Eating in the Absence of Hunger: A Genetic Marker for Childhood Obesity in Prepubertal Boys?

Authors

  • Myles S. Faith,

    Corresponding author
    1. Department of Psychiatry, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania;
    2. Department of Child and Adolescent Psychiatry, The Children's Hospital of Philadelphia, Philadelphia, Pennsylvania
    Search for more papers by this author
  • Robert I. Berkowitz,

    1. Department of Psychiatry, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania;
    2. Department of Child and Adolescent Psychiatry, The Children's Hospital of Philadelphia, Philadelphia, Pennsylvania
    Search for more papers by this author
  • Virginia A. Stallings,

    1. Department of Pediatrics, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania
    2. Department of Pediatrics, The Children's Hospital of Philadelphia, Philadelphia, Pennsylvania.
    Search for more papers by this author
  • Julia Kerns,

    1. Department of Psychiatry, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania;
    Search for more papers by this author
  • Megan Storey,

    1. Department of Psychiatry, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania;
    Search for more papers by this author
  • Albert J. Stunkard

    1. Department of Psychiatry, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania;
    Search for more papers by this author

  • The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Weight and Eating Disorders Program, 3535 Market Street, Third Floor, Philadelphia, PA 19104. E-mail: mfaith@mail.med.upenn.edu

Abstract

Objective: Eating in the absence of hunger (EAH) may be a behavioral trait through which obesity-promoting genes promote positive energy balance. The primary aim of this study was to compare children born at high vs. low risk for obesity with respect to EAH at 5 years of age.

Research Methods and Procedures: This was an observational investigation of families enrolled in the University of Pennsylvania and The Children's Hospital of Philadelphia's Infant Growth Study. Five-year-old children born at high (N = 28) or low (N = 25) risk for obesity on the basis of maternal prepregnancy body weight were evaluated at a hospital-based laboratory. Children consumed 11 snack foods ad libitum after consuming an ad libitum dinner and reporting fullness. Parents reported on snack foods at home and their own eating styles. Nutritive sucking at 3 months of age was evaluated by computerized apparatus.

Results: EAH in high-risk boys (mean ± standard error = 326 ± 66 kJ] was more than twice that of low-risk boys (mean ± standard error = 151 ± 39 kJ), p = 0.03. Among girls, there was a trend for EAH to be associated with increased parental limitations on daughter snack food consumption at home (p = 0.06). EAH was unrelated to 3-month sucking behavior.

Discussion: Genes that promote childhood obesity may partially exert their influence through EAH, an effect that was limited to boys born at risk for obesity. The unique influences of genes and home environment on this trait should be disaggregated in subsequent studies.

Introduction

Obesity is a polygenic disorder that results from the influence of multiple genetic and environmental factors (1). Genes may influence the development of body habitus through intermediary metabolic, growth, and behavioral pathways such as eating or physical activity. It was proposed that obesity results from a lowered total energy expenditure or resting metabolic rate in children (2,3), but larger studies failed to confirm this association (4,5,6). Later studies suggest that obesity may be better characterized as a behavioral than a metabolic perturbation (7) and that the genes for obesity may at least partially operate through behavioral traits (8).

Several behavioral pathways underlying the development of obesity in children have been proposed (8). One that may be informative is eating in the absence of hunger (EAH),1 a form of eating disinhibition initially described by Fisher and Birch (9). In this paradigm, children consume ad libitum from a protocol meal in the laboratory setting, after which they rate themselves as feeling hungry, half-full, or full using age-appropriate silhouette scales. Immediately afterwards, children consume ad libitum from a variety of snack foods that vary in macronutrient content and taste properties. Fisher and Birch (9) reported that 3- to 5-year-old boys and girls who reported feeling full after the protocol meal ate an additional 904 ± 59 kJ (or 216 ± 14 kcal) from the protocol snack food selection. However, there was notable variability in this trait, a finding that was observed in a subsequent report of 5-year-old girls who consumed an average of 523 ± 389 kJ (range 0 to 1820 kJ) in an EAH study (10).

