Effect of Fluoride on Gut Microbiota: A Systematic Review.

Abstract

Context

Fluoride can prevent dental caries by inhibiting demineralization and promoting remineralization of teeth while affecting the physiology of oral microbiota, thus inhibiting cellular enzymes. However, the effect of systemic fluoride on gut microbiota is unknown.

Objective

To explore the impacts of systemic fluoride on gut microbiota composition and abundance and associated functions such as gene and metabolic regulation.

Data Sources

A systematic database search was conducted of MEDLINE, Web of Science, Scopus, PubMed, CINAHL, and Embase to find articles on studies reporting the effects of fluoride on gut microbiota.

Data Extraction

Forty-nine studies were included (n = 42 in animals, 4 of humans, 3 in vitro studies) after screening for title, abstract, and full text using Covidence to check against eligibility criteria. Data were extracted using Covidence and study quality was assessed using the Mixed Method Appraisal Tool by 2 reviewers independently.

Data Analysis

Two human studies of dental fluorosis and 1 of patients with breast cancer (intestinal fluorine-18 fluorodeoxyglucose uptake) showed significant differences in gut microbial composition, with increased relative abundance of Acidobacteria and Proteobacteria, and decreased abundance of Firmicutes and Bacteroidetes. An ex vivo study of human feces indicated that <2 mg L^–1 NaF might boost “health-associated” taxa, but concentrations (>10 mg L^–1 NaF) could increase the ratio of some unhealthy microbes after 24 hours. The animal studies examined the effects of high fluoride doses in water and diet (50-1200 mg L^–1 NaF) for long-term (1-6 months) and short-term (6 hours to 7 days) exposure, with all showing a significant disturbance in the Firmicutes to Bacteroidota ratio.

Conclusion

In humans, high doses potentially may be detrimental to the microbiome, whereas , <2 mg L^–1 NaF had positive effects. Similarly, in animals, >50 mg L^–1 NaF was unsafe, whereas <25 mg L^–1 NaF had harmless effects.

Systematic Review Registration

PROSPERO registration No. CRD42022347357.

EXCERPTS: Links toTables. No References are included – see Online full-text study

INTRODUCTION

The human body has distinct habitats colonized by microbes, forming a functional and dynamic interface between our genes and the environment.² Over recent decades, considerable attention has been given to the gut microbiota or microbiome that resides within the gastrointestinal tract.³ The Human Microbiome Project discovered that up to 70% (10¹³-10¹⁴ bacterial cells) of the body’s total microbiome (>39 trillion) reside in the gastrointestinal tract.^4–6 These bacteria interact in commensal, symbiotic, or parasitic ways and are a source of competition for nutrition and adhesion toward some external pathogenic bacteria.⁷ Advances in molecular sequencing and computational methods have provided an unprecedented understanding of how the gut microbiota functions in symbiotically with the host, contributing to nutrition, metabolism, immune response, and intestinal architecture.³ Furthermore, targeting the gut microbiota with lifestyle interventions can result in significant changes in bacterial composition aligned with improvements in health conditions.⁸ Therefore, the relationship between the gut microbiota and human health appears to be bidirectional rather than consequential.⁹^,¹⁰

There is not yet a single definition of a healthy microbiota. Nevertheless, it is mostly accepted that greater diversity, richness, and stability, and a higher relative abundance of species associated with the production of short-chain fatty acids (SCFAs) (eg, Faecalibacterium prausnitziii, Bacteroides spp, Roseburia spp, Bifidobacterium spp, and Lactobacillus spp) are hallmarks of a microbial community that is associated with better health outcomes.¹¹ Specific bacteria shown to be more abundant in health compared with disease states include Bifidobacterium and Lactobacillus spp.¹²

Fluoride is well recognized for its role in preventing and reversing dental caries.¹³ Fluoride is naturally found in soil, water, and almost all food and drink items in different concentrations.¹⁴ Fluoride, often in the form of sodium fluoride (NaF), is added to dental products, as well as to water, salt, and milk, in many countries globally to prevent dental caries.¹³^,¹^,¹⁵

Although fluoride has been widely used as a preventive agent for dental caries, the effects of fluoride on the gut microbial communities still need to be understood.¹⁶ Fluoride can affect microbial community abundance, which may enhance its ability to prevent the growth of harmful bacteria and fungi or stimulate the growth of useful bacteria that reside in the gastrointestinal tract and vice versa, as summarized by Moran et al¹⁷ in a mini-review in which they discuss recent studies in which the effects of ingested fluoridated water on human and animal microflora were observed, mainly focusing on oral microbiomes.¹⁷ When gut microbiome was discussed, it was mainly in animal models, with a focus on how acute toxicity of fluoride can perturb normal gut microbiomes. To our knowledge, no previously published systematic reviews have highlighted the impacts of fluoride with a specific target on gut microbiota and its associated functions. Fluoride may be beneficial or harmful to the gut microbiome depending upon the doses ingested, transit time, and duration of exposure.¹^,¹⁸ Levels of daily fluoride intake mainly depend on the geographic region and sources of exposure, with most people exposed to a range between 0.46 and 3.6-5.4 mg day^–1.¹⁹ The daily adequate intakes of fluoride suggested by the US Environmental Protection Agency and World Health Organization (WHO) are 0.5-0.7 mg for infants, 1-2 mg for children, and 3-4 mg for adults males and females.²⁰ The Environmental Protection Agency allows a range of 0.8-1.7 mg L^–1 in the drinking water, with a tolerable upper intake level of 4 mg L^–1.²¹

However, the optimal scale of fluoride in drinking water recommended by the WHO is 1.5 mg L^–1, with the high or upper limit being 4 mg L^–1 for humans to prevent dental caries and skeletal disorders.¹⁷^,²² In 1984, WHO discovered that teeth mottling is associated with fluoride levels > 1.5 mg L^–1 in water, and skeletal fluorosis can ensue if fluoride levels exceed 10 mg L^–1. Therefore, 1.5 mg L^–1 was recommended as a safe dose by WHO. However, this value is not fixed and should be adjusted by considering local conditions (eg, diet, water consumption).²³^,²⁴ In diet, the dose of <5 mg kg^–1 fluoride in the form of salt is regarded as the minimum but is dependent on age and sex.²³^,²⁴ Doses effective for gut microbiota still need to be recognized; human intervention trials are relatively limited.

