Fluoride Action Network



Fluoridation of drinking water has been a subject of controversy for decades. Over the past 50 years, the incidence of dental caries (cavities) has declined considerably in the United States, an important health advance that most scientists attribute principally to increased access to fluoridated water and dental products. According to the U.S. Centers for Disease Control and Prevention, approximately 132 million Americans now receive drinking water that contains fluoride, either naturally occurring or added, at concentrations of 0.7 milligrams per liter (mg/L) or higher. Since the 1960s, the U.S. Public Health Service (PHS) has recommended an ”optimal” fluoride concentration of 0.7-1.2 mg/L to prevent dental caries and minimize dental fluorosis. However, there has been an increase in the prevalence of dental fluorosis a mottling of tooth enamel that ranges from barely discernible enamel flecks in its mildest forms to staining and pitting in its severest forms; the severest forms are rare in the United States. EPA considers dental fluorosis to be a cosmetic effect and not an adverse health effect.

Recent findings have renewed long-standing concerns of those who oppose water fluoridation, claiming that ingestion of fluoride can lead to a variety of unwanted effects. One animal study reported an equivocal increase in osteosarcomas (malignant bone tumors) in male rats, but not in female rats, at very high concentrations (100-175 mg/L). However, that result was not substantiated in a subsequent study in rats at even higher doses.

In the United States, a few cases of crippling skeletal fluorosis (an increase in the thickness of bones, which in advanced stages, is associated with decreased mobility of joints and pain) have been reported in humans only when fluoride concentrations in drinking water exceeded 8 mg/L over many years. That concentration is much higher than the PHS recommendation of 0.7-1.2 mg/L. To prevent skeletal fluorosis, the U.S. Environmental Protection Agency (EPA) set the maximum contaminant level (MCL) for fluoride at 4 mg/L of drinking water. In addition to fluoride in drinking water, however, people also can ingest fluoride in toothpaste, mouth rinse, and dietary fluoride supplements or in beverages and foods prepared with fluoridated water. As a result, many Americans might ingest more “incidental” fluoride than was anticipated by PHS and EPA in recommending standards for drinking water.

EPA responded to the concerns by requesting that the National Research Council’s Board on Environmental Studies and Toxicology (BEST) review the current toxicological and exposure data on fluoride and determine whether EPA’s current MCL of 4 mg/L is acceptable for protecting the public from potential adverse health effects of fluoride. EPA also asked BEST to identify gaps in the fluoride toxicity data and to make recommendations for future research.

In response to EPA’s request, BEST’s Committee on Toxicology established the Subcommittee on Health Effects of Ingested Fluoride. The subcommittee based its evaluation on a detailed examination of current data in the following areas:

• Intake, metabolism, and disposition of fluoride.
• Dental fluorosis.
• Bone strength and the risk of bone fracture.
• Effects on the renal, gastrointestinal, and immune systems.
• Reproductive effects in animals.
• Genotoxicity.
• Carcinogenicity in animals and humans.

This report deals with the possible toxic effects of ingested fluoride in humans. It does not attempt to weigh fluoride’s well-documented health benefits against its possible adverse health effects.


The major sources of fluoride intake are water and other beverages, food, and fluoride-containing dental products. Groundwater, a major source of drinking water in many communities, can contain fluoride at concentrations ranging from less than 0.1 mg/L to more than 100 mg/L in some parts of the world (e.g., Africa and China). Coffee, tea, and soft drinks made with that water might have corresponding concentrations of fluoride. Concentrations in food depend on the levels in the soil, but they can increase or decrease depending on the fluoride concentrations in the water used in food preparation. Dental products available in the United States and meant to be used topically (and not ingested) contain fluoride at concentrations ranging from 230 to over 12,000 parts per million (ppm).

The average intake of dietary fluoride by young children whose water supply contains fluoride at 0.7-1.2 mg/L is approximately 0.5 mg per day, although substantial variation occurs. Because it is associated with optimal fluoridation, this level of fluoride intake (0.5 mg per day, or 0.04-0.07 mg per kilogram (kg) of body weight per day) generally has also been accepted as optimal. Nursing infants receive negligible fluoride from breast milk, but intake from formulas can range from less than 0.4 to over 1.0 mg per day when reconstituting powder concentrate with fluoridated water, a range that can exceed the optimal level. Dietary intake by adults in areas with fluoridated water has been estimated at 1.22.2 mg per day, although intake by some individuals, such as outdoor laborers in warm climates or those with urinary disorders, can be considerably higher.

