Periodontal diseases and all-cause dementia risk: Genetic instrument analyses in half a million UK Biobank participants

Background Numerous studies suggest that periodontal diseases might be associated with the development of dementia, but the causality is inconclusive. Objective This study aims to explore the casual effect of periodontal diseases on all-cause dementia. Methods The UK Biobank (UKB) data (n = ∼500,000) has been implemented, where participants were divided into two independent groups (2/3train and 1/3test). The exposure is the self-reported periodontal diseases, and the outcome is all-cause dementia measured by both clinical diagnoses based on ICD10 and ICD9 codes, and self-reported dementia. Four sets of genetic instruments were developed based on four different thresholds (main approach: p < 5 × 10−8; alternative approach I: p < 5 × 10−6; alternative approach II: p < 10−4; and alternative approach III: the best-fit p-value threshold calculated by polygenic risk score). The causal association between periodontal diseases and dementia was assessed by inverse-variance weighted (IVW), MR-Egger regression, weighted median, and mode-based estimate models. Results The number of genetic instruments included in these four approaches varied from 3 to 1020, after passing the MR assumption checks. Most MR results suggested no causal association between periodontal diseases and dementia except the IVW model from main approach (coefficient beta: −0.816, 95% confidence interval, CI (−1.617, −0.015)) and the weighted median model from alternative approach II (beta: 0.077 95%CI (0.006, 0.149)) suggested potential causal relationship between periodontal diseases and dementia. Conclusions The results showed inconsistent evidence of causal link between periodontal diseases and dementia using UKB. Future studies are needed with clinically defined periodontal diseases to better understand the causal link.