In this systematic review, we provide an overview of all recent studies that investigated the association between fluoride exposure from different sources, doses, and duration with changes in gut microbiome composition in animals, in vitro models, and humans. The primary outcome was to define any fluoride-induced changes in the composition of the gut microbiome, measured by (1) abundance and/or (2) diversity or richness of the microbes. The secondary outcome measured the significant impacts of dose and duration of fluoride on microbiota-associated functions (ie, metabolites and gene expression) in both animals and humans.

METHODS

Protocol and Registration

This systemic review was conducted according to the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) statement and the guidelines of the Cochrane Handbook for Systematic Reviews of Interventions.²⁵^,²⁶ The full details of the processes including the eligibility criteria, search strategy, extraction process, and data analysis were prespecified and documented in a protocol that was registered in the PROSPERO database (registration no. CRD42022347357).

Search Strategy

The search strategy was developed according to the research question, using the population, exposure, outcome, and study design (PEOS) approach listed in Table 1. Open in new tab

Research question based on population, exposure, outcome, and study design terms.

… A literature search was undertaken in July 2022 and a rerun of the search was conducted in July 2023 and then in 2024, using the following 6 electronic databases: MEDLINE, CINAHL, Embase, Scopus, PubMed, and Web of Science. Publications in English were retrieved using a combination of key terms, subject index, and Medical Subject Heading terms. A search of the reference lists of relevant articles was also performed to identify other potentially relevant sources. The terms used were “gut,” “gastrointestinal,” “intestine,” “colon,” “bowel,” “microbiota,” “micro biota,” “microbiome,” “micro biome,” “flora,” “microflora,” “micro flora,” “bacteria,” “fluorid,” “fluoride,” and “fluoridation.” All keywords and search strategies were adapted according to the specifics of each database and are presented in Tables S1 and S2.

There were no restrictions based on the date of publication. All published articles that met the stated eligibility criteria were included. Studies reported only in the English language were included, due to lack of translation facilities.

Inclusion and Exclusion Criteria

The eligibility criteria were based on the PEOS framework (ie, population, exposure, outcome, study design). Based on the study design, all observational or analytical studies that attempted to identify the relationship between fluoride exposure to gut microbiota and an associated outcome, such as change in the composition of gut microbiota and its associated effects, were included. Quantitative studies using all interventional trials or designs, such as randomized controlled trials and nonrandomized controlled trials, along with laboratory-based studies (eg, fluoride exposure to in vitro models) were also considered. Before and after studies, also called pre-post studies, and multiarm studies were also considered for inclusion.

All studies of humans or other animals of any age and sex that were exposed to different forms and doses of fluoride, in which gut microbiome speciation analysis was conducted using DNA analysis and gene sequencing to determine the effect of systemic fluoride on participants’ gut microbiome, were selected. Additionally, studies using in vitro models created from fecal or gut samples to evaluate the effect of fluoride were also included.

Exposure or intervention included fluoride in all forms and sources: topical (eg, dental products) and systemic (eg, diet, dietary and nondietary supplements, air). The outcome was the evaluation of microbiota composition after fluoride exposure, including the assessment of diversity or richness, prevalence of bacterial taxa, and their corresponding functions, such as metabolic and gene expressions.

Diversity or richness was evaluated by considering alpha and beta diversity values. Three diversity scales or indices were estimated for alpha diversity: Shannon diversity index, Simpson diversity index, and Chao 1 index. The Shannon diversity index is used to measure both the operational taxonomic units of abundance or richness, and evenness of species.²⁷ The Simpson diversity index is another measure of diversity; it measures the relative abundance of each species and gives more weight to the more abundant species in the sample.²⁸ Chao 1 is an estimator for total species richness that gives more importance to unique species.²⁹

Qualitative studies such as case reviews, case series, expert opinions, and reviews, focus groups and interviews, narrative reports, abstracts without full text, and conference papers, as well as other systematic reviews and study protocols, were excluded.

Study Selection

After the search stage, all the citations were exported to Endnote 21.2 (Clarivate, London, UK) and deduplicated. The remaining studies were then exported to Covidence 2022 software (Melbourne, VIC, Australia) for screening. The Covidence software was also useful for removing duplicates that were not identified in Endnote.

The abstract and title of each paper were screened independently for eligibility by 2 individuals (M.Y. and M.P.S.) according to the eligibility requirements. In case of disagreement, the conflicts were resolved by discussion with another reviewer (F.V.Z., C.H.O., and P.D.). Full text of seemingly relevant articles was then screened by 2 reviewers independently (M.Y. and M.P.S.) to determine their eligibility for inclusion in the final review. Again, any conflicts were resolved by discussion with a third reviewer. All missing full texts were requested via ResearchGate and/or the university library to allow access to the full text. Corresponding authors were contacted to request full-text articles not otherwise available. Full-text articles that could not be accessed after following these steps were excluded.

Manual searching was also done to exhaust all possibilities and to reduce the risk of bias, as follows: (1) reference lists from the included studies were searched to identify any relevant study; (2) performing citation tracking in which all the articles that cited each of the included articles were tracked; and (3) related articles in Google Scholar and Science Direct were searched for to prevent any chance of missing a relevant study. The relevant articles underwent further scrutiny against the inclusion criteria, after title and abstract and full-text screening.

Data Extraction

A bespoke tool was developed in Covidence that included the study identifier, author, year of publication, journal, title, study location, study design, study period, details of exposure, study participants, number of participants analyzed, outcomes (methods for measuring outcome), statistical analysis method, and software used. Data extraction was conducted by 1 reviewer (M.Y.) and cross-checked by another reviewer (E.A.K.). Differences in judgment between the 2 reviewers were settled by discussion and consensus.

Quality Assessment

The Mixed Methods Appraisal Tool was used to assess the quality of the included studies.³⁰ The Mixed Methods Appraisal Tool contains 5 and 6 evaluation elements/questions each for the various study designs, with a judgement of “yes,” “no,” and “unclear” for each question.

Data Synthesis and Analysis

All the data are presented in a narrative form using figures and tables to assist in data presentation by following PRISMA guidelines. The supplementation and control groups’ pre- and postintervention mean values and SDs were aggregated in the form of tables. Subgroup analysis was conducted based on dose and duration.

