Journal of Orthopaedic Reseurch
10581-587 Kaven Press, Ltd., N c y York C 1992 Orthopaedic Research Soclety
The Effects of Fluoridated Water on Bone Strength Charles H. Turner, *M. P. Akhter, and *Robert P. Heaney Orthopciedic Surgery Department, Indiana University, Indianapolis, Indicina; und 'Crnter for Hard Tissue Research, Creightnn University, Omaha, Nebraska, U.S.A.
Summary: Fluoride from fluoridated water accumulates not only in the enamel of teeth but also in the skeleton. The effects of fluoridated water on the skeleton are not well understood. yet there is qome evidence that fluoridated water consumption increases the incidence of fractures. In the present study, femoral bending strength was measured in rats on fluoride intakes that ranged from low levels to levels well above natural high fluoride drinking water. Bone strength followed a biphasic relationship with bone fluoride content. Fluoride had a povitive effect on bone strength for lower fluoride intakes and a negative influence on bone strength for higher fluoride intakes. The vertebral fluoride content at which femoral strength was maximum was between 1,100 and 1,500 ppm. The increase in femoral strength at this fluoride level was not accompanied by an increase in femoral bone density. The optimal fluoride content is within the range of bone fluoride contents found in persons living in regions with fluoridated water (1 ppm) for > l o years. Key Words: Bone-FluorideBone strength.
Simonen and Laitinen (21) reported a decrease in hip fracture incidence in a fluoridated community ( I ppm of fluoride) in Finland compared with a neighboring nonfluoridated community. Conversely, Arnala et al. (4) reported no difference in hip fracture incidence in individuals from the same two communities studied by Simonen and Laitinen. Recent epidemiological studies of elderly populations in the United States and England showed a positive ecological association between water fluoride levels and hip fracture (7,15). A prospective trial by Sowers et al. (22) showed an increase in fracture incidence in a high fluoride community (4 ppm) compared with a fluoridated community (1 ppm). In sum, these studies demonstrate that the effects of fluoridated water on bone strength require further investigation.
Fluoride affects bone strength in at least two different ways: (a) Fluoride ions replace the hydroxyl ions in bone crystals to form fluorapatite (14) and (b) high serum levels of fluoride cause osteoblast activity to increase (11). It is commonly thought, though not rigorously proven, that a minimum serum fluoride level of 95 ng/ml must be achieved before bone cells will be stimulated (24). Ekstrand and Spak (9) have shown, in humans, that a bolus of 250 ml of 80-ppm fluoride solution (or 20 mg of fluoride) causes a peak serum fluoride level of 700 ngiml. Linear extrapolation of this result predicts peak serum fluoride levels of 8.75 ng/ml after consumption of 250 ml of fluoridated drinking water (1 ppm) and 35 ng/ml after consumption of 250 ml of high fluoride well water (4 ppm). In each case, the expected serum fluoride levels are less than the threshold for bone cell activation (95 ngiml). One might contrive a situation in which high fluoride water consumption could lead to bone mass changes: For instance, consumption of 1 L of high fluoride well water would cause a peak serum fluo-
Received May 15, 1991; accepted December 13, 1991. Address correspondence and reprint requests to Dr. C. H. Turner at Orthopaedic Surgery, Indiana University, 541 Clinical Dr., Rm. 600. Indianapolis, IN 46202, U.S.A.
581
582
C . H . TURNER ET AL.
