A two-sample bidirectional Mendelian randomization study on gut microbiota, dental caries, and toothache

Abstract

We aimed to evaluate whether gut microbiotas have an effect on caries and toothache and explore whether dental caries and toothache have any impact on the identified significant bacterial genera. We first used genetic instruments of gut microbiota from the MiBioGen consortium to investigate the relationship with toothache and dental caries. The genome-wide association study summary data were from the UK Biobank and the FinnGen consortium, respectively. Mendelian randomization (MR) Egger regression, weighted median, inverse-variance weighted, simple mode, and weighted mode were used in the analysis. MR-PRESSO and Cochrane Q statistics were used to detect pleiotropy and heterogeneity. Furthermore, we performed a bidirectional MR analysis to examine the direction of the relationship. In our study, we identified 22 gut microbiota that were associated with toothache and dental caries. In the reverse MR analysis, the analysis suggested a strong correlation between the gut microbiome (Eubacterium nodatum group) and toothache. Our MR analysis suggested that the gut microbiota influences the occurrence of dental caries and toothache, while dental caries and toothache exert an impact on the identified significant bacterial genera.

1. Introduction

Caries is a bacterial infectious disease and the most prevalent disease that causes chronic, progressive destruction of the hard tissues of the tooth.^[1–3] When decay progresses to the dentin layer, it can cause many symptoms, such as sensitivity and pain that often require restoration to resolve. Toothache is a common predicament that significantly disrupts normal daily activities, posing a serious challenge for affected individuals. There are many causes of toothache, such as deep dental caries, cracks, and periodontal disease.^[4] However, in the final analysis, bacterial infection is the core.^[5] When bacteria enter the dentinal pulp, diseases such as pulpitis and apical periodontitis can occur, causing chronic or acute pain that often requires root canal treatment or tooth extraction to resolve. It is worth mentioning that a higher quantity of bacteria does not necessarily result in a greater prevalence of dental caries or toothache, even though the oral cavity has a complex microbial landscape and is composed of 700 species of bacteria.^[6–8] Some studies have shown that different bacterial flora infections or the body’s immune response and infection status will have corresponding effects.^[9]

Intestinal flora is undoubtedly one of the factors that has the greatest impact on the state of the body. The gut microbiome is an essential part of the human holobiont. In recent years, many studies have highlighted this bacterial tribe with immunological, metabolic, neurological, psychiatric features, drug metabolism, and cancer.^[10] Relevant literature has shown that there is a potential relationship between the occurrence and development of periodontitis and intestinal flora at the genetic level.^[11,12] Caries, a disease recognized as one of the most common human diseases mediated by bacteria,^[13] may also be influenced by gut microbiota, similarly to toothache.^[9] The literature shows that pathogenic bacteria found in acute apical periodontitis, such as Peptostreptococcus and Eubacterium,^[7] also play an important role in the gut microbial landscape.^[14] At present, prevention methods for dentinal hard tissue diseases focus more on reducing bacterial adhesion on the tooth surface. Those approaches hinge on individual awareness, making it challenging to anticipate the outcomes. Gut microbes exhibit complexity and diversity among individuals. In addition, an interactive influence between intestinal flora and oral flora has existed for a long time.^[15] Studies have shown that ectopic colonization of oral microbes in the gut is a common and widespread phenomenon, and the oral cavity is considered to be an endogenous reservoir of the gut microbiota.^[8,16] At the same time, in healthy individuals, more than half of the microbial species, such as Streptococcus and Veillonella, are often detected in both the oral cavity and the intestine, which shows evidence of translocation in addition to the oral-intestinal microbiota.^[16]

If genetic relationships between intestinal microbiota and caries or toothache can be explored, they will undoubtedly provide new insights for the prevention of dentinal hard tissue diseases.

Mendelian randomization (MR) provides an alternative approach to making causal inferences possible when conducting an RCT is either impractical or infeasible.^[17] MR uses genetic variation to establish cause-and-effect relationships between exposures and outcomes. Genetic variants are randomly assigned at conception and often remain independent of environmental risk factors.^[18] Furthermore, these genetic variants precede disease outbreaks. This allows the MR analysis to eliminate the influence of unmeasured confounding factors.^[19–21]

Thus, a two-sample bidirectional MR study was conducted using publicly available genome-wide association studies (GWASs) databases to assess the potential impact of gut microbiota on the occurrence of dental caries and toothache. Additionally, the study aimed to investigate any relationship between dental caries, toothache, and the identified significant bacterial genera.