RESULTS

Identification and Selection of Studies

Study identification and selection are specified in the PRISMA flow chart (Figure 1). A total of 1004 articles were identified from the database search. After deduplication 590 records were screened at the title and abstract level, and 63 articles were included in the full-text screening based on the preset inclusion and exclusion criteria. An additional 10 articles were identified from reference lists and citation chaining of included studies. Of the 73 remaining articles, 49 were eligible for and included in this review. All 49 articles reported on studies of the effect of different forms and doses of fluoride on the gut microbiome. These studies were 39 randomized controlled trials,¹⁰^,^31–67 1 nonrandomized controlled trial,⁶⁸ 6 laboratory-based or experimental studies,¹⁴^,^69–73 1 cohort study,⁷⁴ and 2 case-control studies.⁷⁵^,⁷⁶ The principal aim differed among the studies. The details of studies are given in Tables 2 and 3¹⁰^,¹⁴^,^31–76 and Table S3.

… [See online full-text study for Tables]

The total sample size across all the included animal studies was 3249, of which at least 45% were male and 19% were female; 12 (29%) studies were unable to report the sex of participants. The 4 included human studies reported a total sample size of 217 participants, with 2 studies indicating participants’ sex (Table 2).

Intervention Characteristics

Most of the included studies (n = 35; 73%) used NaF as a source of fluoride in water or diet as an intervention; 3 studies used perfluoroalkyl fluorides (organic fluoride; ie, sodium fluoroacetate⁴⁴^,⁶⁹ and perfluorooctanoic acid⁶⁶); and 1 study looked at the effect of polyfluorinated ether sulphonate on the gut microbiota of mice.⁶⁷ The in vivo NaF doses used in water and diet in the animal studies were 0.5,⁴⁹ 5,⁴⁹ 24,⁵⁸ 25,¹⁰^,⁵⁴^,⁶⁵ 50,¹⁰^,⁴⁸^,⁴⁹^,⁵⁴^,⁶³^,⁶⁵ 80,⁵⁵^,⁵⁷ 100,¹⁰^,^31–33^,^35–37^,⁴⁰^,^45–48^,^53–55^,⁵⁷^,⁵⁹^,⁶²^,⁶⁴^,⁶⁷^,⁶⁵^,⁷⁶ 150,⁵⁴^,⁶⁵ 200,³¹^,³⁵^,³⁹^,⁶⁰^,⁶⁸ 300,³⁵ 400,³⁵^,^41–43 500,³⁵ 750,⁴¹, 800,⁴¹ and 1200 mg L^–1.^41–43 Three studies indicated only fluoride levels and used 4, 10, 15, 45, 50, and 75 mg L^–1 fluoride.³⁴^,⁵³^,⁶² Three more studies used 0.5, 1, and 3 mg kg^–1 perfluorooctanoic acid⁶⁶; 0.57 and 5.7 mg L^–1 polyfluorinated ether sulphonate⁶⁷; and 0.2 mg kg^–1 sodium fluoroacetate.⁴⁴ The in vitro doses used were 0.1,⁷¹ 1,⁷¹ 4.76,⁷⁰ 10,^71–73 20,⁷²^,⁷³ 30,⁷²^,⁷³ 40,⁷²^,⁷³ 50,⁷²^,⁷³ and 100⁷¹ mM NaF and 20 mM sodium fluoroacetate.⁶⁹ The ex vivo study of human fecal samples used 1, 2, 10, and 15 mg L^–1 F.¹⁴ The details of doses are listed in Table S4.

Most in vivo animal studies (83%) reported the long-term association (1-6 months) between fluoride exposure and gut microbiome composition; 19% of animal and 1 human study reported short-term outcomes (1 day). Two animal studies compared participants at baseline before the inception of any intervention (ie, pre-post studies).⁵³^,⁶⁹

Results from all the 49 studies are presented narratively because they were unable to be statistically pooled.

Quality Assessment

The methodological quality of the included studies was assessed based on the Mixed Methods Appraisal Tool using criteria specific to observational studies (including cohort studies and case-control studies), quantitative studies (including randomized and nonrandomized controlled studies), and laboratory-based studies. In this systematic review, the included studies were mainly of high quality (79% of studies). As shown in Figure S1, 39 studies (79%)¹⁰^,¹⁴^,^31–43^,⁴⁵^,⁴⁶^,^49–51^,^54–56^,^58–67^,⁷⁰^,⁷¹^,⁷⁴^,⁷⁶ met all the quality assessment criteria, and 4 studies (8%)⁵²^,⁵⁷^,⁶⁸^,⁷² met 7 of 8 assessment criteria. Three studies (6%)⁴⁴^,⁵³^,⁷³ met 6 criteria, and 3 studies (6%) met 5 assessment criteria.⁵⁸^,⁶⁹^,⁷⁵ The risk of selection bias remained low because most of the studies were deemed of good quality.

Effect of Fluoride on Gut Microbiota (Primary Outcome)

Shifts in Microbial Diversity/Richness (Alpha and Beta Diversity)

An increase or decrease in alpha diversity indices,⁷⁷ with Chao 1 used to estimate the number of species in a community and the Shannon and Simpson indices used to estimate the abundance of each species,²⁸^,⁷⁸ in fluoride-treated groups when compared with control groups indicates fluoride is affecting the microbial community structure. Alpha diversity differences were estimated in 17 of the included studies (35%) and computed using the aforementioned 3 diversity indices.¹⁰^,³¹^,³⁸^,³⁹^,⁴³^,⁴⁵^,⁵⁰^,⁵²^,⁵⁴^,⁵⁷^,⁵⁸^,^61–63^,⁶⁷^,⁷⁴^,⁷⁶ Table 3 and Table S5 show an increase and decrease in Chao 1, Simpson, and Shannon diversity indices at different doses and durations of fluoride. (Table 3 and Table S5). Eight of the included animal studies¹⁰^,³⁸^,⁴⁵^,⁵⁴^,⁵⁶^,⁵⁷^,⁶²^,⁷⁶ reported a decrease in alpha diversity and richness (by all 3 indices) after exposure to high doses of fluoride (ie, 50, 80, 100, 150, and 750 mg L^–1 NaF in both water and diet) compared with the control group for 1-4 months. Conversely, some studies³¹^,³⁹^,⁴³^,⁵⁰^,⁵²^,⁵⁴^,⁶¹^,⁷⁶ reported increased alpha diversity and richness at 100, 200, 400, and 1200 mg L^–1 NaF in water and diet for 16 hours to 2 months. Two studies found no significant differences in alpha diversity.⁵⁸^,⁶⁷ Zhong et al⁵⁴ and Zhu et al¹⁰ found that microbial diversity and richness were negatively correlated with fluoride dose.

