Genome-wide association study for osteoarthritis – Authors’ reply

We welcome this opportunity to expand further on two of the points we raised in our paper.

Obesity is a well established risk factor for osteoarthritis.

Large-scale, prospective studies consistently show a strong relation between weight gain or being overweight and future risk of osteoarthritis.

^, 

This finding indicates that obesity is more likely to be on the causal pathway to osteoarthritis rather than the converse, although clearly osteoarthritis-induced inactivity is also expected to have an adverse effect.

Our genetic analyses offer evidence in support of this body of epidemiological data: we find the FTO significantly-associated signal to be attenuated when adjusted for body-mass index.

We therefore maintain our assertion: our data suggest that the FTO gene exerts its effect on osteoarthritis through obesity. Extended detailed investigations are warranted. It remains possible that obesity and osteoarthritis have further shared genetic risk loci,

as shown by overlapping association signals at additional loci, which could be acting through a different mechanism.

Hao Pang and colleagues recapitulate our Discussion point that it would be of great interest to assess whether the identified loci have global relevance to osteoarthritis risk through similar studies of non-European cohorts with osteoarthritis. Our focus on European-ancestry populations for this study has helped minimise allelic and locus heterogeneity and has thus guarded against diminishing power caused by potential differences inherent in the genomic architecture of disease across diverse populations. This approach led to the identification of eight novel loci.

It will undoubtedly be useful for follow-up studies to investigate the association between these variants and osteoarthritis in populations of non-European descent. We postulate that it will be scientifically more valuable to combine genome-wide case-control data for osteoarthritis across different populations to empower novel locus discovery and transethnic fine-mapping of established loci. Differences in levels of correlation between variants across populations can narrow down association intervals, helping identify the causal loci. Collaborative efforts will be required to enable these studies and thereby help progress our understanding of the biological underpinnings of disease.

We declare that we have no conflicts of interest.