Reasons for interchild variability in EAH are poorly understood. Parental restriction of child food choices predicted increased EAH by girls who were tracked from ages 5 to 7 years (10). However, these studies did not address the question of whether there may be genetic influences on this behavior and, if so, whether there is a gender difference. There is reason to believe that EAH may be a genetically loaded behavioral phenotype. First, twin studies document a heritable component to eating attitudes and behavior (12,13). Second, eating disinhibition measured by the adult Three-Factor Eating Inventory (14) was linked to genetic markers on chromosomes 7 (maximum logarithm of the odds ratio = 1.5, p = 0.0033) and 16 (maximum logarithm of the odds ratio = 1.5, p = 0.0033) in an extended Amish family study (15). Finally, a preliminary study of siblings and cousins suggested a heritable component to EAH in Hispanic youth (16).

The present study tested the hypothesis that children born at high risk for obesity, based on maternal prepregnancy weights, would display greater EAH than low-risk children. Analyses were conducted with and without statistical adjustment for measures of parental eating style and home provision of protocol snack foods because these environmental variables could confound any association between genetic risk and EAH. We also explored whether the children's nutritive sucking behavior at 3 months of age was associated with EAH, which might suggest a common mechanism underlying the two behavior traits. In previous analyses of this cohort, increased nutritive sucking rate was associated with increased weight gain during 2 years of life (5,6).

Research Methods and Procedures

Participants

Participating families were enrolled in a prospective Infant Growth Study (5,6,17,18,19) designed to assess the metabolic and behavioral determinants of obesity by a high risk design. Children included in this study were enrolled at 3 months of age and had been born to either obese or lean mothers, classified as a prepregnancy body weight either greater than the 66th percentile or less than the 33rd percentile (20). The mean ± standard deviation (SD) BMI (kilograms per meter squared) of high-risk mothers was 30.2 ± 3.6, whereas that of low-risk mothers was 19.6 ± 1.1. Infants in the Infant Growth Study were enrolled at birth and were followed from 3 months of age to the present age of 5 years. With the exception of the evaluation of nutritive sucking behavior completed at 3 months of age, all other laboratory procedures were evaluated at the 5-year visit in the Growth and Nutrition Laboratory and General Clinical Research Center of The Children's Hospital of Philadelphia.

Procedures

Children arrived at the laboratory between 4 pm and 4:30 pm to consume ad libitum from a standardized dinner meal protocol that included excess quantities of eight foods and beverages: pizza (260 grams), cheeseburger on a bun (130 grams), french fries (70 grams), ketchup packet (20 grams), carrots (60 grams), 2% milk (260 grams), fruit punch (130 grams), and apple juice (130 grams). Parents were instructed to behave as they would during a typical meal with their child in an out-of-home setting, and children were told they could request additional servings. The accompanying parent was allowed to sit near his or her child during the dinner but was not allowed to eat food from the child's plate. Before serving the dinner, the child was encouraged to “eat until s/he felt full.” We confirmed that children felt full using cartoon figures that quantify child satiety level (9) and by child verbal report. Children who reported feeling half full or hungry were encouraged to return to the dinner area to continue eating; however, children were neither encouraged to overeat nor forced to continue when they wanted to stop eating.

The EAH evaluation followed dinner in a nearby playroom that contained toys and activities, as well as generous servings of the 11 snack foods (see below) that were described as leftovers. The child was told that s/he could eat as much of the snack foods as desired, and s/he was then left alone unsupervised for 10 minutes. During this time, the child was free to play with the games and books and to eat the snack foods while being monitored, without his/her knowledge of the video camera. The investigator stopped in the playroom after 5 minutes to say hello and answer any questions. The experimental snack food session was completed in 10 minutes, after which the family's visit was formally completed and they left the laboratory. The foods were weighed before and after the EAH assessment. Using the nutrient information from the manufacturers, the child's total energy intake from the snack foods was calculated, and this amount represented his/her EAH.

