Inorganic Fluorides - PSL1

3.0 Assessment of "Toxic" under CEPA

3.1 CEPA 11(a) Environment

Inorganic fluorides are produced in Canada and emitted into the Canadian environment both from anthropogenic (estimated at 23 500 tonnes/year) and from natural sources (amounts not known). Gaseous inorganic fluoride compounds (e.g., hydrogen fluoride and sulphur hexafluoride) are primarily released into the atmosphere whereas particulate compounds (e.g., sodium fluoride and calcium fluoride) are released into aquatic and terrestrial environments.

For the purpose of assessing "toxic" under paragraph 11(a) of CEPA, this assessment focuses on the environmental compartments that have the highest concentrations and the biota considered to be at greatest risk from fluoride exposure and effects in Canada (e.g., phytotoxicity of atmospheric levels, impact on aquatic biota from water exposure, and wildlife exposure through plant consumption).

Hydrogen fluoride is believed to be the most important inorganic fluoride species affecting terrestrial plants, and its phytotoxicity has been studied extensively (Weinstein, 1977). The most sensitive plant species tested (gladioli, Gladiolus grandiflorus) elicited significant tissue necrosis at fluoride levels as low as 0.35 µg/m³fluoride. In addition, other controlled greenhouse and field studies have reported similar effect thresholds for a variety of species, including 0.44 µg/m³fluoride for apples (Malis domestica borkh), 0.54 µg/m³fluoride for pole beans (Phaseolus vulgaris Linné), and 0.9 µg/m³fluoride for balsam fir (Abies balsamea), black spruce (Picea mariana), and larch (Larix laricina). In all cases the extent of fluoride phytotoxicity was positively correlated with increased concentration and increased exposure duration. Mean inorganic fluoride levels (principally hydrogen fluoride) measured in air at various locations in Canada range from 0.01 to 1.0 µg/m³fluoride, with the higher values occurring near known industrial sources. Much of the available concentration data from the vicinity (within a few km) of industrial sources in Canada are within the same range of effects thresholds for several sensitive terrestrial plants.

The clam (Musculium transversum) and rainbow trout (Oncorhynchus mykiss) are among the freshwater aquatic species most sensitive to the effects of inorganic fluorides, based on an 8-week ^LC50 of 2.8 mg/L fluoride and 20-day LC₅₀ of 3.7 mg/L fluoride (13°C, 3mg CaCO₃ /L), respectively. The brine shrimp ( Artemia salina) and the amphipods Grandidierella lutosa and G. lignorum were among the most sensitive marine species tested with lowest-observed-effect-concentrations of 5.0 mg/L fluoride and 6.9 mg/L fluoride (based on growth impairment over 12 days and decrease in egg production over 90 days, respectively). Dividing the lowest-observed-effect-levels by a factor of 10 to account for differences in inter-species sensitivity and to extrapolate laboratory findings to the field yields estimated effects thresholds of 0.28 mg/L fluoride for freshwater species and 0.5 mg/L fluoride for marine aquatic species. Most of the available data on inorganic fluoride levels in surface waters of aquatic ecosystems in Canada were identified in relation to known anthropogenic sources. Available data on mean inorganic fluoride levels in freshwater in Canada indicate a range between 0.1 and 3.8 mg/L. Mean inorganic fluoride levels in marine surface waters in Canada range between 1.0 and 3.0 mg/L. Therefore, the majority of freshwaters and virtually all of the marine waters sampled for inorganic fluorides in Canada in the vicinity of known anthropogenic sources have mean concentrations that are equal to, or exceed the lowest-estimated-effects-threshold-concentrations for freshwater and marine species.

In order to determine the potential impact of fluoride exposure on wildlife in Canada, estimates were made of the uptake of fluoride by white-tailed deer (Odocoileus virginianus) as a result of consumption of contaminated browse. Deer were chosen as the model species, as field data demonstrating effects in the wild suggest that they are a sensitive indicator. As well, a controlled feeding study is available to help interpret exposure estimates in the field. Cornwall Island, Ontario, was chosen as an exposure site as it is adjacent to an aluminum smelter, a known source of fluoride emission. For many years cattle in the area ingesting fluoride-contaminated forage have suffered from a number of fluorotic lesions, and studies suggest that deer are as sensitive as or more sensitive than cattle to fluoride exposure (Suttie et al., 1985). From available data on levels of fluoride in air and plants, this site is representative of several of the more contaminated areas in Canada.

