BACKGROUNDEffective and specific biomarkers are warranted for the management of vascular dementia. We aimed to systematically screen the human blood metabolome to identify potential mediators of vascular dementia via a two-sample Mendelian randomization (MR) design.METHODSWe selected 93 unique blood metabolites from 3 metabolome genome-wide association studies (GWASs) with a total of 147,827 participants of European ancestry. Summary statistics for vascular dementia originated from a European-descent GWAS dataset released by the FinnGen Study, involving 859 cases and 211,300 controls. We applied the inverse-variance weighted MR method in the main analysis to examine the causal roles of blood metabolites in vascular dementia, followed by several sensitivity analyses for robustness validation.RESULTSGenetically determined glycoproteins (OR per 1-SD increase, 0.75; 95 % CI, 0.68-0.83, P = 1.08 × 10-8) and O-methylascorbate (OR per 1-SD increase, 0.08; 95 % CI, 0.02-0.32; P = 3.74 × 10-4) levels had negative associations with the risk of vascular dementia, whereas genetically determined total cholesterol (OR per 1-SD increase, 1.77; 95 % CI, 1.32-2.38; P = 1.39 × 10-4) and low-density lipoprotein (LDL) cholesterol (OR per 1-SD increase, 1.94; 95 % CI, 1.48-2.55; P = 1.61 × 10-6) levels had positive associations with the risk of vascular dementia. MR-Egger regression suggested no directional pleiotropy for the identified associations, and sensitivity analyses with different MR models further confirmed these findings.CONCLUSIONGlycoproteins, O-methylascorbate, total cholesterol, and LDL cholesterol might be promising blood markers of vascular dementia, which may provide novel insights into the prevention of vascular dementia. Further studies are warranted to replicate our findings and elucidate the potential mechanistic pathways.