2. Methods

2.1. Study design

We performed a total of two MR analyses using summary statistics from a GWAS to investigate the bidirectional causal association between dental caries, toothache, and gut microbiota. We used the human gut microbiota as the exposure and dental caries and toothache as the outcomes to perform the forward MR analysis. Meanwhile, a reverse MR analysis was conducted as well. In general, single-nucleotide polymorphisms (SNPs) are often used as genetic instruments in MR studies.^[22] These genetic variants have a reliable association with the exposure but must not be associated with outcome variables or confounders. The study flowchart is shown in Figure 1.

Figure 1.

The study flowchart in our MR analysis. GWASs = genome-wide association studies, IVs = instrumental variables, LD = linkage disequilibrium, MR = Mendelian randomization, SNPs = single-nucleotide polymorphisms.

2.2. Exposure data

We chose SNPs associated with the human gut microbiota composition from the MiBioGen consortium’s GWAS datasets as instrumental variables (IVs). MiBioGen has collected 16S ribosomal RNA gene sequencing profiles and genome-wide SNP data from 18 population-level cohorts and 19,000 participants to investigate the influence of human genetics on human gut microbiota (with a focus on Europeans as the major population). The gut microbiota were classified into 5 levels according to phylum, class, order, family, and genus.^[23,24]

2.3. Outcome data

The GWAS summary-level data for the associations between genetic variants and toothache were obtained from the UK Biobank, which included 461,113 European individuals with 18,964 cases. The GWAS dataset for dental caries used in our MR analysis was obtained from the FinnGen consortium, which included 7680 cases and 259,234 controls.^[25] We aggregated these data to obtain all the data related to dental caries and toothache. Summary data were finally obtained for genetic variants associated only with dental caries and toothache.

2.4. IV selection

In order to ensure the validity and accuracy of the results regarding the study of gut microbiota with dental caries or toothache, the SNPs significantly correlated with gut microbiota were selected as IVs.^[26] We selected SNPs that were below the genome-wide significance threshold of 1 × 10⁻⁵. The F-statistic in the regression model was used to evaluate the effectiveness of the IVs, and any bias caused by weak IVs could be effectively avoided with an F-statistic > 10.^[27] We eliminated the IVs with an F-statistic < 10. The presence of linkage disequilibrium (LD) can lead to correlations between genotypes, which may interfere with association signals in association studies. To address this issue, we examined the LD between the remaining IVs and eliminated those showing significant LD.^[28] We employed the PLINK clumping technique to assess LD (with R² set to <0.1 and clumping distance = 500 kb). Moreover, it is crucial to verify the congruence of effect alleles between the exposure and outcome, ensuring alignment in the direction of SNP alleles for both. An integral aspect of the analysis involved a pivotal test known as the horizontal pleiotropy test, which was conducted on the retained IVs. IVs showing evidence of pleiotropy were subsequently excluded from the analysis.^[29] The inspection methods we used to detect pleiotropy included MR-PRESSO^[30] and MR-Egger.^[31] Firstly, we tested for global horizontal pleiotropy, and if it was present, outliers were excluded. Following this method, outlier SNPs were removed one by one until the P-value for the global test became insignificant (P > .05), ensuring that the remaining IVs could be utilized for subsequent MR analysis.

2.5. MR analysis

We employed MR analysis to investigate the relationship between human gut microbiota and dental caries or toothache. Various methods, such as the inverse-variance weighted (IVW) test,^[32,33] MR-Egger,^[31] weighted mode (WMO),^[34] weighted median estimator (WME), and simple model, were utilized to assess causality in the study. IVW was considered the primary method, with other methods serving as auxiliary analyses in the study.^[35] IVW is characterized by not considering the intercept term in the regression and using the reciprocal of the outcome variance as a weight for fitting the model.^[36] However, MR-Egger must consider the existence of the intercept term.^[37] IVW does not account for pleiotropy between IVs, whereas MR-Egger can be utilized to assess and correct for pleiotropy. The MR-Egger closely resembles IVW when the intercept term is very close to 0. In the WMO method, the causal effects of IVs on traits are weighted, and then the mode is taken as the final causal effect estimate. On the other hand, the WME method calculates the median as the result of the causal effect estimation.^[34–38] Both the WMO and WME methods can help reduce bias resulting from estimation errors in certain genetic variations. In order to measure the strength of the association between exposure and outcome, odds ratio (OR) values were calculated, with P < .05 considered statistically significant.

