2.4.2 Heterogeneity and pleiotropy analyses
To evaluate heterogeneity, we used Cochran heterogeneity test. If there is heterogeneity, random-effects IVW models are applied; otherwise, the fixed-effect IVW model is applied. To evaluate pleiotropy, we used MR-Egger regression (assessing whether the MR-Egger intercept was different from zero). If causality was estimated in the same direction among the various MR methods, the conclusions will be considered as relatively robust causal inferences. Analyses were carried out using MendelianRandomization version 0.9.0, TwoSampleMR version 0.5.7 and MVMR version 0.4 in the R environment version 3.6.3 (The R Foundation for Statistical Computing, Vienna, Austria).
Results