6
Reproductive and Developmental Effects of Fluoride
This chapter provides an update on studies of the reproductive and developmental effects of fluoride published since the earlier NRC (1993) review. Studies on reproductive effects are summarized first, primarily covering structural and functional alterations of the reproductive tract. This is followed by a discussion of developmental toxicity in animal and human studies.
REPRODUCTIVE EFFECTS
More than 50 publications since 1990 have focused on the reproductive effects of fluoride. Most of the studies used animal models, primarily rodents, and evaluated structural or functional alterations in the male reproductive tract associated with fluoride. Fewer animal studies evaluated the effects of fluoride on female reproductive tract structure or function. In this section, reports of fluoride effects on reproduction in animal models are reviewed first, followed by a discussion of the available studies of humans.
Animal Studies
The large number of studies gleaned from a search of the literature since 1990 that evaluated reproductive tract structure or function in animal models are outlined in Table 6-1, listing the fluoride dosing regimens and main observations. Most of the studies were conducted for the purpose of hazard identification and involved high doses of fluoride to reveal potentially sensitive reproductive-tract targets and pathways. A few selected
TABLE 6-1 Reproductive Toxicity Studies
Species, Sex, Number |
Exposure Route |
Concentration/Dose |
Exposure Duration |
Effects |
Reference |
Mice, F, 15/group |
Gavage |
10 mg/kg/day (NaF) |
30 days |
Decreased protein in liver, muscle, and small intestine were observed. Significant accumulation of glycogen in gastrocnemius muscle and liver. Decline in succinate dehydrogenase activity in pectoralis muscle of treated mice. Administration of ascorbic acid and calcium to NaF-treated mice caused significant recovery from fluoride toxicity. |
Chinoy et al. 1994 |
Mice, F, 25/group |
Orally, feeding tube attached to hypodermic syringe |
5 mg/kg/day (NaF) |
45 days |
Fluoride concentrations were increased in the urine, serum, and ovary compared with controls. In the ovary, there was impaired production of gluthathione and impaired function of the protective enzymes—namely, gluthathione peroxidase, superoxide dismutase, and catalase. There was increased ovarian lipid peroxidation. Enhanced concentrations of potassium and sodium were observed in the serum. The concentrations of serum calcium showed significant depletion. Withdrawal of NaF for 45 days showed partial recovery. Recovery was enhanced by treatment with ascorbic acid, calcium, vitamin E, and vitamin D. |
Chinoy and Patel 1998 |
Mice, F, 20/group |
Gavage |
10 mg/kg/day (NaF) |
30 days |
Significant decline of ovarian protein and 3β- and 17β-hydroxysteroid dehydrogenase activities. Hypocholesterolemic effect in serum detected. Accumulation of glycogen in uterus. |
Chinoy and Patel 2001 |
Mice, M, 40/group |
Drinking water |
10, 20 mg/kg/ day (NaF) |
30 days |
Epithelial-cell pyknosis and absence of luminal sperm were observed. Disorganization of germinal epithelial cells of seminiferous tubules with absence of sperm in the lumina. Reduction in denudation of cells, epithelial cell height, nuclear pyknosis, and absence of sperm observed in the cauda epididymis. The vas deferens epithelium showed clumped sterocilia, nuclear pyknosis, and cell debris but no sperm in the lumen and an increase in the lamina propria. Marked recovery was observed with withdrawal of treatment. No effects observed in the prostate gland or seminal vesicles. |
Chinoy and Sequeira 1989 |
Mice, M, 20/group |
Gavage |
10, 20 mg/kg/day (NaF) |
30 days |
NaF caused lessened fertility rate when normal cycling female mice were mated with treated mice. Large numbers of deflagellated spermatozoa with acrosomal, midpiece, and tail abnormalities were observed. Significant recovery in sperm count, sperm motility, and fertility rate was observed after withdrawal of treatment for 2 months. |
Chinoy and Sequeira 1992 |
Mice, M, 20/group |
Gavage |
10 mg/kg/day (NaF) |
30 days |
Alterations in epididymal milieu as elucidated by the significant decrease in concentrations of sialic acid and protein as well as activity of ATPase in epididymides. Significant decrease in body and epididymis weight. Weight of vas deferens and seminal vesicle were not affected. Sperm maturation process was affected, leading to decline in cauda epididymal sperm motility and viability. Significant reduction in fertility rate and cauda epididymal sperm count. Treatment induced substantial metabolic alterations in the epidymides, vas deferens, and seminal vesicles of mice. Supplements of vitamin D and E during the withdrawal period enhanced recovery of all NaF-induced effects. |
Chinoy and Sharma 1998 |
Species, Sex, Number |
Exposure Route |
Concentration/ Dose |
Exposure Duration |
Effects |
Reference |
Mice, M, 20/group |
Gavage |
10 mg/kg/day (NaF) |
30 days |
