To ensure the reliability of the identified causal effect of lipids and apolipoproteins on GDM, we carried out a thorough set of sensitivity analyses. Cochran’s Q statistic was utilised to assess potential heterogeneity within the data [22 (link)]. The MR-Egger intercept analysis was employed to investigate horizontal pleiotropy [23 (link)]. We also conducted a Leave-one-out analysis to examine if any single SNP substantially affected the outcomes by systematically removing SNPs individually. Additionally, reverse MR analyses were performed to explore the potential reverse causal link between lipids and apolipoproteins (as seen in the forward MR analysis) and GDM.
For multivariable MR analysis, we applied two models to further understand the connection between lipid-related traits and GDM risk. In Model 1, five lipid-related traits (apoA-I, apoB, LDL cholesterol, HDL cholesterol, and triglycerides) were included in multivariable analysis.
In Model 2, we included BMI for analysis, along with the three traits that showed positive associations in univariable analysis: apoA-I, HDL cholesterol, and triglycerides.
All analyses were performed using R (version 4.2.0) and RStudio, employing the R packages “TwoSampleMR” and “MR-PRESSO”.
For multivariable MR analysis, we applied two models to further understand the connection between lipid-related traits and GDM risk. In Model 1, five lipid-related traits (apoA-I, apoB, LDL cholesterol, HDL cholesterol, and triglycerides) were included in multivariable analysis.
In Model 2, we included BMI for analysis, along with the three traits that showed positive associations in univariable analysis: apoA-I, HDL cholesterol, and triglycerides.
All analyses were performed using R (version 4.2.0) and RStudio, employing the R packages “TwoSampleMR” and “MR-PRESSO”.
Free full text: Click here
