Elevated BMI vs Severe Obesity: Different Genetic Diseases?

Ali A. Torkamani, PhD


June 11, 2013

Genome-wide Meta-analysis Identifies 11 New Loci for Anthropometric Traits and Provides Insights Into Genetic Architecture

Berndt SI, Gustafsson S, Mägi R, et al
Nat Genet. 2013;45:501-512

Genome-wide SNP and CNV Analysis Identifies Common and Low-frequency Variants Associated With Severe Early-Onset Obesity

Wheeler E, Huang N, Bochukova EG, et al
Nat Genet. 2013;45:513-517


Obesity is a largely heritable trait -- twin studies estimate that 40%-70% of variability in body mass index (BMI) is explained by genetics.[1] However, although some genome-wide association studies have identified more than 35 genetic loci associated with BMI, they explain less than 2% of interindividual variability in BMI.[2] Also, a few specific loci, such as PCSK1[3] or rare copy-number variants,[4] have been associated with severe, usually early-onset, obesity, but they have not been convincingly associated with BMI in general. These observations raise the question of whether genetic findings in severe forms of obesity are relevant to the population at large.

A pair of recent publications attempted to answer this question by comparing genetic associations in population extremes vs the general population. In one study, Wheeler and colleagues and a team of British investigators compared genetic changes associated with severe early-onset obesity vs those associated with variability in BMI; through a variety of analyses, they concluded that there is incomplete overlap between the genetic loci influencing the 2 obesity categories. In contrast, in a second, multinational study, Berndt and colleagues compared genetic changes associated with extreme clinical classes of obesity vs those associated with variability in BMI and concluded there is near-complete overlap in genetic variants across all subclassifications of obesity and BMI.


These 2 studies produced apparently conflicting results as to whether genetic findings from population extremes of obesity are relevant to the population at large. Is obesity a single continuing spectrum, or is severe obesity a genetically distinct phenomenon?

The answer to this question probably lies in the inclusion criteria used to define severe obesity. The way in which extreme obesity was defined in these 2 studies is materially different and more than likely drives the difference in their ultimate conclusions.

In Wheeler and colleagues' study, severe obesity was defined as highly elevated BMI in the young, whereas Berndt and colleagues defined it using a more inclusive cutoff at the extreme ends of the population distribution of BMI. This suggests that the specific genetic architecture underlying early-onset vs extreme obesity is quite different, even though the genes influenced by those genetic changes are probably the same.

That is, just as different genetic changes in the same TP53 gene predispose patients to either Li-Fraumeni syndrome or to a modestly elevated risk for cancer, so too different changes in the same genes influence the development of early-onset vs general yet severe obesity. Indeed, it is likely that the genetic architecture of early-onset severe obesity mimics a severe disease syndrome, whereas "extreme" obesity (defined as the top 5th percentile in BMI) simply comprises individuals enriched for obesity variants that influence BMI in the population at large.

Ultimately, then, it is not surprising that the genes -- but not the specific involved variants -- would be similar, if not identical, across these 2 groups. What remains important, then, is that findings from these studies inform the biological processes underlying both forms of obesity and thus point to potential avenues for therapeutic intervention.