Nine studies¹⁰^,¹⁴^,³⁴^,³⁶^,³⁷^,⁵⁴^,⁵⁵^,⁵⁷^,⁷⁶ also assessed the beta diversity differences between control and treated groups by unconstrained principal coordinate analysis. Of these, 2 studies¹⁴^,³⁴ indicated no differences in beta diversity after the administration of fluoride compared with a control. In the remaining studies, the numeric data were not available to be displayed in tables. Four good-quality studies¹⁰^,⁵⁴^,⁵⁵^,⁵⁷ indicated large differences in beta diversity after fluoride exposure. Beta diversity estimates the similarity and dissimilarity (uniqueness) between the populations (samples).⁷⁹

Fluoride and Abundance of Gut Microbial Species

The abundance of gut microbiota was estimated and compared between the control and interventional groups in all included studies. Data were analyzed by fluoride dose and duration.

Effects of fluoride on the abundance of human gut microbiota

Limited studies (n = 4) conducted with human participants considered the impact of ingested fluoride on gut microbiota. One of the ex vivo studies of human feces, which used a fermenter, indicated a low dose (1-2 mg L^–1) of NaF did not show a significant difference in the abundance at phylum and genus levels but may promote taxa associated with health, including Faecalibacterium and Lactobacillus. However, high doses (10-15 mg L^–1 NaF) significantly increased the relative abundance of Proteobacteria (synonym Proteobacteriota) (Table 3).¹⁴ Two population studies explored the correlation between dental fluorosis and gut microbiota change by analyzing the fecal samples of children and adult patients with fluorosis. The patients had significant variations in microbiota composition and abundance compared with healthy participants. This was characterized by an increase in the relative abundance of Verrucomicrobiota, Desulfobacterota, Acidobacteriota, and Proteobacteria, and a significant decline in the relative abundance of Firmicutes and Bacteroidetes at the phylum level. At the genus level, the abundance of Bifidobacterium, and Faecalibacterium was reduced significantly.⁷⁵^,⁷⁶ In another case-control study, the effect of fluorine-18 fluorodeoxyglucose (¹⁸F-FDG) positron emission tomography on the gut microbiota of patients with breast cancer was investigated. The researchers used ¹⁸F-FDG to determine the breast cancer stage and found a negative correlation between intestinal ¹⁸F-FDG uptake and the abundance of members of the Ruminococcaceae, which lead to mucosal inflammation in patients due to the increase of proinflammatory cytokines.⁷⁴

The results from all the human studies suggest high fluoride exposure changed the taxonomic composition of gut microbiota in humans by disturbing the balance between useful microorganisms (namely, Ruminococcaceae, Firmicutes, Faecalibacterium, Lactobacillus, and Bacteroidetes) and pathogenic microorganisms (eg, Proteobacteria, Acidobacteriota). This disturbance can lead to IBD, colitis, and even cancers, due to lack of microbial metabolites (eg, SCFAs) that control the metabolic pathways. Moreover, the low dose can stimulate the growth of beneficial bacteria. However, the population studies did not mention the dose that was responsible for fluorosis.^74–76

Chronic and acute effects of fluoride on abundance change patterns at phylum and genus levels in animal models

Overall, data were classified by different doses and duration of fluoride. Across animal species, we found consistent exposure to high doses of fluoride may result in disturbance in gut microflora. However, the phyla or genera most affected differ between different animal types.

Effect of fluoride on gut microbiota abundance and composition in rodents

In rodents, most studies (72%) reported a dramatic decrease in the abundance of Firmicutes at the phylum level and lactobacilli at the genus level upon exposure to a high dose of fluoride (25-100 mg L^–1 NaF) compared with a control.³¹^,³⁶^,³⁷^,⁴⁶^,^50–52^,⁵⁶^,⁷⁵ Similarly, the abundance of Bacteroidetes could also be affected by fluoride exposure. Among the rat and mice studies, 10 reported an increase in Bacteroidetes level and a decrease in Parabacteroides and Bacteroidales at 100 mg L^–1 NaF.³¹^,³⁶^,^49–52^,⁵⁶

Of the Actinobacteria, Bifidobacterium was also influenced by fluoride intervention, with an increased abundance indicated in 5 studies³¹^,⁴⁷^,⁵⁰^,⁵¹^,⁷⁵ and a decrease in 3 studies when animals were exposed to 50-200 mg L^–1 NaF.³⁶^,⁴⁹^,⁵⁹

Rarer gut phyla, such as Proteroidetes,⁴⁹ Cyanobacteria,⁵⁰ Proteobacteria,³¹ Atoprostipes, Lachnospiraceae,⁶¹ Erysipelotrichaceae,⁶³ and Tenericutes (synonym Tenericuteota)³⁶ increased upon high fluoride (25-150 mg L^–1 NaF) exposure, but populations of Saccharibacteria³¹^,⁴⁹ and Verrucomicrobia (synonym Verrucomicrobiota)³⁶^,⁴⁹^,⁵⁶ decreased. Four studies investigated the dose-response (low to high dose) effects of fluoride on the gut microbiota of mice. Three of the studies used doses of 25, 50, 100, and 150 mg L^–1 NaF for 70 days¹⁰ and 12 weeks.⁵⁴^,⁶⁵ Fluoride led to significantly lower levels of Actinobacteria, Bacteroidetes, and Verrucomicrobia and significantly higher levels of Firmicutes, whereas Turicibacter, Lachnospiraceae, Roseburia, and Clostridium sensu stricto were negatively associated with fluoride exposure. Ruminococcaceae populations increase with increasing fluoride exposure doses.¹⁰^,⁵⁴^,⁶⁵ On the contrary, 1 study used 15, 45, and 75 mg L^–1 NaF for 3 months and found that low fluoride exposure upregulated the abundance of Lactobacillus and Lachnospiraceae; bacilli, Lactobacillus, and Lactobacillaceae were also dominant in the 45 mg L^–1 NaF group; and Bifidobacteriales, Bacteroidetes, and Actinobacteria were dominant taxa in the 75 mg L^–1 NaF group.⁶²

However, in 1 study of mice, by Yasuda et al,⁵³ applicable fluoride concentrations of 4 mg L^–1 and 4 mg L^–1 plus gavage resulted in the treated groups having a reduction in acidogenic oral bacteria, but there was no impact on the gut microbiome. The researchers concluded that most of the fluoride may be absorbed in the upper part of the gastrointestinal tract.⁵³

