Pediatric Highlight

International Journal of Obesity (2009) 33, 42–45; doi:10.1038/ijo.2008.174; published online 7 October 2008

The FTO gene and measured food intake in children

J Wardle1, C Llewellyn1, S Sanderson2 and R Plomin3

  1. 1Department of Epidemiology and Public Health, Health Behaviour Research Centre, University College London, London, UK
  2. 2Social and Behavioral Research Branch, National Human Genome Research Institute, National Institutes of Health, Bethesda, MD, USA
  3. 3King's College London, Social, Genetic and Developmental Psychiatry Centre, Institute of Psychiatry, London, UK

Correspondence: Professor J Wardle, Health Behaviour Research Centre, Department of Epidemiology and Public Health, University College London, Gower Street, 2-16 Torrington Place, London WC1E 6BT, UK. E-mail:

Received 3 August 2008; Revised 13 August 2008; Accepted 6 September 2008; Published online 7 October 2008.





Polymorphisms in the obesity-associated gene, FTO, have been linked with sensitivity to satiety in children, indicating FTO may be influencing one of the regulatory drivers underlying food intake. In this study, we tested the hypothesis that food intake in a standard eating behaviour paradigm in which palatable food is offered under conditions of satiety would be associated with FTO genotype status, after controlling for differences in body mass index (BMI).



Participants were 131 children aged 4–5 years, taking part in a behavioural study of food intake for whom DNA was available for genotyping. The phenotypic indicator of intake was the child's consumption of palatable food presented after having eaten a meal. We also assessed physical activity using parental reports of the child's enjoyment of active games, their level of activity relative to other children and a standard measure of fidgetiness. Associations between polymorphisms of the intronic FTO single nucleotide polymorphism (rs9939609) and behaviour (food intake and activity) were assessed by analysis of variance controlling for sex, age and BMI s.d. scores.



The distribution of AA (homogenous for A allele), AT (heterogeneous T and A alleles) and TT (homogenous for T allele) genotypes was 18, 50 and 32%, respectively. As predicted, TT homozygotes ate significantly less than heterozygotes (P=0.03) or AA homozygotes (P=0.02). The effect was not diminished by controlling for BMI s.d. scores. There were no significant associations between FTO genotype and any marker of physical activity.



We showed that children with two copies of the lower-risk FTO alleles ate less than those with one or two higher-risk alleles. We conclude that the T allele is protective against overeating by promoting responsiveness to internal signals of satiety.


FTO, diet, children, appetite, eating behaviour



The discovery that common variants in the FTO gene are consistently associated with higher body weight1, 2, 3, 4, 5 has turned attention to the gene's functional effects. FTO expression in the brain is greatest in areas of the hypothalamus associated with feeding,6 and also varies with acute food deprivation,7 suggesting it may act through effects on appetite or satiety. This is consistent with observations of the phenotypic abnormality in many monogenic obesity disorders, where abnormalities of appetite are common.8

A recent report from a sample of adult men who had completed 7-day food diaries and fitness assessments9 found differences in energy intake by FTO genotype but no differences in fitness, which was consistent with FTO's effects being related to appetite rather than energy expenditure. Further evidence for an association between FTO and appetite has been found in a study on children, which used a standardized questionnaire measure of appetite completed by the mother. Children with two higher risk FTO alleles scored significantly lower on the satiety responsiveness scale (which includes items such as ‘my child cannot eat a meal if he/she has had a snack just before’), and the association was independent of body mass index (BMI).10

The primary aim of this study was to test the hypothesis that higher risk FTO alleles would be associated with greater food intake in a standardized eating behaviour task, reflecting sensitivity to satiety, in a sample of 4-year-old children. This methodology avoids the problems of validity associated with self-reported or parent-reported food intake. Testing children at an early age also limits the influence of self-consciousness about weight and eating. We measured food intake using the ‘eating in the absence of hunger’ (EAH) paradigm, in which children are offered an array of palatable foods after having eaten to satiety, with observed food intake as the outcome.11 EAH has been shown to be associated with obesity in paediatric samples.11, 12 A secondary aim was to examine associations between FTO genotype and parents' ratings of their childrens' habitual physical activity.