Significant decline in sperm acrosomal acrosin and hyaluronidase. Acrosomal damage and deflagellation observed. Sperm nuclear integrity not affected. Structural and metabolic alterations and reduced activity of the enzymes in sperm resulted in a significant decrease in sperm count and poor fertility rate. Cessation of NaF treatment for 30 days did not bring about complete recovery. Administration of ascorbic acid or calcium enhanced recovery and was more pronounced in groups treated with both ascorbic acid and calcium. |
Chinoy and Sharma 2000 |
Mice, M, 10/group |
Drinking water |
100, 200, 300 mg/L (NaF) Mean doses during 4-week treatment: 12.53, 21.80, 39.19 mg/kg/day Mean doses during 10-week treatment: 8.85, 15.64, and 27.25 mg/kg/day |
4 and 10 weeks |
Fertility reduced significantly at 100, 200, and 300 mg/L after 10 weeks but not after 4 weeks. Implantation sites and viable fetuses were significantly reduced in females mated with males that had ingested NaF at a concentration of 200 mg/L for10 weeks. Relative weights of seminal vesicles and preputial glands were significantly increased in animals exposed to NaF 200 and 300 mg/L for 4 weeks but not in animals exposed for 10 weeks. |
Elbetieha et al. 2000 |
Rat, F, 25 (treated), 18 (control) |
Drinking water |
150 mg/L (NaF) |
From 60 days before mating and through pregnancy and lactation |
There was inhibition of lactation in rats with chronic fluorosis, as measured by slower rates of body weight gain in pups and lower amount of milk suckled in 30 minutes compared with control pups. Prolactin concentration was decreased in serum but increased in the pituitary gland. Microscopic examination showed accumulation of large mature secretory granules and appearance of extremely large abnormal secretory granules in lactotroph cytoplasma. |
Yuan et al.1994 |
Rat, F, 33-35/group |
Drinking water |
10, 25, 100, 175, 250 mg/L (NaF) Mean doses: 1.4, 3.9, 15.6, 24.7, and 25.1 mg/kg/day (NaF) |
From day of sperm detection to gestation day 20. |
Significant reductions in maternal water consumption in the two highest dose groups and a significant reduction in maternal feed consumption in the high-dose group. Body weights of dams were reduced in the higher-dose groups. No significant effect on any reproductive end points. Developmental effects of fluoride were minimal, with 250 mg/L (25.1 mg/kg/day being the lowest observed effect level due to skeletal variations). |
Collins et al. 1995 |
Rat, F, 10/group |
Drinking water |
200, 400, and 600 mg/L (NaF) Mean doses: 22.58, 18.35, and 28.03 mg/kg/day (NaF) |
30 days, before mating |
None of the rats in the 28.03 mg/kg/day group survived the study period, and only three survived from the 18.35 mg/kg/day group. Clinical signs of toxicity (dehydration, lethargy, hunched posture) were observed in these groups. All the rats exposed to 22.58 mg/kg/ day survived, and showed no signs of toxicity. Fetotoxicity observed at 22.58 mg/kg/day. Reduced number of viable fetuses, increased number of pregnant rats with resorptions, and increased total number of resorptions. |
Al-Hiyasat et al. 2000 |
Rat, F, 10/group |
Gavage |
40 mg/kg/day (NaF) |
Days 6 to 19 of gestation |
Significant reductions in body weight, feed consumption, absolute uterine weight, and number of implantations. Significantly higher incidence of skeletal and visceral abnormalities. When NaF was administered with vitamin C, the total percentage of skeletal and visceral abnormalities was significantly lower compared with the group treated with NaF only. Vitamin E also had that effect but was not as great as vitamin C. |
Vermaand Guna Sherlin 2001 |
Species, Sex, Number |
Exposure Route |
Concentration/ Dose |
Exposure Duration |
Effects |
Reference |
Rat, M, 15-20/group |
Single microdose injection into the vasa deferentia |
50 µg/50 µL (NaF) |
Single dose injection |
Arrest of spermatogenesis and absence of spermatozoa in the lumina of the seminiferous tubules of the testes. This resulted in a decline in sperm count in caudae epididymides. Deflagellation and tail abnormalities were observed. |
Chinoy et al. 1991a |
Rat, M, 12/group |
Drinking water |
5 and 10 mg/kg/day (NaF) |
30 days |
Succinate dehydrogenase activity in the testes, adenosine triphosphatase activity, and sialic acid concentrations in epididymides in testes were inhibited. A more pronounced effect was observed on the cauda epididymis. Testicular cholesterol and serum testosterone concentrations were not affected. Significant decline in fertility attributed to decreased sperm motility and count. |
Chinoy et al. 1992 |
Rat, M, 14/group |
Drinking water |
100 and 200 mg/L (NaF) |
6 and 16 weeks |
Severalfold increase in fluoride concentrations in the testes and bone at both test concentrations compared with controls. Fifty percent of the rats in both treatment groups exhibited histopathologic changes in the germinal epithelium of the testes after 16 weeks. Concentrations of copper and manganese in the testes, liver, and kidneys were not changed. Iron concentrations in the testes and plasma were not affected by fluoride but were increased in the liver, kidneys, and bone. Concentrations of zinc in the testes, plasma, liver, and kidneys decreased significantly, particularly in the 16-week groups. Zinc tended to increase in the bone. |
Krasowska and Wlostowski 1992 |