Data on fluoride levels in native plants were used to estimate seasonal rates of uptake in white-tailed deer. Since plant tissues were not collected during winter, concentrations of fluoride in winter twigs were calculated from air and green leaf levels (Sidhu, 1992). The diet composition of deer in Eastern Canada was estimated from Crawford (1982), Brown and Doucet (1991) and Hiebsch and Boersma (1990). The mean fluoride content of deer browse calculated on an annual basis for the Cornwall Island area is 38 µg/kg dry weight (Table 2). In a captive study, white-tailed deer exposed to 35 and 60 µg/kg fluoride (wet weight) in their diet for 2 years demonstrated lesions of the incisors at both doses; as well, at the higher dose, a mild increase in wear of the molars and mild hyperostoses of the long bones of the leg were observed. Levels in the mandibles were equivalent to levels observed in some wild deer obtained near sources of industrial pollution, where effects have been seen. Assuming equal bioavailability of the fluoride in the browse as in the controlled study, levels of fluoride in browse are at a level where fluorosis would occur in deer populations. Exposure to predatory mammals and birds capable of digesting bone would be higher, although data to quantitatively estimate this are unavailable.

Table 2 Estimated Diet of White-tailed Deer Living on Cornwall Island, Ontario by Season

Diet Component	Seasonal Variation of Inorganic Fluoride content of White-tailed Deer (% of diet)aµg/kg (dry weight)
	Inorganic Fluoride Concentration µg/kg (dry weight)	Spring	Summer	Fall	Winter
Leaves	68b	(65) 44	(65) 44	(35) 24	(0) 0
Twigs	20c	(10) 2	(0) 0	(10) 2	(65) 13
Other	14.5d	(25) 4	(35) 5	(55) 8	(35) 5
Mean diet by season (µg/kg [dry weight])		50	49	34	18
Mean Annual Inorganic Fluoride Content of Deer Diet: 38 µg/kg (dry weight)

1. Proportion of diet items by season (Hiebsch and Boersma, 1990).
2. Mean level of inorganic fluorides in red maple leaves (Acer rubrum) collected in summer, 1988 from Cornwall Island (Ontario Ministry of the Environment, 1990b).
3. Based on a 1:1 diet of red maple twigs (Acer rubrum) and balsam fir (Abies balsamea). Inorganic fluoride levels in maples leaves were converted to twig levels by multiplying by 0.2 (Sidhu, 1992). Balsam fir levels were calculated from average ambient air levels on Cornwall Island during summer, 1991 (Environment Canada, 1991) and the regression equation Y = 4.12 + 0.982X + 0.00175X², where X is the air concentration in mg fluoride/dm²/7 days and Y is the Balsam fir current year foliar concentration in µg/kg (dry weight) [Sidhu, 1992].
4. Mean 1988 fluoride levels in foliage plants for cattle at Cornwall Island (Ontario Ministry of the Environment, 1990b).

On the basis of available information, inorganic fluorides are entering the Canadian environment from anthropogenic sources in quantities resulting in concentrations in some Canadian waters, plants, and air, that may cause long-term harmful effects to biota in aquatic and terrestrial ecosystems. It has been concluded that inorganic fluorides have the potential to cause harm to the environment.

3.2 CEPA 11(b): Environment on Which Human Health Depends

Inorganic fluoride compounds (except sulphur hexafluoride) are not expected to remain in the troposphere very long or migrate to the stratosphere. Even though sulphur hexafluoride is long-lived enough to migrate into the stratosphere, its contribution to stratospheric ozone depletion is considered minimal, because fluorine is much less efficient and less available than chlorine in the catalytic destruction of ozone in the stratosphere (Chu, 1991).

The global atmospheric concentration of sulphur hexafluoride is estimated to range from 0.006 to 0.3 µg/m³(average approximately 0.0091 µg/m³). This is about 300 to 3 000 times less than the combined CFCs and about 10⁸times less than carbon dioxide global concentrations. Due to the long tropospheric residence time of sulphur hexafluoride and its strong absorption potential in the infrared region, the estimated Global Warming Potential of sulphur hexafluoride is 6 800 to 12 000 times higher than for carbon dioxide, and 2 to 8 times higher than CFC-11, the standard example of the CFCs known to contribute to global climate change (Environment Canada, 1993). Furthermore, there is considerable uncertainty regarding global and Canadian release rates of sulphur hexafluoride, its Global Warming Potential, and its potential contribution to global climate change. Chu (1991) estimates the potential contribution of sulphur hexafluoride to the global climate change as less than 0.01%, based on estimated air concentrations of sulphur hexafluoride (0.01 µg/m³) and CFCs (10 µg/m³), assuming a comparable infrared absorption efficiency between these 2 compounds, and attributing 10% to 12% of the global climate change effect to CFCs. Using the same equation, but assuming air concentrations ranging from 0.006 to 0.3 µg/m³for sulphur hexafluoride and from 2.8 to 23.9 µg/m³for CFCs (Ramanathan et al., 1985), infrared absorption efficiencies for sulphur hexafluoride of 2 to 8 based on the Global Warming Potential, and attributing 10% of the global climate change effect to CFCs, results in potential contributions of 0.02% to 1% by sulphur hexafluoride. These two estimates differ by two orders of magnitude.

