Introduction

Childhood obesity is portrayed by the accrual of excessive body fat as well as the growth of excess adipocytes [1]. It is widely accepted as has been reported that overweight or obese children are at greater risk of becoming overweight or obese adults; who will face a life-time of increased risk for various diseases, including diabetes mellitus, cardiovascular disease, liver disease and certain cancers [2, 3]. Even during childhood, obesity has been reported to increase the risk for various medical problems, such as prediabetes and diabetes, metabolic syndrome, cardiovascular, pulmonary, orthopedic and gastrointestinal diseases, as well as psychological problems [3–5].

The incidence of overweight and obesity has great implications for public health. Globally, the prevalence of overweight and obesity has increased at an alarming rate. According to the World Health Organization (WHO), it is estimated that more than 40 million children under five years of age were overweight in year 2010, with close to 35 million of these children in developing countries [6]. Based on secular trends using the International Obesity Task Force (IOTF) criteria, Wang and Lobstein [7] estimated overweight to occur in 46% of school-aged children in the Americas, 41% in the Eastern Mediterranean region, 38% in the European region, 27% in the Western Pacific, and 22% in the Southeast Asia.

At the individual level, determinants of obesity include genetics, biology, behavior, and environments that foster an adverse balance between energy intake and energy expenditure [8]. However, there are also other relevant factors related to childhood obesity, including dietary status, physical activity, sleep duration, socio-economic status and intrauterine factors. It is therefore necessary to further explore the causal factors that contribute to obesity in children. Table 1 summarized the findings of recent studies on childhood obesity, while Fig. 1 presents the conceptual framework identifying several major modifiable and unmodifiable risk factors that may exert substantial influence on the etiology of childhood obesity.

Table 1 Summary of recently reported studies on childhood obesity
Full size table
Fig. 1
figure 1

Conceptual framework describing the etiology of childhood obesity

Full size image

Unmodifiable Risk Factors in Childhood Obesity

Genetic Influences

Genetic susceptibility to obesity is recognized as a highly heritable condition that begins at an early age. Genetic factors also determine whether one is susceptible to other obesity-related diseases [9]. Monogenic obesities have been examined extensively in children given that these conditions are rare, very severe, and generally begin in childhood. Candidate gene approaches have been used to identify the association between a variant or mutation within or near the candidate gene and a trait of interest (for example: obesity). With respect to single-gene mutation, a number of genes have been identified in individuals who were severely obese, namely pro-opiomelanocortin (POMC), proprotein convertase subtilisin/kexin 1 (PCSK1), melanocortin-4 receptors (MC4R), corticotrophin-releasing hormone receptor (CRHR), leptin (LEP), leptin receptor (LEPR), cocaine- and amphetamine-regulated transcript (CART), brain-derived neurotrophic factor (BDNF) and single-minded homolog 1 (SIM-1) [10–16]. The defects in POMC produced altered peptide which is different from α-melanocyte-stimulating hormone (α-MSH) through PCSK1 and without α-MSH, MC4R is unable to play a role in weight regulation [10]. Deficiency of MC4R was found to associate with obesity by accelerated linear growth and increased final height as well as excess in insulin and partially suppress growth hormone [11]. Meanwhile, CRHR shows anorexigenic hypothalamic response by activation of urocortin to potently suppress food intake [12]. LEP and LEPR suppress the increase of CART [13] and prevent overeating by controlling the appetite [14]. BDNF have been implicated in the energy regulation by increasing gene expression for LEPR and POMC [15]. SIM1 are key factors interacting with the central melanocortin pathway in the control of appetite by decreasing MC4R [16]. However, cases of single-gene obesity cannot explain many others with latent genetic predisposition that is expressed only upon exposure to an obesogenic environment.

In most individuals, obesity results from the interaction of multiple genes that encode peptides that transmit hunger and satiety signals, regulate adipocyte growth and differentiation and control energy expenditure [17]. The Human Obesity Gene Map reported 253 loci from 61 genome-wide linkage scans, of which 15 loci have been replicated in at least three studies [18]. However, a meta-analysis by Saunders et al. showed that genome-wide linkage analysis may not be the best method for identifying genetic variants for obesity [19].