ride level of 120 ngiml. However, Sowers et al. (23) found no difference between bone mass of people who live in a high fluoride community (4 ppm of F in the water) and people who live in a low fluoride community (1 ppm). Therefore, it is reasonable to conclude that, with respect to drinking water fluoride contents up to 4 ppm, the effect of fluoride is limited to the formation of fluorapatite with negligible effects on bone cells. However, mineral waters with higher levels of fluoride have been shown to affect bone cells. For instance, drinkers of St.Yorre mineral water, which contains 8.5 ppm fluoride, have increased bone mineral density and increased serum levels of osteocalcin (18). It should be noted that the individuals studied by Meunier et al. (18) consumed large amounts of St.-Yorre water (a minimum of 750 ml/day). Our concerns in this article are directed toward the effects of drinking water fluoride levels up to 4 ppm, so the pertinent question must be: Does the incorporation of fluoride into bone mineral have an effect on bone strength? To answer this question, we raised rats on fluoride intakes that ranged from low levels to levels well above natural high fluoride drinking water and measured the strengths of rat femoral bone at varying bone fluoride contents. METHODS
Animal Care Adult rats, with different levels of fluoride incorporated uniformly throughout their skeletons, were raised by the following procedure: Fluoridedeprived rat pups were provided by Amitech (Omaha, N E , U.S.A.). Pregnant rats were given a fluoride-free diet throughout pregnancy and nursing to produce fluoride-deprived pups. After the pups were weaned (21 days of age), 61 were randomly divided into nine experimental groups. A fluoridefree group (1 3 rats) received fluoride-free rat chow and distilled water. The rat diet was prepared by Teklad (Madison, W1, U.S.A.). The diet was nutritionally adequate, containing a standard mineral and vitamin mix (0.6% calcium, 0.4% phosphorus, and 2 U vitamin D,/g) but was limited to 1,216; see Fig. 41. The regression was significant ( r = 0.394, p = 0.0097). The model showed bone strength to increase 18% as bone fluoride content increased from 100 to 1,216 ppm. Bone strength then decreased 31% as bone fluoride content increased from 1,216 to 10,000 ppm. Separate regressions were done for bone fluoride contents of 1,216 ppm. The regression for low bone fluoride contents (I ,216 ppm) was significant and negative ( r = -0.49, p = 0.02). The biphasic relationship between femoral bone strength and bone fluoride content was affirmed by regression analysis to cubic and quadratic equa-
ppm was significantly greater than that of the 64ppm group (p = 0.04) and the 128-ppm group (p = 0.002). There was some change in bone density over the range of fluoride intakes (p = 0.05 by ANOVA). A Fisher LSD test indicated that the bone density of the 128-ppm group was significantly less than that of the 0-ppm group (p = 0.03), the I-ppm group (p = 0.005), the 4-ppm group (p = 0.03), the 16-ppm group (p = 0.02), and the 32-ppm group (p = 0.02) (see Fig. 3). A Newman-Keuls test indicated that the bone density of the 128-ppm group was significantly less than that of the 1-ppm group (p = 0.03). However, neither the Fisher LSD test (p > 0.4) nor the Newman-Keuls test (p > 0.8) indicated a significant increase in bone density for the 16-ppm compared with the 0-ppm group. Therefore, the increase in femoral strength at a fluoride level of 16 ppm was not accompanied by an increase in femoral bone density. A segmented regression model showed a biphasic
25
4
2.1. 21.
22:.
.
21[
FIG. 3. Mean values and standard deviations of femoral bone density as a function of fluoride intake.
’
2.
.
IY.
in11
l
J
~
,
,
,
,
,
Orthnp Rex, Vol. 10, N O . 4, 1992
,
,
.
,
.
,
,
,
,
,
.
,
,
~
585
WATER FLUOHiDATION AND BONE STRENGTH
FIG. 4. Femoral bone strength was related to fluoride content in bone mineral using a segmented regression model: y = 69.4 + 17.8 x1 - 60.8(x, - 1,216)x2, where y is the femoral bone strength and x1 is the logarithm of fluoride content in bone mineral. The maximum bone strength occurs at 1,216 ppm of fluoride in the bone ash. The variable x, = 0 if bone fluoride content 4 , 2 1 6 and 1 if bone fluoride content >1,216. Superimposed upon the plot are ranges of bone fluoride content that have been measured in human bone for intakes of 1 ppm fluoridated water (1,3,28), 4 ppm fluoridated water (28), and skeletal fluorosis (6). 100
I [xxi
I(XMX1
Fluoride in Bone Mineral (ppm) tions. The cubic curve fit gave the equation y = -600 + 693x - 214x’ + 21x3(p = 0.01, r = 0.42). The quadratic curve fit gave the equation y = -69 133x - 2 3 . 5 (p ~ ~= 0.01, r = 0.38). In each case y is femoral bone strength and x is the logarithm of bone fluoride content. Each equation predicts increasing bone strength values at low bone fluoride contents (0 to -700 ppm) and decreasing bone strength values at high bone fluoride contents (>700 PPm).
+
DISCUSSION
Post hoc analyses using a Fisher LSD test indicated that a fluoride intake of 16 pprn had a positive effect on bone strength (cornpared with the 0-ppm group) and fluoride intakes of 64 and 128 ppm had negative influences on bone strength (compared with the 0-ppm group). However, a Newman-Keuls test did not detect a significant increase in bone strength at a 16-ppm intake level, but it did show a decrease in bone strength for fluoride intakes of 64 and 128 ppm. Of these two statistical tests, the Newman-Keuls is the more conservative. This means that the Newman-Keuls test may have an increased probability of Type I1 error compared to the Fisher LSD test. For a positive effect of fluoride intake to be demonstrated by the Newman-Keuls test, a larger number of rats would be necessary in each fluoride intake group. Regression analyses affirmed the positive effect of low fluoride intakes
(
10581-587 Kaven Press, Ltd., N c y York C 1992 Orthopaedic Research Soclety
The Effects of Fluoridated Water on Bone Strength Charles H. Turner, *M. P. Akhter, and *Robert P. Heaney Orthopciedic Surgery Department, Indiana University, Indianapolis, Indicina; und 'Crnter for Hard Tissue Research, Creightnn University, Omaha, Nebraska, U.S.A.
