A Commentary on

Ma C, Wu M, Gao J et al.

Periodontitis and stroke: A Mendelian randomization study. Brain Behav. 2023; https://doi.org/10.1002/brb3.2888.

GRADE Rating:


The risk of ischemic stroke has been shown to be higher among the patients with periodontitis compared to those without periodontitis in numerous observational studies1,2,3. These studies indicate association but not causality4. The causal effect of periodontal disease in atherosclerotic cardiovascular diseases including stroke is controversial due to the major limitation of confounding factors5. Even after adjusting for confounding factors, study results may still bear the effects of residual confounding as well as other inherent biases in the observational study designs. Interventional studies to evaluate whether periodontal treatment would lower the risk of atherosclerotic cardiovascular diseases have not been able to provide sufficient evidence4.

To overcome the effects of confounding factors, researchers have started utilizing study design referred to as Mendelian randomization. Individuals randomly inherit or do not inherit a genetic variant. Mendelian randomization is a type of research study that uses these genetic variants to study possible causal relationships between risk factors (exposure) and disease (outcomes). Because genetic variants are fixed at conception, mendelian randomization studies reduce the effect caused by confounding factors and reverse causation that are typically seen in traditional observational studies. Mendelian randomization studies could be considered as interface studies between conventional observational studies and interventional studies6,7. Mendelian randomization study designs operate under three major assumptions—(1) the genetic variant is associated with the risk factor (relevance assumption); (2) there is no other confounder that exists between genetic variant and outcome (independence assumption); and (3) the genetic variant exerts its effect on the outcome through the effect of the risk factor only (exclusion restriction). Readers of mendelian randomization studies should ensure that these assumptions are valid. Other items to consider while appraising mendelian randomization study designs include justification for selection of SNPs, accounting for weak instrument bias when using two different samples, clinically meaningful effect size, consistency of association with previous studies, among others7,8.

In this study, the authors have used the mendelian randomization design to study the causal relationship between periodontitis (CP and AgP) and ischemic stroke and its subtypes (CE stroke, LAA, and SVO). They utilised large sample data from national and multinational databanks in Europe to select the genetic variants of risk factors of CP and AgP and for the outcome of stroke. They conducted statistical tests to ensure the robustness of their results including checking for heterogeneity and pleiotropy that explains effects of genetic variant on the outcome via alternate biological pathways. Despite these strengths, the study had some limitations. Due to the paucity of genetic variants, the authors increased the statistical threshold from the conventional p value <5 × 10−8 to <1 × 10−5 to select the SNPs. In addition, risk factors and outcome data were derived from different datasets. It is unclear if these approaches of selection of genetic variants could have been associated with introduction of potential confounders or shared risk factors and affected the causal inference. The authors found no causal inference between AgP and all subtypes of ischemic stroke. They also found no causal inference between CP and two subtypes of ischemic stroke, LAA and SVO. However, they found a causal inference between the combined genetic variants of CP and CE stroke (OR, 1.052; 95% CI, 1.002–1.104; p = 0.042). This must be interpreted with caution as the effect size was small (1.052). In addition, the lower end of the 95% confidence interval was very close to the null (1.002) and could be interpreted as being influenced by weak instrument or having unknown confounding effects. In other words, the main conclusion of the study on the causal association between CP and CE stroke is based on weak association and may not be clinically meaningful. In addition, the conclusion of this study is inconsistent with previous study using mendelian randomization and that these associations may not be accurate and could be attributed to shared risk factors such as smoking9. The genetic variant data was obtained from European databases and hence may not represent the entire global population. Additional mendelian randomization studies from large databases representing different populations across the globe is necessary to support or refute the findings of this study. The authors mainly focused their discussion on the causal association between CP and CE stroke. Perhaps the interesting finding of this study was that there was no causal inference between AgP and CP to most subtypes of stroke which contradicts the findings of most conventional observational studies. Perhaps, the results of positive association between periodontal disease and atherosclerotic cardiovascular disease from conventional observational studies could have been confounded by shared risk factors. Additional investigations that compare the estimates of the current and other mendelian randomization studies to the estimates of conventional observational studies investigating the association between periodontal disease and stroke. Lastly, as majority of the readers would be aware that AgP and CP have been combined into one diagnosis of periodontitis in the latest classification of periodontal diseases10. Future studies should also clarify the case definitions of periodontitis in such studies that include data from large databanks.