Accurate estimates of fluoride exposure cannot be based simply on the concentration of fluoride in drinking water. For example, dental products can be an important source of ingested fluoride, especially for young children who have poor control over the swallowing reflex. Even for older children, intake from toothpaste and mouth rinse can still equal the daily intake from food, water, and other beverages. Soft drinks, which often contain fluoride at 0.3-0.5 mg per 12-ounce serving, are an important additional source of fluoride. Studies of dental fluorosis indicate that children’s exposure to fluoride has increased since the 1970s.

Approximately 75-90% of the fluoride ingested each day is absorbed from the alimentary tract. Because of its chemical affinity for calcium compounds, about half of that fluoride becomes associated with teeth and bones within 24 hours of ingestion. In growing children, even more of the ingested fluoride is retained because of the large surface area provided by numerous and loosely organized bone crystallites. The remaining fluoride is eliminated almost exclusively by the kidneys, and the rate of renal clearance is directly related to urinary pH. As a result, diet, drugs, metabolism, and other factors can affect the extent to which fluoride is retained in the body.

Recommendations for further research are (1) to determine and compare intake of fluoride from all sources (this recommendation has implications for research design in several of the areas that follow); and (2) to determine the metabolic characteristics of fluoride in infants, young children, and the elderly, as well as in patients with progressive renal disease.


Fluoride prevents tooth decay by enhancing the remineralization of enamel that is under attack, as well as inhibiting the production of acid by decaycausing bacteria in dental plaque. Fluoride is also a normal constituent of the enamel itself, incorporated into the crystalline structure of the developing tooth and enhancing its resistance to acid dissolution. One side effect of too much fluoride ingested 0.7-1.2 mg/L, was designed to maximize prevention of dental caries while
limiting the prevalence of dental fluorosis to about 10% of the population, virtually all of it mild to very mild.

A 1991 report from PHS of the U.S. Department of Health and Human Services compiled the results of independent investigations conducted during the 1980s on dental fluorosis in 24 cities and compared them with a series of PHS surveys conducted during the late 1930s and early 1940s in 21 cities. That comparison showed that the prevalence of dental fluorosis, most of it very mild to mild, had increased. The 1980s data showed that the mean prevalence of dental fluorosis in four cities with optimally fluoridated water supplies was around 22% (17% very mild, 4% mild, 0.8% moderate, and 0.1% severe). In another city with a water fluoride concentration in the range of 1.8-2.2 mg/L, dental fluorosis prevalence was 53% (23% very mild, 17% mild, 8% moderate, and 5% severe). In two other cities with water fluoride concentrations greater than 3.7 mg/L, prevalence was around 84% (25% very mild, 27% mild, 19% moderate, and 14% severe). The data in the PHS report also showed that the greatest relative increase in fluorosis prevalence since the early studies was in communities with very low water fluoride concentrations, demonstrating the influence of sources of fluoride other than water. Those sources make it difficult to estimate fluoride exposure; they represent a source of possible error in estimating fluoride intake in studies of the relation between fluoride exposure and dental fluorosis. Moreover, there is disagreement on whether dental fluorosis (even moderate-to-severe dental fluorosis, in which substantial tooth enamel is affected and dental treatment might be required) is a cosmetic problem or an adverse health effect.

In general, the evidence supports the conclusion that fluoridation at the recommended concentrations, in the absence of fluoride from other sources, results in a prevalence of mild-to-very-mild (cosmetic) dental fluorosis in about 10% of the population and almost no cases of moderate or severe dental fluorosis. At 5 or more times the recommended concentration, the proportion of moderate-to-severe dental fluorosis is substantially higher. The most effective approach to controlling the prevalence and severity of dental fluorosis, without jeopardizing the benefits of fluoride to oral health, is likely to come from more judicious control of fluoride in foods, processed beverages, and dental products, especially those items used by young children.

Recommendations for further research are to identify sources of fluoride during the critical stages of tooth development in childhood and evaluate the contribution of each source to dental fluorosis. Further research should be aimed at the goal of minimizing exposure to fluoride concomitant with maintaining effectiveness in preventing caries.


The effect of fluoride on bone strength, hip fractures, and skeletal fluorosis in humans has been addressed in two types of studies. The first type involves clinical trials of the effectiveness of high concentrations of fluoride supplements in strengthening bones and preventing further fractures in patients with osteoporosis; this treatment has been used primarily in Europe for almost 30 years. When conducted using proper control groups, these studies showed little or no benefit even at dosages of 20-32 mg per day, well over 10 times the exposure from fluoridated drinking water. If anything, the treated groups experienced a greater number of new fractures, including painful stress fractures in bones other than the vertebrae.

