Genetic influence on appetite in children



The modern environment is ubiquitously ‘obesogenic’, yet people vary enormously in weight. One factor contributing to weight variation could be genetically determined differences in appetite that modulate susceptibility to the environment. We assessed the relative contribution of genes and environment for two aspects of appetite that have been implicated in obesity.


Parents of a population-based sample of 8- to 11-year-old twins (n=5435 pairs) completed validated, questionnaire measures of responsiveness to satiety and responsiveness to food cues for both children.


Quantitative genetic model fitting gave estimates of 63% (95% confidence interval: 39–81%) for the heritability of satiety responsiveness and 75% (52–85%) for food cue responsiveness. Shared and non-shared environmental influences were 21% (0–51%) and 16% (10–21%) for satiety responsiveness, and 10% (0–38%) and 15% (10–18%) for food cue responsiveness, respectively.


The high heritability of appetitive traits that are known to be related to weight suggests that genetic vulnerability to weight gain could operate through behavioural as well as metabolic pathways. Intervention strategies aimed at improving satiety responsiveness and reducing food cue responsiveness in high-risk individuals could help in preventing the development of obesity.


Modern environments provide ample opportunities for sedentary pastimes and excessive food consumption;1 yet there is enormous individual variation in adiposity, suggesting that individuals differ in their susceptibility to ‘obesogenic’ environments. Evidence for the high heritability of weight both in adults and children indicates that environmental susceptibility may be genetically determined.2, 3, 4

The search for genetic markers associated with higher body weights is making progress,5, 6, 7, 8 and this raises the issue of the mechanisms through which genetic variations affect weight. Most attention has been paid to energy expenditure and fat storage as mechanisms underlying the genetic effects.9, 10 However, genes could also affect weight through appetite-related behavioural traits.11 There is evidence for genetic influence on circulating peptides and steroids that affect appetite in humans,12 but much of the endocrinological research derives from studies of rare monogenic obesity syndromes, which may or may not apply to population weight variation.6 If appetitive traits that are known to be related to weight could be shown to be heritable, this would strengthen the argument that genetically determined variations in appetite could contribute to weight variation in the population.

To find candidate appetitive traits, we revisited Schachter's psychological theory of obesity,13 which proposed that impaired sensitivity to internal satiety cues and excessive responsiveness to external food cues contribute to the development of obesity. Clusters of behaviours related either to satiety responsiveness (for example, failure to downregulate intake after a preload,14 lack of deceleration during an eating episode15) or to hyperresponsiveness to food cues (for example, eating in the absence of hunger,16 enhanced salivation at the presentation of food,17 upregulation of intake with palatable foods,18 higher reinforcing value of food19) have been observed in obese children and adults.

The aim of this study was to test the hypothesis that weight-related appetitive characteristics are heritable. Several studies have indicated moderate heritability for eating-disordered traits such as disinhibited eating,20, 21 and others have demonstrated moderate heritability for measures of eating behaviour.22, 23, 24 However, few studies have investigated conceptually defined appetitive traits that are known to be associated with common obesity, and no large-scale studies have assessed the heritability of appetite early in life when eating is more spontaneous and less susceptible to social and cognitive factors. We have shown that satiety responsiveness and food cue responsiveness not only distinguish obese from normal-weight children, but are also associated with the full range of variation in adiposity.25 This study examined the heritability of these traits, using a validated psychometric instrument in a large sample of young twins from a population-representative cohort of families. Ethical approval for this study was granted by the UCL Committee for the Ethics of non-National Health Service (NHS) Human Research.



The sampling frame for this study was the Twins’ Early Development Study, a population-based cohort of twins born in the UK in 1994–1996. The Twins’ Early Development Study has been shown to be reasonably representative of the UK population and is described in more detail elsewhere.26 Families defined as ‘active participants’ in the Twins’ Early Development Study (that is, parents had participated since the twins were 7 years old; n=8978) were sent the Child Eating Behaviour Questionnaire (CEBQ), of whom 5543 families (61.7%) completed and returned it. A further 3234 ‘inactive’ families were contacted, of whom 359 responded. The two groups did not differ in demographic characteristics or key outcome variables, and they were combined to give a total sample of 5902 families.