Rat, M, 25-30/group |
Gavage |
10 mg/kg/day (NaF) |
50 days |
After 50 days of treatment, sperm acrosomal hyaluronidase and acrosin were reduced. Other observations included acrosomal damage and deflagellation of sperm, decline in sperm motility, decreased cauda epididymal sperm count, and reduced fertility. Incomplete recovery observed at withdrawal of NaF treatment for 70 days. Ascorbic acid and calcium produced significant recovery of NaF-induced effects. |
Narayana and Chinoy 1994a |
Rat, M, 10/group |
Drinking water, administered before feeding |
10 mg/kg/day (NaF) |
50 days |
No significant change in testicular cholesterol concentrations. Testicular 3β-HSD and 17β-HSD activities were modestly decreased by NaF ingestion. Histomorphometric analyses indicated a significant change in the Leydig cell diameter in correlation with androgen concentrations. |
Narayana and Chinoy 1994b |
Rat, M, 10-30/group |
Gavage |
10 mg/kg/day (NaF) |
30 and 50 days |
Significant elevation in serum fluoride concentrations (3.6 ± 0.11 ppm) with a simultaneous rise in sperm calcium. Treatment resulted in structural and metabolic alterations in sperm, leading to low sperm motility, low sperm mitochondrial activity index, reduced viability, and changes in sperm membrane phospholipids. A significant reduction in electrolyte concentrations of sperm was observed. Protein concentrations in cauda epididymal sperm suspension, vas deferens, seminal vesicle, and prostate significantly decreased after treatment. Glycogen accumulated in vas deferens and fructose decreased in seminal vesicles and vas deferens. |
Chinoy et al. 1995 |
Rat, M, 18/group |
Drinking water |
100, 200 mg/L (NaF) |
2, 4, 6 weeks |
Serum testosterone concentration decreased with time in exposed rats. Testis cholesterol concentration was significantly decreased in the liver of rats exposed 4 and 6 weeks. |
Zhao et al. 1995 |
Rat, M, 24/group |
Injection, left testis |
50, 175, 250 ppm (NaF) |
Single injection |
Seminiferous tubule damage observed in vehicle-injected control and exposed testes; no damage was observed in noninjected testes. Polymorphonuclear leukocyte infiltration was observed at injection site in both vehicle- and fluoride-injected groups after 24 hours. No effect on Leydig cells. |
Sprando et al. 1996 |
Species, Sex, Number |
Exposure Route |
Concentration/ Dose |
Exposure Duration |
Effects |
Reference |
Rat, M, 12/group |
Drinking water |
25, 100, 175, 200 mg/L (NaF) |
14 weeks (10 weeks pretreatment, 3 weeks mating, 1 week postmating) |
No effects were observed within the P generation males and the F1 generation groups in testis weights, prostate/seminal vesicle weights, nonreproductive organ weights, testicular spermatid counts, sperm production per gram of testis per day, sperm production per gram of testis, lutenizing hormone, follicle-stimulating hormone, or serum testosterone concentrations. No histological changes were observed in testicular tissues from either the P or the F1 generation. |
Sprando et al. 1997 |
Rat, M, 25 |
Drinking water |
25, 100, 175, 250 mg/L (NaF) |
In utero, during lactation, 14-weeks post-weaning |
No significant effect on absolute volume of the seminiferous tubules, interstitial space, Leydig cells, blood vessel boundary layer, lymphatic space, macrophages, tubular lumen or absolute tubular length and absolute tubular surface area, mean Sertoli cell nucleoli number per tubular cross-section, mean seminiferous tubule diameter, and mean height of the seminiferous epithelium. Statistically significant decrease in the absolute volume and volume percent of the lymphatic endothelium was observed in NaF-treated groups (175 and 250 mg/L) and in the testicular capsule in the NaF-treated group (100 mg/L). |
Sprando et al. 1998 |
Rat, M, F, 36-48/group 3 generations |
Drinking water |
0, 25, 100, 175, 250 mg/L (NaF) |
10 weeks |
Decreased fluid consumption observed at 175 and 250 mg/L attributed to decreased palatability; no effect on reproduction. No cumulative effects were observed in any generation. Mating, fertility, and survival, organ-to-body weight ratios, and organ-to-brain ratios were not affected. Treatment up to 250 mg/L did not affect reproduction. |
Collins et al. 2001a |
Rat, M, 6/group |
Gavage |
20 mg/kg/day(NaF) |
29 days |
Testicular 3β-HSD and 17β-HSD activities were decreased significantly. Substantial reduction in plasma concentrations of testerosterone in the exposed group. Decreased epididymal sperm count and fewer mature luminal spermatozoa in the exposed group. NaF treatment was associated with oxidative stress, as indicated by an increased concentration of conjugated dienes in the testis, epididymis, and epididymal sperm pellet. Significant reduction in peroxidase and catalase activities in the sperm pellet in exposed group as compared with controls. |
Ghosh et al. 2002 |
Rat, M, F, 10/group |
Gavage |
40 mg/kg/day (NaF) |
Day 6 of gestation to day 21 of lactation |