2.6. Sensitivity analysis

The sensitivity analysis primarily included tests for horizontal pleiotropy, data heterogeneity, and leave-one-out analysis. MR-PRESSO^[30] and MR-Egger^[31] were used to test horizontal pleiotropy. The data heterogeneity test served as a method to quantitatively evaluate the extent of heterogeneity, with the Cochran Q test^[33,39] utilized to ascertain the combinability of results from individual studies.

Since meta-analyses were employed to merge homogeneous results, they could not be effectively utilized in cases where there existed significant heterogeneity among studies. Significant Q statistics, with a P-value < .05, indicate the presence of heterogeneity. If there was heterogeneity, a random-effects model was used instead of a fixed-effect model. Forest, scatter, leave-one-out, and funnel plots were produced to further assess heterogeneity. The “leave-one-out” stepwise culling method examined the change in effect size of the remaining SNPs after each SNP was culled.

2.7. Reverse MR analysis

To evaluate any reverse causation effects, we also conducted a study in reverse to examine whether dental caries and toothache had a causal effect on the composition of the human gut microbiome. This involved using the SNPs corresponding to dental caries and toothache as IVs, with the gut microbiota serving as the outcome variable.

2.8. Statistics

All data were statistically analyzed using “TwoSampleMR” and “MR-PRESSO” in R software v.4.3.1 (R Foundation for Statistical Computing). We did not conduct multiple tests to adjust our results for the exploration of additional potential significance to the fullest extent possible.^[40] A significance level of P < .05 was considered statistically significant. The IVW analysis method was used as our primary analysis.

3. Results

3.1. SNPs selection

Through screening the genome-wide significance threshold (P < 1 × 10⁻⁵), conducting LD tests, harmonizing, and verifying F-statistics, we identified 8 bacterial taxa causally associated with dental caries and 14 bacterial taxa associated with toothache, resulting in 75 and 197 SNPs as their IVs, respectively. The detailed characteristics of the GWAS datasets were presented (Supplementary Table A1, Supplemental Digital Content 1). The F-statistics of all retained SNPs are over 10, indicating sufficient correlation strength between IVs and the corresponding bacterial taxa (Supplementary Table A2–A8, Supplemental Digital Content 2). Thus, our study is free from weak instrument bias.

3.2. Dental caries

In the set of IVs (P < 1 × 10⁻⁵), we found potential causal relationships between the 8 gut microbiota and dental caries, as the IVW analysis results for all 8 bacterial taxa show significant differences (P < .05; Table 1, Fig. 2). It was determined that 6 gut microbiota have a potential positive causal effect on dental caries, including: the class Betaproteobacteria (IVW OR = 1.159, 95% confidence interval [CI] = 1.037–1.297, P = .010), the genus Coprococcus3 (IVW OR = 1.137, 95% CI = 1.014–1.274, P = .028), the genus Lachnoclostridium (IVW OR = 1.167, 95% CI = 1.055–1.291, P = .003), the genus RuminococcaceaeUCG009 (IVW OR = 1.080, 95% CI = 1.006–1.159, P = .033), the genus Terrisporobacter (IVW OR = 1.123, 95% CI = 1.010–1.248, P = .032), and the order Burkholderiales (IVW OR = 1.133, 95% CI = 1.015–1.266, P = .026; Fig. 2). Meanwhile, it also suggests that 2 gut microbiota are negatively associated with dental caries, including: the family Bacteroidaceae (IVW OR = 0.874, 95% CI = 0.772–0.990, P = .034) and the genus Bacteroides (IVW OR = 0.874, 95% CI = 0.772–0.990, P = .034; Fig. 2).

Table 1.

The Mendelian randomization analysis results of gut microbiotas on the dental caries and toothache.

