Potassium fluoride

Endpoint summary

Description of key information

Fluoride is not an essential element for humans, but it has been shown that fluoride is beneficial in the prevention of dental caries and also supporting bone strength at moderately high levels. However, effects on the skeletal system have been identified as the key, leading to adverse effects caused by excessive exposure to fluorides, and include reduced bone strength, increase risk of fractures and/or skeletal fluorosis (stiffness of joints, skeletal deformities).
Daily intakes below ca. 14 mg F/day can be considered safe for humans, based on a weight-of-evidence approach. To protect the workforce, a workplace exposure limit (=DNEL) has been established at 1 mg F/m³. 
The DNEL of 1 mg F/m³ is adequately protective, even when considering that workers can - in addition to the inhalation route - be further exposed to fluorides via other routes, i. e. dermal and/or oral (hand-to-mouth-transfer) at the workplace, and also due to a background intake via diet, drinking water and dental care products.
In addition to maintaining workplace air concentration below the DNEL of 1 mg F/m³, urine biomonitoring is also recommended to control worker exposure, using the following threshold values:
4.0 mg F / g creatinine, when sampled before the work shift, and/or
7.0 mg F / g creatinine, when sampled after an exposure period (e.g. end-of-shift).

Additional information

Introductory remarks and read-across justification

This CSR is for the substance potassium fluoride (KF), which is a readily soluble salt. It is common practice in the scientific and regulatory community to assess the physiological function and the toxicity of inorganic fluorides in general. Several national and international bodies have assessed fluoride in the past, for example the US DHHS (1991), SCOEL (1998), WHO (2002), MAK (2006 with update 2007), EFSA (2006). The two most recent reviews by MAK (focussed on worker exposure via inhalation) and EFSA (focussed on oral exposure for the general population) have been considered extensively in the preparation of this CSR chapter on repeated dose toxicity and the resulting derivation of DNELs (see also Chapter 5.11 below).

Read-across: particularly in animal studies, the most common inorganic fluoride salt sodium fluoride is used as test item. Other soluble salts of fluoride, such as potassium fluoride, ammonium fluoride or also potassium hydrogen difluoride are also used occasionally. Unrestricted read-across is possible between these substances with regards to systemic toxicity, which is of key concern for fluorides. All mentioned fluorides are readily soluble (water solubilities of several tens or hundreds of grams per litre) and release the fluoride ion F^-. The respective counter-ions are not assumed to contribute significantly to any toxicity.

A Hägg-Diagram shows the speciation (as molar fractions) of the HF/F- system. Under neutral, physiologically relevant conditions, F- is the only relevant species.

<diagramm attached to technical dossier as jpg file in endpoint summary for section 7.5>

 

When assessing the toxicity of fluorides, effect levels such as NOAELs or other dose descriptors are often expressed in terms of “fluoride”, e.g. mg F/kg/day or mg F/m³. However, under REACH the DNELs ultimately need to be reported on a substance basis. Care has been taken throughout this chapter to clearly express the units of reported figure, for example as mg F/kg/day or as mg KF/m³. The atomic weight of F is 18.998, the atomic weight of K is 39.098 and the molecular weight of KF is 58.097. Thus, the conversion factor from F to KF is 58.097/18.998 = 3.058, which is rounded to 3.

Dietary intake of fluoride via drinking water and food

A baseline intake of fluoride is unavoidable because of the ubiquitous presence of fluoride in the environment and thus in food and drinking water. Further sources are fluoride-containing dental products and dietary supplements (incl. fluorinated table salt). EFSA (2005) has summarised fluoride intake from these sources as follows:

“The total daily intake of fluoride from all sources can range from the low intake of 0.5 mg/day from solid foods, milk, beverages and low-fluoride water reported for Germany, when no fluoridated salt is used, no fluoride containing dentifrice is used and no supplements are taken, to the moderate amount of 1.2 mg/day reported for the United Kingdom. If fluoridated salt would be used 0.5-0.75 mg fluoride would be added, if fluoridated water was drunk (1 mg/L) and used for the preparation of food and tea (1-2 L of water/day; 500 mL of tea with a fluoride concentration of 5 mg/L) 3.5 to 4.0 mg fluoride would be added. The sum could be 6.0 mg fluoride per day, without fluoride from toothpaste taken into account. Even more extreme scenarios are possible and not completely unrealistic, when fluoridated drinking water is replaced by the regular use of mineral water with fluoride concentrations above 1 mg/L.”(Source: EFSA (2005), references removed).

Additionally, MAK (2005) cite Bergmann and Bergmann (1995) who reported that an adult in Germany consumes 0.7 - 1.2 mg fluoride per day via drinking water and food, including fluorinated table salt.