Participants and methods

Participants were from 214 families from the Twins Early Development Study (TEDS) who had taken part in an intensive study of eating behaviour and weight.13 Families with overweight/obese parents were oversampled. This study used data on only one child per family; selected at random. Children and mothers were weighed and measured at the home visit. Childrens' adiposity was indexed with BMI s.d. scores using UK 1990 reference values.14 Overweight and obesity were defined using International Obesity Task Force (IOTF) criteria. FTO genotype status was available for 133 children from this sample, of whom 131 had completed the EAH task and 127 had parental activity ratings.

The FTO single nucleotide polymorphism rs9939609 was genotyped using a TaqMan assay that incorporates minor groove binding probe technology for allelic discrimination. The call rate was 98% and the single nucleotide polymorphism was in the Hardy–Weinberg equilibrium (P=0.729).

The EAH task was carried out within an hour of the children having had a meal served by their mother at home. The child was given a plate containing three varieties of biscuits (two sweet and one savoury) and invited to eat as much as they liked for 10min. The plate of biscuits was weighed before and after consumption to assess the amount consumed. A children's video (not involving food or eating) was shown to provide an alternative activity. The child was observed unobtrusively by the researcher to ensure that food was not removed.

During the home visit, mothers completed three measures designed to give a subjective evaluation of their child's level of physical activity. The child's enjoyment of a selection of active pastimes (for example, bike riding, playing ball and climbing) was rated using 5-point scales anchored with ‘like’ and ‘dislike’, from which a mean enjoyment score was calculated (adapted from Epstein et al.15). Mothers also rated the child's activity level ‘compared to other children of the same age’ on a 5-point scale from ‘much less active’ to ‘much more active’. Finally, a score for ‘fidgetiness’ was obtained using four items from a standardized measure of hyperactivity16 (‘my child runs about and climbs on things excessively’; ‘my child has difficulty sitting still and fidgets excessively’; ‘my child has difficulty staying seated’ and ‘my child is always on the go and acts as if driven by a motor’) rated on a 5-point scale from ‘never’ to ‘always’.

We certify that all applicable institutional and governmental regulations concerning the ethical use of human volunteers were followed during this research. The University College London Committee for the Ethics of non-NHS Human Research granted approval for the study.

Associations between genotype and behaviour were analyzed using analysis of variance. We also carried out analysis of covariance including BMI s.d. scores to test whether associations between FTO and eating or activity were independent of adiposity. Pairwise differences were assessed using least significant difference tests. Linear regression and bootstrapping analyses were used to infer the best-fitting genetic model for the data on the basis of the R2 values. Analyses were carried out using SPSS 14.0 and Intercooled Stata 9.2 for Windows.



Sample characteristics are shown in Table 1. The average age of the children at the time of testing was 4.4 years. There were more girls than boys in the sample (60 vs 40%) but this was the same over the three genotypic groups. The rs9939609 genotype distribution (AA (homogenous for A allele)=18.0%, AT (heterogeneous; T and A alleles)=49.6%, TT (homogenous for T allele)=32.3%) was similar to the full sample.10

Maternal BMI did not differ significantly across the three groups. Anthropometric characteristics of the full TEDS sample at age 10 years, and significant associations between adiposity and FTO at 7 and 10 years but not at 4 years, have been reported earlier.17, 18 In this subsample, genotypic differences in childrens' adiposity (BMI s.d. scores) were not significant, although as expected, the trend was for the highest BMI s.d. score in the AA genotype group. The percentage of children who met IOTF criteria for obesity was also in the expected direction (8% in the AA group, 3% in AT and 2% in TT) but did not reach statistical significance.

Food intake (in grams) in the EAH test (see Table 2 and Figure 1) differed significantly across the three genotype groups (F (2, 128)=3.540, P=0.032), with 25% higher intake in the AA group (39.9g) than in the TT group (30.0g), whereas the AT group (37.9g) was in between. Pairwise differences tests showed that the TT group ate significantly less than AT or AA groups (P=0.023 and P=0.027, respectively), but AT and AA groups were not significantly different from one another (P=0.628). The main effect of genotype remained significant after controlling for BMI s.d. score (P=0.036).

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Means (and standard errors) for food intake in eating in the absence of hunger (EAH) task by genotype group.

