In a high-risk design of obesity, children who are born at ‘high-risk’ are compared to children who are born at ‘low-risk’ for obesity on specific eating behaviors or food preferences. In this design, child risk status for obesity is commonly defined on the basis of parental, typically maternal, weight status 15
. Comparing eating behaviors of children whose parents differ in weight status (e.g., normal-weight versus overweight/obese) can be a useful strategy for testing whether familial vulnerability for obesity may express itself through specific eating traits. Further, data from a prospective cohort study can be used to test whether specific eating traits mediate the relationship between obesity risk status and subsequent adiposity gain. The design is not suited for discriminating between genetic and environmental influences on children’s eating behavior and weight development.
The ‘Infant Growth Study’ (IGS) is an example of a high-risk design of obesity. It is a prospective cohort study which assesses the growth and development of children born with different familial predispositions to obesity. Child participants’ obesity risk status was based on maternal pre-pregnancy body mass index (BMI; kg/m2
). Children who were born to mothers with a BMI of less than the 33rd
percentile (mean maternal BMI 19.5 ± 1.1 kg/m2
) were classified as at ‘low-risk’ for obesity, while children who were born to mothers with a BMI greater than the 66th
percentile (mean maternal BMI 30.3 ± 4.2 kg/m2
) were classified as at ‘high-risk’ for obesity. Details of parental characteristics and study design were reported previously 16–18
Children’s weight and adiposity were assessed every 3 months during the first year, every 6 months through year 4, and yearly thereafter. These assessments showed that by age 4, high-risk children were significantly heavier than low-risk children as indicated by a significantly greater weight, BMI, lean body mass, and waist circumference 19
. By age 6, high-risk children also showed a significantly greater fat mass and percentage body fat than low-risk children. Given these early differences in the weight development and adiposity, it is important to identify behavioral mediators related to children’s eating or physical activity patterns that may underlie the diverging growth patterns in the two obesity risk groups. The energy density of food is one dietary component that has been identified as an important predictor of energy intake in both children and adults.
3.1. Dietary energy density
Energy density is the amount of energy provided in a particular weight of food (kcal/g). The energy density of a food is influenced by its water content and macronutrient composition. Water has the greatest impact on energy density because it adds weight but no calories. Of the macronutrients, fat is the most energy dense, providing 9 kcal/g, followed by carbohydrates and protein which both provide 4 kcal/g. In experimental studies, the energy density of foods has been identified as an important predictor of short-term energy intake in both adults 20–27
and children 28–31
. To date, few randomized behavioral treatment studies with children and their parents have tested the effects of reducing dietary energy density on longer-term energy intake and weight development 32,33
. Data from cross-sectional studies showed that dietary energy density was independently and significantly associated with higher BMI 34
, elevated fasting insulin levels, and the metabolic syndrome in U.S. adults 35
. Similarly, data from prospective studies in children indicated that higher dietary energy density was associated with excess adiposity during childhood 36–38
To date, little is known about the genetic and learned familial influences on the relationship between dietary energy density and energy intake in children. There is evidence for a genetic contribution to daily and meal-specific dietary energy density 39
. Therefore, it is conceivable that children may seek energy-dense foods by way of their genetic predisposition to the taste or higher energy content of those foods.
Home environmental factors, such as availability and easy accessibility of specific foods, have also been shown to influence the formation of food preferences and consumption patterns in children 40,41
. At the same time, parental feeding practices may enhance innate predispositions for the liking of foods higher in energy density, consequently promoting eating habits that favor an increased consumption of energy-dense foods. For example, repeated exposure to foods has been associated with increased liking and consumption of those foods 42–44
. On the other hand, it is also possible that parents of obesity-prone children restrict access to more energy-dense foods in an effort to moderate their children’s energy intake and weight gain. The use of overly restrictive feeding practices, however, has been associated with greater energy intake and increased child weight status 45–47
. It is important to note that several previous studies were either conducted in relatively narrowly defined study samples, namely Caucasian children 46
, or have focused only on girls 47
. The lack of diversity in these study samples limits generalizability of the findings to children as a whole. It thus would be desirable to examine child feeding practices in families from more diverse racial and ethnic backgrounds and to also include boys.
