Ambient Water Quality Criteria for Fluoride

Water Quality

4.0 Drinking Water

Water Treatment

Some water supplies may have excessive fluoride levels which need to be reduced before delivery to the consumer. There are a number of processes which will do this, but all are expensive and few jurisdictions carry them out (Benefield et al., 1982; and Sawyer and McCarty, 1967). Precipitation of CaF₂ by adding calcium salts such as Ca(OH)₂, CaS0₄ or CaCl₂ is one method and adsorption to the insoluble compound Al(OH)₃ which is produced by adding alum, Al₂(S0₄)₃.14H₂0, to the water is another method. Ion exchange and sorption on bone char, synthetic ion exchange media and activated alumina (Al₂0₃) are also practiced (Benefield et al., 1982). Activated alumina defluoridation can reduce fluorides from 8 to 1 mg/L (Sorg, 1978; Bishop and Sansoucy, 1978; and Choi and Chen, 1979). Caustic soda is a better regenerant than alum in ion exchange methods. Silicate and hydroxyl compete for exchange sites when the pH is over 7, but between pH 5 and pH 6 fluoride is preferentially adsorbed. The acidic water resulting from this process needs to be neutralized with limestone to reduce its corrosiveness and a 96% water recovery is possible (Anon, 1985). Reverse osmosis can also reduce fluoride from 2.2 to less than 1.0 mg/L (Naylor and Dague, 1975).

Fluoride is added to water as sodium fluoride (NaF), sodium silicofluoride (Na₂SiF₆) or fluorosilicic acid (H₂SiF₆) where water fluoridation is practiced (Anon, 1971; Sawyer and McCarty, 1967; and Gunther and Gray, 1988). In 1987, 53.7% of Alberta's population received fluoridated water amounting to 186 X 10⁶ m³ of water with a mean fluoride level of 1.03 mg/L (Anon, 1985). In 1985, only 11.1% of British Columbia's population received fluoridated water. The fluoride content of natural water supplies in Canada varies between 0.01 and 4.5 mg/L. Ground water infiltration is suspected of being the major source of fluoride in surface water with high fluoride concentrations (Anon, 1980). Since some natural supplies exceed the fluoride drinking water objective of 1.0 mg/L, they need to be treated to remove excess fluoride. Other supplies are below the objective and need to have fluoride added since too little has detrimental effects on teeth.

As of 1985 there were 22 communities in British Columbia where fluoridation of the drinking water was carried out. These were all smaller communities with a total population of about 330 000. The start dates of these water treatments ranged from 1955 to 1975. The raw water supplies in these communities, before fluoridation, had natural fluoride levels ranging from 0.01 to 0.95 mg/L. Fluoride was added as NaF, H₂SiF₆ or Na₂SiF₆. With never more than 13% of the provinces' population drinking fluoridated water, British Columbia has traditionally had the lowest percentage of any provincial population in Canada being served by fluoridated water supplies.

The two largest population centres in British Columbia, Greater Victoria and Greater Vancouver, do not fluoridate their water. They draw water from large watershed reserves and the water is virtually all recent rainfall or snowmelt and low in fluoride. Fluoridation would likely decrease dental caries in children living in these communities.

Studies done in several cities where fluoridation occurs have shown the expected 60% reduction in tooth decay. A comparison of 13-year-old students in British Columbia, exclusive of those in Victoria and Vancouver areas, showed a significant decrease in the dental caries index of up to 19% in students living in communities with fluoridated water as opposed to those in communities which did not practice fluoridation. This difference showed up in spite of the complexities, described in the next paragraph, which were not accounted for in a study designed for other purposes. If a study was designed specifically to determine the effects of water fluoridation and these variables were controlled, one would expect to see a better percentage improvement.

Much of the population uses fluoride toothpastes, fluoride rinses, fluoride supplements and topical fluoride applications. Few students were born and remained in either a fluoridated or non-fluoridated community; mobility is quite high, estimated at around 50% for 13-year-old students. Thus 13-year-old students may well have been brought up in a different community than that in which they were tested. Even communities classified as fluoridated had children in their schools who came from outlying areas not under fluoridation. The differences are smaller in younger children and become more pronounced with age as cumulative effects begin to appear (Gunther and Gray, 1988; and Gray and Gunther, 1987).