Full figure and legend (8K)

Linear regression analyses showed significant results for both additive and dominant genetic models (P=0.014 and P=0.010, respectively), but not for the recessive model (P=0.195). The genetic effect in the additive and dominant models remained significant after controlling for BMI s.d. score. Although R2 values indicated that the dominant genetic model explained a greater proportion of the variance (R2=0.051) than the additive model (R2=0.046), bootstrapping analysis showed that the 95% confidence intervals for the two models overlapped (0.006–0.125 and 0.003–0.127 for dominant and additive models, respectively), giving no evidence that one model was a better fit for the data than the other.

There were no significant differences between genotype groups in the mother's ratings of the child's enjoyment of physical activity, their habitual level of physical activity or their fidgetiness.



In this sample of children who took part in a home-based study of eating behaviour at the age of 4 years, higher-risk FTO alleles were associated with significantly higher consumption of a highly palatable food and the effect was independent of BMI. Eating behaviour was assessed using the EAH paradigm, in which palatable food is offered to satiated individuals (similar to being offered chocolates after a meal). EAH is usually construed as a measure of (lack of) responsiveness to internal satiety signals, with higher intake indicating lower responsiveness to satiety, although it is not possible to rule out differential responsiveness to the palatable food cues as contributing to the amount consumed. The findings of this study are consistent with results obtained using a psychometric measure of appetite in a larger sample at the age of 10 years,10 which found significant differences in satiety responsiveness related to FTO genotype status.

The pattern of means across the three genotype groups, the pairwise differences and the R2 values indicated that the best-fitting genetic model for the effect of FTO on EAH may be a dominance model rather than a linear dosage effect for the A allele, in keeping with Speakman et al.'s8 conclusion that a dominance model was a better fit for their data.9 This was the same pattern as shown in our earlier psychometric study.10 However, bootstrapping analysis showed that the 95% confidence intervals for possible R2 values for the two genetic models were too wide to draw firm conclusions. Given that the width of confidence intervals using bootstrapping is a function of sample size, a larger sample is required to draw clear conclusions about the most appropriate model.

We found no evidence for differences by genotype in the enjoyment of active pastimes, habitual physical activity or fidgetiness. Similar negative results were reported in a study of associations between FTO and levels of fitness in Scottish adults,9 and in a study assessing both fitness and metabolic rate in Swedish men.19 None of the studies indicated even a trend towards differences, but they were all under-powered to detect small differences in activity, and further research is needed with larger samples before ruling out an effect on energy expenditure. There is also a possibility of interactions between genotype and habitual activity,20 which could be tested in a larger sample with better measures of physical activity.

There were limitations to this study other than the relatively small sample size. The EAH task is a single instance of food intake, in a very specific situation (just after a meal) but this context was selected to reflect differences related to satiety responsiveness. We did not have control over childrens' intake in the prior meal, so there could be differences in prior consumption, but it seems unlikely that the AA children had systematically smaller prior meals. The home context limited experimental control over the situation, but at the same time increased ecological validity. The activity measures were neither objective nor sophisticated and were based entirely on maternal reports. It was also unfortunate that we did not have any other data on adiposity except BMI, although we have shown data on associations between FTO and both waist circumference and BMI in a much larger sample of twins from TEDS at the age of 10 years,10 as well as associations between eating behaviour and waist circumference at 10 years.21

In summary, these results, in combination with our earlier findings from parental reports of childrens' eating behaviour, support the idea that FTO could influence responsiveness to satiety signals. In environments with multiple opportunities to eat highly palatable foods, those with higher satiety responsiveness are likely to be relatively protected from overeating, whereas individuals with a less responsive satiety system will be more likely to overeat and will therefore be at higher risk for weight gain.




The authors state no conflict of interest.