Exposure	Outcome	Method	Number of SNPs	β	SE	P	OR	95% CI	P Heterogeneity	P Pleiotropy
Class Betaproteobacteria	Dental caries	IVW	10	0.148	0.057	.010	1.159	1.037–1.297	.934	.581
Family Bacteroidaceae		IVW	8	−0.135	0.063	.034	0.874	0.772–0.990	.761	.186
Genus Bacteroides		IVW	8	−0.135	0.063	.034	0.874	0.772–0.990	.761	.186
Genus Coprococcus3		IVW	10	0.128	0.058	.028	1.137	1.014–1.274	.534	.423
Genus Lachnoclostridium		IVW	13	0.155	0.051	.003	1.167	1.055–1.291	.431	.521
Genus RuminococcaceaeUCG009		IVW	11	0.077	0.036	.033	1.080	1.006–1.159	.969	.472
Genus Terrisporobacter		IVW	5	0.116	0.054	.032	1.123	1.010–1.248	.869	.921
Order Burkholderiales		IVW	10	0.125	0.056	.026	1.133	1.015–1.266	.863	.485
Class Coriobacteriia	Toothache	IVW	18	−0.004	0.002	.037	0.996	0.992–1.000	.627	.676
Class Verrucomicrobiae		IVW	12	−0.004	0.002	.048	0.996	0.993–1.000	.792	.647
Family Coriobacteriaceae		IVW	18	−0.004	0.002	.037	0.996	0.992–1.000	.627	.676
Family Verrucomicrobiaceae		IVW	12	−0.004	0.002	.048	0.996	0.993–1.000	.792	.649
Genus Eubacterium nodatum group		IVW	11	−0.003	0.001	.021	0.997	0.995–1.000	.370	.431
Genus Akkermansia		IVW	12	−0.004	0.002	.048	0.996	0.993–1.000	.792	.652
Genus Clostridiumsensustricto1		IVW	7	−0.006	0.002	.019	0.994	0.990–0.999	.393	.641
Genus FamilyXIIIAD3011 group		IVW	13	0.005	0.003	.047	1.005	1.000–1.010	.081	.590
Genus Intestinimonas		IVW	17	0.003	0.002	.046	1.003	1.000–1.006	.671	.152
Genus Phascolarctobacterium		IVW	11	−0.005	0.002	.021	0.995	0.992–0.999	.698	.675
Genus Slackia		IVW	6	0.005	0.002	.011	1.005	1.001–1.009	.418	.305
Genus unknowngenus		IVW	12	−0.005	0.002	.020	0.995	0.991–0.999	.130	.995
Order Coriobacteriales		IVW	18	−0.004	0.002	.037	0.996	0.992–1.000	.627	.676
Order Verrucomicrobiales		IVW	12	−0.004	0.002	.048	0.996	0.993–1.000	.792	.647

β = beta, CI = confidence interval, IVW = inverse-variance weighted, OR = odds ratio, SE = standard error, SNPs = single-nucleotide polymorphisms.

Figure 2.

Mendelian randomization estimates for the relationship between genetically instrumented gut microbiota and dental caries. CI = confidence interval, IVW = inverse-variance weighted, OR = odds ratio.

3.3. Toothache

In the set of IVs (P < 1 × 10⁻⁵), the figure shows there are a total of 14 gut microbiota that are involved in the occurrence of toothache (Table 1, Fig. 3). We found that the genus FamilyXIIIAD3011group (IVW OR = 1.005, 95% CI = 1.000–1.010, P = .047), the genus Intestinimonas (IVW OR = 1.003, 95% CI = 1.000–1.006, P = .046), and the genus Slackia (IVW OR = 1.005, 95% CI = 1.001–1.009, P = .011) were positively associated with toothache (Table 1, Fig. 3). On the contrary, we identified 11 gut microbiota negatively related to toothache, including the class Coriobacteriia (IVW OR = 0.996, 95% CI = 0.992–1.000, P = .037), the class Verrucomicrobiae (IVW OR = 0.996, 95% CI = 0.993–1.000, P = .048), the family Coriobacteriaceae (IVW OR = 0.996, 95% CI = 0.992–1.000, P = .037), the family Verrucomicrobiaceae (IVW OR = 0.996, 95% CI = 0.993–1.000, P = .048), the genus Eubacterium nodatum group (IVW OR = 0.997, 95% CI = 0.995–1.000, P = .021), the genus Akkermansia (IVW OR = 0.996, 95% CI = 0.996–1.000, P = .048), the genus Clostridiumsensustricto1 (IVW OR = 0.994, 95% CI = 0.990–0.999, P = .019), the genus Phascolarctobacterium (IVW OR = 0.995, 95% CI = 0.992–0.999, P = .021), the genus unknowngenus (IVW OR = 0.995, 95% CI = 0.991–0.999, P = .020), the order Coriobacteriales (IVW OR = 0.996, 95% CI = 0.992–1.000, P = .037), and the order Verrucomicrobiales (IVW OR = 0.996, 95% CI = 0.993–1.000, P = .048; Fig. 3).

