We conducted a two-sample Mendelian Randomization (MR) study. As instruments, we used 105 genetic variants known to be associated (p<5x10^-8) with a composite of caries, dentures and missing teeth in the Gene-Lifestyle Interactions in Dental Endpoints (GLIDE) Consortium. In stroke-free participants enrolled in the UK Biobank, we tested for associations between these genetic variants and white matter hyperintensity volume (natural log-transformed), fractional anisotropy and mean diffusivity. For the last two neuroimaging traits, we evaluated composite scores defined by the first principal component of measurements obtained across 48 brain regions.

Our primary analysis using the inverse variance-weighted MR method indicated that genetically-increased risk of poor oral health was associated with: (1) higher burden of silent cerebrovascular disease, as represented by a 30 % increase in white matter hyperintensities volumes (beta=0.30, SE=0.06 p-value=3x10^-6), and (2) increased microstructural damage, as represented by a 43% change in composite FA scores (beta=0.43, SE=0.06; p=6x10^-11) and a 10% change in composite MD scores (beta=0.10, SE=0.03; p=0.005). Sensitivity analyses identified horizontal pleiotropy in our primary results, but an outlier-corrected analysis confirmed each of our findings (all p-values <0.001).