NaF treatment associated with significant reductions in body weight, feed consumption, concentration of glucose, and protein in the serum. Administration of vitamins C, D, and E helped to restore body weight loss as well as glucose, protein, sodium, and potassium concentrations in the serum of exposed rats. Withdrawal of NaF treatment during lactation caused significant amelioration in feed consumption and in serum sodium, potassium, glucose, and protein concentrations. Additional treatment with vitamin E caused substantial improvements in body weight reductions and in serum concentration of sodium, potassium, glucose, and protein. |
Verma and Guna Sherlin 2002a |
Rabbit, F, 10/group |
Subcutaneous injection |
5, 10, 20, 50 mg/kg/day (NaF) |
100 days |
Abnormal accumulation of lipids in testes observed in treated rabbits. Hyperphospholipidemia, hypertriglyceridemia, and hypercholesterolemia indicated enhanced lipid biosynthesis was observed in response to fluoride toxicosis. Significant (P < 0.001) increase in amount of free fatty acids observed in testes of treated animals. |
Shashi 1992a |
Rabbit, M, 5/group |
Feed |
20, 40 mg/kg/day (NaF) |
30 days |
Decline in fertility related to reduced sperm motility and count and changes in morphology and metabolism. No recovery after withdrawal for 30 days from treatment. With administration of ascorbic acid and calcium, marked recovery occurred. |
Chinoy et al. 1991b |
Species, Sex, Number |
Exposure Route |
Concentration/Dose |
Exposure Duration |
Effects |
Reference |
Rabbit, M, 10/group |
Drinking water |
10 mg/kg/day (NaF) |
18 or 29 months |
Loss of cilia on the epithelial cells lining the lumen of the ductuli efferentes of the caput epididymidis and of stereocilia on the epithelial cells lining the lumen of the vas deferens was observed. The boundaries of cells peeled off and were not clear in some regions of the epithelial lining of the lumen of the ductuli efferentes and vas deferens. Cessation of spermatogenesis was noted only in rabbits treated for 29 months. |
Susheela and Kumar 1991 |
Rabbit, M, 8/group |
Drinking water |
10 mg /kg/day (NaF) |
18 months |
Structural defects in the flagellum, the acrosome, and the nucleus of the spermatids and epididymal spermatozoa were observed in the treated rabbits. Absence of outer microtubules, complete absence of axonemes, structural and numeric aberrations of outer dense fibers, breakdown of the fibrous sheath, structural defects in the mitochondria of the middle piece of the flagellum, and detachment and peeling of the acrosome from the flat surfaces of the nucleus was observed. |
Kumar and Susheela 1994 |
Rabbit, M, 12/group |
Drinking water |
10 mg/kg/day (NaF) |
20 and 23 months |
Fluoride concentrations in the sera of treated animals were significantly increased. Loss of stereocilia, significant decrease in the height of the pseudostratified columnar epithelium, and significant increase in the diameter of the caput and cauda ductus epididymis observed in the 23-month treatment group. Weights of the cauda epididymis and caput were significantly reduced in the 23-month-treated animals; the number of secretory granules in those organs was reduced. |
Kumar and Susheela 1995 |
Rabbit, M, 12/group |
Drinking water |
10 mg/kg/day (NaF) |
18 and 23 months |
Fluoride concentrations in the sera were significantly increased in treated rabbits (P < 0.001). There was dilation of the smooth endoplasmic reticulum and mitochondrial cristae of the Leydig cells. Leydig cells had lower numbers of lipid droplets and smooth endoplasmic reticulum compared with Leydig cells of unexposed rabbits. Intranuclear filamentous inclusions observed in treated rabbits. Interstitial tissue of the testis was degenerated. |
Susheela and Kumar 1997 |
Guinea pig, M, 10/group |
Gavage |
30 mg/kg/day (NaF) |
30 days |
Structural and metabolic alterations of the cauda epididymal spermatozoa led to substantial decreases in sperm mitochondrial activity index, motility, live/dead ratio. Increases in sperm membrane phospholipids were observed. ATPase, succinate dehydrogenase, and glutathione concentrations were decreased in testis of treated animals. Administration of ascorbic acid led to recovery in these parameters. |
Chinoy et al. 1997 |
Sheepdog, F, M, 5/group |
Feed |
460 ppm (fluoride) |
2 years |
No adverse effect on reproduction attributable to treatment. Bony exostoses was observed in 4 of 10 treated dogs. |
Schellenberg et al. 1990 |
ABBREVIATIONS: F, female; HSD, hydroxysteroid dehydrogenase; M, male. |
examples illustrate the results of the many hazard identification studies: (1) cessation of spermatogenesis and alterations in the epididymis and vas deferens were observed in rabbits administered sodium fluoride (NaF) at 10 milligrams per kilogram (mg/kg) of body weight for 29 months (Susheela and Kumar 1991); (2) effects on Leydig cells and decreased serum testosterone were observed in rats exposed to NaF at 10 mg/kg for 50 days (Narayana and Chinoy 1994b); and (3) decreased protein in the ovary and uterus and decreased activity of steroidogenic enzymes (3β-hydroxysteriod dehydrogenase [HSD] and 17β-HSD) was found in mice treated with NaF at 10 mg/kg for 30 days (Chinoy and Patel 2001). In general, the hazard identification studies show that the reproductive tract is susceptible to disruption by fluoride at a concentration sufficiently high to produce other manifestations of toxicity.