Figure 3.

Mendelian randomization estimates for the relationship between genetically instrumented gut microbiota and toothache. CI = confidence interval, IVW = inverse-variance weighted, OR = odds ratio.

3.4. Bidirectional effects between gut microbiota and dental caries, toothache

We employed a reverse MR analysis to further assess the causal relationship between gut microbiota and dental caries, as well as toothache. Therefore, dental caries and toothache were used as exposures, respectively, and the previous gut microbiota were taken as outcomes (Table 2, Figs. 4 and 5). We found a bidirectional effect between toothache and the genus E nodatum group (IVW OR = 14,504.42, 95% CI = 6.819–30,851,353.680, P = .014; Fig. 5), which indicates a bidirectional effect between toothache and the genus E nodatum group.

Table 2.

The Mendelian randomization analysis results of dental caries and toothache on the identified significant bacterial genera.

Exposure	Outcome	Method	Number of SNPs	β	SE	P	OR	95% CI	P Heterogeneity	P Pleiotropy
Dental caries	Class Betaproteobacteria	IVW	13	−0.028	0.088	.754	0.973	0.818–1.157	.064	.715
	Family Bacteroidaceae	IVW	13	0.022	0.076	.775	1.022	0.880–1.187	.191	.612
	Genus Bacteroides	IVW	13	0.022	0.076	.775	1.022	0.880–1.187	.191	.612
	Genus Coprococcus3	IVW	13	−0.006	0.071	.934	0.994	0.865–1.143	.472	.777
	Genus Lachnoclostridium	IVW	13	−0.005	0.08	.947	0.995	0.850–1.164	.139	.552
	Genus RuminococcaceaeUCG009	IVW	12	−0.116	0.105	.269	0.891	0.726–1.094	.617	.521
	Genus Terrisporobacter	IVW	12	0.080	0.112	.471	1.084	0.871–1.349	.924	.385
	Order Burkholderiales	IVW	13	−0.033	0.088	.712	0.968	0.814–1.151	.065	.623
Toothache	Class Coriobacteriia	IVW	15	0.256	1.425	.858	1.291	0.079–21.085	.748	.726
	Class Verrucomicrobiae	IVW	15	2.149	1.736	.216	8.576	0.285–257.77	.811	.455
	Family Coriobacteriaceae	IVW	15	0.256	1.425	.858	1.291	0.079–21.085	.748	.726
	Family Verrucomicrobiaceae	IVW	15	2.153	1.736	.215	8.612	0.287–258.833	.811	.451
	Genus Eubacterium nodatum group	IVW	14	9.582	3.909	.014	14,504.42	6.819–30,851,353.679	.184	.018
	Genus Akkermansia	IVW	15	2.125	1.736	.221	8.371	0.278–251.637	.809	.471
	Genus Clostridiumsensustricto1	IVW	15	1.765	1.631	.279	5.843	0.239–142.89	.406	.989
	Genus FamilyXIIIAD3011 group	IVW	15	1.242	1.636	.448	3.463	0.140–85.455	.392	.654
	Genus Intestinimonas	IVW	15	1.558	1.726	.367	4.751	0.161–140.019	.482	.569
	Genus Phascolarctobacterium	IVW	15	1.505	1.803	.404	4.504	0.132–154.130	.670	.866
	Genus Slackia	IVW	15	0.034	2.452	.989	1.034	0.008–126.424	.425	.494
	Genus unknowngenus	IVW	15	4.255	2.742	.121	70.479	0.327–15,208.687	.007	.170
	Order Coriobacteriales	IVW	15	0.256	1.425	.858	1.291	0.079–21.085	.748	.726
	Order Verrucomicrobiales	IVW	15	2.149	1.736	.216	8.576	0.285–257.77	.811	.455

β = beta, CI = confidence interval, IVW = inverse-variance weighted, OR = odds ratio, SE = standard error, SNPs = single-nucleotide polymorphisms.

Figure 4.

Reverse Mendelian randomization estimates for the relationship between genetically instrumented dental caries and gut microbiota. CI = confidence interval, IVW = inverse-variance weighted, OR = odds ratio.

Figure 5.

Reverse Mendelian randomization estimates for the relationship between genetically instrumented toothache and gut microbiota. CI = confidence interval, IVW = inverse-variance weighted, OR = odds ratio.