Effects

The effects of excessive fluoride have been covered in Chapter 3 and can be found in more detail in the review articles referenced in Chapter 1. High dental caries levels may occur if fluoride levels are below 0.5 mg/L (Anon, 1982). Small amounts of fluoride reduce dental caries, especially in children, while excessive levels cause mottling of teeth (McNeely et al., 1979; Anon, 1982; Anon, 1983; and Anon, 1986). Dental fluorosis is not considered an adverse health effect, but due to cosmetic effects, fluoride should not be allowed to rise above 2.4 to 4.0 mg/L (Anon, 1958). Water with less than 0.9 to 1.0 mg/L fluoride will seldom cause mottled tooth enamel in children, and there is abundant literature to show the advantages of maintaining a fluoride level of 0.8 to 1.5 mg/L. In adults, less than 3.0 to 4.0 mg/L will not cause endemic cumulative skeletal effects (McKee and Wolf, 1963; McClure et al., 1945; and McClure and Kinser, 1944); fluorides up to 5.0 mg/L cause no effects except mottling of tooth enamel (Smith and Cox, 1952; Heyroth, 1952; and Hillboe and Ast, 1951).

Adverse effects of fluoride at concentrations of 5 to 8 mg/L are essentially limited to effects on tooth enamel, although some individuals are reported to experience bone density changes at these concentrations. At 12 to 20 mg/L crippling fluorosis occurs (Anon, 1958).

Radiologic surveys of people who lived more than 15 years in Bartlett, Texas, where water had 8 mg/L fluoride, showed minimal increase in bone density in only 12% of the people, but no interference with the use of bones or joints (Anon, 1971). Mortality rates from kidney disease, heart disease or cancer, in high and low fluoride areas, show no association with fluoride levels (Smith and Fox, 1952; and Heyroth, 1952). It is estimated that daily levels of 15 to 20 mg fluoride for several years would be required to induce chronic fluorosis in adult man (Mitchell and Edman, 1953). Polio incidence has been shown to be lower in areas where surface water contains over 1.0 mg/L fluoride (Shay, 1947; and Shay, 1948), but no clear-cut conclusion could be reached regarding a correlation between fluoride and goiter (Fellenberg, 1938).

Literature Criteria

Criteria from the literature are summarized in Table 4.1 and 4.2. Criteria for drinking waters are usually based on the most sensitive users: young children whose teeth are still growing. The criteria are set to avoid dental fluorosis (mottling), since teeth are one of the most sensitive tissues (Anon, 1980; and Anon, 1977). In some jurisdictions, the fluoride criterion is based on the total amount of water likely to be consumed, and is thus temperature dependent as shown in Table 4.2 (Anon, 1969; Anon, 1962; Anon, 1968; Anon, 1975; Anon, 1978; Rose and Marier, 1977; and Anon, 1980). In Canada, 1.0 mg/L is used everywhere except the arctic and subarctic zones where 1.2 mg/L is permitted (McNeeley et al., 1979; and Anon, 1979). These are objective levels, the maximum acceptable level is 1.5 mg/L (Anon, 1979). The arctic and subarctic are defined as areas where the annual mean daily maximum temperature is less than 10°C (Anon, 1979). In British Columbia, the optimum concentration is set at 1.2 mg/L (Anon, 1982; and Anon, 1969), while the maximum acceptable concentration is 1.5 mg/L.

Class C water in Manitoba, suitable for domestic consumption after a full treatment, should not exceed 1.2 mg/L (Anon, 1980). The maximum acceptable concentration in Ontario drinking water is 2.4 mg/L (Anon, 1984; and Anon, 1982). Where fluoridation is practiced, the recommended fluoride level is 1.2 + 0.2 mg/L (Anon, 1983). WHO recommends a limit of 1.5 mg/L (Anon, 1961) or 1.0 mg/L (Anon, 1958) in drinking water. Water with fluoride in the range of 0.7 to 1.5 mg/L is acceptable for drinking water supplies after any kind of water treatment from simple filtration and disinfection to full and complete treatment (Anon, 1975).

Recommended Criteria

The recommended total fluoride level in raw drinking water is 1.0 mg/L as a 30-day average with a maximum value of 1.5 mg/L (Anon, 1979; Anon, 1982; Anon, 1969; Anon, 1987; Anon, 1962; and Anon, 1987). This is consistent with British Columbia, CCREM and United States Public Health guidelines.