For risk evaluation, a comprehensive multigenerational study of fluoride effects on reproduction using standard guidelines and adequate numbers of animals has been conducted in rats (Collins et al. 2001a). Rats were administered drinking water with NaF at 0, 25, 100, 175, and 250 mg/L over three generations. No compound-related effects were found on mating or fertility; gestation or lactation; or F1 survival, development, and organ weights. No alterations in the teeth were seen except for mild whitening observed in rats exposed to fluoride at 100 mg/L or greater. That well-conducted study concluded that NaF at concentrations up to 250 mg/L in the drinking water did not alter reproduction in rats (Collins et al. 2001a).
Human Studies
The few studies gleaned from a search of the literature since 1990 that evaluated reproductive effects of fluoride ingestion in humans are outlined in Table 6-2, listing the estimated fluoride exposure and main observations. In highly exposed men with and without skeletal fluorosis (fluoride at 1.5-14.5 mg/L in the drinking water), serum testosterone concentrations were significantly lower than in a control cohort exposed to fluoride at less than 1.0 mg/L in drinking water (Susheela and Jethanandani 1996). Although there was a 10-year difference in the mean ages between the skeletal fluorosis patients (39.6 years) and control subjects (28.7 years), this study suggests that high concentrations of fluoride can alter the reproductive hormonal environment.
In an ecological study of U.S. counties with drinking-water systems reporting fluoride concentrations of at least 3 mg/L (Freni 1994), a decreased fertility rate was associated with increasing fluoride concentrations. Because methods for analyzing the potential amounts and direction of bias in ecological studies are limited, it is possible only to discuss some of the strengths and weaknesses of this complicated study (see Chapter 10 and
Appendix C for a more in-depth discussion of ecologic bias). Freni’s study is actually partially ecologic; the outcome (fertility) is age-standardized at the individual level, while exposure to fluoride and covariates are measured at the group level. Controlling for age of the mother is a strength of the study, but to avoid bias all ecological variables should be standardized in the same fashion (Greenland 1992). The model adjusted for a number of important socioeconomic and demographic variables at the group level, but these might not adequately control for individual-level determinants of fertility such as family income and use of contraceptives. For example, median income (a group-level variable) and family income (an individual-level variable) may have independent and interactive effects on outcome. One of the two ecologic exposure measures examined the percentage of the population served by water systems with fluoride concentrations of at least 3 mg/L. That has the potential advantage of not assuming an effect at lower fluoride concentrations (as does the mean fluoride concentration, the other exposure measure), but it has the disadvantage that, unlike individual-level studies, nondifferential misclassification of dichotomous exposures within groups tend to bias ecologic results away from the null (Brenner et al. 1992). While the results of the Freni study are suggestive, the relationship between fertility and fluoride requires additional study.
A study of workers in Mexico, who were occupationally exposed to fluoride (estimated to range from 3 to 27 mg/day) producing hydrofluoric acid and aluminum fluoride, found alterations in serum hormone concentrations with normal semen parameters (Ortiz-Perez et al. 2003). However, that study involved a comparison of a high-fluoride-exposed group and a low-fluoride-exposed group with poorly defined exposures and overlapping exposure characteristics.
Overall, the available studies of fluoride effects on human reproduction are few and have significant shortcomings in design and power, limiting inferences.
DEVELOPMENTAL EFFECTS
There is wide variation with some correlation between fluoride concentration in maternal serum and cord blood, indicating that fluoride readily crosses the placenta. In general, average cord blood concentrations are approximately 60% of maternal serum concentrations, with proportionally lesser amounts present as higher maternal serum concentrations (Gupta et al. 1993; Malhotra et al. 1993; Shimonovitz et al. 1995). Therefore, potential toxicity to the developing embryo and fetus in the setting of high maternal ingestion of fluoride has been a concern evaluated in both animal and humans.
