2.4 Statistical Methods
We conducted the primary MR analysis using the fixed effects inverse-variance weighted (IVW) method to assess the causal relationships between herpes viruses and leukocytoclastic angiitis. This method provides robust causal estimates, although it is somewhat sensitive to pleiotropy. In addition, we performed supplementary analyses using the weighted median method and MR-Egger regression(14, 15). In the main analyses, we calculated odds ratios (ORs) and 95% confidence intervals (CIs) for the IVs.
The heterogeneity was evaluated in the odds ratio using Cochran’s Q test and visually inspected a funnel plot to assess symmetry, which indicates the absence of systematic bias (16). We also conducted an MR-Egger intercept analysis to assess the presence of directional pleiotropy. Furthermore, we performed a leave-one-out sensitivity test to evaluate the impact of individual SNPs on the results, mitigating potential bias (14, 16, 17).
All statistical analyses were carried out using R Studio (v.4.3.1), with the MR analysis conducted using the ‘TwoSampleMR’ package(18). (Figure. 1)