Summary: Fluoride from fluoridated water accumulates not only in the enamel of teeth but also in the skeleton. The effects of fluoridated water on the skeleton are not well understood. yet there is qome evidence that fluoridated water consumption increases the incidence of fractures. In the present study, femoral bending strength was measured in rats on fluoride intakes that ranged from low levels to levels well above natural high fluoride drinking water. Bone strength followed a biphasic relationship with bone fluoride content. Fluoride had a povitive effect on bone strength for lower fluoride intakes and a negative influence on bone strength for higher fluoride intakes. The vertebral fluoride content at which femoral strength was maximum was between 1,100 and 1,500 ppm. The increase in femoral strength at this fluoride level was not accompanied by an increase in femoral bone density. The optimal fluoride content is within the range of bone fluoride contents found in persons living in regions with fluoridated water (1 ppm) for > l o years. Key Words: Bone-FluorideBone strength.
Simonen and Laitinen (21) reported a decrease in hip fracture incidence in a fluoridated community ( I ppm of fluoride) in Finland compared with a neighboring nonfluoridated community. Conversely, Arnala et al. (4) reported no difference in hip fracture incidence in individuals from the same two communities studied by Simonen and Laitinen. Recent epidemiological studies of elderly populations in the United States and England showed a positive ecological association between water fluoride levels and hip fracture (7,15). A prospective trial by Sowers et al. (22) showed an increase in fracture incidence in a high fluoride community (4 ppm) compared with a fluoridated community (1 ppm). In sum, these studies demonstrate that the effects of fluoridated water on bone strength require further investigation.
Fluoride affects bone strength in at least two different ways: (a) Fluoride ions replace the hydroxyl ions in bone crystals to form fluorapatite (14) and (b) high serum levels of fluoride cause osteoblast activity to increase (11). It is commonly thought, though not rigorously proven, that a minimum serum fluoride level of 95 ng/ml must be achieved before bone cells will be stimulated (24). Ekstrand and Spak (9) have shown, in humans, that a bolus of 250 ml of 80-ppm fluoride solution (or 20 mg of fluoride) causes a peak serum fluoride level of 700 ngiml. Linear extrapolation of this result predicts peak serum fluoride levels of 8.75 ng/ml after consumption of 250 ml of fluoridated drinking water (1 ppm) and 35 ng/ml after consumption of 250 ml of high fluoride well water (4 ppm). In each case, the expected serum fluoride levels are less than the threshold for bone cell activation (95 ngiml). One might contrive a situation in which high fluoride water consumption could lead to bone mass changes: For instance, consumption of 1 L of high fluoride well water would cause a peak serum fluo-
Received May 15, 1991; accepted December 13, 1991. Address correspondence and reprint requests to Dr. C. H. Turner at Orthopaedic Surgery, Indiana University, 541 Clinical Dr., Rm. 600. Indianapolis, IN 46202, U.S.A.
581
582
C . H . TURNER ET AL.