The study protocol made three modifications from prior studies. First, all children were encouraged to eat until feeling full, rather than studying only those children who reported feeling full after eating the preload protocol dinner. Second, an evening dinner rather than lunch was provided. This meal time was more practical for families, most of whom preferred to come in the late afternoon for evaluation in the longitudinal study. Third, families visited our hospital-based laboratory rather than a university-based preschool laboratory. However, families were very familiar and comfortable with our setting, having made multiple visits to our center since the children were 3 months of age and knowing our investigative staff and environment. All procedures were approved by the Institutional Review Boards of the University of Pennsylvania and The Children's Hospital of Philadelphia. Informed consent was obtained from all families.

Measures

Obesity Risk Status

Age and gender were recorded for all children. All children were white. At study enrollment, children were classified as at high or low risk for obesity based on prepregnancy maternal weight, i.e., >66th percentile or <33rd percentile for their age group (20). At the 5-year visit, child weight was measured on a digital scale (model 4800; Scaletronix, Carol Stream, IL) and height on a wall-mounted stadiometer (Holtain, LTD, Crymych, United Kingdom) using standard techniques (21). Child BMI was computed and converted to z scores for analyses (22).

Nutritive Sucking

At 3 months of age, infants’ nutritive sucking behavior was measured at a midday test meal in the laboratory using an automated nutritive sucking apparatus (23,24). To maximize acceptance, nipples were constructed from several commercially available baby bottle nipples (Evenflo Products, Canton, GA; Gerber Products Co., Fremont, MI; and Playtex Products Inc., Westport, CT), adapted to deliver identical flow rates. The total number of sucks in the feeding interval was used in the present analyses (5,6).

Child EAH

At the 5-yr visit, after completion of the laboratory-based protocol dinner, children were given the opportunity to consume ad libitum the following foods: popcorn, low-fat popcorn, peanuts, potato chips, pretzels, fruit candy (Skittles), chocolate bars (Hershey), chocolate chip cookies, fig bars (Fig Newtons), jelly beans, and chocolate candy (M&Ms). Children were first asked to taste and rate very small portions of each snack food, using a rating scale of yummy, OK, or yuk. Afterwards, children were given free access to eat as much as they wanted from the snack foods, being told that the foods were leftovers. A child's total energy consumption (kilojoules) during the snack food session represented his/her EAH score. Parents were in a different room during this procedure.

Home Availability of Experimental Snack Foods

Parents completed a questionnaire that documented the extent to which they made available at home each of the EAH snack foods. This questionnaire was modeled after one previously used in the literature (9), although no validity data are available in the present sample. For each snack food, parents responded to the following six questions: How often s/he buys each snack food (0, never; 1, once per month; 2, twice per month; 3, once per week; 4, more than once per week); whether s/he deliberately limits how often each snack food is purchased (0, no; 1, yes); whether s/he tries to keep each snack food out of the child's reach (0, no or never; 1, yes); whether s/he limits when, during the day, the child is allowed to eat each snack food (0, no; 1, yes); in general, how much s/he limits the child's allowance of each snack food (0, never; 1, rarely; 2, sometimes; 3, usually; 4, always); and whether s/he limits how often the child may have each snack food (0, no; 1, yes). For each of the six questions, parents’ responses across the 11 snack foods were summed to create the following composite scales: parental frequency of snack food purchases, parental restriction of snack food purchases, parents keep snack foods out of reach, parents limit time of eating, parents limit allowance of snack foods, and parents limit snack food frequency. It should be noted that, for the first scale only, higher scores reflect a greater tendency of the parent to permit snack foods into the home, whereas for the remaining scales, higher scores generally reflect greater restriction of snack foods at home.

