Mendelian Randomization Study Reveals No Causal Link Between Periodontitis and Intracranial Aneurysms

doi:10.21203/rs.3.rs-4286240/v1

Objective: Previous studies have suggested an association between periodontitis (PD) and intracranial aneurysms (IAs). However, it remains unclear whether the association is causal. This study systematically investigated the potential genetic link between PD and IAs, including the formation and rupture of IAs.

Methods: This study utilized publicly available genome-wide association study (GWAS) summary statistics data for a two-sample bidirectional Mendelian randomization (MR) analysis. The main statistical analysis method used was inverse variance weighting (IVW). The reliability of the results was verified through sensitivity analysis and assessment of the strength of genetic instrumental variables.

Results: There was no causal relationship between genetically determined PD and uIAs (OR=1.02, 95%CI: 0.89-1.17, P=0.76) or aSAH (OR=0.96, 95%CI: 0.85-1.08, P=0.50). In the reverse MR analysis, the increased risk of uIAs or SAH was not statistically significant for an increased risk of PD (uIAs: OR=0.99, 95%CI: 0.94-1.05, P=0.78; aSAH: OR=1.02, 95%CI: 0.96-1.09, P=0.51). The results from MR-Egger regression and weighted median method were consistent with the IVW method. Sensitivity analysis indicated that horizontal pleiotropy was unlikely to distort the causal estimates.

Conclusion:Our study does not support a causal relationship between PD and the formation or rupture of IAs, and vice versa.

periodontitis

intracranial aneurysms

Mendelian randomization

causal relationship

Intracranial aneurysms (IAs) are severe cerebrovascular diseases characterized by pathological dilation of the intracranial artery wall. IAs are relatively common, occurring in 3 to 6 percent of the adult population^[1]. Clinically, IAs are present in unruptured intracranial aneurysms (uIAs) and aneurysmal subarachnoid hemorrhagea (aSAH), which is a catastrophic complication with a mortality rate of 25%-50%^[2,3]. Despite significant improvements in treatment methods and intensive care management, aSAH is still associated with substantial disability, death, and socioeconomic impact. Each year, approximately one-third of individuals with IAs die from aSAH; In half of survivors, additional support is required to continue normal daily activities^[4]. Furthermore, although the adoption of interventional techniques, novel materials, and treatment strategies have decreased mortality and disability rates in treating uIAs, the outcome is still not optimistic^{[5, 6]}. Therefore, the mechanisms underlying the formation and progression of IAs, especially preventive and therapeutic measures, have garnered considerable attention from scholars.

Periodontitis (PD), inflammation of the supporting tissues of the teeth, accompanied by a gradual loss of periodontal attachment and bone. Its characteristics include the formation of periodontal pockets and/or gingival recession, with an overall prevalence of 45%-50%^[7,8]. Several recent observational studies have supported the association between periodontitis and IAs and suggested that periodontal bacteria may be involved in the pathology of IAs by spreading through gingival tissue into circulation^[9-12]. As such, PD appears to be a modifiable non-traditional risk factor for IAs. However, this association is still lacking in validation from randomized controlled trials, and the evidence for a causal relationship is insufficient in strength.

Factors such as reverse causation and selection bias are inherent flaws in these observational studies. To address these limitations of observational studies, we employed Mendelian randomization (MR) analysis, using genetic variations as instrumental variables (IVs), to investigate the causal relationship between the exposure phenotype and the outcome phenotype^[13]. In this study, we performed a two-sample MR analysis to assess the causal relationship between PD and the risk of IAs formation and rupture. Additionally, we also performed reverse MR to investigate the causal relationship of IAs formation and rupture on PD.

Data Sources

The genome-wide association study (GWAS) summary statistics data used in this study for PD were obtained from the FinnGen project (https://www.finngen.fi/en), led by the Finnish Biobank, which draws on longitudinal health registry data from the entire population of Finland since 1969.The GWAS summary data for PD contained 3,046 diagnosed cases and 195,395 control cases.