From the analysis, we excluded 467 families in which either twin had a specific medical syndrome or severe perinatal problems such as extreme low birth weight, or for whom sex, zygosity or age data were unavailable, leaving a sample of 5435 families. We additionally excluded 45 families in which either twin scored more than three standard deviations below or above the mean for the CEBQ scales, to avoid disproportionate influence from outliers, leaving a sample of 5390. We repeated the analyses using a range of alternative exclusion criteria and the findings remained the same; we therefore do not report these results.


Appetite was assessed with two scales from the CEBQ, a validated, psychometric, parent-report instrument.27 Satiety responsiveness was assessed using a shortened version of the Satiety Responsiveness/Slowness in Eating scale (SR/SE; six items), and food cue responsiveness was indexed with the Enjoyment of Food (EF; four items) scale. The SR/SE scale assesses the child's responsiveness to internal satiety cues (for example, ‘my child cannot eat a meal if he/she has had a snack just before’), together with slowness in eating, a trait that has been thought to reflect heightened sensitivity to satiety (for example, ‘my child eats more and more slowly during the course of a meal’). The EF scale assesses the child's general interest in food and eating (for example, ‘my child loves food’) and indicates greater responsiveness to food cues. Scale scores were generated by calculating item means and were standardized and residualized for age and sex effects using a regression procedure.


Independent samples t-tests were used to test for differences in CEBQ scores by sex or zygosity. Intraclass correlations within twin pairs were calculated for each sex and zygosity group to compare within-pair resemblance in monozygotic (MZ) twins (who are genetically identical) and dizygotic (DZ) twins who share on average half of their segregating genes. If MZ twin pairs are more similar than DZ pairs, genetic influence is implicated, with the difference in the degree of resemblance providing an estimate of the heritability of the trait. The extent that twin pairs (MZ or DZ) are more similar than would be expected from the heritability of the trait gives an indication of the influence of the shared family environment (aspects of the environment that make the members of the pair more alike). The extent that MZ (genetically identical) twins differ from one another implicates aspects of the individual environment of each twin (non-shared environment) as determinants of the trait.

Mx structural equation modelling software was used to perform standard ACE model-fitting analyses to gain accurate estimates of additive genetic (A), shared environment (C) and non-shared environment (E) contributions to variation on each scale.28, 29 We used a sibling interaction model because the DZ intraclass correlation was much lower than the MZ intraclass correlation. This can suggest assimilation effects, contrast effects or non-additive genetic effects.30 However, in this sample, there was also greater DZ than MZ variance, which is indicative of contrast effects in the data.31 Contrast and assimilation effects (collectively called sibling interaction effects) refer to either genuine interaction between twins’ phenotypes or rating biases that artificially inflate differences between MZ and DZ twin correlations, and therefore overestimate genetic influence. Under these circumstances, more accurate A, C and E parameter estimates are obtained by modelling the contrast effects in a sibling interaction model.

Although the low DZ correlation may also suggest genetic dominance, we did not fit an ADE model because it does not take account of variance differences between the MZ and DZ twins. In addition, additive genetic and dominant genetic influences are highly correlated and therefore require very large samples to have the power to discriminate them.

Figure 1 shows a path diagram of a sibling interaction model. The rectangular boxes refer to observed phenotypes and the circles represent latent genetic and environmental factors. The single-headed arrows represent partial regressions of the variable on the latent factor (that is, the relative influence of the latent variable (for example, A) on the phenotype) and the curved connectors represent correlations between the connected factors. In this model, the genetic and shared environment correlations are fixed at 1.0 and 1.0 for MZ twins and 0.5 and 1.0 for DZ twins. In addition, there is a path (s) that directly links the phenotypes of the twins and allows the modelling of variance differences between MZ and DZ twins. Therefore, each twin's phenotype is a function of additive genetic influences, shared and non-shared environmental influences and their co-twin's phenotype. This model provides estimates for seven parameters: male A, C, E estimates (am, cm, em) and female A, C, E estimates (af, cf, ef) and an estimate of the interaction (s).