No prospective study had assessed developmental changes in daily energy density in young children, while also taking into account children’s familial predisposition to obesity. It is possible that obesity-promoting genes exert their influence through consumption of more energy-dense diets. If true, children born at high risk for obesity would be expected to consume diets of higher energy density compared to children born at low risk for obesity. Hence, as part of the IGS, we characterized 4-year trajectories in dietary energy density across both high-risk and low-risk children at ages 3–6 years; we also sought to examine the relationship between dietary energy density and child weight status within each year. Dietary energy density (with and without beverages) was computed from 3-day weighed food records which parents completed at each of the respective years 48
depicts the mean daily dietary energy density, based on solid foods only (all beverages excluded), consumed by low-risk and high-risk children at ages 3, 4, 5, and 6 years. The findings from this study indicated that dietary energy density increased as children grew older, but did not significantly differ for low-risk or high-risk children at any of the years. The lack of a significant difference in dietary energy density between high-risk and low-risk children suggests that the types of foods consumed by children at an early age were similar. Because young children often have a more limited exposure to different types of foods than do older children, it is possible that differences in dietary energy density between obesity risk groups may start to emerge later in life when children have access to a wider array of foods both inside and outside the home.
Mean (± SE) energy density (Panel A) and energy intake (Panel B) consumed by low-risk and high-risk children from ages 3 to 6 years
With respect to daily energy intake (calories consumed from foods only, all beverages excluded), a significant risk group-by-time interaction was found, indicating that high-risk children consumed significantly more energy from food at age 6 than did low-risk children (). The finding of a significant difference in energy intake from foods at age 6 in the absence of a significant difference in dietary energy density suggests that children in the two risk groups may have consumed differential quantities of foods at that age. Child BMI z-scores, waist circumference, and percent body fat were not significantly correlated with dietary energy density in any of the 4 years.
When the IGS subjects were 12 years of age, they were invited to participate in a dietary assessment as part of their annual study visit. The aim of this assessment was to determine if, when given a choice, the dietary energy density and energy intake at a single, self-selected meal differed for adolescents in this study as a function of their familial predisposition to obesity and their sex 49
. For lunch, subjects were served a buffet meal consisting of a variety of foods and beverages ranging in energy density. The meal provided approximately 5200 calories, and subjects could freely choose the types and amounts of foods and beverages they wanted to consume. The energy density of the meal was based on subjects’ intake and was computed by dividing the total calories consumed by the total weight of food consumed excluding all beverages. High-risk subjects consumed a significantly more energy-dense meal than low-risk subjects (1.84 vs. 1.42 kcal/g). This difference remained significant when adjusting for subjects’ current BMI z-score. There was a trend (p = 0.16) for energy intake, when expressed as a percentage of subjects’ daily estimated energy requirement, to be different between high-risk (42% ± 4) compared to low-risk (33% ± 4) subjects. The susceptibility of high-risk subjects to consume a more energy-dense diet, if sustained, may predispose this youth to negative long-term health consequences. Data from a cross-sectional, nationally representative survey in adults indicated that diets that are higher in energy density tend to be of lower diet quality 50
, which in the long term, can lead to adverse health consequences, such as the metabolic syndrome 35
In contrast to the findings from earlier years 48
, this study showed that, during a single multi-item meal, high-risk subjects self-selected a more energy-dense meal compared to low-risk subjects when they were 12 years of age. It is possible that differences in the methodologies used to assess subjects’ intake (maternal self-report vs. laboratory measure) may have contributed to the difference in findings with respect to dietary energy density. It is also conceivable that differentiations in children’s dietary patterns may develop over time, thus explaining why differences in dietary energy density between low-risk and high-risk children did not yet appear at a younger age. Home food environments, which help shape eating behaviors and food preferences in children 14
, may change over time or may exert their influence at critical developmental periods. For example, in households with a family history of obesity, adolescents may have had easier access to more energy-dense foods during their upbringing, which in turn may have helped shape their long-term preferences for those foods. Parents also may have used different feeding strategies at different stages of their children’s lives. It is possible that, out of concern for their children’s increasing adiposity, parents of high-risk youth may have restricted access to energy-dense foods more than parents of low-risk youth. In turn, this restriction may have enhanced high-risk adolescents’ preference for more energy-dense foods. Lastly, parents, mothers in particular, serve as important role models for eating behavior and food choices for their children 15,51–55
. It is conceivable that mothers, through many years of role modeling, may have passed on to their children some of their own food preferences.