The GWAS summary data for IAs were obtained from a meta-analysis conducted by Bakker^[14]et al. , which included uIAs and aSAH cases. The former included 7,495 diagnosed cases and 71,934 controls cases, while the latter included 5,140 diagnosed cases and 71,952 control cases. To mitigate population stratification bias, all GWAS summary data were included only from individuals of European ancestry. The research workflow was illustrated in Figure 1.

Selection of IVs

To estimate causal effects using genetic instruments, it is essential to meet the three key assumptions of IVs. Therefore, quality control measures were implemented. Firstly, since the number of independent SNPs with a strong association with PD at p < 5×10^-8 was limited, we set the threshold at p < 5×10^-5 to select a sufficient number of IVs. Secondly, to exclude SNPs in strong linkage disequilibrium (LD), the clump function was performed with default parameters. Thirdly, SNPs with a minor allele frequency (MAF) less than 0.01 were eliminated. Fourthly, the F statistic for each SNP was calculated using the following equation: F = R²× (N - 2) / (1 - R²). R²represents the exposure variance for each IV interpretation. The filtering criterion was an F-test value > 10^[15]. Fifthly, SNPs with allele inconsistency between exposure and outcome samples, as well as palindromic alleles were excluded. Lastly, the MR-PRESSO global test was applied to detect potential horizontal pleiotropy of SNPs, and SNPs with a p< 0.05 were excluded^[16]. Finally, we obtained a set of high-quality SNPs.

Statistical Analysis

In this study, we employed multiple statistical analysis methods, including IVW, MR-Egger regression, and Weighted Median, to estimate the causal effects of exposure phenotypes on the outcome. The IVW method was the primary statistical approach used in this study. When all selected SNPs are valid IVs, the IVW method provides the most accurate results. MR-Egger regression method allows for consistent estimation of causal effects even in the presence of pleiotropy effects^[17]. The Weighted Median method is applicable to some or as many as 50% of SNPs that are invalid IVs and gives consistent estimates^[18].

Pleiotropic and sensitivity analysis

In this study, we employed various methods to detect the presence of pleiotropy in IVs. Firstly, the intercept term of MR-Egger regression can effectively indicate whether horizontal pleiotropy drives the results of the MR analysis^[17]. Secondly, the asymmetry of the funnel plot can reflect the presence of horizontal pleiotropy in IVs^[19]. Lastly, we conducted the MR-PRESSO test to assess the presence of pleiotropy in IVs^[16]. To identify heterogeneity in IVs, we used both the IVW and MR-Egger regression to quantify heterogeneity by Cochran's Q statistic. Additionally, we performed leave-one-out analyses to examine the robustness and consistency of the MR analysis results.

In this study, the Bonferroni method was used to correct for p-values for multiple comparisons, i.e. P<0.006 (0.05/8) to show convincing evidence of causation. All analyses were conducted using the "TwoSampleMR"^[19] and "MRPRESSO"^[20]packages in R software version 4.3.1.

Causal Relationship of PD with uIAs and SAH

In the forward MR analysis, PD was included as an exposure factor for uIAs and aSAH. A total of 18 independent SNPs with p<5×10^-5 were selected as IVs for uIAs, while 13 SNPs were selected for aSAH. The detailed information of these IVs can be found in Supplementary Table 1. These SNPs explained 2.13% and 2.72% of the variance for uIAs and aSAH, respectively. The F-statistic for each SNP was greater than 10, indicating that the explained phenotype variance and high F-statistic confirmed the MR assumption (1). The MR estimates for the different methods were shown in Figure 2. The main results from the IVW method indicated that an increased risk of PD was not statistically significant associated with an increased risk of uIAs (OR=1.02, 95%CI: 0.89-1.17, P=0.76) or aSAH (OR=0.96, 95%CI: 0.85-1.08, P=0.50). Additionally, the MR-Egger regression and weighted median method showed consistent negative results (Figure 2).