Figure 1

Path diagram of a sibling interaction model.

A series of nested models were fitted to test the significance of the interaction parameters and to test for quantitative sex differences (that is, different levels of influence between the sexes). The first nested model dropped the interaction parameter(s) and provides estimates for six parameters: am, cm, em, af, cf, ef. A significant worsening of fit suggests that keeping the interaction term gives a better approximation of the data. The second nested model tested for sex differences in A, C and E parameters, with the interaction term included in the model. It provides estimates for four parameters: a, c and e (for the sexes combined) and s. Two-fit indices (χ2), and Akaike's information criterion32 were used to establish the best-fitting model.


Sample characteristics

Of the 5435 families with complete data, there were 1932 MZ pairs (902 male and 1030 female pairs) and 3503 DZ pairs (885 male, 914 female, 1704 opposite sex). The twins were aged 9.9 years (range: 8.3–11.6) when the questionnaires were returned. There were equal numbers of boys and girls, and the numbers of MZ and DZ twin pairs reflected expected rates in the population (36% MZ; 64% DZ). Most parental respondents were the mothers (97%). Most families were white (93%), reflecting the ethnic distribution in the UK population, and there was a range of educational levels, also reflecting the UK population.

Means and standard deviations for CEBQ scores are presented in Table 1. There were significant effects of zygosity on both scores, but the effect sizes were less than 1% of the variance. Girls had slightly higher SR/SE scores than boys, but again the effect size was very small. For SR/SE, the variance for DZ twins exceeded the variance among MZ twins, indicating a contrast effect in DZ twins. There were no significant interactions between sex and zygosity.

Table 1 Means (standard deviations) for CEBQ subscale scores by zygosity and sex

Intraclass twin correlations for each CEBQ scale are shown in Table 2. For both scales, MZ correlations greatly exceeded the DZ correlations, indicating strong genetic effects. Across zygosity categories, correlations in male and female pairs were similar. Correlations for opposite-sex DZ twins were similar to those for same-sex DZ twins. DZ correlations for SR/SE were less than half the MZ correlations.

Table 2 Intraclass correlations between twins for CEBQ subscale scores, by zygosity and sex

Tables 3a,b show the results of model-fitting analyses. For SR/SE, the likelihood ratio χ2 tests identified the third model (no quantitative sex differences and a sibling interaction parameter) as the best-fitting model. The results indicated substantial heritability and modest shared environmental influence. Estimates from the best-fitting models were 0.63, 0.21 and 0.16 for genetic, shared environment and non-shared environment influences, respectively, and −0.25 for the interaction term. For EF, the best-fitting model was the full model including an interaction parameter, and allowing male and female A, C and E estimates to differ, although the actual values for males and females were very similar. This gave estimates of 0.78, 0.10 and 0.13 for boys and 0.70, 0.13 and 0.17 for girls, with estimates of 0.75, 0.10 and 0.15 for the sexes combined, respectively. The sibling interaction term was significant and negative, indicating contrast effects.

Table 3 Model fit and parameter estimates (95% confidence intervals) for sex-limitation model fitting for CEBQ satiety responsiveness/slowness in eating


The results of this study indicate substantial genetic influence on two potentially obesogenic appetitive traits among 9- to 11-year-old children. Heritability was 0.63 for satiety responsiveness and 0.75 for food cue responsiveness. Shared environment effects were modest, indicating that being reared in the same home had a relatively small effect on making the twins similar to one another in terms of appetite. These results resemble the results for adiposity itself in this sample,4 which found heritability to be at the upper end of the typical estimates of 0.5–0.833 with a shared environment effect below 0.15. In combination with evidence that lower satiety sensitivity and higher responsiveness to food cues are associated with higher weight,25 the present results are consistent with the hypothesis that genetically determined differences in appetite are functional steps on the pathway from genes to obesity. This implies that the correct answer to the question ‘is obesity due to genes or behaviour?’ is ‘both’, because weight-related behavioural traits are under genetically influence. We propose that heritable differences in appetite contribute to differences in sensitivity to the obesogenic effects of the environment.