The second type of human study involves epidemiological investigations. These studies compared the rate of bone fracture in populations of the elderly that differed in their exposure to natural or added fluoride in drinking water. Geographic and time-trend analyses were made; time-trend analysis is considered the stronger methodology because there is less opportunity for confounding by other risk factors. Of the six epidemiological studies that used geographic comparisons (where no actual intake data were available), four found a weak association between fluoride in drinking water and the risk of hip fracture. Two additional studies examined time trends in bone fracture before and after water fluoridation: one found no association and the other a negative association. Only two additional studies collected information on individual exposure: one (essentially a geographic comparison) found an increased risk of hip fracture at water fluoride concentrations of 4 mg/L, and the other observed no difference in risk.

Studies with several species of experimental animals have yielded various outcomes. Most of the studies indicated little or no effect on bone strength, even with very high fluoride intake and very high concentrations of fluoride in bone. The subcommittee identified many potential problems in the experimental design of the animal studies, including the lack of suitable control groups with reasonably low fluoride exposures. However, the subcommittee concluded that the weight of evidence indicates that bone strength is not adversely affected in animals that are fed a nutritionally adequate diet unless there is long-term ingestion of fluoride at concentrations of at least 50 mg/L of drinking water or 50 mg/kg in diet.

In view of the conflicting results and limitations of the current data base on fluoride and the risk of hip or other fractures, the subcommittee concludes that there is no basis at this time to recommend that EPA lower the current standard for fluoride in drinking water for this end point. However, the subcommittee recommends additional research to improve the current data base.

A recommendation for further research is to conduct additional studies of hip and other fractures in geographical areas with high and low fluoride concentrations in drinking water and to make use of individual information about water consumption. These studies should also collect individual information on bone fluoride concentrations and intake of fluoride from all sources, as well as reproductive history, past and current hormonal status, intake of dietary and supplemental calcium and other cations, bone density, and other factors that might influence risk of hip fracture.


Renal excretion is the major route of elimination for inorganic fluoride from the body. As a result, kidney cells are exposed to relatively high fluoride concentrations, making the kidney a potential site for acute fluoride toxicity. Animal studies have shown that very high water fluoride concentrations of 100-380 mg/L can lead to necrosis of proximal and renal tubules, interstitial nephritis, and dilation of renal tubules. However, human epidemiological studies have found no increase in renal disease in populations with long-term exposure to fluoride at concentrations of up to 8 mg/L of drinking water.

The subcommittee concludes that available evidence shows that the threshold dose of fluoride in drinking water for renal toxicity in animals is approximately 50 mg/L. The subcommittee therefore believes that ingestion of fluoride at currently recommended concentrations is not likely to produce kidney toxicity in humans.


In the acid environment of the stomach, fluoride and hydrogen ions can combine to form hydrogen fluoride, which, at sufficiently high concentrations, can be irritating to the mucous membranes of the stomach lining. Experimental studies with several animal species have shown dose-dependent adverse effects, such as chronic gastritis and other lesions of the stomach, at fluoride concentrations of 190 mg/L and higher. Reports of gastrointestinal effects in humans often involve workers exposed to unknown concentrations of fluoride in the workplace, so that the contribution of fluoride exposure to the risk of adverse health effects is unknown. The subcommittee noted that these workers could also be exposed to other toxic substances present in the work environment. There have been few studies of the gastrointestinal effects of fluoride at low concentrations.

The subcommittee concludes that the available data show that the concentrations of fluoride found in drinking water in the United States are not likely to produce adverse effects in the gastrointestinal system.


Few animal and human data on sodium fluoride-related hypersensitivity reactions are found in the literature. In animal studies, excessively high doses, inappropriate routes of administration of fluoride, or both were used. Thus, the predictive value of those data, in relation to human exposures at accepted exposure levels, is questionable. Reports of hypersensitivity reactions in humans resulting from exposure to sodium fluoride are mostly anecdotal.

The literature pertaining to immunological effects of fluoride is limited. Although direct exposure to high concentrations of sodium fluoride in vitro affects a variety of enzymatic activities, the relevance of the effects in vivo is unclear. Standardized immunotoxicity tests of sodium fluoride at relevant concentrations and routes of administration have not been conducted. The weight of evidence shows that fluoride is unlikely to produce hypersensitivity and other immunological effects.