Table 1.Pleiotropic and sensitivity analysis in Bidirectional MR Analysis

Heterogeneity
Exposure	Outcome	Q¹	Df²	pval
periodontitis	uIAs	9.65	17.00	0.92
periodontitis	SAH	9.90	12.00	0.62
uIAs	periodontitis	28.02	46.00	0.98
SAH	periodontitis	35.95	48.00	0.90

MR.PRESSO. results
Exposure	Outcome	Global.Test.RSSobs	Global.Test.pval
periodontitis	uIAs	11.00	0.90
periodontitis	SAH	11.69	0.64
uIAs	periodontitis	29.32	1.00
SAH	periodontitis	37.40	0.91

MR-Egger regression for directional pleiotropy
Exposure	Outcome	Intercept	Se³	pval
periodontitis	uIAs	-0.01	0.03	0.95
periodontitis	SAH	0.03	0.03	0.40
uIAs	periodontitis	-0.02	0.02	0.35
SAH	periodontitis	-0.01	0.02	0.96
¹Q, heterogeneity statistic Q; ²Df, degree of freedom; ³se, standard error

Pleiotropic and sensitivity analysis

Although the exposure and outcome came from different samples, the p-values of the heterogeneity tests were all greater than 0.05, indicating no heterogeneity. The results of the MR-PRESSO global test and the horizontal pleiotropy test are not significant, suggesting that the IVs were only related to the exposure and not related to other confounding factors (Table 1). The funnel plot showed a roughly symmetrical distribution of SNPs on both sides of the IVW method (Figure 3). Low or no heterogeneity, no indication of potentially confounding SNPs, and pleiotropic robust analysis results ensured the validity of the MR assumptions (2) and (3). By performing leave-one-out analyses, no significant impact of individual SNPs on the test results was found, which reflects the robustness of the aforementioned conclusions (Supplementary Table 1-2).

Causal relationship of uIAs/SAH on PD

Finally, this study also explored the causal relationship of uIAs/SAH on PD. After screening, a total of 47 qualified IVs for uIAs and 49 qualified IVs for aSAH were obtained. The detailed information of IVs was listed in Supplementary Table 2. The primary results of IVW showed that there was no significant association between increased prevalence of uIAs or aSAH and increased risk of PD (uIAs: OR=0.99, 95% CI: 0.94-1.05, P = 0.78; aSAH: OR=1.02, 95%CI: 0.96-1.09, P = 0.51).In addition, MR-Egger regression and weighted median showed consistent results (Figure 2).The p-values of the heterogeneity tests, MR-PRESSO global test, and horizontal pleiotropy test were all greater than 0.05 (Table 1), indicating the robustness of this conclusion.

This study utilized the genetic correlation of SNPs to investigate the association between PD and IAs, including their formation or rupture. Our two-sample MR analysis did not observe a causal relationship between PD occurrence and the formation or rupture of IAs in the European population sample. Similarly, the reverse MR analysis also found no evidence of an association between the formation or rupture of IAs and PD. To the best of our knowledge, this is the first study to explore the bidirectional causal relationship between PD and the formation or rupture of IAs through the MR method.

Previous epidemiological research has suggested a potential association between PD and the formation of IAs. A case-control study found a significant correlation between poor periodontal health and the formation of IAs^[21]. A cohort study revealed an association between PD and the formation of IAs (OR: 5.99, 95%CI: 2.6-13.8, p < 0.001)^[9]. Another large-scale cohort study conducted in Korea also showed a significant correlation between the presence of PD and an increased risk of IAs (HR: 1.21, 95%CI: 1.09-1.34, p < 0.001)^[22]. Furthermore, there may be a potential association between PD and aSAH. In a 14-year follow-up study of PD patients, baseline periodontitis increased the risk of aSAH (HR: 14.3, 95%CI: 1.5-135.9, p=0.020)^[9]. Additionally, the KUH cohort demonstrated that periodontitis was associated with an increased risk of aSAH diagnosis during the follow-up period (OR: 5.3, 95%CI: 1.1-25.9, p < 0.000)^[10]. Despite the relatively consistent findings, it is still challenging to establish a causal relationship between PD and the formation or rupture of IAs due to limitations of observational studies, such as selection bias and confounding factors.