TABLE 6-2 Human Reproductive Studies
Subjects |
Exposure Route, Duration |
Concentration/Dose |
Pregnant women (n = 25) |
Drinking water |
Maternal blood fluoride concentrations ranging from 0.1 to 2.4 ppm |
Pregnant women (n = 25) |
Drinking water |
Maternal plasma fluoride concentrations ranging from 0.12 to 0.42 µg/mL |
Pregnant women undergoing amniocentesis (n = 121, divided into 6 exposure groups) |
Oral doses, 24 hours and 3 hours before amniocentesis |
0.56, 1.12, 1.68, 2.30, or 2.80 mg of NaF corresponding to 0.25, 0.50, 0.75, 1.00, or 1.25 mg of F- |
Men (ages 28-30; n = 8) |
In vitro with spermatozoa, intervals of 5, 10, and 20 minutes |
25, 50, 250 mM (NaF) |
30 regions spread over nine states |
Drinking water |
≥ 3 mg/L (fluoride) |
Pregnant women (n = 22) |
Drinking water |
Maternal serum fluoride concentrations ranging from 0.003-0.041µg/ml |
Men with skeletal fluorosis (n = 30) |
Drinking water |
1.5-14.5 mg/L (fluoride) |
Male workers in Mexico (ages 20-50; n = 126) , who produce fluorohydric acid and aluminum fluoride |
Drinking water |
3-27.4 mg/day (fluoride) |
ABBREVIATIONS: FSH, follicle-stimulating hormone. |
Results |
Reference |
Fairly positive correlation (r = 0.736) between cord blood values and maternal blood fluoride concentrations. On average, the cord blood fluoride concentration was about 60% that in maternal blood. At a maternal fluoride concentration greater than 0.4 ppm, the cord blood fluoride concentration increased by only about 12%. The placenta was found to serve as an effective barrier within this range. |
Gupta et al. 1993 |
Cord plasma fluoride concentrations ranged from 0.11-0.39 µg/ml. In 8% of the cases, cord plasma concentrations were higher than maternal plamsa concentrations. Positive correlation (r = 0.97) in fluoride concentrations between maternal and cord plasma indicates that the placenta allowed passive diffusion of fluoride from mother to fetus. |
Malhotra et al. 1993 |
F-concentration in amniotic fluid was significantly higher than controls in the 1.25 mg/day F-group but not in any of the other exposure groups. No significant correlation between F-concentration in maternal plasma and in aminotic fluid. |
Brambilla et al. 1994 |
Substantial enhancement of acid phosphatase and hyaluronidase activities after 5 and 10 minutes (P < 0.001). Decrease in lysosomal enzyme activity after 20 minutes. Analysis of sperm revealed elongated heads, deflagellation, splitting, loss of the acrosome, and coiling of the tail. Glutathione concentrations exhibited time-dependent decrease with complete depletion after 20 minutes (P < 0.001). Suppressed sperm motility after 20 minutes at a dose of 250 mM (P < 0.001). |
Chinoy and Narayana 1994 |
In this ecological study, there was an association between decreasing total fertility rate and increasing fluoride concentrations in most regions. Combined result was a negative total fertility rate/fluoride association with a consensus combined P value of 0.0002-0.0004. Association was based on population means rather than individual women. |
Freni 1994 |
Cord serum fluoride concentrations ranged from 0.003-0.078 µg/ml, and neonatal serum concentrations ranged from 0.017-0.078 µg/ml. No correlation in fluoride concentrations found between maternal and cord sera, maternal and neonatal sera, or maternal and neonatal sera. |
Shimonovitz et al. 1995 |
Serum testosterone concentrations in patients were significantly lower than controls (P < 0.01). |
Susheela and Jethanandani 1996 |
In the high-fluoride exposure group, a significant increase in FSH (P < 0.05) and a reduction of inhibin-B, free testosterone, and prolactin in serum (P < 0.05) were observed. Decreased sensitivity was found in the FSH response to inhibin-B (P < 0.05) when the high-exposure group was compared with the low-exposure group. Significant partial correlation was observed between urinary fluoride and serum concentrations of inhibin-B (P < 0.028). No abnormalities were found in the semen parameters in either the high- or low-fluoride exposure groups. |
Ortiz-Perez et al. 2003 |
Animal Studies
Studies gleaned from a search of the literature since 1990 that evaluated developmental toxicity in animal models are outlined in Table 6-3, listing the fluoride dosing regimens and main observations. High-dose hazard identification studies, such as a recently reported Xenopus embryo development study using the FETAX assay (Goh and Neff 2003), suggest that developmental events are susceptible to disruption by fluoride.
