In the present study, we investigated associations of 11 BMI-associated loci with risk of central obesity in the Chinese children. The results indicated that four SNPs (FTO rs9939609, MC4R rs17782313, GNPDA2 rs10938397, BDNF rs6265) and GRS calculated from these 4 SNPs significantly predicted the risk of central obesity. However, we failed to observe the associations of seven other BMI-associated loci with central obesity.
Frayling et al. firstly reported that FTO
gene was significantly associated with generalized obesity in European adults and children 
. Since then, many studies have examined the association among different ethnic populations 
. Although contradictory results were reported, three recent meta-analyses have confirmed the significant association of FTO
gene with BMI and generalized obesity 
. Two other obesity associated genes, MC4R
were also identified by GWAS in Europeans and replicated in other studies 
. A 12-year longitudinal study showed that FTO
rs8050136 and GNPDA2
rs10938397 SNPs, rather than MC4R
rs17782313, predicted persistent central obesity in the Chinese adults 
. However, it should be noted that only 354 subjects with central obesity were recruited in the abovementioned study and hence was underpowered to detect the association. Another study from Japan also demonstrated significant association between FTO
rs1558902 SNP and central obesity in adults 
. However, to our knowledge, there is no published study on the association analysis of these BMI-associated loci with central obesity in pediatric population. Our study presents the new evidence that FTO
rs10938397 and BDNF
rs6265 are statistically significantly associated with risk of central obesity in the Chinese children.
The mechanisms of how BMI associated SNPs influence central obesity are unclear. FTO and MC4R proteins are highly expressed in hypothalamus, which regulates the energy balance 
. Many studies have indicated that FTO
SNPs influence energy-dense food intake rather than regulation of energy expenditure 
. Other studies have suggested that FTO
SNPs are associated with metabolic traits (higher fasting insulin, glucose, triglycerides and lower HDL cholesterol) that are mediated through BMI. Future studies focusing on these phenotypes might help in clarifying the mechanisms through which these loci and central obesity are related.
The joint effect of four strongly associated SNPs (FTO rs9939609, MC4R rs17782313, GNPDA2 rs10938397, BDNF rs6265) on the risk of central obesity was further investigated. It is worth mentioning that the age- and sex-adjusted OR of GRS for central obesity was only 1.25, which points to their limited predictive effect on risk of central obesity in the Chinese children. However, the genetic information may be useful to identify high-risk children who may need early interventions such as lifestyle modifications and individualized management strategies, in order to reduce the incidence of obesity related diseases.
Our study has two strengths; firstly, we achieved sufficient power (>0.75) for each SNP and secondly, the findings are novel for several SNPs since very few studies have focused on the associations between BMI associated SNPs and central obesity. However, few limitations should also be noted. First, the case-control design does not allow interpreting causality of the associations. Second, recent GWASs have identified a large number of SNPs to be associated with BMI/generalized obesity 
and hence, future studies would need to take more SNPs into account, while conducting such association analyses. Finally, the findings of our study may need to be replicated in other pediatric cohort.
In conclusion, our study confirmed the significant association of four SNPs (FTO rs9939609, MC4R rs17782313, GNPDA2 rs10938397, BDNF rs6265) with risk of central obesity in the Chinese children. Large-scale longitudinal studies with consideration of gene-gene and gene-environment interactions should be conducted to investigate the causality of these associations in future.