Genome-wide association studies (GWAS) have implicated many genetic loci for obesity in the past 5 years [20••], and of particular interest, the fat mass and obesity-associated protein (FTO) gene [21]. FTO gene expression in the hypothalamic arcuate nucleus was determined to contribute significantly to obesity as a consequence of hyperphagia, seemingly in the absence of changes in energy expenditure [22, 23]. Two other genes, TNNI3K and POMC, were identified by GWAS as being associated with childhood obesity [24]. In 2010, a study discovered two additional obesity loci (TNKS-MSRA and SDCCAG8) in extremely obese children and adolescents, with odds ratios of approximately 1.10 per risk allele for both loci for early-onset obesity [25].

Furthermore, gene-environment interactions also play an important role in childhood obesity [26, 27]. Differences in body composition is also a likely factor; specifically the lower abdominal adiposity found in African-American children compared with European or Hispanic American children [28, 29], and the higher trunk skinfold thickness in Chinese girls compared to Malaysian and Lebanese girls [30]. These differences in body composition phenotype suggest that the genetic makeup of individuals interacts with environmental factors from early developmental stages, as has been demonstrated by Fernandez et al. [26] and the twin and adoption study [31]. With the advent of next-generation sequencing techniques and advances in the field of exposomics, sensitive and specific tools to predict the obesity risk as early as possible are the challenges in the coming decade [32].

An environmental factor of major importance is nutrition. The risk of obesity is higher when an individual with a high-risk genetic profile is exposed to high-risk environment. For example, the Pima Indians have an inherent susceptibility to obesity, and obesity is more likely to develop among those living in Arizona as they are exposed to obesity-promoting environments, such as ‘Western diet’ and less exercise, compared with Pima Indians living in Mexico, who survive on traditional food and manual labor [33]. Moreover, obesity genes might also play a role in the choice and preference of dietary intake of saturated fat, carbohydrates, mono- and disaccharides, and polysaccharides [34]. A longitudinal study had suggested that high consumption of saturated fat increase the obesity risk associated with FTO gene [35]. To date, the genes responsible for individual differences in sensitivity to changes in energy balance have not all been identified. In view of the complexity of the biological systems involved in body weight regulation, these genes are likely to be numerous [36].

Di Castelnuovo et al. [37] suggested that positive assortative mating by BMI may assemble obesity-promoting risk alleles and probably undergo interaction between gene and environment in a family [38]. A recent paper revealed that weight changes over a period of two years was associated with marital status; partly due to shared environment, such as the stimulus to eat when dining together or the motivation for weight control to increase attractiveness [39]. Comparing children born to normal weight parent, children had higher obesity risk if they had an obese father (OR=2.11), an obese mother (OR=7.66), or two obese parents (OR=8.05) [40]. These observations suggested that assortative mating is related to the epidemic of childhood obesity.

Ethnic Differences

Ethnic disparities have been widely discussed as a contributing factor to adiposity. In adults, very large differences between ethnic groups, especially among women, have been observed, with non-Hispanic black individuals exhibiting the highest prevalence of obesity [41]. Asian Indian individuals were observed to have significantly greater total abdominal fat and intra-abdominal adipose tissue and truncal subcutaneous (SC) adipose tissue than Caucasians [42]. Similarly in children, South Asian Indians reportedly had smaller body mass index and waist circumference than Caucasians, and yet both boys and girls from India had higher percent body fat [43]. Hispanic Americans exhibit the greatest trunk, intra-abdominal and SC adipose tissue, followed by European Americans and African American [44•].

In comparison with Caucasian children of the same age, Asian children have lower BMIs by 3-6 units for a given percentage of body fat [45]. Ethnic differences in adiposity had been reported in adolescence, with greater central adiposity being observed in women of Asian ancestry compared with Caucasians [46]. For a given BMI, children and infants of South Asian origin have higher adiposity compared with White Europeans, suggesting that either genetic factors or exposure to maternal physiology contribute to obesity rather than behaviors or diet during childhood or later in life [47].