For risk evaluation, several comprehensive studies of fluoride effects on development using standard guidelines and adequate numbers of animals have been conducted in rats and rabbits (Collins et al. 1995; Heindel et al. 1996; Collins et al. 2001b). Those high-quality studies evaluated fluoride concentrations in drinking water of 0-300 mg/L in rats and 0-400 mg/L in rabbits. Across the studies, there was a trend toward lower maternal body weights and lower maternal intake of food and water at the higher concentrations in both rats and rabbits (250-400 mg/L). Overall, developmental effects of fluoride were minimal, with 250 mg/L in rats being the lowest-observed-adverse-effect level due to skeletal variations (Collins et al. 1995, 2001b). For rabbits, the no-observed-adverse-effect level was >400 mg/L for administration during gestation days 6-19, the period of organogenesis (Heindel et al. 1996).
Human Studies
The few studies gleaned from a search of the literature since 1990 that evaluated developmental effects of fluoride ingestion in humans are outlined in Table 6-4, listing the type of study, estimated fluoride exposure, and main observations. These studies have focused on examining an association between fluoride and three different human developmental outcomes—spina bifida occulta, sudden infant death syndrome, and Down’s syndrome. Two small studies have raised the possibility of an increased incidence of spina bifida occulta in fluorosis-prone areas in India (Gupta et al. 1994, 1995); larger, well-controlled studies are needed to evaluate that possibility further. Studies from New Zealand (Mitchell et al. 1991; Dick et al. 1999) found no association between fluoride and sudden infant death syndrome. In one of those studies (Dick et al. 1999), a nationwide case-control database of sudden infant death syndrome was evaluated for fluoride exposure status and controlled for the method of infant feeding (breast or reconstituted formula) with the conclusion that exposure to fluoridated water prenatally or postnatally at the time of death did not affect the relative risk of sudden infant death syndrome.
A small number of ecologic studies have examined Down’s syndrome (trisomy 21) prevalence among populations in municipalities with differ-
ences in water fluoride concentrations. The possible association of cytogenetic effects with fluoride exposure (see Chapter 10) suggests that Down’s syndrome is a biologically plausible outcome of exposure. There are other indications in the literature that environmental exposures could contribute to an increased incidence of Down’s syndrome births among younger mothers (Read 1982; Yang et al. 1999; Hassold and Sherman 2000; Peterson and Mikkelsen 2000).1 There are many difficulties with analyzing the available data on Down’s syndrome and fluoride. First, the source of the data on Down’s syndrome births must be considered. Sources have included birth certificates, hospital records, and reports from parents. Birth certificates are not an ideal source of data because signs of Down’s syndrome are not always readily apparent at birth and the condition, even when diagnosed early, is not always recorded on the birth certificate. Thus, considerable differences can be expected in the data collected when different sources are used to determine the incidence of the disorder. At the present time, the only firm diagnosis of Down’s syndrome comes from examination of chromosomes or DNA. Second, the mother’s history of exposure to fluoride is difficult to determine. The fact that a woman has a baby in one city does not mean she is from that city or indicate how long she has been in the region. Third, the age of the mother is an important risk factor in the occurrence of children with Down’s syndrome; the rates rise exponentially with age.
TABLE 6-3 Developmental Toxicity Studies
Species, Sex, Number |
Exposure Route |
Concentration/Dose |
Exposure Duration |
Rat, F, 33-35/group |
Drinking water |
0, 10, 25, 100, 175, 250 mg/L (NaF) Mean doses: 0, 1.4, 3.9, 15.6, 24.7, and 25.1 mg/kg/day (NaF) |
From day of sperm detection to gestation day 20 |
Rat, F, 10/group |
Drinking water |
40 mg/kg/day (NaF) |
From day 6 to 19 of gestation |
Rat, M, F, 40-50 animals/group from 4 or 5 litters at each age |
Intraperitoneal injection |
0, 30 and 48 mg/kg (NaF) |
Single injection on postnatal day 1, 8, 15, or 29 |
Rat, M, F, 26/group Rabbit, M, F, 26/group |
Drinking water |
Rat: 0, 50, 150, 300 mg/L (NaF) (mean doses 6.6, 18.3, and 27.1 mg/kg/day) Rabbit: 0, 100, 200, 400 mg/L (NaF) (mean doses 10.3, 18.1, and 29.2 mg/kg/day) |
Rat: from gestational day 6 to 15 Rabbit: from gestational day 6 to 19 |
Rat, M, F, 3 generations (F0, F1, F2), F0: 48 M, 48 F/group; F1: 36 M, 36 F/group; F2: 238 fetuses |
Drinking water |
0, 25, 100, 175, 250 mg/L (NaF) Mean doses: (F0): 3.4, 12.4, 18.8, 28.0 mg/kg/ day (NaF) (F1): 3.4, 13.2, 19.3, 25.8 mg/kg/day (NaF) |
