Supplementary Material for: Causal associations of gestational and perinatal factors with the risk of childhood obesity: A Mendelian Randomization study

Introduction: This study was to explore the causal association between gestational and perinatal factors and the risk of childhood obesity based on two-sample Mendelian Randomization (MR).
Methods: The single-nucleotide polymorphisms (SNPs) associated with exposure were selected. MR Egger regression was used for the evaluation of the horizontal pleiotropy in SNPs. F-statistic reflects the strength of instrumental variables. Cochran’s Q statistic was used to investigate the heterogeneity of SNPs. The primary MR method exploring the causal association between gestational and perinatal factors and the risk of childhood obesity was inverse variance weighted (IVW). Leave-one-out analysis was employed to identify whether the MR results were affected by single SNPs. Odds ratio (OR) and 95% confidence interval (CI) were imputed.
Results: IVW indicated that birth weight was a risk factor for childhood body mass index (BMI) (OR=1.347, 95%CI: 1.212-1.498) and childhood obesity (OR=1.793, 95%CI: 1.300-2.473). Increased risk of childhood obesity was identified in mother with gestational diabetes (OR=1.059, 95%CI: 1.006-1.116). Gestational diabetes was causally related to elevated childhood BMI. Gestational diabetes was causally related to elevated risk of childhood obesity (OR=1.249, 95%CI: 1.044-1.495).
Conclusions: Birth weight and gestational diabetes was causally related to childhood BMI and childhood obesity.

引言：本研究基于两样本孟德尔随机化（two-sample Mendelian Randomization, MR）方法，探讨妊娠与围产期因素和儿童肥胖风险之间的因果关联。
方法：筛选与暴露因素相关的单核苷酸多态性（single-nucleotide polymorphisms, SNPs）。采用MR埃格回归（MR Egger regression）评估SNPs的水平多效性。F统计量用于反映工具变量的强度。使用科克伦Q统计量（Cochran’s Q statistic）检验SNPs的异质性。本研究探讨妊娠与围产期因素和儿童肥胖风险因果关联的主要MR方法为逆方差加权（inverse variance weighted, IVW）。采用留一法分析以验证MR结果是否受单个SNPs的影响。计算得到比值比（Odds ratio, OR）与95%置信区间（95% confidence interval, CI）。
结果：逆方差加权分析显示，出生体重是儿童体质指数（body mass index, BMI）升高（OR=1.347，95%CI：1.212~1.498）与儿童肥胖（OR=1.793，95%CI：1.300~2.473）的危险因素。妊娠期糖尿病母亲所生子女的儿童肥胖风险升高（OR=1.059，95%CI：1.006~1.116）。妊娠期糖尿病与儿童体质指数升高存在因果关联。妊娠期糖尿病与儿童肥胖风险升高存在因果关联（OR=1.249，95%CI：1.044~1.495）。
结论：出生体重与妊娠期糖尿病均与儿童体质指数及儿童肥胖存在因果关联。