There have been reports of adverse effects on reproductive outcomes associated with high levels of fluoride intake in many animal species. In most of the studies, however, the fluoride concentrations associated with adverse effects were far higher than those encountered in drinking water. The apparent threshold concentration for inducing reproductive effects was 100 mg/L in mice, rats, foxes, and cattle; 100-200 mg/L in minks, owls, and kestrels; and over 500 mg/L in hens.

Based on these findings, the subcommittee concludes that the fluoride concentrations associated with adverse reproductive effects in animals are far higher than those to which human populations are exposed. Consequently, ingestion of fluoride at current concentrations should have no adverse effects on human reproduction.


Fluoride has been tested extensively for its genotoxicity. It does not damage DNA or induce mutations in microbial systems, but it has produced mutations and chromosomal damage in several in vitro tests with mammalian cells. Sodium fluoride, in particular, inhibits protein and DNA synthesis and has been reported to cause chromosomal aberrations in human cells. The lowest effective dose in these cell-culture studies was a fluoride concentration of approximately 10 ug/mL, whereas the normal concentration in human plasma is 0.02-0.06 ug/mL, even in areas where drinking water is fluoridated, which means that there is a large margin of safety.

Sodium fluoride and other fluoride salts also have been tested for genotoxicity in the fruit fly Drosophila, as well as in mice and rats. The subcommittee’s review of the results of these in vivo studies was inconclusive, however, because of differences in protocols and insufficient detail to support a thorough analysis. There are no published studies on the genetic or cytogenetic effects of fluoride in humans.

The subcommittee concludes that the genotoxicity of fluoride should not be of concern at the concentrations found in the plasma of most people in the United States.


More than 50 epidemiological studies have examined the relation between fluoride concentrations in drinking water and human cancer. Most studies compared geographic or temporal patterns of cancer occurrences with distributions of fluoride in drinking water. These studies provide no credible evidence for an association between fluoride in drinking water and the risk of cancer. The existence of such an extensive epidemiological data base on fluoride with no consistent evidence of carcinogenic effects suggests that, if there is any increase in cancer risk due to exposure to fluoride, it is likely to be small. However, most of these studies used geographic and temporal comparisons of cancer rates and hence are of limited sensitivity. Further analytical studies with accurate information on individual fluoride exposures and disease diagnoses are therefore desirable.

The subcommittee also reviewed the literature on the potential carcinogenic effects of fluoride in animals. Although the results of earlier animal studies were largely negative, the studies were not conducted using current bioassay techniques and are thus of limited value. The subcommittee placed greater weight on two recent studies. The first, conducted by the National Toxicology Program (NTP), administered fluoride at concentrations of up to 175 mg/L of drinking water. Although the results were negative for male and female mice and female rats, there was some evidence of a dose-related increase in the incidence of osteosarcomas in male rats. However, these results were not confirmed by a second study conducted by Procter & Gamble, in which fluoride was administered in the diet at doses higher than those in the NTP study. The Procter & Gamble study did produce a significant dose-related in crease in the incidence of osteomas (benign bone tumors) in male and female mice. However, these lesions were not considered to be neoplastic and, in any event, have no known counterpart in human pathology.

The subcommittee concludes that the available laboratory data are insufficient to demonstrate a carcinogenic effect of fluoride in animals. The subcommittee also concludes that the weight of the evidence from the epidemiological studies completed to date does not support the hypothesis of an association between fluoride exposure and increased cancer risk in humans.

Nonetheless, the subcommittee recommends conducting one or more carefully designed analytical epidemiological (case-control or cohort) studies to more fully evaluate the relation between fluoride exposure and cancer, especially osteosarcomas, at various sites, including bones and joints. In conducting such studies, it is important that individual exposure to fluoride from all sources be determined as accurately as possible.


Based on its review of available data on the toxicity of fluoride, the subcommittee concludes that EPA’s current MCL of 4 mg/L for fluoride in drinking water is appropriate as an interim standard. At that level, a small percentage of the U.S. population will exhibit moderate or even severe dental fluorosis. However, the question of whether to consider dental fluorosis a cosmetic effect or an adverse health effect and the balancing of the health risks and health benefits of fluoride are matters to be determined by regulatory agencies and are beyond the charge or expertise of this subcommittee.

The subcommittee found inconsistencies in the fluoride toxicity data base and gaps in knowledge. Accordingly, it recommends further research in the areas of fluoride intake, dental fluorosis, bone strength and fractures, and carcinogenicity. The subcommittee further recommends that EPA’s interim standard of 4 mg/L should be reviewed when results of new research become available and, if necessary, revised accordingly.