The levels of sulphur hexafluoride in the Canadian atmosphere

are unknown. Furthermore, the detailed relationship between environmental levels, Global Warming Potential, and contribution to global climate change are not known. Thus, it is not possible to determine the contribution of inorganic fluorides to global climate change.

Therefore, based upon the available data, there is insufficient information to conclude whether sulphur hexafluoride is entering the environment in quantities or under conditions that may constitute a danger to the environment on which human life depends.

3.3 CEPA 11(c): Human Life or Health

Population Exposure

The average daily intake of inorganic fluoride has been estimated based on the levels in air, drinking water, soil, and food and the amounts consumed by various age groups of the population of Canada (Table 3) [Environmental Health Directorate, 1992]. The total average daily intake of inorganic fluoride via ingestion by exclusively breast-fed infants was estimated to range from approximately 0.5 to 2.6 µg/kg bw/day, while the intake in exclusively formula-fed infants was estimated to range from 13.6 to 93.0 µg/kg bw/day. The total average daily intake of inorganic fluoride via ingestion for individuals consuming "non-fluoridated" drinking water was estimated to range from approximately 17.2 to 96.4 µg/kg bw/day. The total daily intake of inorganic fluoride via ingestion for individuals consuming "fluoridated" drinking water was estimated to range from 32.8 to 160.4 µg/kg bw/day. On the basis of available data, it is evident that for individuals consuming "non-fluoridated" drinking water, the greatest source of exposure to inorganic fluoride occurs via the ingestion of food; for individuals consuming "fluoridated" drinking water, the greatest contribution to the total intake of inorganic fluoride comes from the water itself, as well as food.

Generally, exposure to airborne fluoride makes only a minor contribution to the total intake of inorganic fluoride; the average daily intake via inhalation was approximately 0.01 µg/kg bw/day. Dental products that contain fluoride, such as toothpaste, have been identified as significant sources of inorganic fluoride for children and adolescents (Drummond et al., 1990).

The intake of inorganic fluoride could possibly be higher in communities near point sources such as aluminum smelters, and brick and phosphate plants; however, the estimated daily intake of inorganic fluoride by individuals residing near point sources was not substantially greater than that observed for the general population (estimated intakes of inorganic fluoride via ingestion by individuals consuming "non-fluoridated" and "fluoridated" drinking water ranged up to 102 and 166 µg/kg bw/day, respectively). The estimated intakes of inorganic fluoride via inhalation ranged from 0.02 to 0.38 µg/kg bw/day. These estimates were based on monitoring data from various locations, and represent a "worst case scenario," in which an individual is assumed to be exposed to contaminated air and soil. Based on comparisons of data (Kumpulainen and Koivistoinen, 1977; Selikoff et al., 1983; Jones et al., 1971; Muramoto et al., 1991; Sakurai et al., 1983) on the concentrations of (total) fluoride in a limited number of foodstuffs obtained in close proximity to point sources with those obtained elsewhere, it is likely that intake in food may be elevated for populations residing in the vicinity of industrial sources; however, available data were considered insufficient to quantitatively estimate the intake of inorganic fluoride in food for populations in Canada under such conditions and it was considered, therefore, to be similar to that for the general population.

Table 3 Estimated Daily Intake of Inorganic Fluoride by the General Population of Canada

Route of Exposure	Estimated Intake of Inorganic Fluoride by Various Age Groups (µg/kg bw/day)
	0 - 6 moa	7 mo - 4 yrb	5 - 11 yrc	12 - 19 yrd	20+ yre
Ambient Airf	0.01	0.01	0.01	0.01	0.01
Foodg	13.6 - 91.5	22.30	16.44	13.64	30.08
Breast Milkh	0.47 - 1.05	-	-	-	-
Soili	0.03 - 1.55	0.02 - 1.19	0.01 - 0.40	0.002 - 0.11	0.002 - 0.09
"Fluoridated" Drinking Water j	-	44.92 - 76.92	24.33 - 41.67	16.65 - 28.51	15.64 - 26.79
"Non-Fluoridated" Drinking Waterk	-	3.08 - 12.92	1.67 - 7.00	1.14 - 4.79	1.07 - 4.50
Household Productsl	-	20.00 - 60.00	8.15 - 20.00	2.46	1.14
Total IntakemBreast-Fed Infants	0.51 - 2.61	-	-	-	-
Total IntakenFormula-Fed Infants	13.64 - 93.06	-	-	-	-
Total Intake "Fluoridated" Watero	-	87.25 - 160.42	48.94 - 78.52	32.76 - 44.73	46.87 - 58.11
Total Intake "Non-Fluoridated" Waterp	-	45.41 - 96.42	26.28 - 43.85	17.25 - 21.01	32.30 - 35.82