3.5. Sensitivity analyses

The MR-PRESSO analysis revealed the absence of any noteworthy outliers, as indicated by a global test with a P-value exceeding .05. Additionally, no significant horizontal pleiotropy was identified with P > .05, as detailed in Tables 1 and 2. The MR-Egger intercept approach, outlined in Supplementary Table A9–A16, Supplemental Digital Content 3, further supported the lack of evidence for horizontal pleiotropy in the association between gut microbiota and both toothache and dental caries, with P > .05. Moreover, the Cochrane Q statistics displayed no substantial heterogeneity, with P > .05, as illustrated in Tables 1 and 2.

4. Discussion

We conducted a two-sample MR analysis involving 211 gut microbiota taxa (9 phyla, 16 classes, 20 orders, 35 families, and 131 genera) causally associated with caries and toothache. The results showed that 22 microorganisms were significant, with 8 potentially related to caries and 14 related to toothache. Meanwhile, there is a correlation between toothache and the E nodatum group. Although several associations were observed at the nominal significance level, none remained statistically significant after correction for multiple testing.

It is generally recognized that dental caries and toothache are associated with bacterial infections.^[41,42] The oral microbiota, as the second-largest microbial habitat for humans, is similar to the intestinal flora, and the connection between them is referred to as the oral-gut axis.

Recent literature suggests that intestinal dysbiosis can lead to low-grade chronic inflammation.^[43] In this subset of patients, there is an impact on the production of cytokines and antibodies in saliva, leading to altered immune function.^[44,45] During the active phase of digestive tract disease, the overproduction of pro-inflammatory molecules may predispose individuals to apical periodontitis.^[46,47] Poyato-Borrego et al have shown that patients with inflammatory bowel disease (IBD) are more likely to have radiolucent periapical lesions than the general population.^[48] This may be one of the reasons why the gut microbiota is associated with toothache. Apart from apical periodontitis, pulpitis also contributes to toothache, although research in this area remains limited. Investigating the correlation between pulpitis and intestinal diseases or intestinal flora may emerge as a promising avenue for future research.

With the advancement of oral microecology research, the dynamic balance of oral microbiology may be broken due to changes in local, systemic, and environmental factors.^[49] This imbalance can lead to a continual decrease in the pH level of saliva below the critical value (pH = 5.5),^[50] thereby causing an imbalance between tooth demineralization and remineralization, ultimately resulting in tooth decay.^[51] Gut flora is a crucial factor influencing overall body health. Patients with IBD exhibit greater instability in their microbiome, immune response, and metabolite pool (such as acylcarnitines, bile acids, and short-chain fatty acids) compared to healthy individuals. These factors can undergo significant alterations within a short period, even just a few weeks.^[52,53] The potential association between intestinal homeostasis and saliva pH may become one of the future research directions.

Verrucomicrobiae play a role in glucose homeostasis in the human gut and possess anti-inflammatory properties that can enhance gut health. Studies have shown a positive correlation between this bacterium and the Foxp3 gene, which is involved in anti-inflammatory responses and immunity in humans.^[54] Some diseases, including obesity, IBD, sleep disorders, and type 2 diabetes, have been linked to a decreased proportion of Verrucomicrobiae in the gut microbiota.^[55,56] The abundance of Verrucomicrobiae also varies in individuals with other conditions, such as asthma and autism.^[57,58] From previous research, we can conclude that this microorganism may be a beneficial bacterium. The results of this MR also support the conclusion that this bacterium is a protective factor against toothache at the genetic level.

According to previous studies, the composition of the gut microbiota can vary among different populations. Enterotype is associated with long-term dietary habits, and there are statistically significant differences in the way people metabolize and store energy between different enterotypes.^[59] The composition and species of microorganisms are related to enterotypes. This group of people who have the Ruminococcus enterotype tends to eat a diet that is high in carbohydrates.^[60] From the point of view of the causative factors of caries,^[61] a high-carb diet is undoubtedly risky. In this MR study, the same conclusion was reached, indicating that Ruminococcus is a risk factor for caries. Fat-soluble vitamins are closely related to the metabolism of calcium, and the body needs vitamins A and D to make 2 different proteins involved in calcium transport and deposition.^[62] Therefore, fat-soluble vitamins are important for maintaining healthy teeth. To enhance the absorption of fat-soluble vitamins, it is essential to consume a certain amount of fat. Bacteroides are ubiquitous in the gut of people living in Western countries (North America and Europe) because Western diets are usually high in fat and protein.^[63] Bacteroides belongs to the Family Bacteroidaceae. Coincidentally, the Family Bacteroidaceae is a protective factor against tooth decay. With these similar results, we have to wonder if there is some mysterious internal logic connecting dietary preferences, intestinal flora, and caries. It is difficult to draw conclusions about the relationship from one analysis, but this is at least a starting point that hopefully inspires scholars to get more decisive results.