  1. 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. | Article | PubMed | ISI | ChemPort |
  2. Dina C, Meyre D, Gallina S, Durand E, Korner A, Jacobson P et al. Variation in FTO contributes to childhood obesity and severe adult obesity. Nat Genet 2007; 39: 724–726. | Article | PubMed | ISI | ChemPort |
  3. Chang YC, Liu PH, Lee WJ, Chang TJ, Jiang YD, Li HY et al. Common variation in the FTO gene confers risk of obesity and modulates body mass index in the Chinese population. Diabetes 2008; 57: 2245–2252. | Article | PubMed | ChemPort |
  4. Cha SW, Choi SM, Kim KS, Park BL, Kim JR, Kim JY et al. Replication of genetic effects of FTO polymorphisms on BMI in a Korean population. Obesity (Silver Spring) 2008; 16: 2187–2189. | Article | PubMed | ChemPort |
  5. Loos RJ, Lindgren CM, Li S, Wheeler E, Zhao JH, Prokopenko I et al. Common variants near MC4R are associated with fat mass, weight and risk of obesity. Nat Genet 2008; 40: 768–775. | Article | PubMed | ChemPort |
  6. Stratigopoulos G, Padilla SL, LeDuc CA, Watson E, Hattersley AT, McCarthy MI et al. Regulation of Fto/Ftm gene expression in mice and humans. Am J Physiol Regul Integr Comp Physiol 2008; 294: R1185–R1196. | PubMed | ChemPort |
  7. 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. | Article | PubMed | ChemPort |
  8. Farooqi IS, O'Rahilly S. Monogenic obesity in humans. Annu Rev Med 2005; 56: 443–458. | Article | PubMed | ISI | ChemPort |
  9. Speakman JR, Rance KA, Johnstone AM. Polymorphisms of the FTO gene are associated with variation in energy intake, but not energy expenditure. Obesity (Silver Spring) 2008; 16: 1961–1965. | Article | PubMed | ChemPort |
  10. Wardle J, Carnell S, Haworth C, Farooqi S, O'Rahilly S, Plomin R. Obesity-associated variation in FTO is associated with diminished satiety. J Clin Endocrinol Metab 2008; 93: 3640–3643. | Article | PubMed | ChemPort |
  11. Fisher JO, Birch LL. Eating in the absence of hunger and overweight in girls from 5 to 7 y of age. Am J Clin Nutr 2002; 76: 226–231. | PubMed | ISI | ChemPort |
  12. Faith MS, Berkowitz RI, Stallings VA, Kerns J, Storey M, Stunkard AJ. Eating in the absence of hunger: a genetic marker for childhood obesity in prepubertal boys? Obesity (Silver Spring) 2006; 14: 131–138. | Article | PubMed |
  13. 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. | Article | PubMed | ChemPort |
  14. Cole TJ, Freeman JV, Preece MA. Body mass index reference curves for the UK, 1990. Arch Dis Child 1995; 73: 25–29. | PubMed | ISI | ChemPort |
  15. Epstein LH, Valoski A, Wing RR, Perkins KA, Fernstrom M, Marks B et al. Perception of eating and exercise in children as a function of child and parent weight status. Appetite 1989; 12: 105–118. | Article | PubMed | ChemPort |
  16. Conners CK. A teacher rating scale for use in drug studies with children. Am J Psychiatry 1969; 126: 884–888. | PubMed | ChemPort |
  17. 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. | PubMed | ChemPort |
  18. Haworth C, Carnell S, Meaburn E, Davis O, Plomin R, Wardle J. Increasing heritability of body mass index and stronger associations with the FTO gene over childhood. Obesity (Silver Spring) 2008 (in press).
  19. Berentzen T, Kring SI, Holst C, Zimmermann E, Jess T, Hansen T et al. Lack of association of fatness-related FTO gene variants with energy expenditure or physical activity. J Clin Endocrinol Metab 2008; 93: 2904–2908. | Article | PubMed | ChemPort |
  20. Andreasen CH, Stender-Petersen KL, Mogensen MS, Torekov SS, Wegner L, Andersen G et al. Low physical activity accentuates the effect of the FTO rs9939609 polymorphism on body fat accumulation. Diabetes 2008; 57: 95–101. | Article | PubMed | ChemPort |
  21. 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. | PubMed | ChemPort |


We are very grateful to all the families who made this research possible. The research was funded by a grant from the Biotechnology and Biological Sciences Research Council (D19086).

Supplementary Information accompanies the paper on International Journal of Obesity website (



These links to content published by NPG are automatically generated


A blast from the past: clearance of apoptotic cells regulates immune responses

Nature Reviews Immunology Review (01 Dec 2002)

Extra navigation