1. Assumed to weigh 7 kg, breathe 2 m³air, drink 750 mL of breast milk or infant formula (as food), and consume 35 mg soil per day (Environmental Health Directorate, 1992).
2. Assumed to weigh 13 kg, breathe 5 m³air, drink 0.8 litres of water, and consume 50 mg soil per day (Environmental Health Directorate, 1992).
3. Assumed to weigh 27 kg, breathe 12 m³air, drink 0.9 litres of water, and consume 35 mg soil per day (Environmental Health Directorate, 1992).
4. Assumed to weigh 57 kg, breathe 21 m³air, drink 1.3 litres of water, and consume 20 mg soil per day (Environmental Health Directorate, 1992).
5. Assumed to weigh 70 kg, breathe 23 m³air, drink 1.5 litres of water, and consume 20 mg soil per day (Environmental Health Directorate, 1992).
6. Based on the mean concentration of inorganic (gaseous and particulate) fluoride in ambient air of 0.03 µg/m³, reported for Toronto, Ontario (McGrath, 1983), and assuming the concentration in indoor air is identical to (outdoor) ambient air (Environmental Health Directorate, 1992).
7. Formula-fed infants (0 - 6 months): based on the mean concentrations of inorganic fluoride in infant formulas purchased in the United States of 0.127 and 0.854 mg/L reported for ready-to-use, milk-based formula and soy-based powdered formula (prepared with drinking water containing 1 ppm fluoride), respectively (McKnight-Hanes et al., 1988), and assuming infants are exclusively formula-fed and consume 750 mL formula per day (Environmental Health Directorate, 1992). General Population (7 months and older): based on levels of inorganic fluoride detected (Dabeka and McKenzie, 1993; Taves, 1983) in 109 individual foods from Canada (and the United States), in the following food groups (Environmental Health Directorate, 1992): 0.01 - 0.80 µg/g in dairy products; 0.12 - 1.02 µg/g in cereal products; 0.01 - 0.58 µg/g in fruit; 0.01 - 0.68 µg/g in vegetables; 0.04 - 4.57 µg/g in meat/fish/eggs; 0.05 - 0.13 µg/g in fats; 0.11 - 0.35 µg/g in nuts/legumes; 0.02 - 0.86 µg/g in foods containing primarily sugar; 0.41 - 0.84 µg/g in soup; 4.97 µg/g in tea; and the daily intake of each food item by the various age groups of the general population of Canada (Environmental Health Directorate, 1992).
8. Based on the mean concentrations of inorganic fluoride of 4.4 and 9.8 ng/g reported for samples of breast milk from mothers living in communities served by "non-fluoridated" and "fluoridated" drinking water, respectively (Dabeka et al., 1986), assuming the density of breast milk is equal to 1.0 g/mL.
9. Based on a range of concentrations of total inorganic fluoride of 6 ppm (µg/g) reported by Sidhu (1982) for soil collected in Newfoundland, to 309 ppm (µg/g) [mean concentration in Canadian surface soil (0 - 130 cm depth)] (Schuppli, 1985).
10. Based on a range of mean concentrations of inorganic fluoride in "fluoridated" drinking water of 0.73 mg/L, determined from fluoride levels in 3 communities in Newfoundland and Labrador, to 1.25 mg/L, determined from 2 communities in the Yukon (Droste, 1987).
11. Based on a range of mean concentrations of inorganic fluoride in "non-fluoridated" drinking water of (at least) 0.05 mg/L (reported for 3 communities in British Columbia) [Greater Vancouver Water District, 1990], to 0.21 mg/L (reported for an unspecified number of communities in the Yukon) [Health and Welfare Canada, Yukon Territory, 1989, cited in Hill and Hill, 1991].
12. Based on a mean concentration of inorganic fluoride in most dentifrice products of 1 000 ppm (µg/g) (Beltran and Szpunar, 1988; Whitford, 1987) and an estimated intake of dentifrice of 0.26 - 0.78 g/day for children 7 months to 4 years of age, 0.22 - 0.54 g/day for children 5 to 11 years of age, 0.14 g/day for adolescents 12 to 19 years of age, and 0.08 g/day for adults 20+ years of age (Levy, 1993), assuming an average of 2 brushings per day (Bruun and Thylstrup, 1988).
13. Estimated total daily intake of inorganic fluoride by exclusively breast-fed infants in Canada.
14. Estimated total daily intake of inorganic fluoride by exclusively formula-fed infants in Canada.
15. Estimated total daily intake of inorganic fluoride by individuals consuming "fluoridated" drinking water in Canada.
16. Estimated total daily intake of inorganic fluoride by individuals in Canada consuming drinking water that is not "fluoridated".