The study conducted a bidirectional MR analysis, which resulted in consistent non-significant findings. The OR values for both hypotheses were slightly above and below 1, with highly overlapping confidence intervals. This indicates a weak clinical significance of the observed effects. MR analysis typically provides strong evidence in situations where the effect size is particularly small or nonexistent^[23]. Therefore, it is unlikely that there is any causal relationship between PD and IAs formation or rupture. One possible explanation is that the previously observed association between PD and IAs formation or rupture was coincidental or confounded by factors such as smoking^{[14, 24, 25]}. Meanwhile, we adopt a rigorous IVs screening program, MR-PRESSO global test, and horizontal pleiotropy tests ensured the robustness of the MR analysis results. However, it is important to note that our conclusions are limited to the European population and cannot be extrapolated to the general population. Additionally, considering the intermediate position of MR analysis in the evidence pyramid, future randomized controlled trials may still be worth considering^[26].

In conclusion, our study does not support a causal relationship between PD and the formation or rupture of IAs, nor vice versa. Further randomized controlled trials are needed to validate these findings and obtain more reliable and conclusive results.

Authors’ contributions

Yao chen wrote the first draft of the manuscript, conducted statistical analysis, curated the data, and performed writing review and editing. Yuanbao Kang and Jianhuang Huang contributed to the investigation, statistical analysis, and editing. Quanming Zhou critically reviewed the manuscript, provided editing assistance, and partially funded the study. The final version was approved by all authors.

Funding

None.

Data Availability Statement:

The data that support the findings of this study are available from the corresponding author upon request.

Ethics statement

The publicly available GWAS Summary data used in this study generally does not require ethical review.

Conflicts of Interest Statement:

All authors disclose no conflict of interest.

Acknowledgments

The authors acknowledge and thank the investigators of the original GWAS studies for sharing summary data used in this study.

