BACKGROUNDThe aim of this study was to explore the causal association between immune cells and VaD based on a two-sample bidirectional Mendelian randomization study.METHODSBidirectional two-sample MR analyses based on pooled datasets from publicly available genome-wide association studies were performed using inverse variance weighted (IVW), weighted median (WE), and MR-Egger regressions to evaluate the causal relationships between immune cells and vascular dementia. Heterogeneity was assessed using Cochran's Q statistic. The reliability of the MR analysis results was verified by using the MR-PRESSO method for outlier detection, the MR-Egger method for horizontal multivariate analysis, and the leave-one-out method for sensitivity analysis.RESULTSSpecifically, 27 immunophenotypes were associated with VaD pathogenesis, including Sw mem %lymphocyte (P = 0.043), CD38 on CD20- (P = 0.039), CD11c+ monocyte AC (P = 0.024), DC AC (P = 0.002), CCR2 on CD62L+ myeloid DC (P = 0.039), Resting Treg %CD4 (P = 0.042), Activated & resting Treg %CD4+ (P = 0.038), CD28+ CD45RA- CD8br %CD8br (P = 0.047), NK %CD3- lymphocyte (P = 0.042), CD45 on B cell (P = 0.029), FSC-A on NKT (P = 0.033), CD45 on CD33br HLA DR+ CD14- (P = 0.039) were significantly correlated with increased VaD risk. Additionally, four immune phenotypes, namely, CD19 on CD20-, Resting Treg %CD4, Activated & resting Treg %CD4+, and CD11c+ monocyte AC, showed bidirectional effects on VaD.CONCLUSIONSMR analysis revealed potential causal relationships between certain immune cells and VaD. Our preliminary exploration through immune cell infiltration analysis highlights the significant value of immune cells in VaD. Therefore, this study may provide a new perspective for the prevention and treatment of VaD.