Two studies observed the effect of perfluoroalkyl compounds on mice gut microbiota.⁶⁶^,⁶⁷ Shi et al⁶⁶ detected that exposure to perfluorooctanoic acid increased Bacteroidetes and decreased Firmicutes populations significantly. At the genus level, exposure to perfluorooctanoic acid decreased the abundance of Bifidobacterium and Ruminococcus.⁶⁶ Feng et al⁶⁷ found that 5.7 mgL^–1 polyfluorinated ether sulphonate decreased the abundance of lactobacilli but increased that of Staphylococcus.⁶⁷

High doses of NaF (range, 25-200 mg L^–1) with long-term exposure (1-6 months) most often were administered in studies using rat models, and these demonstrated that excessive fluoride exposure impacts gut microbiota composition, although variations in specific phyla and genera abundances were observed in different studies using the same dose and duration of fluoride.

Impacts of fluoride dose and duration on the microbial abundance in other animal models

In silkworm³⁹^,⁶⁸^,⁷⁰ and Bufo gargarizans tadpole⁴⁹ animal models, increased exposure to a high dose of fluoride (200 mg L^–1 and 5-50 mg L^–1 NaF, respectively) led to a decrease in Enterococcus and Bacillus in silkworms and Firmicutes, Synergistetes (synonym Synergistota), and Bacteroides in Bufo gargarizans but an increase in Staphylococcus in silkworm and in Deltaproteobacteria, Endomicrobia, and Spirochaetes (synonym Spirochaetota) in B gargarizans.

Studies of bird models such as laying hens, broiler chickens, and ducks usually involved high concentrations of fluoride (400-1200 mg kg^–1). Findings showed that at high doses of fluoride (given as NaF, 1200 mg kg^–1 and 800 mg kg^–1 in diet) the abundance of Lactobacillus and Bifidobacterium was lower and that of Escherichia-Shigella, Gammaproteobacteria, Enterobacteriaceae, Streptococcus, and Staphylococcus was higher compared with controls after both short- (16 hours) and long-term (21-59 days) exposure.⁴²^,⁴³ On the contrary, in ducks, an increased richness was observed for Bacteroidetes and Firmicutes, but lower richness for Verrucomicrobia and Proteobacteria upon intake of 750 mg kg^–1 NaF for 28 days.³⁸

Fish models (common carp, zebrafish)⁵⁵^,⁵⁷ indicated that exposure to high doses of fluoride (ie, 80 mg L^–1 NaF for 90 days) can affect the abundance of Fusobacteria (synonym Fusobacteriota), Firmicutes, Verrucomicrobia, Proteobacteria, and Actinobacteria.

Overall, when comparing specific bacterial phyla and genera within the gut microbiota of animals treated with fluoride relative to the control, the most consistent findings were a lower abundance of Firmicutes and Verrucomicrobia³¹^,³⁵^,^42–44^,⁴⁶^,⁴⁷^,⁴⁹^,⁵⁰ followed by an increase in abundance of Bacteroidetes, Actinobacteria, Streptococcus, Staphylococcus, Desulfobacterota, and Proteobacteria³¹^,³⁶^,⁴²^,⁴⁹^,⁵¹^,⁵²^,⁷⁵^,⁷⁶ in the populations receiving high doses of fluoride (eg, 50-1200 mg L^–1) in the form of NaF. Some studies also showed an increase in the abundance of Firmicutes and a decrease in Bacteroidetes.¹⁰^,³²^,³⁷ Many genera were implicated across the control and treated groups, including the lower abundance of lactobacilli, Faecalibacterium, and Bifidobacterium³¹^,³²^,³⁶^,³⁷^,⁴⁰^,⁴⁶^,⁴⁹^,⁵⁰^,⁵² (important in maintaining intestinal function), as well as higher abundance of E coli.⁴² On the other hand, the included articles in this review also indicated that low fluoride doses (0.5-25 mg L^–1 NaF) might upregulate the abundance of probiotics (eg, Lactobacillus, Lachnospiraceae, bacilli, Lactobacillaceae) in animals.

In Vitro Models

All 3 in vitro studies used fluoride concentrations of 0.1-100 mM NaF. Findings indicated a biphasic response with fluoride-inducing bacterium growth at a low dose (0.1 mM NaF)⁷¹ and a dramatic increase in bacterium growth and enzyme production, especially of enolase. However, the growth of the microorganisms, especially lactobacilli, was inhibited as the dose increased from 10 mM to 100 mM NaF (Table 3).^71–73

Secondary Outcomes

Effect of Fluoride Dose and Duration on Microbiota-Associated Functions (Transcriptomic and Metabolomic Profile)

Human studies

The ex vivo human study observed metabolite concentrations after fluoride exposure. Both low (1-2 mg L^–1 NaF) and high doses (10-15 mg L^–1 NaF) of fluoride had no significant effects on the production of SCFAs¹⁴ (Table 3 and Table S6). The population-based studies revealed the depletion of the relative absence of functional genes associated with pentose and glucuronate interconversion (a step-in carbohydrate metabolism) in children with fluorosis. These findings were correlated with the increase in Acidobacteriota.⁷⁶ Acidobacteriota contain the genes that can degrade polysaccharides and are involved in pentose and glucuronate interconversion (carbohydrate metabolism). An abnormal increase or decrease in the expression of these genes can result in several diseases, such as familial tumoral calcinosis.⁸⁰^,⁸¹ Additionally, 3 tryptophan metabolites (namely, 5-hydroxy indole acetic acid, tryptamine, and indole acetaldehyde) that are directly produced by the gut microbiota were significantly decreased in adult patients with fluorosis. This decrease was correlated with the decreased ratio of Firmicutes to Bacteroidetes and increased abundance of Proteobacteria, Acidobacteriota, and Verrucomicrobiota.⁷⁵ Tryptophan metabolites are important for regulating host physiology and an aberrant decrease in their production is associated with IBD and colorectal cancer.⁸²