Rationale

The fluoride criteria are temperature dependent. The lowest isotherm used is 12°C (Anon, 1962), or 10°C (Anon, 1979; and Anon, 1987), and in both these cases the optimum fluoride level is 1.2 mg/L. In British Columbia the mean annual daily maximum temperature is below the 10°C isotherm, except for two areas which may reach 10°C or 11°C some years, but never reach the 12°C isotherm. The areas where the mean may be this high are, however, the most highly populated areas of southeastern Vancouver Island, the lower Fraser Valley and the Okanagan Valley. In these areas, the optimum level of fluoride would be 1.0 mg/L (Anon, 1979; and Anon, 1987), or 1.2 mg/L (Anon, 1962) but in all other areas of British Columbia the optimum level would be 1.2 mg/L (Anon, 1979; Anon, 1982; and Anon, 1962). The maximum acceptable level is 1.5 mg/L (Anon, 1979; Anon, 1982; and Anon, 1987), or 1.7 mg/L (Anon, 1962).

Only a few areas in British Columbia reach the 10°C isotherm which, according to two references, Anon, 1979 and Anon, 1987, would require water with 1.0 mg/L fluoride, but which according to another reference, Anon, 1962, still qualifies them for 1.2 mg/L fluoride in the water. These isotherms are not always reached and these areas are thus marginal cases. It is not judged that any harm would result from using the 1.2 mg/L criterion uniformly throughout British Columbia especially since the peak summer temperatures in the Victoria and Vancouver areas are relatively low and the main reason the mean reaches 10°C is due to relatively high winter temperatures. Thus, summer water consumption would not be inordinately high. However, the 1.0 mg/L fluoride concentration is recommended since it should adequately protect teeth and allows a little more safety margin between the therapeutic dose and harmful levels.

The fluoride criterion is unique in this regard since most other criteria have at least, and usually more than, a ten-fold safety factor between the criterion and any known effect levels. This is not possible with fluoride and in any case the first "harmful" effects to appear are cosmetic and not functional.