Parental Eating Styles

The Three-Factor Eating Inventory (14) was completed by mothers and fathers, respectively, at either the Year 4 or 5 annual assessment. This 51-item scale assesses three factors: dietary restraint (21 items), disinhibition (16 items), and perceived hunger (14 items). It has been used extensively in the literature to quantify individual differences in adult eating patterns.

Data Analysis

Descriptive statistics are presented as means ± SD. A two (risk status, high vs. low) × two (gender, boy vs. girl) ANOVA tested whether EAH differed as a function of risk status, gender, and a possible interaction. Similar two × two ANOVAs were also conducted for anthropometric measures and 3-month sucking behavior. Next, a series of two × two ANOVAs tested whether the measures of parental eating patterns and provision of experimental snack foods at home differed as a function of child risk status and gender. These analyses were conducted to identify parenting variables that might confound any association between risk status and EAH. That is, genetic risk might be associated with certain parenting variables evaluated in this report (i.e., a gene-environment correlation). If so, it would be unclear whether any association between risk status and EAH was due to potential genetic influences per se or parental eating attitudes and home snack food provisions. Therefore, additional analyses were conducted that tested the association between risk status and EAH adjusting for the Three-Factor Eating Inventory subscales and other feeding style variables. Finally, correlation analyses tested the relationship between EAH (kilojoules) and measures of the home availability of experimental snack foods as well as the total number of sucks/event at the 3-month nutritive sucking evaluation.

We excluded four subjects from an initial sample of 57 children who consumed >200 kcal during the EAH procedure and were statistical outliers from the rest of the sample per Tukey's (25) criteria (their scores exceeded 1.5 times the interquartile range for EAH scores). With respect to power for the primary study aim, the SD of EAH energy intake was estimated at ∼377 kJ (90 kcal) (10). Assuming an SD of 90 kcal, α = 0.05, and a two-tail significance test, a sample size of 53 children would have 80% power to detect a mean difference in EAH as small as ∼293 kJ (∼70 kcal) between high- and low-risk children for the full sample (or ∼140 kcal within each gender). Based on past research (11), we did not anticipate having sufficient power to detect an association between EAH levels and child weight status; therefore, this was not a study aim.

Results

Effects of Risk Status and Gender on EAH

Table 1 presents descriptive statistics on child anthropometry and ingestive behaviors. ANOVA revealed no significant effects of risk status (p = 0.31) or gender (p = 0.71) on EAH but a significant risk status-by-gender interaction (p = 0.05). EAH was significantly greater among high- than low-risk boys [mean = 326 vs. 151 kJ (78 vs. 36 kcal), p = 0.03]. EAH was less among high- than low-risk girls [mean = 188 vs. 246 kJ (45 vs. 59 kcal)] and was not significant (p = 0.50) (see Figure 1). The interaction p value did not change (p = 0.05) when BMI z score was included as a covariate. The two-way interaction was also observed (p = 0.05) when EAH was expressed per kilogram of body weight, with high-risk boys consuming approximately twice as much energy intake in absence of hunger than low-risk boys (p = 0.03). High-risk girls consumed non-significantly less energy in the absence of hunger per kilogram of body weight than low-risk girls (p = 0.50).

Table 1. . Means and SDs of child anthropometric and energy intake measures
 BoysGirls
VariableLow riskHigh riskLow riskHigh risk
  • *

    Obesity risk status main effect significant at p < 0.05.

  • Trend for obesity risk status main effect, p = 0.07.

  • Obesity risk status-by-gender interactions significant at p ≤ 0.05.

