Obesity, genetic risk, and environmentBMJ 2014; 348 doi: http://dx.doi.org/10.1136/bmj.g1900 (Published 19 March 2014) Cite this as: BMJ 2014;348:g1900
- Alexandra I F Blakemore, professor of human molecular genetics,
- Jessica L Buxton, research associate
- 1Section of Investigative Medicine, Division of Diabetes, Endocrinology, and Metabolism, Faculty of Medicine, Imperial College, London W12 0NN, UK
- Correspondence to:
The alarming global rise in prevalence of obesity is caused by unhealthy obesogenic environments. In westernised societies we are all exposed to calorie dense food, sedentary lives, stress, and sleep deficit. Some people seem relatively insensitive to these environmental pressures, while others are severely affected and become obese.
In a linked paper, Qi and colleagues (doi:10.1136/bmj.g1610) examined the interaction between common genetic variants associated with body mass index (BMI) and frequency of fried food consumption in over 37 000 people.1 The large study size was necessary to obtain adequate statistical power because the individual effects of these variants (single nucleotide polymorphisms, SNPs) on BMI are relatively subtle.2 The authors combined the weighted individual effects of 32 SNPs into a single “genetic risk score” for each participant.
As expected, participants who ate fried food more often tended to have higher BMI, and, independently, those with the highest genetic risk score also had higher BMI. The novel finding in this study is the observed interaction between genetic risk and fried food consumption: people in the highest risk groups for both had the highest BMI overall. Eating fried food more than four times a week had twice the effect on BMI for those in the highest third of genetic risk score as those in the lowest third.
This work provides formal proof of interaction between a combined genetic risk score and environment in obesity. Similar studies have already shown interaction between genetic risk scores for obesity and physical activity3 and dietary calcium.4 These results are unlikely to have a direct impact on personal healthcare because, though such genetic risk scores are statistically robust at the population level, they have poor predictive power for any given individual. Similarly, the results of the latest study are unlikely to influence public health advice, as most of us should be eating less fried food.
It would be a great shame, however, to assume that genetics can be ignored in the management of obesity. This widespread misconception arises from confusion of the common SNPs that have relatively subtle effects on BMI (the strongest of which, close to the FTO gene, raises body weight by only a mean of 3 kg even in those who inherit a double dose of the variant) with more dramatic genetic alterations that cause “Mendelian” forms of obesity.
There are, in fact, at least 15 single gene forms of obesity, with new ones reported almost monthly with advances in DNA sequencing technology.5 6 7 8 There are also chromosome microdeletions that cause Mendelian obesity.9 10 These forms of obesity are inherited in the same way as other genetic conditions such as cystic fibrosis or Huntington’s disease. Mutations in just one of the obesity genes (MC4R) cause around one in 20 cases of severe childhood obesity,11 and the total number of obese adults with Mendelian obesity is likely to be substantial.
Where molecular mechanisms are understood, obesity-causing genetic mutations disrupt appetite control systems in the brain, so affected people are unlikely to be able to maintain long term dietary restraint. It might also be unwise to offer them some types of weight loss surgery, such as adjustable gastric banding.
Why is this important? Our current options for management of morbidly obese people are limited, though several new therapeutic approaches are in development. Different types of obesity might require different management. For example, lifestyle interventions are likely to be much more successful in people whose BMI is not too far from the desirable range. The UK’s National Institute for Health and Care Excellence (NICE) already implicitly recognises that lifestyle and medical approaches are unlikely to help those with the most severe forms of obesity. Weight loss surgery is recommended as the first line treatment in people with BMI >50.
As researchers, policy makers, and healthcare professionals, we should adopt an evidence based approach. Unfortunately, clinical logic is not always applied in the consideration of obesity—health professionals commonly recommend the same lifestyle based interventions to those with overweight or mild obesity as to those with more severe problems. We would not adopt this approach with hypertension or type 2 diabetes, where it is well recognised that diet and appropriate physical exercise might be sufficient for mild cases, but more severely affected people need drugs (or even surgery) and a lifelong care plan. There is a danger that issues surrounding blame and “personal responsibility” in obesity might lead to lack of appropriate support for extremely obese people, who often have complex care needs. This can be the case in families with undiagnosed inherited obesity: genetic screening and counselling is rarely offered in those with morbid obesity.
In summary, use of combined genetic risk scores in gene-environment interaction studies allows joint analysis of factors influencing obesity. Inclusion of multiple genetic variants into a single risk score is helpful in terms of power for such investigations. It would be useful if similar studies could be carried out in individuals with Mendelian forms of obesity, who might show even more striking gene-environment interactions. Complex analyses integrating thousands of common variants have the potential to explain yet more of the genetic contribution to traits such as BMI.12 Predictive power at the individual level, however, remains low when only common genetic variants are considered. Integration of rare genetic alterations that cause Mendelian forms of obesity mighty improve this, providing clinically useful predictions for individuals and enabling stratification of patients for appropriate care and treatment.
Cite this as: BMJ 2014;348:g1900
Competing interests: We have read and understood the BMJ Group policy on declaration of interests and declare the following interests: none.
Provenance and peer review: Commissioned; not externally peer reviewed.