Effects

There has been no consistent evidence of an association between the consumption of "fluoridated" drinking water and the incidence of, or mortality due to cancer in a large number of ecological studies performed in many countries. Although these results do not support the hypothesis of an association, their considerable limitations preclude firm conclusions regarding the carcinogenicity of fluoride in humans. For example, cancer of the bone was not assessed in the majority of these studies. The incidence of, or death due to cancer has also been investigated in a number of historic cohort studies of workers exposed to fluoride predominantly in the aluminum smelting industry. While excesses of cancers of different types have been reported in various studies, the only site for which there was excess risk in several investigations is lung cancer, for which the weight of evidence is considered moderate. Due to possible confounding by concomitant exposure to other substances in analytical studies of occupationally exposed workers, however, the observed excesses cannot confidently be attributed to fluoride. Moreover, bone cancer was not assessed in the majority of these studies. Available information is considered inadequate, therefore, to assess the carcinogenicity of inorganic fluoride in humans.

In early carcinogenicity bioassays conducted by Kanisawa and Schroeder (1969), Taylor (1954), and Tannenbaum and Silverstone (1949), the incidence of tumours in mice administered sodium fluoride (in either the diet or drinking water) was, in general, not markedly greater than that observed in controls. Owing to inadequate documentation and to numerous methodological shortcomings, however, the results of these investigations do not contribute meaningfully to an assessment of the weight of evidence of the carcinogenicity of (sodium) fluoride.

The administration of drinking water containing sodium fluoride (in amounts estimated to provide intakes ranging from 0.6 to 9.1 µg/kg bw/day fluoride) to male and female B6C3F₁ mice over a period of 2 years produced a marginal (statistically insignificant) increase in the incidence of hepatoblastoma, compared to the incidence in groups of "controls" administered drinking water without added fluoride; however, this minor increase was not considered "biologically significant," since the overall incidence of hepatic tumours (adenoma, carcinoma, hepatoblastoma) was not increased in animals receiving sodium fluoride, and the incidence of all hepatic tumours in these groups of mice was higher than that in previous NTP carcinogenicity bioassays (NTP, 1990). The marginal increase in the incidence of malignant lymphoma in female B6C3F₁ mice administered drinking water containing sodium fluoride was considered not to be compound-related, since the incidence in the high-dose group was similar to that observed in historical controls (NTP, 1990). No other increases in tumour incidence were observed; however, failure to attain the maximum tolerated dose may have reduced somewhat the sensitivity of this study in mice. In a less extensive carcinogenicity bioassay in which the incidence of osteomas in male and female CD-1 mice receiving 25 µg/kg bw/day sodium fluoride in the diet was increased compared to controls (Maurer et al., 1993), the specific role of fluoride in the etiology of the tumours cannot be determined with certainty, owing to the infection of these animals with Type C retrovirus (U.S. DHHS, 1991; U.S. NRC, 1993; Maurer et al., 1993).

The administration of drinking water containing sodium fluoride (in amounts estimated to provide intakes ranging from 0.2 to 4.5 µg/kg bw/day fluoride) to F344/N rats produced a marginal increase in the incidence of oral cavity neoplasms (in males and females) and tumours in the thyroid gland (in males) [NTP, 1990]. The squamous cell tumours of the oral cavity were not considered to be compound-related, since the incidence of tumours in the high-dose group was not significantly different from the controls, the incidence of these neoplasms was within the range observed in historical controls, and there was no supporting evidence of focal hyperplasia of the oral mucosa (NTP, 1990). The marginal increase in thyroid (follicular cell) tumours was also considered not to be compound-related, since the incidence of these tumours in the high-dose group was not significantly different from the controls, the incidence of these neoplasms was within the range observed in historical controls, and the incidence of follicular cell hyperplasia was not increased in fluoride-exposed animals (NTP, 1990). The incidence of osteosarcoma in groups of male and female F344/N rats administered drinking water containing sodium fluoride (in amounts estimated to provide intakes ranging from 0.8 to 4.5 µg/kg bw/day fluoride) over a period of 2 years was not significantly increased, compared to controls receiving approximately 0.2 µg/kg bw/day fluoride (NTP, 1990); however, for the male F344/N rats, it was reported that "the osteosarcomas occurred with a significant dose response trend (P = 0.027, by logistic regression)" [NTP, 1990]. In a more limited carcinogenicity bioassay conducted by Maurer et al. (1990), the administration of diets containing sodium fluoride (in amounts estimated to provide intakes ranging from 1.8 to 11.3 µg/kg bw/day fluoride) to male and female Sprague-Dawley rats over a period of 95 to 99 weeks produced no significant increase in the incidence of any types of tumours, compared to groups of controls receiving approximately 0.1 µg/kg bw/day fluoride, although a small number of malignant tumours of the bone was observed.