N14131115
Weight (kg)*17.8 ± 1.620.2 ± 3.918.5 ± 1.620.3 ± 4.2
Height (cm)107.5 ± 4.5110.3 ± 3.5109.1 ± 3.3108.8 ± 4.0
Child BMI (kg/m2)*15.4 ± 0.616.6 ± 3.015.5 ± 0.917.0 ± 2.7
Child BMI z score0.0 ± 0.60.5 ± 1.50.2 ± 0.70.8 ± 1.2
Pre-meal/lunch intake (kJ)1979 ± 9121974 ± 7401644 ± 8991690 ± 987
EAH, snacks (kJ)151 ± 146326 ± 246246 ± 238188 ± 192
Weight-adjusted eating in absence of hunger (kJ/kg)8.7 ± 9.216.3 ± 13.013.4 ± 13.09.2 ± 10.0
3-month nutritive sucks (no.)*550 ± 241845 ± 591736 ± 3691052 ± 555
Figure 1.

Mean ± standard error of EAH (kilojoules) as a function of obesity risk status in boys and girls. The difference in intake between low- and high-risk children was statistically significant among boys (p = 0.03) but not among girls (p = 0.50).

ANOVA for anthropometric outcomes revealed that, compared with low-risk children, high-risk children weighed more (p = 0.02), had a higher BMI (p = 0.03), and had a trend toward a higher BMI z score (p = 0.07). There were no gender differences or gender-by-risk interaction for any of these measures. ANOVA for 3-month nutritive sucks revealed a significant main effect only of risk group, with high-risk children exerting more total sucks than low-risk children (956.2 vs. 609.0 sucks, p = 0.02).

Association among Risk Status, Gender, and Parent Eating Behaviors and Provision of Protocol Snack Foods

Table 2 presents descriptive statistics for the Eating Inventory subscales and for parental provision of the experimental snack foods at home. Disinhibition was significantly greater among mothers of high- than low-risk children (p < 0.001), although this differed by gender (interaction p = 0.05). Specifically, the effect of risk status on maternal disinhibition was more pronounced for boys (10.2 vs. 2.6, p < 0.001) than for girls (8.1 vs. 3.6, p = 0.003). Hunger levels were significantly elevated among mothers of high- compared with low-risk children (p = 0.03), and this effect differed by gender (interaction p = 0.04). The effect of risk status on maternal hunger was significant for boys (6.4 vs. 2.8, p = 0.001) but not for girls (4.2 vs. 4.1, p = 0.93). With respect to paternal eating styles, there was a gender-by-risk status interaction for restrained eating (p = 0.03). Dietary restraint was significantly lower among the fathers of high-risk boys than fathers of low-risk boys (3.9 vs. 7.3, p = 0.06), and there was no significant difference among the fathers of high- vs. low-risk girls (6.4 vs. 4.6, p = 0.18). Finally, high-risk parents were more likely than low-risk parents to limit how often they purchased the 11 experimental snack foods for the home environment (p = 0.04).

Table 2. . Means and SDs for measures of parent eating patterns and the home availability of experimental snack foods
 BoysGirls
VariableLow riskHigh riskLow riskHigh risk
  • *

    The sample sizes for these analyses were N = 14 (low-risk boys), N = 13 (high-risk boys), N = 11 (low-risk girls), and N = 14 (high-risk girls) due to parents who did not complete certain questions.

  • The sample sizes for these analyses were N = 12 (low-risk boys, restraint scale), N = 13 (low-risk boys, eating disinhibition and hunger subscales), N = 11 (high-risk boys), N = 11 (low-risk girls), and N = 12 (high-risk girls) due to parents who did not complete certain questions.

  • The sample sizes for these analyses were N = 14 (low-risk boys), N = 12 (high-risk boys), N = 10 (low-risk girls), and N = 14 (high-risk girls) due to parents who did not complete certain questions.

  • §

    Obesity risk status main effect significant at p < 0.05.

  • Obesity risk status-by-gender interactions significant at p ≤ 0.05.