	Water Quality Ambient Water Quality Criteria for Fluoride 4.0 Drinking Water Water Treatment Some water supplies may have excessive fluoride levels which need to be reduced before delivery to the consumer. There are a number of processes which will do this, but all are expensive and few jurisdictions carry them out (Benefield et al., 1982; and Sawyer and McCarty, 1967). Precipitation of CaF₂ by adding calcium salts such as Ca(OH)₂, CaS0₄ or CaCl₂ is one method and adsorption to the insoluble compound Al(OH)₃ which is produced by adding alum, Al₂(S0₄)₃.14H₂0, to the water is another method. Ion exchange and sorption on bone char, synthetic ion exchange media and activated alumina (Al₂0₃) are also practiced (Benefield et al., 1982). Activated alumina defluoridation can reduce fluorides from 8 to 1 mg/L (Sorg, 1978; Bishop and Sansoucy, 1978; and Choi and Chen, 1979). Caustic soda is a better regenerant than alum in ion exchange methods. Silicate and hydroxyl compete for exchange sites when the pH is over 7, but between pH 5 and pH 6 fluoride is preferentially adsorbed. The acidic water resulting from this process needs to be neutralized with limestone to reduce its corrosiveness and a 96% water recovery is possible (Anon, 1985). Reverse osmosis can also reduce fluoride from 2.2 to less than 1.0 mg/L (Naylor and Dague, 1975). Fluoride is added to water as sodium fluoride (NaF), sodium silicofluoride (Na₂SiF₆) or fluorosilicic acid (H₂SiF₆) where water fluoridation is practiced (Anon, 1971; Sawyer and McCarty, 1967; and Gunther and Gray, 1988). In 1987, 53.7% of Alberta's population received fluoridated water amounting to 186 X 10⁶ m³ of water with a mean fluoride level of 1.03 mg/L (Anon, 1985). In 1985, only 11.1% of British Columbia's population received fluoridated water. The fluoride content of natural water supplies in Canada varies between 0.01 and 4.5 mg/L. Ground water infiltration is suspected of being the major source of fluoride in surface water with high fluoride concentrations (Anon, 1980). Since some natural supplies exceed the fluoride drinking water objective of 1.0 mg/L, they need to be treated to remove excess fluoride. Other supplies are below the objective and need to have fluoride added since too little has detrimental effects on teeth. As of 1985 there were 22 communities in British Columbia where fluoridation of the drinking water was carried out. These were all smaller communities with a total population of about 330 000. The start dates of these water treatments ranged from 1955 to 1975. The raw water supplies in these communities, before fluoridation, had natural fluoride levels ranging from 0.01 to 0.95 mg/L. Fluoride was added as NaF, H₂SiF₆ or Na₂SiF₆. With never more than 13% of the provinces' population drinking fluoridated water, British Columbia has traditionally had the lowest percentage of any provincial population in Canada being served by fluoridated water supplies. The two largest population centres in British Columbia, Greater Victoria and Greater Vancouver, do not fluoridate their water. They draw water from large watershed reserves and the water is virtually all recent rainfall or snowmelt and low in fluoride. Fluoridation would likely decrease dental caries in children living in these communities. Studies done in several cities where fluoridation occurs have shown the expected 60% reduction in tooth decay. A comparison of 13-year-old students in British Columbia, exclusive of those in Victoria and Vancouver areas, showed a significant decrease in the dental caries index of up to 19% in students living in communities with fluoridated water as opposed to those in communities which did not practice fluoridation. This difference showed up in spite of the complexities, described in the next paragraph, which were not accounted for in a study designed for other purposes. If a study was designed specifically to determine the effects of water fluoridation and these variables were controlled, one would expect to see a better percentage improvement. Much of the population uses fluoride toothpastes, fluoride rinses, fluoride supplements and topical fluoride applications. Few students were born and remained in either a fluoridated or non-fluoridated community; mobility is quite high, estimated at around 50% for 13-year-old students. Thus 13-year-old students may well have been brought up in a different community than that in which they were tested. Even communities classified as fluoridated had children in their schools who came from outlying areas not under fluoridation. The differences are smaller in younger children and become more pronounced with age as cumulative effects begin to appear (Gunther and Gray, 1988; and Gray and Gunther, 1987). Effects The effects of excessive fluoride have been covered in Chapter 3 and can be found in more detail in the review articles referenced in Chapter 1. High dental caries levels may occur if fluoride levels are below 0.5 mg/L (Anon, 1982). Small amounts of fluoride reduce dental caries, especially in children, while excessive levels cause mottling of teeth (McNeely et al., 1979; Anon, 1982; Anon, 1983; and Anon, 1986). Dental fluorosis is not considered an adverse health effect, but due to cosmetic effects, fluoride should not be allowed to rise above 2.4 to 4.0 mg/L (Anon, 1958). Water with less than 0.9 to 1.0 mg/L fluoride will seldom cause mottled tooth enamel in children, and there is abundant literature to show the advantages of maintaining a fluoride level of 0.8 to 1.5 mg/L. In adults, less than 3.0 to 