Etminan N, Rinkel GJ. Unruptured intracranial aneurysms: development, rupture and preventive management[J]. Nat Rev Neurol. 2016;12(12):699-713.
Etminan N, Buchholz BA, Dreier R, et al. Cerebral aneurysms: formation, progression, and developmental chronology[J]. Transl Stroke Res. 2014;5(2):167-73.
Daou BJ, Koduri S, Thompson BG, et al. Clinical and experimental aspects of aneurysmal subarachnoid hemorrhage[J]. CNS Neurosci Ther. 2019;25(10):1096-112.
Ferreira M, Nahed BV, Babu MA, et al. Trapped fourth ventricle phenomenon following aneurysm rupture of the posterior circulation: case reports[J]. Neurosurgery. 2012;70(1):E253-8; discussion E8.
Algra AM, Lindgren A, Vergouwen MDI, et al. Procedural Clinical Complications, Case-Fatality Risks, and Risk Factors in Endovascular and Neurosurgical Treatment of Unruptured Intracranial Aneurysms: A Systematic Review and Meta-analysis[J]. JAMA Neurol. 2019;76(3):282-93.
Etminan N, Dörfler A, Steinmetz H. Unruptured Intracranial Aneurysms- Pathogenesis and Individualized Management[J]. Dtsch Arztebl Int. 2020;117(14):235-42.
Kassebaum NJ, Bernabé E, Dahiya M, et al. Global burden of severe periodontitis in 1990-2010: a systematic review and meta-regression[J]. J Dent Res. 2014;93(11):1045-53.
Hegde R, Awan KH. Effects of periodontal disease on systemic health[J]. Dis Mon. 2019;65(6):185-92.
Hallikainen J, Pessi T, Vehkalahti M, et al. Unlike severe periodontitis, caries does not associate with intracranial aneurysms or aneurysmal subarachnoid hemorrhage[J]. Acta Neurochir (Wien). 2023;165(1):169-75.
Hallikainen J, Lindgren A, Savolainen J, et al. Periodontitis and gingival bleeding associate with intracranial aneurysms and risk of aneurysmal subarachnoid hemorrhage[J]. Neurosurg Rev. 2020;43(2):669-79.
Hallikainen J, Pyysalo M, Keränen S, et al. Systemic immune response against the oral pathogens Porphyromonas gingivalis and Aggregatibacter actinomycetemcomitans is associated with the formation and rupture of intracranial aneurysms[J]. Eur J Neurol. 2021;28(9):3089-99.
Pyysalo MJ, Pyysalo LM, Pessi T, et al. The connection between ruptured cerebral aneurysms and odontogenic bacteria[J]. J Neurol Neurosurg Psychiatry. 2013;84(11):1214-8.
Emdin CA, Khera AV, Kathiresan S. Mendelian Randomization[J]. Jama. 2017;318(19):1925-6.
Bakker MK, van der Spek RAA, van Rheenen W, et al. Genome-wide association study of intracranial aneurysms identifies 17 risk loci and genetic overlap with clinical risk factors[J]. Nat Genet. 2020;52(12):1303-13.
Burgess S, Thompson SG. Avoiding bias from weak instruments in Mendelian randomization studies[J]. Int J Epidemiol. 2011;40(3):755-64.
Verbanck M, Chen CY, Neale B, et al. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases[J]. Nat Genet. 2018;50(5):693-8.
Bowden J, Davey Smith G, Burgess S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression[J]. Int J Epidemiol. 2015;44(2):512-25.
Bowden J, Davey Smith G, Haycock PC, et al. Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator[J]. Genet Epidemiol. 2016;40(4):304-14.
Hemani G, Zheng J, Elsworth B, et al. The MR-Base platform supports systematic causal inference across the human phenome[J]. Elife. 2018;7.
Verbanck M, Chen C-Y, Neale B, et al. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases[J]. Nature Genetics. 2018;50(5):693-8.
Liu K, Sun J, Shao L, et al. Correlation of periodontal diseases with intracranial aneurysm formation: novel predictive indicators[J]. Chin Neurosurg J. 2021;7(1):31.
Woo HG, Chang YK, Lee JS, et al. Association of Periodontal Disease with the Occurrence of Unruptured Cerebral Aneurysm among Adults in Korea: A Nationwide Population-Based Cohort Study.[J]. Medicina (Kaunas). 2021;57(9).
Dudbridge F. Polygenic Mendelian Randomization[J]. Cold Spring Harb Perspect Med. 2021;11(2).
Karhunen V, Bakker MK, Ruigrok YM, et al. Modifiable Risk Factors for Intracranial Aneurysm and Aneurysmal Subarachnoid Hemorrhage: A Mendelian Randomization Study[J]. J Am Heart Assoc. 2021;10(22):e022277.
Leite FRM, Nascimento GG, Scheutz F, et al. Effect of Smoking on Periodontitis: A Systematic Review and Meta-regression[J]. Am J Prev Med. 2018;54(6):831-41.
Davies NM, Holmes MV, Davey Smith G. Reading Mendelian randomisation studies: a guide, glossary, and checklist for clinicians[J]. Bmj. 2018;362:k601.

No competing interests reported.

Mendelian Randomization Study Reveals No Causal Link Between Periodontitis and Intracranial Aneurysms

Status:

Version 1

Abstract

Figures

Introduction

Material and Methods

Results

Discussion

Conclusions

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1