 

Toxicity of fluorides (excess intake)

The effects of excessive fluoride intake have been studied extensively in animal studies, as well as in investigations involving human populations. For the sake of clarity and in consideration of several existing reviews in this regards, only the following key data is summarised and discussed below:

(i) animal studies that are suitable to fulfil the formal data requirements for repeated dose toxicity under REACH and

(ii) relevant studies involving humans

 

(i) animal studies

Subacute Toxicity:

A 28-day repeated oral toxicity study (Proctor & Gamble, Co., 1991) was conducted to characterize the systemic toxicity of sodium fluoride. Weanling male and female Sprague-Dawley rats were dosed by gavage at concentrations of 2.5, 25 or 250 ppm sodium fluoride. The NOEL was 25 ppm. The following effects were observed at 250 ppm; hematology - significant depression in mean cell volume of males and females and mean cell hemoglobin of males; clinical chemistry - significant decrease in total protein of males and females and significant increase in alanine aminotransferase, potassium and chloride of females only; several changes in the mineral analysis of teeth and bone; and increased absolute and relative stomach weights.

Subchronic Toxicity:

The U. S. National Toxicology Program (NTP, 1990) evaluated the toxicological effects of continuous exposure to 0, 30, 100 or 300 ppm sodium fluoride in drinking water on F344 male and female rats for a 6-month period. Sodium fluoride caused weight loss at 300 ppm, fluorosis of the teeth at 100 and 300 ppm, minimal hyperplasia of the gastric mucosa of the stomach at 100 and 300 ppm (however, one high dose rat of each sex had an ulcer), a dose-related increase in fluoride content of bone and urine with increasing fluoride concentration in the drinking water, and a significant increase in fluoride content in the plasma at 300 ppm. No significant signs of toxicity were observed at concentrations of 10 or 30 ppm.

The U. S. National Toxicology Program (NTP, 1990) evaluated the toxicological effects of continuous exposure to 0, 10, 50, 100, 200, 300 or 600 ppm sodium fluoride in drinking water of male and female B6C3F1 mice for a 6-month period. Sodium fluoride caused death in some animals at 600 ppm and in a single male animal at 300 ppm, weight loss at 200 to 600 ppm, fluorosis of the teeth at 100, 200, 300 and 600 ppm, acute nephrosis and/or lesions in the liver and myocardium in mice that died early, minimal alterations in bone growth/remodeling in the long bones at 50 to 600 ppm, a dose-related increase in fluoride content of bone and urine with increasing fluoride concentration in the drinking water, and a possible dose-related increase in fluoride content in the plasma. No signs of toxicity were observed at the low dose of 10 ppm sodium fluoride.

Turner, C.H. et al. (2001) administered fluoride in drinking water to male Sprague-Dawley rats for 16 or 48 weeks, and at three different calcium intake levels (25, 50 and 100% of normal calcium in the diet). Drinking water was available ad libitum with fluoride concentrations of 5, 15 or 50 mg fluoride/L. The fluoride concentrations chosen in this study were estimated to cause plasma fluoride levels equivalent to those measured in humans consuming water fluoridated at levels of 1, 3 and 10 mg/L. The vertebral biomechanical properties and bone histology were measured. The authors concluded as follows: High fluoride intake had no effect on trabecular structure but increased the amount of unmineralised osteoid, particularly in older rats. The vertebral strength of rats was reduced when fluoride intake is high (50 mg fluoride/L = LOAEC). MAK (2005) has concluded that the middle concentration, corresponding to 3 mg F/L in human drinking water) represents a NOAEC, which can be converted to a fluoride exposure for workers via air of 0.9 mg/m³ (assuming 10 m³ breathing volume and 3 litres of water).

Chronic Toxicity

The U. S. National Toxicology Program (NTP, 1990) evaluated the toxicological and carcinogenic effects of continuous exposure to 0, 25, 100 or 175 ppm sodium fluoride in drinking water on male and female F344/N rats for a 2 -year period. Survival and weight gains of the male and female rats were not affected by fluoride treatment. Rats receiving sodium fluoride developed effects typical of dental fluorosis at 25, 100 and 175 ppm, and female rats had increased osteosclerosis at the high-dose of 175 ppm.

 

(ii) studies involving humans

The summary of the human data presented in this chapter has largely been adopted from the assessment by the German MAK Commission (MAK, 2005).

Hillier et al. (2002) reported a case-control study in which the correlation between fluoride intake and hip fractures was investigated (914 cases and 1196 controls). There was no increase in the incidence of hip fractures at fluoride concentrations of 1 mg/L in drinking water. Several confounders, such as age, sex, body-mass-index and others were considered.