Eating inventory scales (mother)*    
 Restraint7.5 ± 4.38.8 ± 3.710.0 ± 5.012.5 ± 4.4
 Eating disinhibition §,2.6 ± 1.510.2 ± 3.03.6 ± 3.28.1 ± 3.3
 Hunger §,2.9 ± 2.06.4 ± 2.84.1 ± 3.74.2 ± 2.8
Eating inventory scales (father)    
 Restraint7.3 ± 5.83.9 ± 2.74.5 ± 2.86.4 ± 3.7
 Eating disinhibition3.9 ± 3.25.1 ± 2.72.5 ± 1.64.8 ± 3.2
 Hunger3.2 ± 3.14.8 ± 2.43.5 ± 2.63.5 ± 2.3
Parental questionnaire scales    
 Frequency of snack food purchases10.5 ± 4.49.3 ± 5.86.5 ± 3.610.6 ± 5.6
 Restriction of snack food purchases§4.9 ± 3.57.2 ± 3.35.0 ± 3.96.7 ± 2.7
 Keep snack food out of reach2.9 ± 2.44.4 ± 3.43.3 ± 3.73.1 ± 3.4
 Limit time of eating7.0 ± 2.96.5 ± 3.75.6 ± 3.46.3 ± 3.6
 Limit allowance of snack foods29.8 ± 11.634.2 ± 13.729.8 ± 12.531.6 ± 9.9
 Limit snack food frequency4.7 ± 2.97.8 ± 2.46.6 ± 4.15.7 ± 3.0

Effects of Risk Status on EAH Controlling for Parent Behaviors

The initial association between risk status and EAH among boys remained significant, or borderline significant, when controlling for parental Eating Inventory scores or the measures of home snack food provision (p ≤ 0.06 across analyses).

Home Access to Protocol Snack Foods and Child EAH

Among boys, EAH was not significantly associated with any of the six scale scores representing home availability of experimental snack foods (p > 0.20). Among girls, there was a trend for parental limitations on the frequency of daughter consumption of protocol snack foods at home to be associated with increased EAH (r = 0.39, p = 0.06) (see Figure 2). None of the other associations among girls was significant (p > 0.13)

Figure 2.

Association between parental limitations on daughter snacking of protocol snack foods at home (x-axis) and daughter EAH (y-axis). Higher scores represent a greater amount of limitations.

Nutritive Sucking and EAH

Three-month sucking behavior was not associated with EAH in girls (r = 0.06, p = 0.77), boys (r = −0.03, p = 0.90), low-risk children (r = 0.16, p = 0.48), or high-risk children (r = −0.16, p = 0.41).

Discussion

This study provides suggestive evidence that EAH behavior may be a genetically loaded eating phenotype that promotes positive energy balance in 5-year-old boys. Boys born at high risk for obesity consumed over twice as much energy in the absence of hunger as boys born at low risk for obesity. This effect changed minimally, if at all, when controlling for measures of parental eating patterns or the home availability of protocol snack foods. The design could not control for other environmental or parenting style variables that might be correlated with obesity risk status; still, the results are consistent with a potential genetic hypothesis and complement findings from a preliminary study of Hispanic children that reported a heritable component to EAH (16). Thus, EAH may be an informative behavioral phenotype for genetics studies of child nutrition and energy balance, at least among boys.

EAH behavior was unrelated to risk status among girls, suggesting that variability in the trait may be less genetically loaded and more susceptible to environmental influences. That the same eating trait may be under greater genetic influence in males than females has precedence in the literature. van den Bree et al. (26) examined the genetic-environmental architecture of high fat-sugar-salt intake and healthful eating habits in 4500 adult twin pairs who completed a 24-hour food frequency questionnaire. The magnitude of genetic and environmental influences on the eating patterns differed significantly by gender. Heritability estimates for these traits were consistently larger for men, whereas estimates of household environmental influence were significant only among women. A more recent study of 5250 male and female twin pairs, 16 years of age, examined the heritability of reported breakfast eating patterns (27). Results indicated compelling gender differences, with the authors concluding that “Breakfast eating is moderated differently in adolescent boys and girls. Unlike boys, girls are much influenced by the family and pair-specific environment. In girls, environmental influences may override genetically driven factors.”