This is the largest twin study of eating behaviour in the literature and the only one that has focused on children; nevertheless it has a number of limitations. The large sample size had the advantage of providing robust estimates of genetic and environmental effects, but it was ‘over-powered’ for some analyses, for example, producing significant decrements in model fit even when absolute effects were very small. In the present analyses, the sex-limited model produced a better fit for the EF scale despite male and female estimates being only marginally different from one another. The contrast effects were another limitation. They may have resulted from parents of DZ twins rating them more differently from each other than parents of MZ twins,30 and although the model-fitting analyses take account of the contrast effects, estimates of heritability may nonetheless be slightly inflated. It is also possible that dominant genetic effects are important for eating behaviour, but we were not able to assess this statistically. Confirmation of the mode of inheritance will depend on finding the specific genes associated with childhood eating behaviour, where there is already some progress.34

The inherited nature of appetitive traits in children has practical implications. Most importantly, it points out the need for clinicians to be aware that maintaining a healthy weight is harder for some children than others because of the combination of their genetically determined appetite responses and the permissive modern food environment that we all share. It also suggests that placing the blame for childhood obesity on parents may be unfair, because obesogenic eating behaviours can be transmitted genetically, and parents cannot be held responsible for their genes. There is a concern that demonstrating genetic influence promotes fatalistic thinking. But in reality, for complex traits, genetic influence denotes probabilistic risk, and identifying genetic vulnerabilities early in life might make it possible to modify them before their full phenotypic expression becomes difficult to reverse. For example, a child with low satiety responsiveness could be given carefully managed portions and a higher proportion of low-energy-dense foods. A child with high food cue responsiveness may benefit from careful control over environmental cues to eat, for example, by keeping palatable foods out of sight in the home and limiting availability of energy-dense choices.

Establishing genetic influence on these two appetitive traits provides an impetus for exploring other obesogenic behavioural traits. For example, there is some evidence that obese individuals have greater preferences for high fat foods35, 36 and prefer sedentary to active pastimes.37 These may also be heritable traits contributing to adiposity, which could be explored using twin or family designs. It is also important to analyse the relationships among the variety of appetitive traits that are measured in behavioural studies. Satiety sensitivity and food cue responsiveness may—as Schachter13 proposed in his original studies—be different traits with different aetiologies, or they may be different indicators of the same underlying trait. Future research should assess phenotypic associations and investigate the overlap in genetic and environmental aetiological factors to help to distinguish these alternative views of the architecture of human appetite. The findings of this study also suggest that psychometric measures of appetite could be useful phenotypic indicators to incorporate into large-scale genotyping studies, bringing together genetic and behavioural approaches to illuminate the determinants of body weight and obesity.

Table 4 Model fit and parameter estimates (95% confidence intervals) for sex-limitation model fitting for CEBQ enjoyment of food


  1. 1

    Hill JO, Wyatt HR, Reed GW, Peters JC . Obesity and the environment: where do we go from here? Science 2003; 299: 853–855.

    CAS  Article  Google Scholar 

  2. 2

    Stunkard AJ, Sorensen TI, Hanis C, Teasdale TW, Chakraborty R, Schull WJ et al. An adoption study of human obesity. N Engl J Med 1986; 314: 193–198.

    CAS  Article  Google Scholar 

  3. 3

    Maes HH, Neale MC, Eaves LJ . Genetic and environmental factors in relative body weight and human adiposity. Behav Genet 1997; 27: 325–351.

    CAS  Article  Google Scholar 

  4. 4

    Wardle J, Carnell S, Haworth C, Plomin R . Evidence for strong genetic influence on childhood adiposity despite the force of the obesogenic environment. Am J Clin Nutr 2008; 87: 398–404.

    CAS  Article  Google Scholar 

  5. 5

    Barsh GS, Farooqi IS, O’Rahilly S . Genetics of body-weight regulation. Nature 2000; 404: 644–651.