In assessing the evidence for the carcinogenicity of fluoride, some significance has been attributed to the observation of a dose-response trend in the occurrence of osteosarcomas in male F344/N rats administered sodium fluoride in drinking water (NTP, 1990). Such a trend associated with the occurrence of a rare tumour in the tissue in animals and humans in which fluoride is known to accumulate cannot be easily dismissed. Moreover, the level of fluoride in the bones of the high-dose group of male rats in the NTP carcinogenicity bioassay, in which a non-significant increase in osteosarcomas was observed, is similar to that measured in humans with skeletal fluorosis. However, the biological significance of this dose-respond trend is tempered somewhat by the lack of statistical significance of the observed excess in the high-dose males in comparison with controls, as well as by the absence of a comparable statistically significant trend in the incidence of osteosarcomas in female F344/N rats or male and female B6C3F₁ mice receiving comparable amounts of inorganic fluoride (NTP, 1990). Indeed, the levels of fluoride in the bone of (male and female) F344/N rats and B6C3F1 mice administered sodium fluoride in drinking water were similar (NTP, 1990). No dose-response trend in the incidence of osteosarcomas was observed in groups of male and female Sprague-Dawley rats administered diets containing sodium fluoride (Maurer et al., 1990), even though the levels of fluoride in the bone in the high-dose animals were greater than those in the male F344/N rats in which there was an increase in osteosarcomas in the NTP (1990) carcinogenicity bioassay; however, there may be variations in sensitivity of the two strains to the effects of fluoride. Moreover, there were more animals in the high-dose group in the NTP bioassay (n = 100, 42 males at termination) compared to that in the study by Maurer et al. (1990) [n = 70, 26 males at termination], and in the latter study, detailed information on the incidence of tumours in tissues or organs other than the bone and stomach were not presented, and histological examination of bone from the 2 mid-dose groups at terminal sacrifice was limited. Furthermore, a small number of osteosarcomas was observed in the fluoride-exposed Sprague-Dawley rats, compared to none in the controls (Maurer et al., 1990). There is controversy concerning whether or not the osteomas in male and female CD-1 mice observed in the carcinogenicity bioassay conducted by Maurer et al. (1993) should be classified as neoplasms, and a retrovirus (in addition to fluoride) has been implicated in their etiology; however, the significant increase in (the highest dose) fluoride-exposed versus control groups (in animals infected with retrovirus), in a tissue in which fluoride is known to accumulate, adds some weight, albeit weak, to the evidence of carcinogenicity.

There is evidence that fluoride is genotoxic based on the outcome of in vitro and in vivo studies. Sodium fluoride induced recessive lethal mutations in D. melanogaster, and cytogenetic damage after intraperitoneal injection in rodents. Generally, however, in studies in which fluoride was administered to laboratory animals by routes of exposure similar to those by which humans are normally exposed (i.e., orally), it had no effect upon the frequency of chromosomal aberrations, micronuclei, sister-chromatid exchange, DNA strand breaks, or sperm morphology. The mechanism by which sodium fluoride induces genetic alterations is not known; however, it is not likely due to an interaction between the fluoride ion and DNA. Rather, it may be a secondary effect of the actions of fluoride that result from its inhibition of enzymes involved in DNA synthesis and/or repair.

Therefore, although there is some evidence for the carcinogenicity of inorganic fluoride, available data are inconclusive. For such compounds, the assessment of "toxic" under paragraph 11(c) of CEPA is based on a comparison of concentrations in environmental media or estimated daily intakes with those to which it is believed that a person can be exposed daily over a lifetime without developing deleterious non-neoplastic effects. Based on the available data, it is evident that (with the exception of dental fluorosis) following long-term exposure, skeletal effects in humans occur at doses or levels of exposure lower than those associated with other adverse health effects, which is likely a consequence of the accumulation of inorganic fluoride almost exclusively in bone. Therefore, effects on the skeleton are considered to be the most relevant in assessing the toxicological effects of long-term exposure to inorganic fluorides.