Zhang et al (2003) reported morbidities in a Chinese population for dental fluorosis and skeletal fluorosis of 90% and 20.9 % respectively, when drinking water fluoride levels where as high as 3 mg/L. Urinary fluoride levels were increased in patients suffering from both types of fluorosis, when compared to controls. However, important details such as population size and study methodology are lacking in the publication, so that it cannot be used further in the assessment.

In a comprehensive study by Li et al. (2001) in China involving 8266 persons aged >50 years, the prevalence of bone fractures over a 20 year period was investigated in relation to fluoride exposure (different levels in drinking water). Demography, medical history, physical activity, smoking and alcohol consumption were considered. Fluoride exposure via food was also considered, but was equivalent throughout the groups. The authors found a U-shaped relationship between fluoride in drinking water and the prevalence of general bone fractures. There were more general bone fractures in people exposed to low fluoride levels (0.25-0.34 mg/L in their drinking water), as well as in the high exposure group (4.32 - 7.98 mg/L). When restricting the assessment to hip fractures, only the high exposure group, but not the low exposure group, shower a higher prevalence. According to this study, medium fluoride levels in drinking water (0.58 - 3.56 mg/L) do not lead to an increased risk of bone fractures. Authors have estimated that the daily fluoride exposure corresponding with 0.58-5.56 mg/L in drinking water is 1.62 - 7.85 mg/day. This study by Li et al. (2001) - as all ecological studies - has the shortcoming that exposure is assessed on a population basis and not on an individual basis. The associated uncertainties prevent the study to be used for the derivation of a reliable threshold value (such as a DNEL). Nevertheless, the study gives some information on exposure levels that can lead to an increased risk for bone fractures (no increased risk at up to7.85 mg F/day, but increased risk for the 14.13 mg F/day group).

For workers, Derryberry et al. (1963) report an increased risk for fluoride-induced osteosclerosis in workers exposed to 3.4 mg F/m³ for an average duration of 14 years, but not workers exposed to 2.65 mg/m³ for the same average duration.

Kaltreider et al. (1972) reported distinct skeletal fluorosis without physiological effects for workers exposed for 10 years to concentrations between 2.4 and 6.0 mg F/m³.

In an epidemiological study, Chan-Yeung et al. (1983a/b) assessed the influence of working at an Aluminium smelter on the musco-skeletal system, on the haematology, on the liver and kidney and on lung function. Exposure with aluminium, fluoride, benzo(a)pyrene, sulfur dioxide and carbon monoxide was assessed. The average fluoride exposure was 0.48 mg F/m³ (gaseous +particulate) or 0.28 mg F/m³ (particulate only). Haematology, the liver and the kidney were not affected. There were no distinct cases of skeletal fluorosis. X-ray images of some workers that were exposed for more than 10 years showed increased bone density, calcification of ligaments and changes in the periosteum. However, x-rays were separately evaluated by two radiologists, and there was a poor agreement between the assessors.

 

Discussion and conclusion

Effects of excessive exposure to fluorides on the skeletal system have been identified as the most sensitive endpoint. Effects include reduced bone strength, increase risk of fractures and/or skeletal fluorosis (stiffness of joints, skeletal deformities). For quite some time, the occupational exposure limit value for fluorides was set at 2.5 mg F/m³, e.g. SCOEL (1998), as supported by the findings by Derrberry et al. (1963) and Kaltreider et al. (1972). These authors had not reported effects on the skeletal system for workers exposed for 10 years to an average of 2.4 or 2.65 mg F/m³.

However, other and partly more recent human and animal data suggest that a limit value of 2.5 mg F/m³ may not be sufficient to protect humans against effects on the skeletal system: Assuming full absorption, and a shift-breathing volume of 10 m³ (per 8-hours), a worker working in an atmosphere containing 2.5 mg F/m³ would absorb 25 mg F/day. Further, as summarised above, the additional background fluoride intake from food, drinking water and dental care products may add up to a couple of mg F/day or even over 6 mg F/day in worst case scenarios, thus leading to worst case estimates of combined exposure of more than 30 mg F/day.

In contrast, according to US DHHS (1991) and WHO (2002), skeletal fluorosis (clinical phase III) may result when exposed to 20 mg F/day for more than 20 years.

Multiple sources support the conclusion that a safe total daily intake of fluoride should at least be below ca. 14 mg F/day.