    CAS  Article  Google Scholar 

  6. 6

    O’Rahilly S, Farooqi IS, Yeo GS, Challis BG . Minireview: human obesity-lessons from monogenic disorders. Endocrinology 2003; 144: 3757–3764.

    Article  Google Scholar 

  7. 7

    Frayling TM, Timpson NJ, Weedon MN, Zeggini E, Freathy RM, Lindgren CM et al. A common variant in the FTO gene is associated with body mass index and predisposes to childhood and adult obesity. Science 2007; 316: 889–894.

    CAS  Article  Google Scholar 

  8. 8

    Young EH, Wareham NJ, Farooqi S, Hinney A, Hebebrand J, Scherag A et al. The V103I polymorphism of the MC4R gene and obesity: population based studies and meta-analysis of 29 563 individuals. Int J Obes (Lond) 2007; 31: 1437–1441.

    CAS  Article  Google Scholar 

  9. 9

    Froguel P, Boutin P . Genetics of pathways regulating body weight in the development of obesity in humans. Exp Biol Med (Maywood) 2001; 226: 991–996.

    CAS  Article  Google Scholar 

  10. 10

    Mutch DM, Clement K . Unraveling the genetics of human obesity. PLoS Genet 2006; 2: e188.

    Article  Google Scholar 

  11. 11

    Gerken T, Girard CA, Tung YC, Webby CJ, Saudek V, Hewitson KS et al. The obesity-associated FTO gene encodes a 2-oxoglutarate dependent nucleic acid demethylase. Science 2007; 318: 1469–1472.

    CAS  Article  Google Scholar 

  12. 12

    Coll AP, Farooqi IS, O’Rahilly S . The hormonal control of food intake. Cell 2007; 129: 251–262.

    CAS  Article  Google Scholar 

  13. 13

    Schachter S . Obesity and eating. Internal and external cues differentially affect the eating behavior of obese and normal subjects. Science 1968; 161: 751–756.

    CAS  Article  Google Scholar 

  14. 14

    Johnson SL, Birch LL . Parent's and children's adiposity and eating style. Pediatrics 1994; 94: 653–661.

    CAS  PubMed  Google Scholar 

  15. 15

    Barkeling B, Ekman S, Rossner S . Eating behaviour in obese and normal weight 11-year-old children. Int J Obes Relat Metab Disord 1992; 16: 355–360.

    CAS  PubMed  Google Scholar 

  16. 16

    Fisher JO, Birch LL . Eating in the absence of hunger and overweight in girls from 5–7 y of age. Am J Clin Nutr 2002; 76: 226–231.

    CAS  Article  Google Scholar 

  17. 17

    Epstein LH, Paluch R, Coleman KJ . Differences in salivation to repeated food cues in obese and nonobese women. Psychosom Med 1996; 58: 160–164.

    CAS  Article  Google Scholar 

  18. 18

    Hashim SA, Van Itallie TB . Studies in normal and obese subjects with a monitored food dispensing device. Ann N Y Acad Sci 1965; 131: 654–661.

    CAS  Article  Google Scholar 

  19. 19

    Saelens BE, Epstein LH . Reinforcing value of food in obese and non-obese women. Appetite 1996; 27: 41–50.

    CAS  Article  Google Scholar 

  20. 20

    Tholin S, Rasmussen F, Tynelius P, Karlsson J . Genetic and environmental influences on eating behavior: the Swedish young male twins study. Am J Clin Nutr 2005; 81: 564–569.

    CAS  Article  Google Scholar 

  21. 21

    Steinle NI, Hsueh WC, Snitker S, Pollin TI, Sakul H, St Jean PL et al. Eating behavior in the old order amish: heritability analysis and a genome-wide linkage analysis. Am J Clin Nutr 2002; 75: 1098–1106.

    CAS  Article  Google Scholar 

  22. 22

    De Castro JM, Plunkett SS . How genes control real world intake: palatability–intake relationships. Nutrition 2001; 17: 266–268.

