses, although the associations were slightly increased when maternal SCB consumption during pregnancy was added to the multivariate analysis.
This study suggests that the consumption of SCB is associated with obesity in children 4–5 years old. The association seems to be mainly mediated by the consumption of packaged juices, by far the most frequently SCB consumed by the children in our study (3 out 4 SCB drinks were from packaged juices). We have also observed that the consumption of sugar-sweetened soft drinks was associated with obesity although this association was not statistically significant, probably because of the lack of statistical power in part due to the much lower consumption of soft drinks among preschool children.
The direct association between SCB consumption and weight gain, overweight, and obesity in children and adolescents from 6 months to 19 years of age was assessed in an umbrella analysis of systematic literature reviews and meta-analyses published before 2015, mostly based on cross-sectional studies [39
]. Although our results are in part consistent with this review, we should be cautious in making direct comparisons with other studies due to differences in the age range of the children among the studies, the SCB definition, or the criteria used to define obesity in children [40
]. In one cross-sectional study with 9600 children between the ages of 4 and 5 years, the SCB intake was directly associated with BMI [45
]. However, other studies carried out in older children from Mediterranean areas have shown mixed results. In one study among 856 Greek children between the ages of 4 and 7 years, the consumption of sugar-added beverages was positively associated with BMI and a greater risk of being overweight or obese [26
], whereas in another study with 1675 Portuguese children between the ages of 5 and 10 years, no association was found between SCB consumption and overweight and obesity [29
To the best of our knowledge, this is the first study reporting a positive association between the consumption of packaged juices and obesity in preschool children in a Mediterranean country. A positive association between fruit juice intake and overweight has also been reported in among children aged 1–5 years in two longitudinal studies [45
], although not in others [47
]. The lack of consistency among studies may be due to different reasons in part mentioned previously [49
], such as the small sample size of some studies [48
], the differences in child age range [45
], the use of different criteria for defining obesity [45
], or the discrepancies in the methods used to assess and define the fruit juices (e.g., some studies only considered 100% fruit juices) [45
]. It should be noted that children in our study consumed 64 mL/day of packaged juices (including fruit juices and fruit drinks), whereas the consumption reported in other studies for 100% fruit juices in preschoolers was substantially higher, ranging from 127 to 710 mL/day [45
] which may also justify inconsistencies among study results. The consumption of 100% natural orange juice was low among children of our study, it was not considered as a SCB and therefore, we did not explore the independent association with obesity. However, when we included natural orange juice within the category of total SCB, the association observed between SCB and obesity was attenuated (data not shown).
The positive association between the consumption of sugar-sweetened soft drinks and obesity in our study did not reach statistical significance which could be in part due to the lack of statistical power because of the low consumption of sugar-sweetened soft drinks in children of our study (15.1 mL/day), which was much lower than that reported in the studies by Newby (32.5 mL/day) [47
] and Lim (180.4 mL/day) [43
]. Although a cross-sectional study among 2-year-old children has reported a higher proportion of obesity among children that consumed one or more soft drinks a day [43
], other studies in US children under 5 years of age have shown inconsistent results between soft drinks and weight gain, overweight or obesity [47
Regarding possible mechanisms, it has been hypothesized that the association between SCB and obesity might be related to the high sugar content and the low capacity of satiation of this type of drink compared to other food. Therefore, their usual intake can provide extra calories above energy requirements leading to weight gain [12
]. According to our data, the mean daily consumption of SCB was 79.1 mL, mostly from packaged juices (64 mL), thus providing nearly 155 extra kJ per day (230 kJ/day in the case of children with obesity). In this sense, it has been pointed out that small differences in energy intake could produce some weight gain on the long-term if other factors such as socio-economic status or energy expenditure remained stable [54
]. However, we did not find any statistically significant difference in energy intake between children with obesity and non-obese children and therefore, we should be cautious before concluding an effect of energy intake on children obesity. When we did not adjust for total energy intake in multivariate analyses (model 2 in tables), the associations observed were similar or slightly increased, although the main conclusion of the study remain the same.
However, other mechanisms may be involved as obesity is a complex condition influenced by a wide-range of factors and, occasionally, with interactions between them [7
]. In this sense, several modifiable factors such as parents’ lower socioeconomic status or educational level have been associated with higher SCB consumption [55
], and also with childhood obesity [56
]. Children from lower maternal social class and lower educational level showed a higher consumption of SCB, and a higher risk of obesity as well (data not shown), although the significant association between SCB consumption and obesity observed in multivariable analysis in our study was adjusted for the potential effect of these two variables. We also observed higher milk consumption in children with obesity although when milk consumption was controlled in multivariate analyses, the association between SCB and obesity was only slightly increased as shown in sensitivity analyses.
Our study has several limitations. The cross-sectional analysis may limit causal interpretations of study findings owing to potential reverse causation. However, our study was nested in a prospective mother-child cohort study and information for many variables were recorded from parents prior to the outcome at an early age of the children. In addition, we ran different sensitivity analyses to evaluate the robustness of the findings such as those analyses excluding under or over-reporting energy intake children, and the results remained essentially unchanged. Another possible limitation of our study may relate to misclassification of obesity and the inclusion of children with overweight in the reference category. We performed specific sensitivity analyses by excluding children with overweight from the reference category and the association remained significant.
A further limitation of this study may relate to the measurement errors from dietary assessment. The SCB consumption of children was reported by parents and it is possible that some miss-reporting and/or underestimation may have occurred because parents did not know well the food intake of children at the preschool centers or under caregivers. Another limitation of the dietary assessment was that it did not allow us to differentiate between subtypes of packaged juices such as 100% fruit juices, fruit juices with or without sugar added or fruit drinks. In addition, we did not collect information on the way the juices were prepared and we recognize this may be a limitation of our study. Nevertheless, the FFQ captured usual diet during the previous 6–9 months, and showed a good reproducibility for packaged fruits and sugar-sweetened soft drinks and satisfactory overall validity to assess the dietary intake, thereby making less likely a potential differential misclassification. Finally, the use of prevalence odds ratios may have overestimated associations although this potential bias should be small given the low prevalence of the outcome.
The strengths of our study include the large sample of children from a population-based mother-child cohort study in four different regions in Spain, allowing us to explore the effect of a relatively wide range of consumption for different types of SCB, and the use of high quality standardized measurements and protocols to collect all the information [30