Animal studies

In vivo studies showed that fluoride intake by rodents of 100-150 mg L^–1 NaF for 1-6 months could lead to a decrease in levels of SCFAs¹⁰; digestive enzymes such as amylase, trypsin, and lipase⁴⁶; antioxidative enzymes such as superoxide dismutases (SODs); metalloenzymes such as catalase³²; glycoproteins³¹^,⁴⁹; and p62 proteins.⁴⁰ Fluoride could also affect hormones, and exposure to fluoride resulted in a decrease in glutathione (GSH), follicle-stimulating hormone, and gastrotropin levels.³⁴^,⁴⁰^,⁴⁶ Fluoride supplementation stimulated the production of proinflammatory cytokines interleukin (IL)-17A, IL-22, IL-1 , IL-6, tumor necrosis factor-a, and NF-_xB,¹⁰^,³⁷^,⁵⁶ malondialdehyde (MDA) levels, and secretory immunoglobulin A (SIgA) antibody levels,³²^,⁴⁶ which could be due to the disturbance in the Firmicutes to Bacteroidetes ratio (Table 3 and Table S6). However, the low dose of fluoride (15 and 45 mg L^–1 NaF; 0.57 mg L^–1 F35) decreased MDA and elevated SOD and GSH levels.⁶²^,⁶⁷

Similarly, a significant upregulation of mRNA expression of pro-inflammatory cytokines cathelin-related antimicrobial peptide (CRAMP), B-defensin-1, B-defensin-3, and light chain 3 was observed at high doses (80-100 mg L^–1 NaF) and long exposure of up to 70 days.¹⁰^,³⁷^,⁴⁰^,⁵⁶ A sharp decrease in mRNA expression levels of brain-derived neurotrophic factor (BDNF), cAMP-responsive element, stem cell factor, Bclxl, occludin, Zonula occludens-1, and claudin was observed under similar conditions.³⁷^,⁵⁰^,⁶³^,⁶⁴^,⁶⁶ However, a low fluoride dose of 15 mg L^–1 reduced the expression of NQO1 and Nrf2 genes, which are associated with the progression of the cell cycle and triggering immune responses, respectively.

Silkworm³⁹^,⁶⁸ and bird⁴²^,⁴³ models also showed a similar effect with a reduction in energy metabolic enzymes and SCFAs, respectively, at higher fluoride doses (200, 750, and 1200 mg L^–1 NaF), but the production of proinflammatory cytokines was accelerated (Table 3 and Table S6). Silkworms exposed to 200 mg L^–1 NaF had upregulated expression levels of antimicrobial peptides Att2, CecA, Beclin 1, Atg5, Atg7, p62, and Lys, and downregulated expression levels of CecB6, CecD, and Leb1were observed after 36 hours.⁴⁷^,⁶⁸ However, 4.76 Mm NaF inhibited E11 enzyme family genes and energy-synthesizing genes.⁷⁰ Birds exposed to 1200 and 400 mg L^–1 NaF for 59 days had downregulated expression levels of SIgA, mucin 2 (MUC2), Zonula occludens-2, and claudin-4.⁴²^,⁴³

In fish models, a study of common carp focused on the effect of fluoride (80 mg L^–1 NaF) for 90 days on tight junction protein expression. The authors reported a significant decrease in Zonula occludens-1 and occludin levels.⁵⁵ In zebrafish, the effects of the same dose were observed on immune-related enzymes but at different time points. After exposure for 30 days, there was an increase in the levels of MDA, SOD, and catalase but a decrease in GSH. After 60- and 90-day exposure, levels of reactive oxygen species, MDA, and myeloperoxidase were enhanced. However, the levels of catalase, GSH, acid phosphatase, and lysozyme were lowered and were associated with numbers of Firmicutes⁵⁴ (Table 3 and Table S6).

Discussion

Fluoride from food and water is initially absorbed in the stomach (30%) and converted into hydrogen fluoride. The pharmacokinetics of fluoride are largely governed by the pH of the stomach. Its absorption is reduced as the pH increases.²⁴^,⁸³ Hydrogen fluoride can easily diffuse across the cell membrane and gastric epithelium as fluoride ions and enter the gastrointestinal tract. Fluoride is almost completely and rapidly absorbed in the gastrointestinal tract harboring the gut microbiome, with a half-life of about 30 minutes. The gastrointestinal tract bacteria secrete acids and enzymes that assist in fluoride absorption.⁸⁴^,⁸⁵ The absorbed fluoride is transported via the blood to various organs of the body. The excess fluoride is excreted via urine. Unabsorbed fluoride in the gut is excreted through feces.⁸⁴^,⁸⁵

This systematic review shows the possible relationship between fluoride and gut microbiota composition and its associated outcomes at a concentration of >50 mg L^–1 NaF in animals for >1 month and >10 mg L^–1 NaF in humans for 24 hours. However, the studies included were substantially heterogeneous, with different model animals, biomarkers, and methods used to estimate gut microbial composition. Overall, all studies concluded that fluoride in suitable amounts (per the WHO recommendation for each species type) boosts gut microbial growth (probiotics) but, in excess, discourages microbial growth.

The most relevant changes affecting the main phyla of the intestinal microbial community after fluoride exposure include Firmicutes (Firmicuteota), Actinobacteria (Actinobacteriota) and Bacteroidetes (Bacteroidota), with either a decrease or an increase in relative abundance.

Shifts in alpha and beta diversity in gut microbiota

A variation in alpha and beta diversity in the fluoride group compared with the control group indicated that fluoride exposure posed a prominent influence on gut microbiota. However, the effect patterns in different studies varied, which might be due to different exposure levels of fluoride, biomarkers used, and the use of diverse animal models with different genetic makeup, sensitivity, and other characteristics such as age, sex, and weight. The differences in the alpha and beta diversities between the control and intervention groups indicated high doses of fluoride could affect species richness. Although the studies do not provide information on changes in the abundance of specific taxa, they give us access to a broader change or difference in the composition of microorganisms. Positive or negative health implications associated with changes in diversity depend on which species are present abundantly and which are suppressed.