• EPA (1985) reported no skeletal fluorosis when fluoride in drinking waters was 4 mg F/L (i.e. ca. 8 mg F/day assuming a consumption of 2 litres/day).
• The study by Li et al (2001) suggested no increased risk for bone fractures over 20 years at up to7.85 mg F/day, but increased risk for humans exposed to 14.13 mg F/day.
• The study by Hillier et al. (2000) in which no increase in the prevalence of hip fractures was seen at 0.2 - 0.3 mg F/day is supportive.
• The animal study by Turner et al. (2001) suggested a NOAEL for effects on bone density of 0.94 mg F/kg_bw/day with a LOAEL at 3.2 mg F/kg_bw/day. According to the authors, the NOAEL and LOAEL found in the rat study correspond to drinking waters concentrations for humans of 3 mg F/L and 10 mg F/L, respectively. Assuming a water intake of 2 L/day, these can be converted to corresponding daily doses of 6 mg F/day (NOAEL) and 20 mg F/day (LOAEL). Further, assuming a 10 m³ inhalation volume (for an 8 hour shift), these correspond to 0.6 mg F/m³ (NOAEC) and 2 mg F/m³ (LOAEC).

Based on these considerations, a workplace exposure limit (i.e. the DNEL) is derived at 1 mg F/m³. As a worst case assumption, 100% systemic absorption may be assumed, which together with a 10 m³ shift breathing volume results in an estimated maximum daily dose at the workplace of 10 mg F/day. Note: it appears unrealistic to assume that a worker spends a full 8-hour shift in an atmosphere containing 1 mg F/m³. Further, depending for example on the particle size, not all particles will be inhaled (some will not be inhaled at all, others will be exhaled). On the other hand, there may be some contribution of fluorides absorbed via the skin (dermal exposure), or also of inadvertent ingestion (hand-to-mouth transfer). Whereas the latter cannot be quantified reliably and whereas such additional dermal and oral exposure would be expected on under circumstances of poor industrial hygiene, a 100% total systemic absorption is maintained, still presenting a conservative scenario for the risk assessment.

Acknowledging that background intake due to drinking water, food, dental care products etc. may exceed 5-6 mg F/day in highly worst case scenarios, a typical background intake would rather be in the of range of 0.5-2 mg F/day. Adding a worst case workplace exposure of 10 mg/day and a typical background exposure of 2 mg/day will still result in a safe total fluoride intake for workers.

Thus, the DNEL of 1 mg F/m³ is adequately protective, even when considering that workers can - in addition to the inhalation route - be further exposed to fluorides via other routes, i.e. dermal and/or oral (hand-to-mouth-transfer) at the workplace, and also due to a background intake via diet, drinking water and dental care products.

The application of assessment factors is not required in this assessment, for the following reasons:

• Inter-species variability: Not required, since the assessment is based on human data
• Intra-species variability: Not required, since the assessment is based data obtained in studies involving a sufficiently large number of subjects, thus intrinsically addressing intra-species variability
• Exposure duration: Not required, since the studies are based on chronic exposure situations (continuous exposure via drinking waters, or occupational exposure for at least 10 years)
• Dose response: Not required, since no-adverse-effect-levels were identified
• Quality of whole data base: Not required, since the overall quality of the database is good, and findings are consistent

 

Biomonitoring

Urine biomonitoring has been shown to be an effective measure to control for safe systemic fluoride exposure levels in workers. Reference is made to the assessment by the German MAK commission (2006, with addendum 2007). MAK has established biological limit values for fluoride in the urine of workers as follows:

4.0 mg F / g creatinine, when sampled before the work shift

and/or

7.0 mg F / g creatinine, when sampled after an exposure period (or end-of-shift).

These limit values are based on the relationship between internal dose and health effects, and are applicable in addition or in parallel to the workplace air limit value of 1 mg F/m³.

 

References:

EFSA (2005): Opinion of the Scientific Panel on Dietetic Products, Nutrition and Allergies on a request from the Commission related to the Tolerable Upper Intake Level of Fluoride (EFSA-Q-2003-018). EFSA Journal 192, 1-65.

MAK: (2006/2007): Permanent Senate Commission for the Investigation of Health Hazards of Chemical Compounds in the Work Area (MAK Commission) in Germany: MAK and BAT value documentation for fluorides (in German Language). Original document dated 2006, supplemented by addendum in 2007.

SCOEL (1998): Recommendation from Scientific Committee on Occupational Exposure Limits for Fluorine, Hydrogen Fluoride and Inorganic Fluorides (not uranium hexafluoride). Document SCOEL/SUM 56.

US DHHS (1991): Review of Fluoride. Benefits and Risks. Ad Hoc Subcommittee on Fluoride, Committee to Coordinate Environmental Health and Related Programs, Washington DC, USA, Department of Health and Human Services.

US EPA (1985): Drinking water criteria document on fluoride (TR-832-5). NTIS, PB86-118163, ICAIR under EPA contract 68-03-3279.

WHO (2002):Environmental Health Criteria 227 on Fluorides.

 

Justification for classification or non-classification

Based on the results described above, classification for repeated dose toxicity is not warranted.