Parental restriction of the frequency of protocol snack food consumption at home by daughters was associated with increased EAH in girls. Similarly, in a study of 197 girls who were 4 to 6 years of age, daughters whose parents reported greater restriction of the laboratory snack foods at home showed greater EAH in the laboratory (28). Fisher and Birch (9) found that maternal restriction of snack foods was significantly associated with increased EAH in girls (r = 0.59, p < 0.01) but not in boys (r = −0.03, p > 0.05). These findings suggest that parental restriction of child food choices may play a more prominent role in the development of EAH among daughters than sons. It is also likely that parental limitations on home snack food provisions are elicited by daughters’ increased food intake (11,20).

Among boys, obesity risk status was associated with differences in maternal eating disinhibition and hunger levels and differences in provision of the protocol snack foods at home. We were concerned that these experiences, rather than a genetic influence per se, might have influenced the risk group differences in EAH in boys. Therefore, we controlled for these variables as potential confounders and found that the effect of risk status on EAH was changed only minimally. This gives greater assurance to the robustness of the main finding. Twin and adoption designs offer more powerful strategies for disaggregating genetic from home environmental influences on eating phenotypes (8,29,30).

The total number of nutritive sucks at 3 months of age was unrelated to EAH in both genders. Because nutritive sucking significantly differed between high- and low-risk infants at 3 months of age (5), one might have expected an association with EAH, at least among boys. This was not observed. This finding suggests that these two behavioral phenotypes may influence energy balance and obesity through different mechanisms (6).

It is noteworthy that the mean EAH scores for the present sample were less than those reported in a prior study (9). This may relate to the fact that our child participants were served a dinner meal, as opposed to a lunch meal, which may have led to greater total energy intake at the protocol preload meal. Fisher and Birch (10) reported that 5-year-old children in their sample consumed an average of 1556 kJ (372 kcal) at the lunch preceding EAH evaluation, whereas the mean total energy intake at the dinner preceding EAH evaluation in the present study ranged from 1644 kJ (393 kcal) for low-risk girls to 1979 kJ (473 kcal) for low-risk boys. These findings suggest the importance of time of day or meal setting for the design of future studies.

This study has at least three limitations. First, although it was adequately powered to detect at least moderate effect sizes, power to detect small associations was limited. The sample size of this study was much smaller than that of Kreski-Rahkonen et al. (30), although a notable advantage of the present study was its use of weighed food intake rather than self-report. Second, the sample had been selected to include only white families. Ethnically homogenous samples are desirable for genetics studies, and the present findings should not be generalized to other ethnic or racial groups. Third, and most notably, this study cannot formally separate genetic from environmental influences on EAH behavior. Beyond the genes transmitted to children, parents of high-risk children may also foster environments that influence EAH and other eating and physical activity behaviors conducive to the development of obesity in children.

In summary, 5-year-old boys born at high risk for obesity consumed more than twice the amount of energy in the absence of hunger than boys born at low risk for obesity, and this effect changed minimally when controlling for parental eating and food provision characteristics. At least among boys, the genes that predispose certain children to obesity may partially promote positive energy balance through EAH. Additional studies, with larger samples and genetically informative designs, need to follow up the present investigation. There may be important gender differences in the development of EAH tendencies that warrant further investigation. The trait may prove to be a useful target for overweight prevention studies if additional prospective studies link the trait to excess weight gain during childhood (11).

Acknowledgement

This work was supported by NIH Grant DK068899, the General Clinical Research Center (Grant RR 00240), and the Nutrition Center of the Children's Hospital of Philadelphia.

Footnotes

  1. Nonstandard abbreviation: EAH, eating in the absence of hunger; SD, standard deviation.

Ancillary