The risk factors for obesity reportedly exist from as early as the prenatal and early childhood periods [48]. Taveras et al. [49] further suggested that racial or ethnic differences in childhood obesity may be determined by several factors that operate during pregnancy, infancy and early childhood. Compared to Caucasian children, children of Black and Hispanic ancestry had higher odds of rapid infant weight gain, and exhibited lower birth weights for gestational age but had higher BMI z-scores, and a greater prevalence of obesity at the age of three years [49].

Amongst ethnic groups, cultural differences are believed to be one of the contributing factors to the disparities in childhood obesity [50]. For instance, perception on body image occurs in a cultural context and differs by ethnic groups. Women who traditionally feed, educate and take care of their children, also possess their own beliefs about body image, which in turn will have implications toward their children’s own body image [50]. In addition, cultural activities may also be considered as a non-modifiable factor in determining a child’s adiposity. A recent prospective study [51] found negative association between participation in cultural activities with z-scores of waist circumference (WC) and waist-to-height ratio (WHR) in girls; and concluded that cultural activities had a moderating effect on the obesity-susceptibility genes. Although environmental factors pertaining to health-related behaviors or lifestyles and economic disadvantage could contribute to some of the ethnic differences in the prevalence of diseases associated with obesity, these factors cannot explain all the differences in expression and disease patterns due to race or ethnicity. Hence, it is likely that genetics or molecular factors also contribute to the racial disparities in obesity-related comorbidities [52].

Gestational Weight and Intrauterine Factors

The intrauterine environment plays a crucial role in the development of obesity, type 2 diabetes mellitus and metabolic syndrome in the offspring [53, 54]. Shankar and colleagues [55] illustrated that children exposed to maternal obesity in utero were more susceptible to obesity, regardless of birth weight, which seems to indicate that subtle programming of obesity occurs in the absence of clear changes in birth weight. Children of overweight or obese mothers also had more likelihood of being born large for gestational age (LGA) [56•]. These children are at an increased risk of obesity when they also exhibit a high birth weight of more than 4 kg [57].

Pre-pregnancy obesity is the strongest risk factor for childhood obesity [58] and metabolic dysregulation [59]. Higher gestational weight gain (GWG) has been associated with both a greater incidence of pre-eclampsia [60] and a significantly higher risk of gestational diabetes mellitus (GDM) with maternal hyperglycemia as well as fetal hyperglycemia, which can consequently lead to excess fetal insulin and thus fetal overgrowth [61]. Hull and colleagues demonstrated direct relationship between weight gain during pregnancy and birth weight or infant adiposity [62]. Longitudinal data have revealed a strong relationship between gestational weight gain and childhood weight status, regardless of the mother’s pre-pregnancy weight [63] and also independent of genetic factors [64]. It has been suggested that weight status during early human development, especially in the maternally overweight or obese children, can have on-going effects on adiposity and related chronic diseases [65].

In the in utero environment, epigenetic factors may alter gene expression and predispose the fetus to abnormal physical activity and dietary behaviors later in life by compromising the physiological thresholds for energy balance regulation [66]. Moreover, the offspring’s susceptibility to premature chronic diseases may be influenced by constant exposure to excess energy, hormones and growth factors in utero [54]. Human data regarding the effect of maternal lifestyle on epigenetic modifications are scarce; however, animal-model research has demonstrated that the body composition of the offspring changes with maternal diet and is associated with epigenetic alterations in metabolic control genes [67]. Furthermore, the maternal diet may influence the food preferences and feeding responses in the offspring [68]; and, if nutritionally compromised, may promote adiposity as well as early onset of metabolic impairments in the child [69]. The intake of glucose and lipids, such as triglycerides and non-esterified fatty acids, had also been shown to have positive relationship with fetal growth [70, 71•].