Skeletal changes have been observed in a number of animal species (i.e., cattle, sheep, rodents) exposed to inorganic fluoride. Dose-response relationships concerning exposure to inorganic fluoride and effects on the skeleton in rats have been reported in well-designed and adequately conducted long-term toxicological studies (NTP, 1990; Maurer et al., 1990). No adverse effects on the skeleton have been observed in F344/N and Sprague-Dawley rats receiving 2.7 and 1.8 µg/kg bw/day fluoride, respectively (NTP, 1990; Maurer et al., 1990); however, rats are generally regarded as less sensitive to the toxicological effects (on the skeleton) of fluoride than are humans or larger animals (Franke, 1989; Turner et al., 1992). Compared to humans, the lower sensitivity of rats to the toxicological effects of fluoride on the skeleton has been attributed to differences in toxicokinetics (i.e., absorption and elimination) and skeletal development (i.e., in contrast to humans there is little or no bone remodelling in rats) [O'Flaherty, 1991a, 1991b; Chavassieux, 1990; Turner et al., 1992; Franke, 1989; Grynpas, 1990], although Whitford et al. (1991) reported that the ratio of the renal and extra-renal clearance to plasma clearance of fluoride in rats was similar to that in humans. In studies in which effects on the skeleton were observed in dogs (Snow and Anderson, 1986) and pigs (Mosekilde et al., 1987) receiving approximately 0.32 and 2 µg/kg bw/day fluoride, respectively, no overtly pathological effects on the bone or on the general health of the animals were observed, although the animals were exposed for a relatively short duration (i.e., 6 months). Notably, dogs and pigs are considered to be more appropriate models for examination of the potential effects of various agents on the skeleton in humans (Snow and Anderson, 1986; Mosekilde et al., 1987; Chavassieux, 1990). Bone matrix formation (based on histomorphometric analysis) in male C57BL/6 mice receiving 0.8 µg/kg bw/day fluoride in drinking water over a period of 4 weeks was increased 20%, compared to controls; however, no pathological effects were observed (Marie and Mott, 1986). Owing to the availability of data, although limited, on exposure-response relationships in humans, and the interspecies variations in response to fluoride, the results of epidemiological studies on the effects of fluoride in humans have been emphasized in the derivation of an effect level for inorganic fluoride.

Case reports or descriptive ecological studies of skeletal fluorosis or osteosclerosis in the United States (Stevenson and Watson, 1960; Leone et al., 1955; Felsenfeld and Roberts, 1991; U.S. DHHS, 1991) or studies on the occurrence of endemic (crippling) skeletal fluorosis in areas in other countries in which the levels of fluoride are naturally very high, provide little quantitative information useful in establishing an effect level for fluoride, since nutritional intake, the intake of fluoride (and additional minerals) from other sources, potentially confounding concomitant exposures, and the extent of physical labour, all of which have been suggested to play a role in the etiology of this disease (see WHO, 1984, and Singh and Jolly, 1970, cited in U.S. EPA, 1985, for a review), were not adequately documented. There is also limited quantitative information on exposure to inorganic fluoride and the development of skeletal effects (osteosclerosis or fluorosis) in occupationally exposed workers (Kaltreider et al., 1972; Tourangeau, 1944, Boillat et al., 1975, and Schegel, 1974, all cited in Hodge and Smith, 1977; Grandjean, 1982; Chan-Yeung et al., 1983a; Czerwinski et al., 1988; Roholm, 1937, cited in Grandjean, 1982). Inconsistent results of inherently limited cross-sectional studies of small populations exposed to often unspecified concentrations of fluoride in the vicinity of industrial sources (Tsunoda, 1970; Tsiji and Tsunoda, 1970, cited in Hodge and Smith, 1977), contribute little to an assessment of exposure-response for skeletal effects associated with exposure to fluoride. Information obtained from clinical studies, in which sodium fluoride was administered to patients for the treatment of osteoporosis, is inadequate, owing to the limitations of the protocols and characteristics of the patients in these studies. Moreover, these clinical studies were undertaken to assess the considered beneficial effect of fluoride (i.e., its capacity to increase bone mass), rather than its potential to produce adverse effects after long-term exposure.