It is worth mentioning the relationship between toothache and the E nodatum group. This is the only gut microbe we have found that has a positive result in both forward and reverse analyses. The E nodatum group was initially identified in the late 1980s as an obligate anaerobic, Gram-positive microbe that belongs to the phylum Firmicutes and the genus Eubacterium.^[64] The genus Eubacterium is an important gut bacterium found in the colon of healthy people; it is one of the core bacterial genera of the human intestinal microbiota and shows extensive colonization in the intestines and oral cavity, playing an important role in the body’s nutrient metabolism and maintenance of intestinal balance.^[65] Although the previous literature considered the genus Eubacterium to be a beneficial bacterial genus, the E nodatum group is different from the genus Eubacterium. Back in 2013, Tamura et al^[66] found that the E nodatum group is one of the predominant bacteria found in peri-implant sites.^[64] The primary pathological condition associated with the E nodatum group is periodontitis, but this flora acts as a protective factor against toothache at a genetic level.^[67] Thus, more studies are needed to explore the relationship between periodontitis and toothaches in the future.

The GWAS datasets for this MR analysis were obtained from different databases, largely avoiding data dishonesty caused by sample overlap. However, there are still some limitations to our sample selection. Even with data from different databases, the samplesources are generally limited to European ancestry. This limitation makes it challenging to determine the full applicability of our results to other populations or whether they are influenced by race or other factors. We may also need to conduct similar analyses of populations in other places, or even some primitive tribesmen, to draw broader conclusions.

The absence of statistically significant associations after multiple testing correction may reflect several factors. First, the genetic instruments for gut microbiota are relatively weak compared with those for other complex traits, which may limit statistical power. Second, the relationship between microbiota and dental diseases is likely multifactorial and influenced by environmental factors, such as diet, oral hygiene, and host immunity. Therefore, the lack of robust evidence in this study does not necessarily indicate the absence of a biological relationship, but rather suggests that current genetic evidence is insufficient to support strong causal inference.

We were pleasantly surprised by the insights provided through this MR analysis. Many scholars are exploring the potential of using gut microbiota to regulate various systemic diseases, including obesity, high blood pressure, Alzheimer disease, and even insomnia. The results revealed that adjusting the composition of gut microbiota for caries prevention may also be a viable path. However, further biochemical and cell-level studies are necessary before clinical application can be considered.

Our MR analysis suggested that the gut microbiota influences the risk of developing dental caries and toothache, while dental caries and toothache impact identified significant bacterial genera. A bidirectional relationship exists where the E nodatum group exhibits a protective effect against toothache, while toothache increases the abundance of the E nodatum group. This discovery provides an exploratory direction for future basic research on the E nodatum group and toothache.

5. Conclusion

Our MR analysis provides evidence for a bidirectional relationship between the gut microbiota and dental health. Specifically, we found that gut microbiota influences the risk of developing dental caries and toothache, while conditions such as dental caries and toothache, in turn, impact the abundance of specific bacterial genera. Notably, the E nodatum group exhibits a protective effect against toothache, whereas toothache itself increases the abundance of this bacterial group. These findings highlight a potential feedback mechanism and suggest a novel direction for future research into the role of the E nodatum group in oral health. Further basic research is warranted to explore the underlying mechanisms of this interaction, potentially leading to new therapeutic approaches for preventing and managing dental diseases.

Acknowledgments

The authors express their gratitude to the UK Biobank, FinnGen, and the MiBioGen consortium for generously sharing the GWAS summary datasets, which were instrumental in completing this study. The authors also thank Suining Central Hospital for providing relevant resources and financial support for this study.

Author contributions

Conceptualization: Jiwei Xia.

Funding acquisition: Sen Yang.

Data curation: Longyu Zhang, Xianjie Zheng, Jiwei Xia, Lijuan Guo, Sen Yang.

Formal analysis: Longyu Zhang, Xianjie Zheng, Lijuan Guo, Sen Yang.