Modifiable Risk Factors of Childhood Obesity

Socio-economic Status (SES)

Living conditions and societal factors have great impact on a child’s weight status. In Sweden, boys living in semi-urban and rural areas have a greater risk of obesity [72]. In many developed countries, lower socioeconomic groups reportedly had the highest levels of overweight and the lowest levels of physical fitness, and adolescent girls were particularly at risk [73]. In most countries, children in urban areas were more likely to be obese than those in rural areas. Contemporary populations of children were found to exhibit higher rates of obesity than do those from the lowest socioeconomic groups in high-income countries [74]. In India, significantly more children of higher SES status were obese and overweight than those from lower SES [75, 76].

Positive energy balance, high energy intake, and low levels of energy expenditure on physical activity are strongly associated with urbanization and economic affluence [77]. SES groups, usually low-SES in industrialized countries and high-SES in developing countries, with greater access to energy-dense diets are at a higher risk of being obese. However, the obesity–SES association varies with gender, age, and country [78••]. The findings from a longitudinal survey suggest that rather than having a major direct causative role, family income may act primarily as a proxy for other unobserved characteristics that determine children’s weight status [79].

Diet

Breast-feeding is known to have a consistent protective effect against obesity in children [80]; however, a quantitative review by Owen et al. concluded that the precise magnitude of association is still unclear [81]. The macronutrient composition and bioactive substances of breast milk may influence metabolic programming and the regulation of body fatness and growth rate [80]. Formula-fed infants had higher insulin levels and lower leptin levels compared with breast-fed infants; and these proteins could stimulate fat deposition and lead to the early development of adipocytes [82]. Direct breastfeeding during early infancy has been related to better appetite regulation later in childhood; whereby a study showed that children who were fed human milk in a bottle during the first three months of life were 67% less likely to have good response to satiety during their preschool years compared to children who were directly breastfed [83]. The effect of breastfeeding on infant growth may be a significant determinant of early life programming for obesity and chronic diseases later in life, especially for the offsprings of women with diabetes [84].

The adoption of industrialized Western society lifestyles, including urbanization, Western foods, increased sedentariness and car ownership, has been related to increased obesity. Rapid and marked socio-economic advancement has brought about considerable changes to the lifestyles of communities, including among children; especially with respect to dietary patterns, such as the high intake of energy-dense food, which has been identified as an important factor in body weight control in adults, as well as in children and adolescents [85]. Moreover, alterations in meal patterns are also evident with more families eating out, skipping meals especially breakfast and relying too much on fast foods, which are known to be high in saturated and trans fats, energy dense, and served in large portion sizes [86–88]. Children who ate all three main meals, namely, breakfast, lunch and dinner, daily were reported to have 63% lower risk of being overweight or obese than those who did not [89]. Another study had also further confirmed the protective role of consuming three meals per day with a lower likelihood of overweight or obesity, but only if breakfast was not skipped [90]. Skipping breakfast has been shown to affect children's appetite but not their energy intake at subsequent meals [91•]. When children did not consume breakfast, they were hungrier, less full, and could consume more food prior to lunch; hence, increasing daily dietary intake [92].

The rapid development of fast-food outlets and the easy availability of junk food is also a matter of concern. Children with working parents who are cared for by care providers are more likely to receive food that is high in energy and of poor nutritional value, perhaps because care providers are more concerned with placating their wards than with the long-term health of the children [93]. Furthermore, parents who work outside of the home may also serve more high-calorie pre-prepared, convenience, or fast foods due to time constraints. Additionally, unsupervised children tended to make poorer nutritional choices when preparing their own snacks [94]. It has also been reported that children consumed more fast food items and carbonated drinks as compared to fruits and vegetables, as these food items were easily available through vending machines and school canteens [91•].

Changes in lifestyle, dietary habits, and food marketing have brought about undesirable effects, with large proportions of the population afflicted with various non-communicable diseases associated with over-nutrition, including obesity [95–97]. Available evidence suggests that high-energy intake in early infancy and high consumption of energy-dense foods and sweetened drinks during childhood is associated with increased adiposity [85].

