Abstract: Objective To explore the causal association between birth weight and childhood obesity by using the GWAS data and the Mendelian randomization (MR) method.Methods Based on the genetic data published by the Early Growth Genetics Consortium (EGG) (data acquisition time is October 20, 2018), the SNP with statistically significant association with birth weight was screened by MR-base platform. As instrumental variables (screening parameters P<5×10^-8, linkage disequilibrium r²<0.1), birth weight and childhood obesity were determined by inverse variance weighted analysis (IVW), weighted median method and MR-Egger method, respectively (definition of childhood obesity: BMI≥95%). Forest maps and scatter plots were also applied.Results A total of 33 SNPs related to birth weight were screened. A total of 5,530 obese children and 8,318 normal children were included. IVW analysis: OR=1.79, 95% CI: 1.29~2.47, P=4.24×10^-4; weighted median method: OR=1.30, 95% CI: 0.81~2.08, P=0.27; MR-Egger method: OR=1.73, 95% CI: 0.58~5.20, P=0.36. The result of IVW in the forest map showed that the genetic susceptibility of childhood obesity was related to the birth weight level. rs7964361, rs11765649, rs3780573 and rs1351394 showed that birth weight and childhood obesity were statistically significant. The scatter plot showed that the causal correlation estimates for the IVW method, the MR-Egger method, and the weighted median method were similar.Conclusion IVW analysis showed that there was a strong correlation between birth weight and childhood obesity, but the other two statistical methods found that the difference was not statistically significant, so the causal association remained to be supported by further evidence.

Key words: Birth weight, Childhood obesity, Mendeline randomization