Response of specific microbial members to Fluoride in human and animal studies

The majority of human and animal studies showed the effects of fluoride on Firmicutes (synonym Firmicuteota), Bacteroidetes (synonym Bacteroidota) and Actinobacteria (synonym Actinobacteriota). Associated genera were also implicated across the control and treated groups, including the lower abundance of Lactobacilli, Faecalibacterium, and Bifidobacterium³¹^,³²^,³⁶^,³⁷^,⁴⁰^,⁴⁶^,⁵⁰^,⁵² (important in maintaining intestinal function), as well as higher abundance of E coli.⁴² Lactobacilli reinforce the intestinal barrier while inhibiting pathogens’ growth by producing metabolites.⁸⁶ Bifidobacterium and Faecalibacterium maintain the immune system and help maintain T-helper cell and regulatory T-cell balance; a disturbance leads to changes in cellular expression, a potential cause of intestinal inflammation and associated disorders.⁸⁷ Verrucomicrobia include intestinal mucosal bacteria that promote intestinal health and glucose homeostasis and play a role at an interface between the human gut microbiome and host tissues.⁸⁸ Actinobacteria can survive in toxic environments. The opportunistic pathogens of Proteobacteria can cause a major structural imbalance of gut microbiota. At the same time, fusobacteria have been widely recognized as potential inducers of T regulatory cells or carcinogens promoting autophagic activation.⁸⁸ Desulfobacterota are sulphur-reducing bacteria associated with the activation of immune response and inflammation. They can degrade butyrate (an SCFA), which inhibits inflammation. The abrupt increase or decrease in the abundance of members of the Firmcuteota, Verrucomicrobiota Actinobacteriota, Desulfobacterota, Proteobacteriota, and lactobacilli, as well as species of Faecalibacterium and Bifidobacterium) can reduce the production of useful metabolites (eg, SCFAs) involved in maintaining intestinal homeostasis and integrity by controlling the defense mechanism. A deficiency of these metabolites can lead to pathological conditions such as colitis, IBD, and even colorectal cancers.⁸⁹^,⁹⁰ Ruminococcaceae produce butyrate that regulates mucosal immunity and is the best indicator of healthy anaerobic gut microflora. Butyrate limits the production of proinflammatory cytokines like tumor necrosis factor-a, thus reducing the risk of inflammatory bowel disease (IBD).⁹¹

Correlation between the metabolic changes and gut microbiota composition in response to Fluoride

The current review found that the metabolic changes in the participants were closely related to changes in microflora composition after fluoride intake. The increased expression of proinflammatory cytokines and antimicrobial peptides followed by a decrease in antioxidative enzymes and tight junction proteins, for example, can be described by a lower abundance of lactobacilli, Bifidobacterium, Clostridium, and Faecalibacterium, which was responsible for decreased SCFAs levels—specifically, butyrate content that maintains balance in the production of proinflammatory cytokines. These changes resulted in increased expression of proinflammatory cytokines, such as IL-17A and IL-22,¹⁰^,³⁷ and mucosal sIgA antibody,³⁴^,³⁶ that further led to the production of antimicrobial peptides such as CRAMP, B-defensin-1, and B-defensin-3, consequently leading to inflammatory diseases like IBD.¹⁰ The changes also were associated with a reduction in levels of glycoproteins and antioxidant enzymes such as SOD, catalase, and GSH, which may cause colorectal cancer and is linked with impairment in the integrity of the intestinal barrier.¹⁰^,³⁷^,⁴⁹^,⁵⁴ Additionally, a decrease in the mRNA levels of BDNF, cAMP-responsive element, and stem cell factor were observed, which are an important modulators of neuroplasticity, thus leading to neurodevelopmental disorders.⁵⁰ IL-17A and IL-22 are cytokines often produced together at high levels in inflamed tissues. Cytokines provide a protective inflammatory response, but dysregulation can cause autoimmunity and allergies.⁹² Increased cytokine production modulates the secretion of B-defensins that regulate innate immunity, thus inducing inflammatory mediators. B-Defensins maintain the balance between safeguarding from pathogens and tolerance to normal flora. Nevertheless, attenuated expression triggers the activation of transcription factor NF-xB and, in turn, activates the inflammatory pathways.⁹³ Mucosal sIgA antibody serves as the first line of defense against pathogens, and its increased production is an indicator of an existing underlying disease.⁹⁴ The antioxidant enzymes such as SOD protect the cell from oxidative stress, CAT breakdown, or decompose hydrogen peroxide molecules into water, thereby protecting cells and organs from damage caused by oxygen-free radicals. Glutathione is involved in tissue repair. The reduced expression of these enzymes might be due to the reduced levels of probiotics, especially the lactobacilli and Bifidobacterium, which produce SCFAs, that play a key role in reducing oxidative stress by activating the transcription of antioxidative pathways.⁹⁵^,⁹⁶ Similarly, several studies showed that imbalances in the proportion of certain bacteria, such as Clostridium sensu stricto, Romboutsia, Lachnospiraceae, Desulfovibrio, and 2 unidentified genera in the Muribaculaceae and Ruminococcaceae families, after fluoride exposure increased the circulating levels of bacterial metabolites such as trimethylamine-N-oxide, and an upregulation of mRNA expression levels of pro-inflammatory cytokines, CRAMP, B-defensin-1, B-defensin-3, and light chain 3, which were strongly correlated with the higher risk of cardiovascular events. Desulphovibrio spp are associated with the production of propionic acid and hydrogen sulfide, a probable mitochondrion toxin, resulting in both neurotoxic and colonotoxic effects.¹⁰ Furthermore, an increase in the abundance of Parabacteroides increased the expression levels of Beclin 1, light chain 3, autophagy protein 5, and p62 that not only induced autophagy but also altered the levels of luteinizing hormone and follicle-stimulating hormone, thus leading to lower testicular organ functioning and nephrotoxicity.⁴⁰^,⁴⁶

Most of the discussed studies also reported that fluoride may induce obesity and colitis along with IBD. This can be described by the disturbances in the Firmicuteota to Bacteroidota ratio. The abrupt increase or decrease in the abundance of Firmicuteota or Bacteroidota can lead to obesity and colitis.⁴¹^,⁴⁶ Fluoride intervention increased the relative abundance of Firmicuteota and decreased the Bacteroidota population, leading to the elevated Firmicuteota to Bacteroidota ratio. This disturbance can promote the growth of other opportunistic bacteria, such as Proteobacteria, and downregulate the abundance of SCFA-producing bacteria such as members of the Ruminococcaceae, a better indicator of healthy gut microbiota⁴¹^,⁴⁶ and Turicibacter. This disturbance causes the suppression of SCFAs such as butyrate, propionate, and acetate that play an important role in prevention of obesity and colitis by interacting with adipose tissues via G-protein coupled receptors (GPRs) expressed in adipocytes. The upregulation of GPR41 and GPR43 promotes adipocyte formation and inhibits lipolysis, thus inhibiting colitis.⁹⁷ Furthermore, Turicibacter controls the host bile acid profile and is considered a modulator of host fat biology and of glucose and energy metabolism. The depletion of Turicibacter might be associated with obesity and colitis and correlated with decrease in butyrate levels.⁴¹^,⁴⁶