Food marketing to children has been proposed as a means for addressing the global crisis of childhood obesity in recent years. The commercial advertising and marketing of food and beverages influences the diet and health of children and youths. An estimated more than $10 billion a year is spent on marketing of all types of food and beverages to children and youths in America [98]. There is evidence that food marketing to children is (a) massive; (b) expanding in number of avenues, such as product placements, video games, the internet, and cell phones; (c) composed of messages that are almost entirely for nutrient-poor, calorie-dense foods; (d) having harmful effects; and (e) increasing globally and thus difficult to regulate by individual countries [99]. Recently, WHO has produced a set of recommendations to help member countries to either develop or strengthen policy related to marketing of foods and non-alcoholic beverages to children [100].

More recently, vitamin D deficiency has been identified as a new global public health issue. Vitamin D has been suggested to be a potential factor in the prevention of many illnesses, including obesity [101]. Obese individuals were observed to have lower concentrations of 25-hydroxy vitamin D, suggesting that obese individuals may have altered vitamin D and parathyroid hormone physiology [102].

Physical Activity

Physical activity has long been recognized as one of the important determinants of obesity [103] and as a promoter of lifelong positive health behavior in children [104]. The importance of physical activity lies in the basic concept of the Law of Thermodynamics, from which can be derived the fact that human energy expenditure consists of three components; namely, energy that is expended for thermogenesis, energy that is expended for basal metabolism or the basal metabolic rate, and energy that is expended on physical activity [105]. Of these, physical activity is the only factor that can be modified to prevent obesity.

It has been reported that only a third of overweight and obese children engaged in a minimum of 60 minutes of physical activity daily, which suggests that the younger generation led sedentary lifestyles [106]. School children, especially those in Asian countries, have been reported to focus more on academics and are less involved in sports and physical activities [107]. Children also spent much of their leisure time engaged in sedentary activities, such as watching TV or playing computer/video games [108]. Studies have found that normal BMI values were distributed in the higher tertiles of physical activity [109, 110].

A systematic review by Te Velde et al. [111] confirmed that there is an inverse association between total physical activity and overweight; but not, with respect to specific sub-behaviors, such as moderate to vigorous physical activity, aerobic exercise and leisure activity. Jimenez-Pavon et al. also presented evidence that supports negative association between objectively measured physical activity and adiposity, and reported that higher levels of habitual physical activity are protective against child and adolescent obesity [112].

Watching television decreases the amount of time spent on active physical activities [113], and has been associated with increased food consumption either during television viewing or as an indirect result of food advertisements [114]. A study involving students aged 11 to 13 years in California concluded that time spent watching television was significantly associated with obesity [113]. Te Velde et al. [111] also reported moderate evidence for a significant positive association between TV/video/computer screen time and overweight. Increased physical activity and decreased screen-based sedentary behavior were associated with decreased body fat percentages but not decreased BMI [115].

Although the benefits of physical activity have been proven in many studies [103, 104, 111–113], there are a few studies that reported contradicting results regarding physical activity and obesity. A longitudinal study in England found that physical inactivity is the result rather than the cause of obesity, and argued that inactivity does not lead to fatness; hence, explaining why physical activity interventions sometimes fail to prevent excess weight gain in children [116]. Another intervention program conducted among preschool children, reported that physical activity helped to improve motor skills but failed to reduce body mass index [117].

Sleep

Sleep plays an important role in the health of children and adolescents as it allows for the normal diurnal rhythm of hormones that are related to growth, maturation, and energy homeostasis [118•]. Children who sleep for shorter durations have been postulated to have lower energy intake and expenditure; since sleep deprivation is known to lead to alterations in the structure of sleep stage, and hence, giving rise to fatigue, daytime sleepiness, somatic and cognitive problems and low activity levels [119]. Several studies have reported that habitual sleep length is prospectively and independently associated with obesity and mortality [120, 121] with those who sleep for short durations being more likely to be obese [122]. On the other hand, longer sleep duration may reduce the opportunity to eat, and thus prevent over-eating among children and adolescents [123].

