Are the causes of obesity primarily environmental? NoBMJ 2012; 345 doi: http://dx.doi.org/10.1136/bmj.e5844 (Published 11 September 2012) Cite this as: BMJ 2012;345:e5844
- Timothy M Frayling, professor of human genetics
- 1Genetics of Complex Traits, Peninsula College of Medicine and Dentistry, University of Exeter, Exeter EX1 2LU, UK
Genetic variation has not changed appreciably in the past 50 years and therefore cannot explain the secular increases in average body mass index observed over the past few decades. But changes in the environment (decreased need for physical activity and greater availability of cheap food) mean we are all at increased risk of obesity compared with our parents and grandparents. So why do many people remain slim, while others gain weight?
Genetic variation influences our appetites, metabolism, and tolerance of physical activity. This creates a strong genetic component to variation in body mass index in today’s environment. An analogy can be made with smoking—if everyone inhaled the same amount of cigarette smoke every day, the strongest risk factor for lung cancer would be genetic susceptibility to the adverse effects of cigarette smoke (G Davey Smith, personal communication).
Size of genetic effect
Twin and adoption studies show consistently that variation in body mass index has a strong genetic component. One study assessed the heritability of body mass index in over 20 000 young adult twin pairs from eight European countries,1 with data collected from 1963 to 2002 (although mostly from 1980 onwards). The correlation of body mass index between identical twins in the eight countries ranged from 0.65 to 0.83 and was consistently stronger than that between non-identical same sex twins (correlation 0.31 to 0.58). The estimated genetic effects, correcting for age and sex differences, were 60-70%. In a recent systematic review of five adoption studies with several hundred parent-biological child and parent-adoptee comparisons, children’s body mass index was consistently more strongly correlated with that of their biological parents than of their adoptive parents.2 Intrauterine “programming” did not account for the differences because the correlations were similar for father-biological child pairs and mother-biological child pairs.
Strong genetic effects remain even in contemporary environments. In a study of over 5000 twin pairs born in the UK during 1994-97, the correlation of body mass index at age 8-11 years between identical twins (0.86, 95% confidence interval 0.85 to 0.87) was stronger than that between non-identical same sex twins (0.51, 0.47 to 0.53).3 Genetic effects were estimated as 77%. Another study of over 2000 very young twin pairs found that appetite, estimated from parental questionnaires, had a genetic component.4 Correlations between appetite measures in one twin and weight of the other twin were stronger in identical than non-identical twin pairs.
Twin studies may overestimate genetic effects because there is less variation in the environment between children raised in the same household and parents may treat identical twins differently to same sex non-identical twins. But these factors are unlikely to explain the large differences in correlations of body mass index between identical and non-identical twins. In theory, twin and adoption studies could be confounded by epigenetic effects. For example, differences in placentation and response to the maternal intrauterine environment could influence whether genes are permanently switched on or off by processes such as DNA methylation. However, there is little evidence for this causal pathway in humans.
Stronger effects in sedentary people
Genetic factors may have stronger effects in more sedentary individuals or those in more obesogenic environments. Genome-wide studies have identified DNA variation in 32 regions of the human genome associated with body mass index, with a DNA variant in the FTO gene having the strongest association.5 People carrying two copies of the allele associated with obesity are, on average, 0.5 kg/m2 heavier than those carrying two copies of the protective version. Recently, a study of over 200 000 people showed that the FTO variant had a stronger effect in people who were sedentary than in those who were physically active.6 Studies in humans and mice suggest that the FTO gene affects appetite control,7 8 and in inactive people there may be greater scope for genetic factors to influence body mass index through appetite control. Although the DNA variations explain only a small percentage of the variation in body mass index, they provide proof of principle that genetic factors influence it over environmental effects.
Education may not be as important as we think
There is a large amount of evidence that education or other measures of welfare are associated with body mass index, but studies of physical activity in schoolchildren suggest that education may not be as important as we hope. A recent report analysed data from 12 studies consisting of over 8000 children randomised to increased physical activity or normal activity at school. Children were followed up for a median of 18 months. The results showed no evidence that physical activity interventions influenced body mass index.9 This result is consistent with one longitudinal study of 200 children that used actigraph accelerometers to objectively measure physical activity for seven consecutive days, once a year over three years. Inactivity in children preceded increases in percentage body fat, but increased body fat percentage did not precede reduced physical activity.10 In the same study, children from wealthier households were no more active than those from poorer households, despite more out of school structured activity.11
In conclusion, genetic factors influence substantially where you are on the body mass index scale in a given population at a given time. Evidence is accumulating that these genetic factors may operate largely through appetite control. If true, plans based on changing our environment, such as banning the sale of supersized sugary drinks,12 will make genetic factors less important and be more successful than plans to increase awareness through education.
Cite this as: BMJ 2012;345:e5844
Competing interests: The author has completed the ICMJE unified disclosure form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declares no support from any organisation for the submitted work; no financial relationships with any organisation that might have an interest in the submitted work in the previous three years; and no other relationships or activities that could appear to have influenced the submitted work.
Provenance and peer review: Commissioned; not externally peer reviewed