    CAS  Article  Google Scholar 

  23. 23

    Fisher JO, Cai G, Jaramillo SJ, Cole SA, Comuzzie AG, Butte NF . Heritability of hyperphagic eating behavior and appetite-related hormones among Hispanic children. Obesity (Silver Spring) 2007; 15: 1484–1495.

    Article  Google Scholar 

  24. 24

    De Castro JM . Heredity influences the dietary energy density of free-living humans. Physiol Behav 2006; 87: 192–198.

    CAS  Article  Google Scholar 

  25. 25

    Carnell S, Wardle J . Appetite and adiposity in children: evidence for a behavioral susceptibility theory of obesity. Am J Clin Nutr 2008; 88: 22–29.

    CAS  Article  Google Scholar 

  26. 26

    Oliver BR, Plomin R . Twins early development study (TEDS): a multivariate, longitudinal genetic investigation of language, cognition and behavior problems from childhood through adolescence. Twin Res Hum Gen 2007; 10: 96–105.

    Article  Google Scholar 

  27. 27

    Wardle J, Guthrie CA, Sanderson S, Rapoport L . Development of the children's eating behaviour questionnaire. J Child Psychol Psychiat 2001; 42: 963–970.

    CAS  Article  Google Scholar 

  28. 28

    Plomin R, DeFries JC, McClearn GE, McGuffin P . Behavioral Genetics (2001). 4th (edn). New York: Worth Publishers.

    Google Scholar 

  29. 29

    Rijsdijk FV, Sham PC . Analytic approaches to twin data using structural equation models. Brief Bioinform 2002; 3: 119–133.

    CAS  Article  Google Scholar 

  30. 30

    Saudino KJ, Cherny SS, Plomin R . Parent ratings of temperament in twins: explaining the ‘too low’ DZ correlations. Twin Res 2000; 3: 224–233.

    CAS  Article  Google Scholar 

  31. 31

    Eaves L . A model for sibling effects in man. Heredity 1976; 36: 205–214.

    CAS  Article  Google Scholar 

  32. 32

    Akaike H . Factor analysis and the AIC. Psychometrica 1987; 52: 317–332.

    Article  Google Scholar 

  33. 33

    Grilo CM, Pogue-Geile MF . The nature of environmental influences on weight and obesity: a behavior genetic analysis. Psychol Bull 1991; 110: 520–537.

    CAS  Article  Google Scholar 

  34. 34

    Wardle J, Carnell S, Haworth CMH, Farooqi IS, O'Rahilly S, Plomin R . Obesity-associated genetic variation in FTO is associated with diminished satiety. J Clin Endocrin Metab 2008; in press.

  35. 35

    Rissanen A, Hakala P, Lissner L, Mattlar CE, Koskenvuo M, Ronnemaa T . Acquired preference especially for dietary fat and obesity: a study of weight-discordant monozygotic twin pairs. Int J Obes Relat Metab Disord 2002; 26: 973–977.

    CAS  Article  Google Scholar 

  36. 36

    Mela DJ, Sacchetti DA . Sensory preferences for fats: relationships with diet and body composition. Am J Clin Nutr 1991; 53: 908–915.

    CAS  Article  Google Scholar 

  37. 37

    Wardle J, Guthrie C, Sanderson S, Birch L, Plomin R . Food and activity preferences in children of lean and obese parents. Int J Obes Relat Metab Disord 2001; 25: 971–977.

    CAS  Article  Google Scholar 

Download references


This research was supported by a grant from the Biological and Biotechnology Research Council. The Twins’ Early Development Study is funded by the Medical Research Council. SC is funded by an interdisciplinary fellowship from the Economic and Social Research Council and the Medical Research Council, CH and RP by the Medical Research Council and JW by Cancer Research UK.

Author information



Corresponding author

Correspondence to J Wardle.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Carnell, S., Haworth, C., Plomin, R. et al. Genetic influence on appetite in children. Int J Obes 32, 1468–1473 (2008).

Download citation


  • children
  • appetite
  • eating behaviour

Further reading