Sleep deprivation influences the development of obesity through several possible biological pathways. These include increased sympathetic activity, decreased leptin and growth hormone, elevated cortisol and ghrelin levels, and impaired glucose tolerance [124]. Hormonal alterations may contribute to the selection of energy-dense food, excessive energy intake, changes in energy expenditure, insulin resistance, alterations in the basal metabolic rate, modifications in the thermic effect of food and non-exercise activity thermogenesis [123, 125]. As short sleep duration has been clearly associated with increased risk of childhood obesity, sleep could be a vital factor that needs to be considered in childhood obesity prevention [125].

Visceral adiposity is a significant predictor of obstructive sleep apnea (OSA), with the severity being independent of BMI among obese children [126]. This is supported by another study in which obese children exhibited no difference in head, neck and abdominal subcutaneous fat but instead exhibited more abdominal visceral fat [127]. A longitudinal study revealed that children who slept less were more likely to be overweight and had high body fat values [128•]. Besides, sleep duration decreases with age whereas higher body fat mass and BMI both tend to increase with age [129].

Based on a meta-analysis, the recommended sleep durations are 11 hours or more for children aged below 5 years, 10 hours or more for children aged between 5 and 10 years, and 9 hours or more for children aged 10 years and above [125]. Independent of other risk factors, increasing sleep duration may decrease the prevalence of childhood obesity [130]. The current available evidence is unable to support claims of a secular trend [131]. Future randomized intervention trials are necessary to determine the effectiveness of sleep extension for the prevention of obesity among children and adolescents [130].

Parental Determinants

Parental work schedule [132], parental BMI [133], and maternal smoking habits [134] appear to be important in determining children’s health status, particularly with respect to a healthy body weight. Maternal smoking during pregnancy may be a risk factor for childhood obesity, along with low birth weight [135]. A possible explanation for this may be the impact of catch-up growth in the first year of life on childhood obesity [136]. Children whose mothers smoked during pregnancy were at increased risk for overweight at the ages of 3 to 33 years [134]. Although the mechanisms by which maternal smoking influences the weight of the child are not well characterized, they are possibly due to nicotine, which is transported across the placenta, and carbon monoxide, which potentially influences placental vascular function and may cause fetal hypoxia [134]. Nicotine acts to reduce appetite [137] and body weight, while withdrawal results in hyperphagia and weight gain [138]. It has been suggested that the rapid weight gain of the infant during the early postnatal period may be due to the effect of nicotine withdrawal, similar to the increased craving for food [139].

A positive association has also been observed between childhood obesity and parental BMI. Excessive gains in parental BMI during youth and later life were found to be associated with higher BMI and risk of obesity in the offspring [133]. Similarly, child overweight/obesity was also significantly associated with maternal work hours and paternal non-standard work schedules [132, 140]. However, further research is required to determine the manner in which non-standard work schedules disrupt family life, the quality of children’s diet and their opportunity to participate in physical activity [141].

Conclusion

Obesity is considered to be an epidemic given that its prevalence and severity in both adults and children is rising at alarming rates. This increase is related to the interactions between genetic makeup, intrauterine factors, and environmental living conditions. Obesity may result from genetic susceptibility to a single gene or due to the interactions of multiple genes that influence appetite regulation, adipocyte growth and energy expenditure. Ethnic disparities also contribute to childhood adiposity. Moreover, children from low socioeconomic families in industrialized countries and high socioeconomic background in developing countries are at a higher risk of being obese due to easy accessibility of energy-dense food. Maternal weight status prior to or during gestation may also be a key factor related to the short- and long-term risks of childhood obesity. Furthermore, other parental determinants, such as maternal smoking and lengthy work duration, may worsen the risks of childhood obesity. Contemporary lifestyle changes are also important influences leading to childhood obesity; these changes include increasingly sedentary activities, unhealthy dietary habits and lower quality of sleep. Therefore, deeper understanding of the multi-factorial contributors to obesity is of utmost importance in order to aid in the development of effective interventions aimed at reducing the rates of occurrence of global obesity.