BACKGROUND: Extensive perivascular spaces (PVS) burden has been reported to be associated with neurodegenerative diseases and brain structure
however, the causal effects has not been determined yet. Therefore, this study aimed to investigate the causal effect of extensive PVS burden on neurodegenerative diseases and brain structure through Mendelian randomization (MR) analysis. METHODS: Two-sample bidirectional MR was conducted based on publicly available genome-wide association studies (GWAS) summary statistics. Causal estimates of extensive PVS burden on neurodegenerative diseases and brain structure were primarily assessed using the inverse-variance weighted (IVW) method, supplemented by additional methods, including MR-Egger, weighted median, simple mode, and weighted mode. Sensitivity analyses were performed to assess heterogeneity and pleiotropy. In addition, we explored whether brain structure act as a mediating factor in the pathway from extensive PVS burden to neurodegenerative diseases. RESULTS: Our MR study found that extensive PVS burden in white matter (WM-PVS) burden was associated with lower Alzheimer's disease (AD) risk (IVW OR (95 % CI) = 0.963(0.929 to 0.999), P = 0.0428), with no heterogeneity and pleiotropy detected. In addition, following FDR correction, we found bidirectional causal relationships between extensive PVS burden and brain structure. Moreover, our results of the mediated analysis showed that the surface area of parahippocampal, as a mediating variable, plays an important role in the causal relationship between WM-PVS and AD. The mediation effect is 18 %. CONCLUSIONS: Our study provides evidence for the causal associations of different extensive PVS burden phenotypes with neurodegenerative diseases and brain structures, improving our understanding of the complex relationships between different brain injuries.