Estimating an effects threshold for the development of skeletal fluorosis (or related changes) in humans exposed to inorganic fluoride is further complicated by differences in the radiological diagnosis of early stage skeletal fluorosis among health care professionals (Chan-Yeung et al., 1983a), as well as by the multiplicity of factors that may influence the amount of fluoride deposited in the bone, and hence the severity of the disease. When the results are evaluated collectively, however, the available data indicate that potentially adverse effects associated with skeletal fluorosis are likely to be observed at intakes greater than approximately 200 µg/kg bw/day fluoride. In case accounts, the development of crippling skeletal fluorosis was attributed to an intake of approximately 215 to 285 µg/kg bw/day fluoride (in adults) [U.S. DHHS, 1991]. Skeletal fluorosis has been observed in cryolite workers having estimated (occupational) intakes of approximately 285 to 1 142 µg/kg bw/day fluoride (Roholm, 1937, cited in Grandjean, 1982), and in patients receiving approximately 260 to 389 µg/kg bw/day fluoride for the treatment of osteoporosis (Power and Gay, 1986). In clinical studies, minor increases in the incidence of hip fracture have been observed in groups of osteoporotic patients administered (sodium) fluoride at doses equivalent to 260 µg/kg bw/day fluoride or more (Gutteridge et al., 1984, cited in Inkovaara, 1991; Power and Gay, 1986; Mamelle et al., 1988; Hedlund and Gallagher, 1989; Riggs et al., 1990); however, the incidence of hip fracture in these studies, which generally involved small numbers of elderly patients with a clinically detectable disease of the bone, was low. The weight of evidence in ecological studies (Jacobsen et al., 1990, 1992; Cooper et al., 1991; Danielson et al., 1992; Keller, 1991, and May and Wilson, 1991, both cited in Gordon and Corbin, 1992; Suarez-Almazor et al., 1993) indicates that there may be an association between the consumption of "fluoridated" drinking water and an increased incidence of hip fracture (based on hospitalization rates) particularly among the elderly. These results should be interpreted with caution, however, in view of the limitations of epidemiological investigations of this experimental design. Moreover, owing to the lack of data on individual exposure in such studies, it is difficult to derive meaningful conclusions concerning the exposure-response relationship for possible skeletal effects associated with exposure to fluoride. Although the relative risk of hip, wrist, or spinal fracture was increased in some groups of women residing in an elevated-fluoride community (with drinking water containing 4 mg/L, and having an estimated (mean) intake of approximately 72 µg/kg bw/day fluoride) compared to those in a control community (with drinking water containing 1 mg/L fluoride) [Sowers et al., 1986, 1991], the estimated intake of fluoride by women in the elevated-fluoride community was likely underestimated (since it was derived solely on the amount of water-based beverages consumed), and the level of calcium in the drinking water from the elevated-fluoride community was approximately 25% of the level in the control community.

It is concluded, therefore, on the basis of data from several different types of studies, that potentially adverse effects associated with skeletal fluorosis are likely to be observed at intakes of greater than approximately 200 µg/kg bw/day fluoride. Moreover, predicted concentrations of fluoride in bone resulting from a daily intake of 200 µg/kg bw/day are within the range of those reported to be associated with effects on the skeleton (Turner et al., 1993). Confidence in this value is limited, however, due to the limitations of identified case reports, and individual epidemiological and clinical studies in which the relationship between exposure to inorganic fluoride and effects on the skeleton have been examined, as well as those factors which can influence the development of skeletal fluorosis. Based on the limited available data, adverse effects upon haematopoietic, hepatic, or renal function are not expected to occur at such levels of intake, since adverse effects upon the bone marrow, liver, or kidney were not observed following the administration of approximately 389 µg/kg bw/day fluoride to osteoporotic patients over a period of 5 years (Hasling et al., 1987). There are insufficient quantitative data available from studies in humans to conclude unequivocally that exposure to this level of inorganic fluoride would have no adverse effect upon human reproduction and development, or the central nervous and immune systems.

Based on available information, the estimated average daily intakes of inorganic fluoride, which range from approximately 0.5 to 160 µg/kg bw/day by various age groups in the general population (or up to 167 µg/kg bw/day by those residing in the vicinity of point sources of inorganic fluoride) are less than the level at which adverse effects upon the skeleton (the end-point considered most sensitive on the basis of available data) are anticipated (i.e., ³ 200 µg/kg bw/day fluoride).

Therefore, based upon the available data, it has been concluded that inorganic fluorides⁶ are not entering the environment in quantities or under conditions that may constitute a danger to human life or health.

3.4 Conclusion

It has been concluded that inorganic fluorides are entering the environment in quantities or under conditions that may be harmful to the environment. There is insufficient information to conclude whether sulphur hexafluoride is entering the environment in quantities or under conditions that may constitute a danger to the environment on which human life depends. It has been concluded that inorganic fluorides (i.e., the fluoride ion derived from such inorganic substances) are not entering the environment in quantities or under conditions that may constitute a danger to human life or health.

⁶ The assessment of whether

---

inorganic fluorides are entering the environment in quantities or under conditions that may constitute a danger to human life or health is based on the effects of the fluoride ion derived from inorganic substances.