Factors Affecting Exposure Effects of Fluoride

Effects of Fluoride Dose

The dose of fluoride through diet or water has a discerning effect on the stimulation or inhibition of gut microbiomes. Three-fourths of the studies included in this systematic review used high doses of fluoride (>50 mg L^–1 NaF for animals, and 10-15 mg L^–1 NaF for humans). Nevertheless, 1 human (1-2 mg L^–1 NaF) and 4 animal (0.5, 4, 5, and 10 mg L^–1 NaF in water) studies³⁴^,⁴⁴^,⁴⁸^,⁵³ also used applicable doses and observed an increase in the growth of probiotics like lactobacilli and Bifidobacterium. However, when the fluoride supplemental doses were acutely higher, the abundance of pathogenic microorganisms increased, and health-promoting microbial species decreased. Numerous in vitro and in vivo studies also showed dose-dependent effects of fluoride supplementation on gut microbiota compositional changes.¹⁰^,³¹^,³⁴^,³⁵^,⁴¹^,⁴²^,⁴⁶^,⁴⁸^,⁵⁴^,^71–73 Overall, doses of fluoride higher than 25 mg L^–1 NaF in animals and 2 mg L^–1 NaF in humans could disturb microbial homeostasis, which could lead to the upregulation of proinflammatory metabolites expression and downregulation of antioxidative enzymes and junction proteins, thus triggering immune responses and intestinal integrity.

Identifying the crossroads of fluoride dose for these microbes is necessary to reveal underlying biological functions beneficial for health. From the current literature, a fluoride dose of < 50 mg L^–1 for animals and < 10 mg L^–1 NaF for humans was recommended to stimulate the abundance of health-promoting strains and decrease the abundance of possible pathogenic strains without affecting other crucial microbes. However, determining the turning-point dose value of fluoride precisely is challenging, due to varied exposure from diet and drinking water, in addition to the diversity in gut barrier structure and fluoride metabolism of the animals and humans in the included studies.

Effects of Fluoride’s Duration of Exposure

Greater than 70% of the animal studies focused on long-term exposure to high concentrations of fluoride; however, some in vitro studies (8%) examined the acute effects of fluoride on the microbiome after short-term exposure (24-36 hours).³⁵^,⁶⁸^,⁶⁹^,⁷¹ Although the impacts of long-term exposure and low concentrations are not understood, exposure to doses > 25 mg L^–1 NaF in water for longer than 1 month in animals, and > 2 mg L^–1 NaF for > 1 day in humans can lead to severe health conditions. However, the results are not uniform among the studies, due to the diversity of animal species, different biomarkers used for measuring the outcomes, and varying quality of studies.

Limitations and Knowledge Gaps

This systematic review disentangles the effects of fluoride on gut microbiota and its associated functions with good reliability, due to high-quality assessment scores for most studies and specificity due to the focus on gut microbiota within the inclusion criteria.

However, there are a limited number of human studies exploring the effects of fluoride on the gut microbiome, making it difficult to draw conclusions. Most of the included studies were animal based and used very high doses of fluoride (25 mg L^–1 NaF in water) compared with the safe intake recommended by WHO (0.8–1.7 mg L^–1 NaF in water) in humans. Though animals are a useful tool for assessing the impacts of fluoride, extrapolating findings to humans can be difficult due to differences in fluoride sensitivity, intestinal barrier functions, and metabolic rates. Additionally, including all eligible studies led to high inconsistency among the reported outcomes, thus preventing pooling of results.

The mechanisms linking changes in bacterial composition to metabolic dysfunction have not been confirmed, meaning the exact process by which fluoride intervention modulates the gut microbiota and alters SCFAs remains unknown. To further investigate, individual in vitro analyses that includes minimum inhibitory concentrations, minimum bactericidal concentrations, and time-kill curves of strains found in the gut under different fluoride concentrations are needed. The strains under different treatments can be subjected to RNA sequencing and mass spectrometry to identify which genes and proteins are expressed in association with fluoride exposure. This would also help identify the individual response of strains and to obtain a turning-point dose and duration of fluoride exposure. Furthermore, considering the sensitivity of the research area, careful examination of fluoride’s impact in water and diet at low intakes for a short term and long term in human models must be conducted. To better assess the connection, follow-up studies should be considered to investigate whether fluoride-linked microbiome change directly affects the expression of metabolites and human health.

CONCLUSIONS

Current analysis shows that fluoride intervention either in vivo or in vitro may change the abundance of gut microbiota and its associated activities. Fluoride at low levels (<2 mg L^–1 NaF in humans and <25 mg L^–1 NaF in animals) did not affect gut microbiota. Although effects at high doses are inconsistent, impacts could be seen in overall microbial diversity change, change in relative abundance of specific taxa, and changes in the metabolism of present microbiota. Shifts in any of these aspects of the microbial community can lead to health implications. Therefore, more fundamental studies are needed to fully understand the impacts of fluoride at low doses, but for a long duration, as might be expected to be found in diet and water supplies, on key gut microbes such as the lactobacilli and Bifidobacterium. More studies are needed to determine the ideal concentration of fluoride for supplementation, in addition to assessing the durability of this effect. These recent findings should stimulate discussions on the safe use of fluoride in food, water, and dietary supplements.

Author Contributions

M.Y., C.H.O., F. V.Z., and P.D. conceptualized the review; C.H.O., F.V.Z., and P.D. supervised the work; M.Y., C.H.O., F.V.Z., P.D., E.A.K., and M.S. curated the data; M.Y., E.A.K., C.H.O., and F.V.Z. analyzed the data; M.Y. and C.H.O. wrote the original draft of the manuscript; all authors curated the data, reviewed and edited the manuscript, and read and agreed to the published version of the manuscript.

Supplementary Material

Supplementary Material is available at Nutrition Reviews online.

Funding

This work was funded by a Teesside University PhD Studentship.

FULL-TEXT STUDY WITH REFERENCES ONLINE AT

https://academic.oup.com/nutritionreviews/article/83/7/e1853/8068784?login=false

Research Studies

Study Tracker

Effect of Fluoride on Gut Microbiota: A Systematic Review.

View PDF

Abstract