Although the types of foods consumed are important determinants of dietary energy density and energy intake in children, daily energy intake can also be affected by children’s eating behaviors. Eating phenotypes such as the rate of eating or children’s responsiveness to portion size changes can significantly affect the number of calories consumed at a meal. Identifying eating behaviors that distinguish normal-weight from overweight and obese children and that may make overweight and obese children more prone to overeating are important first steps in the prevention and treatment of childhood obesity.
3.2. Sucking and eating rate
When examining the microstructure of eating behavior, the rate of eating has been identified as an indicator of appetitive avidity 56
. Specifically, it has been suggested that a higher eating rate signals a greater motivation to eat. Findings from studies comparing the eating rates of normal-weight adults to those of overweight or obese adults have been mixed; while some studies showed a faster eating rate 57–59
or a less pronounced deceleration in the rate of eating over the course of a meal 60
in obese individuals, others found no differences between weight groups 61–64
. Findings from studies in children are more consistent and indicate that obese children exhibited a higher eating rate during meals and a less rapid deceleration of intake towards the end of the meal compared to normal-weight children 65,66
. More recently, eating rate has been identified as a highly heritable behavioral phenotype (heritability, h2
= 62%) 67
. Together, these data suggest that differences in eating rate, which seem to be in part genetically determined, could be implicated in the development of childhood obesity. Longitudinal studies are needed to examine causal relationships between eating rate and child adiposity.
When IGS participants were 3 months of age, their sucking behavior was measured in the laboratory using an automated nutritive sucking apparatus 17
. In this experiment, infants were fed either expressed breast milk or their customary formula milk with their customary nipple type. High-risk infants, compared to low-risk infants, showed a more vigorous sucking behavior as indicated by a significantly higher total number of sucks (920 ± 89 vs. 620 ± 47 sucks), a higher sucking rate (0.75 ± 0.01 vs. 0.59 ± 0.04 sucks/second), and a larger amount of milk consumed (150 ± 9 vs. 123 ± 8 g), but no difference in feeding time. It remains unknown if energy intake differed between risk groups during the sucking behavior assessment. A prospective analysis showed that the total number of sucks at 3 months of age was a significant predictor of weight and weight-for-length at 12 months of age, accounting for approximately 9% of the variance 17
The assessment of eating rate was repeated when IGS participants were 4 years of age 68
. This time, children consumed a multi-item test meal in the laboratory in the presence of a parent. Eating rate was computed as calories and mouthfuls consumed per minute. Results showed a non-significant trend (p = 0.10) for eating rate to be greater among high-risk (2.60 ± 1.08 mouthfuls/min) compared to low-risk (2.16 ± 0.96 mouthfuls/min) children. Further, energy intake per minute was significantly correlated with meal energy intake and daily energy intake (≥ 0.40). When controlling for maternal BMI, a prospective analysis of eating rate at 4 years of age, expressed as either mouthfuls per minute or calories consumed per minute, was significantly associated with an increased probability of a child being overweight or obese at age 6. These data suggest that an accelerated eating rate may present a behavioral risk factor for increased energy intake and excessive weight gain in children. Interventions designed to slow the eating rate in obesity-prone children may be a promising strategy to moderate their energy intake. For example, Epstein and colleagues 69
tested the effects on eating rate and food intake of instructing 7-year-old children to place their utensils on the table after every ingested bite. The results of the study indicated that both eating rate and the amount of food consumed significantly decreased for both obese and non-obese children as a function of putting eating utensils down between bites combined with praise for compliance.
The microstructure of children’s eating has also been shown to be affected by environmental factors such as the portion size of foods. When the portion size of an entrée was doubled, 2- to 9-year-old children consumed 29% more of the entrée and 13% more energy at the meal than when served a standard entrée portion 70
. The increase in entrée intake was in part attributable to an increase in children’s bite size. Further, children who increased their intake when served the large portion entrée showed increases in both bite size and the total number of bites taken. These data suggest that interventions which combine slowing children’s rate of eating with modifying the amounts and also types of foods served may be a successful strategy to moderate energy intake in children who are prone to excessive weight gain.
Identifying eating traits which promote increased energy intake in children is crucial for the design of successful interventions to prevent and treat childhood obesity. At the same time, however, it is important to study the home environment in which children have early experiences with food and eating. Additionally, because parental behaviors surrounding eating and child feeding are known to also play a crucial role in the formation of food preferences and eating behaviors in young children, the familial transmission of taste preferences, food selections, and eating behaviors must also be studied closely.