ride level of 120 ngiml. However, Sowers et al. (23) found no difference between bone mass of people who live in a high fluoride community (4 ppm of F in the water) and people who live in a low fluoride community (1 ppm). Therefore, it is reasonable to conclude that, with respect to drinking water fluoride contents up to 4 ppm, the effect of fluoride is limited to the formation of fluorapatite with negligible effects on bone cells. However, mineral waters with higher levels of fluoride have been shown to affect bone cells. For instance, drinkers of St.Yorre mineral water, which contains 8.5 ppm fluoride, have increased bone mineral density and increased serum levels of osteocalcin (18). It should be noted that the individuals studied by Meunier et al. (18) consumed large amounts of St.-Yorre water (a minimum of 750 ml/day). Our concerns in this article are directed toward the effects of drinking water fluoride levels up to 4 ppm, so the pertinent question must be: Does the incorporation of fluoride into bone mineral have an effect on bone strength? To answer this question, we raised rats on fluoride intakes that ranged from low levels to levels well above natural high fluoride drinking water and measured the strengths of rat femoral bone at varying bone fluoride contents. METHODS
Animal Care Adult rats, with different levels of fluoride incorporated uniformly throughout their skeletons, were raised by the following procedure: Fluoridedeprived rat pups were provided by Amitech (Omaha, N E , U.S.A.). Pregnant rats were given a fluoride-free diet throughout pregnancy and nursing to produce fluoride-deprived pups. After the pups were weaned (21 days of age), 61 were randomly divided into nine experimental groups. A fluoridefree group (1 3 rats) received fluoride-free rat chow and distilled water. The rat diet was prepared by Teklad (Madison, W1, U.S.A.). The diet was nutritionally adequate, containing a standard mineral and vitamin mix (0.6% calcium, 0.4% phosphorus, and 2 U vitamin D,/g) but was limited to 1,216; see Fig. 41. The regression was significant ( r = 0.394, p = 0.0097). The model showed bone strength to increase 18% as bone fluoride content increased from 100 to 1,216 ppm. Bone strength then decreased 31% as bone fluoride content increased from 1,216 to 10,000 ppm. Separate regressions were done for bone fluoride contents of 1,216 ppm. The regression for low bone fluoride contents (I ,216 ppm) was significant and negative ( r = -0.49, p = 0.02). The biphasic relationship between femoral bone strength and bone fluoride content was affirmed by regression analysis to cubic and quadratic equa-
ppm was significantly greater than that of the 64ppm group (p = 0.04) and the 128-ppm group (p = 0.002). There was some change in bone density over the range of fluoride intakes (p = 0.05 by ANOVA). A Fisher LSD test indicated that the bone density of the 128-ppm group was significantly less than that of the 0-ppm group (p = 0.03), the I-ppm group (p = 0.005), the 4-ppm group (p = 0.03), the 16-ppm group (p = 0.02), and the 32-ppm group (p = 0.02) (see Fig. 3). A Newman-Keuls test indicated that the bone density of the 128-ppm group was significantly less than that of the 1-ppm group (p = 0.03). However, neither the Fisher LSD test (p > 0.4) nor the Newman-Keuls test (p > 0.8) indicated a significant increase in bone density for the 16-ppm compared with the 0-ppm group. Therefore, the increase in femoral strength at a fluoride level of 16 ppm was not accompanied by an increase in femoral bone density. A segmented regression model showed a biphasic
25
4
2.1. 21.
22:.
.
21[
FIG. 3. Mean values and standard deviations of femoral bone density as a function of fluoride intake.
’
2.
.
IY.
in11
l
J
~
,
,
,
,
,
Orthnp Rex, Vol. 10, N O . 4, 1992
,
,
.
,
.
,
,
,
,
,
.
,
,
~
585
WATER FLUOHiDATION AND BONE STRENGTH
FIG. 4. Femoral bone strength was related to fluoride content in bone mineral using a segmented regression model: y = 69.4 + 17.8 x1 - 60.8(x, - 1,216)x2, where y is the femoral bone strength and x1 is the logarithm of fluoride content in bone mineral. The maximum bone strength occurs at 1,216 ppm of fluoride in the bone ash. The variable x, = 0 if bone fluoride content 4 , 2 1 6 and 1 if bone fluoride content >1,216. Superimposed upon the plot are ranges of bone fluoride content that have been measured in human bone for intakes of 1 ppm fluoridated water (1,3,28), 4 ppm fluoridated water (28), and skeletal fluorosis (6). 100
I [xxi
I(XMX1
Fluoride in Bone Mineral (ppm) tions. The cubic curve fit gave the equation y = -600 + 693x - 214x’ + 21x3(p = 0.01, r = 0.42). The quadratic curve fit gave the equation y = -69 133x - 2 3 . 5 (p ~ ~= 0.01, r = 0.38). In each case y is femoral bone strength and x is the logarithm of bone fluoride content. Each equation predicts increasing bone strength values at low bone fluoride contents (0 to -700 ppm) and decreasing bone strength values at high bone fluoride contents (>700 PPm).
+
DISCUSSION
Post hoc analyses using a Fisher LSD test indicated that a fluoride intake of 16 pprn had a positive effect on bone strength (cornpared with the 0-ppm group) and fluoride intakes of 64 and 128 ppm had negative influences on bone strength (compared with the 0-ppm group). However, a Newman-Keuls test did not detect a significant increase in bone strength at a 16-ppm intake level, but it did show a decrease in bone strength for fluoride intakes of 64 and 128 ppm. Of these two statistical tests, the Newman-Keuls is the more conservative. This means that the Newman-Keuls test may have an increased probability of Type I1 error compared to the Fisher LSD test. For a positive effect of fluoride intake to be demonstrated by the Newman-Keuls test, a larger number of rats would be necessary in each fluoride intake group. Regression analyses affirmed the positive effect of low fluoride intakes
(
