Causal Effect of Dietary Patterns on Cerebral Small Vessel Disease: A Mendelian Randomization Study

While emerging observational evidence suggests associations between dietary factors and cerebral small vessel disease (CSVD), the causal nature of these relationships remains unestablished. This study employed a two-sample Mendelian randomization (MR) framework to investigate genetically predicted causal effects of dietary patterns on neuroimaging markers of CSVD. We utilized Genome-Wide Association Study (GWAS) summary statistics from European-ancestry cohorts for 32 dietary exposures and four CSVD phenotypes: white matter hyperintensities (WMH), lacunar strokes (LS), enlarged perivascular spaces (PVS), and brain microbleeds (BMB). Specifically, WMH volume, fractional anisotropy (FA), and mean diffusivity (MD) are indicators related to WMH. Genetic instruments (single-nucleotide polymorphisms, SNPs) were rigorously selected using genome-wide significance thresholds (p < 5 × 10^-8/p < 5 × 10^-6), linkage disequilibrium clumping (r < 0.001), and pleiotropy exclusion criteria. The primary analysis employed inverse-variance weighted (IVW) MR, with validation through four complementary methods: weighted median, MR-Egger regression, weighted mode, and simple mode. Significant associations (p < 0.05) underwent false discovery rate (FDR) correction. Robustness was further assessed through heterogeneity testing (Cochran's Q), horizontal pleiotropy evaluation (MR-Egger intercept), and leave-one-out analysis. Simultaneously perform reverse MR analysis to eliminate reverse causal relationships. Our MR analysis revealed several genetically predicted causal associations between dietary factors and CSVD neuroimaging markers. Higher polyunsaturated fatty acid (PUFA) levels demonstrated a protective effect against WMH volume (β = -0.070, 95% CI: -0.126 to -0.015; se = 0.028; p = 0.013, FDR-adjusted p = 0.033). Similarly, our analysis revealed a protective causal association between monounsaturated fatty acid (MUFA) intake and the risk of lobar BMB (β = -0.298, se = 0.115, 95% CI: -0.523 to -0.073; p = 0.009, FDR-adjusted p = 0.035). Conversely, increased iron intake exhibited detrimental effects on both any BMB (β = 0.247, se = 0.094, 95% CI: 0.064 to 0.430; p = 0.008, FDR-adjusted p = 0.041) and strictly deep BMB (β = 0.414, se = 0.153, 95% CI: 0.116 to 0.713; p = 0.007, FDR-adjusted p = 0.033). Notably, suggestive protective associations of coffee consumption (β = -0.088, se = 0.044, 95% CI: -0.173 to -0.002; p = 0.045, FDR-adjusted p = 0.225) and non-oily fish intake (β = -0.193, se = 0.098, 95% CI: -0.386 to -0.001; p = 0.049, FDR-adjusted p = 0.248) with basal ganglia PVS were observed; however, these associations did not survive multiple testing correction (FDR-adjusted p > 0.05). All significant findings showed no evidence of heterogeneity (PS__{heterogeneity} > 0.05) or horizontal pleiotropy (PS__pleiotropy > 0.05). Reverse MR analyses revealed no causal effects of CSVD features on dietary exposures (PS > 0.05). This study provides novel genetic evidence supporting heterogeneous causal effects of dietary patterns on distinct CSVD phenotypes. The protective effect of PUFA against white matter injury, together with the beneficial role of MUFA in brain microbleeds, contrasted with iron's detrimental impact on brain microbleeds, underscoring the pathophysiological complexity of diet-CSVD interactions. While nominally significant associations between coffee/fish intake and PVS require further validation, our findings emphasize the potential of targeted nutritional interventions in CSVD prevention. Future prospective studies with standardized dietary assessments and longitudinal neuroimaging are warranted to translate these genetic insights into clinical practice.