4.0 mg/L will not cause endemic cumulative skeletal effects (McKee and Wolf, 1963; McClure et al., 1945; and McClure and Kinser, 1944); fluorides up to 5.0 mg/L cause no effects except mottling of tooth enamel (Smith and Cox, 1952; Heyroth, 1952; and Hillboe and Ast, 1951). Adverse effects of fluoride at concentrations of 5 to 8 mg/L are essentially limited to effects on tooth enamel, although some individuals are reported to experience bone density changes at these concentrations. At 12 to 20 mg/L crippling fluorosis occurs (Anon, 1958). Radiologic surveys of people who lived more than 15 years in Bartlett, Texas, where water had 8 mg/L fluoride, showed minimal increase in bone density in only 12% of the people, but no interference with the use of bones or joints (Anon, 1971). Mortality rates from kidney disease, heart disease or cancer, in high and low fluoride areas, show no association with fluoride levels (Smith and Fox, 1952; and Heyroth, 1952). It is estimated that daily levels of 15 to 20 mg fluoride for several years would be required to induce chronic fluorosis in adult man (Mitchell and Edman, 1953). Polio incidence has been shown to be lower in areas where surface water contains over 1.0 mg/L fluoride (Shay, 1947; and Shay, 1948), but no clear-cut conclusion could be reached regarding a correlation between fluoride and goiter (Fellenberg, 1938). Literature Criteria Criteria from the literature are summarized in Table 4.1 and 4.2. Criteria for drinking waters are usually based on the most sensitive users: young children whose teeth are still growing. The criteria are set to avoid dental fluorosis (mottling), since teeth are one of the most sensitive tissues (Anon, 1980; and Anon, 1977). In some jurisdictions, the fluoride criterion is based on the total amount of water likely to be consumed, and is thus temperature dependent as shown in Table 4.2 (Anon, 1969; Anon, 1962; Anon, 1968; Anon, 1975; Anon, 1978; Rose and Marier, 1977; and Anon, 1980). In Canada, 1.0 mg/L is used everywhere except the arctic and subarctic zones where 1.2 mg/L is permitted (McNeeley et al., 1979; and Anon, 1979). These are objective levels, the maximum acceptable level is 1.5 mg/L (Anon, 1979). The arctic and subarctic are defined as areas where the annual mean daily maximum temperature is less than 10°C (Anon, 1979). In British Columbia, the optimum concentration is set at 1.2 mg/L (Anon, 1982; and Anon, 1969), while the maximum acceptable concentration is 1.5 mg/L. Class C water in Manitoba, suitable for domestic consumption after a full treatment, should not exceed 1.2 mg/L (Anon, 1980). The maximum acceptable concentration in Ontario drinking water is 2.4 mg/L (Anon, 1984; and Anon, 1982). Where fluoridation is practiced, the recommended fluoride level is 1.2 + 0.2 mg/L (Anon, 1983). WHO recommends a limit of 1.5 mg/L (Anon, 1961) or 1.0 mg/L (Anon, 1958) in drinking water. Water with fluoride in the range of 0.7 to 1.5 mg/L is acceptable for drinking water supplies after any kind of water treatment from simple filtration and disinfection to full and complete treatment (Anon, 1975). Recommended Criteria The recommended total fluoride level in raw drinking water is 1.0 mg/L as a 30-day average with a maximum value of 1.5 mg/L (Anon, 1979; Anon, 1982; Anon, 1969; Anon, 1987; Anon, 1962; and Anon, 1987). This is consistent with British Columbia, CCREM and United States Public Health guidelines. Rationale The fluoride criteria are temperature dependent. The lowest isotherm used is 12°C (Anon, 1962), or 10°C (Anon, 1979; and Anon, 1987), and in both these cases the optimum fluoride level is 1.2 mg/L. In British Columbia the mean annual daily maximum temperature is below the 10°C isotherm, except for two areas which may reach 10°C or 11°C some years, but never reach the 12°C isotherm. The areas where the mean may be this high are, however, the most highly populated areas of southeastern Vancouver Island, the lower Fraser Valley and the Okanagan Valley. In these areas, the optimum level of fluoride would be 1.0 mg/L (Anon, 1979; and Anon, 1987), or 1.2 mg/L (Anon, 1962) but in all other areas of British Columbia the optimum level would be 1.2 mg/L (Anon, 1979; Anon, 1982; and Anon, 1962). The maximum acceptable level is 1.5 mg/L (Anon, 1979; Anon, 1982; and Anon, 1987), or 1.7 mg/L (Anon, 1962). Only a few areas in British Columbia reach the 10°C isotherm which, according to two references, Anon, 1979 and Anon, 1987, would require water with 1.0 mg/L fluoride, but which according to another reference, Anon, 1962, still qualifies them for 1.2 mg/L fluoride in the water. These isotherms are not always reached and these areas are thus marginal cases. It is not judged that any harm would result from using the 1.2 mg/L criterion uniformly throughout British Columbia especially since the peak summer temperatures in the Victoria and Vancouver areas are relatively low and the main reason the mean reaches 10°C is due to relatively high winter temperatures. Thus, summer water consumption would not be inordinately high. However, the 1.0 mg/L fluoride concentration is recommended since it should adequately protect teeth and allows a little more safety margin between the therapeutic dose and harmful levels. The fluoride criterion is unique in this regard since most other criteria have at least, and usually more than, a ten-fold safety factor between the criterion and any known effect levels. This is not possible with fluoride and in any case the first "harmful" effects to appear are cosmetic and not functional.