F0: 10 weeks |
Frog (Xenopus) embryo, 20/group |
Incubated with NaF solution |
100-1,000 ppm (NaF) |
2, 3, 4, 5, 9, 14.75 hours after fertilization |
ABBREVIATIONS: EC50, median effective concentration; F, female; LC50, median lethal concentration; M, male; NOAEL, no-observed-adverse-effect level. |
Effects |
Reference |
Significant reductions in maternal water consumption in the two highest-dose groups and a significant reduction in maternal feed consumption in the high-dose group. Body weights of dams were reduced in the higher-dose groups. The only significant developmental effect was an increase in the average number of fetuses with three or more skeletal variations in the 25.1-mg/kg/day group. |
Collins et al. 1995 |
NaF caused significantly lowered body weight, feed consumption, absolute uterine weight, and number of implantations. Higher incidence of skeletal (14th rib, dumbbell-shaped 5th sternebrae, incomplete ossification of skull, wavy ribs) and visceral abnormalities (subcutaneous hemorrhage) in fetuses. Vitamin D treatment improved reductions in body weight, feed consumption, and uterine weight. |
Guna Sherlin and Verma 2001 |
Changes in renal function included decreased body weight after NaF treatment at 30 or 48 mg/kg; increased kidney/body weight ratio in the 48-mg/kg group; decreased urinary pH; decreased chloride excretion in the 48 mg/kg group, and increased urinary volume 120 hours after treatment. Renal toxicity was observed in postweaning day 29 rats. NaF exposure resulted in increased kidney/ body weight ratio and kidney weight, profound diuresis, decreased urinary osmolality, and decreased ability to concentrate urine during water deprivation. Decrease in urinary chloride excretion was observed for the first 2 days after exposure; it was increased in water-deprived rats 120 hours after treatment. Hematuria and glucosuria were observed for 2 days after treatment with 48 mg/kg. Renal sensitivity noted after weaning in day 29 rats. Histological lesions noted in proximal tubules of treated day 29 rats. |
Datson et al. 1985 |
In high-dose group, initial decreased body weight gain (recovered over time) and decreased water consumption. No clinical signs of toxicity were observed. In both the rabbit and rat, maternal exposure to NaF during organogenesis did not substantially affect frequency of postimplantation loss, mean fetal body weight/ litter, and visceral or skeletal malformations. The NOAEL for maternal toxicity was 18 mg/kg/day (NaF) in drinking water for rats and rabbits. The NOAEL for developmental toxicity was greater than 27 mg/kg/day (NaF) for rats and greater than 29 mg/kg/day for rabbits. |
Heindel et al. 1996 |
No dose-related feed consumption or mean body weight gain in either F0 or F1 females. Statistically significant decreases in fluid consumption for F0 at 250 mg/L and F1 at 175 and 250 mg/L. Corpora lutea, implants, fetal morphological development, and viable fetuses were similar in all groups. No dose-related anomalies in internal organs were observed in F2 fetuses. Ossification of the hyoid bone was significantly decreased among F2 fetuses at 250 mg/L. |
Collins et al. 2001b |
Reduction in head-tail lengths and dysfunction of the neuromuscular system of the tadpoles. EC50 for malformation in growth after exposure to NaF 5 hours after fertilization is 184 ppm. Calculated LC50 is 632 ppm. Values for EC50 and LC50 met the limits established for a teratogen in frog embryos. |
Goh and Neff 2003 |
TABLE 6-4 Human Developmental Studies
Subjects |
Exposure Route, Duration |
Concentration/Dose |
Results |
Reference |
Pregnant women (mean age 29; n = 91), routine examination at 6th month of pregnancy, 4 groups |
Oral doses, taken during final trimester of pregnancy |
0, 1.5 mg of F (CaF2) per day; 1.5 mg of F (NaF) per day; 0.75 mg of F (NaF) twice per day |
Significant difference between cord plasma fluoride concentrations of newborns in untreated group (mean 27.8 µg/L) and of combined supplemented groups (mean 58.3 µg/L). |
Caldera et al. 1988 |
Pregnant women (n = 25) |
Drinking water |
1.2 mg/L, continuous fluoride concentration in drinking water |
Fluoride in maternal plasma varied from 12.00 µg/100 mL to 41.8 µg/100 mL. In cord blood, it ranged from 11.20 µg/100 mL to 38.8 µg/100 mL; 8% of cases showed cord plasma fluoride concentrations higher than that of maternal concentrations. A highly significant correlation was found between the plasma fluoride concentration of maternal and fetal blood (P < 0.001). |
Malhotra et al. 1993 |
Children (ages 4-12; n = 30) |
Drinking water |
4.5-8.5 mg/L (fluoride) |
Blood fluoride concentrations of children were 0.9 ppm and 1.1 ppm. Serum fluoride concentrations ranged from 1.6 to 1.9 ppm. Of 30 skiagrams of the lumbosacral region, 14 (47%) showed spina bifida occulta. |
Gupta et al. 1994 |
Pregnant women (n = 22) |
Drinking water |
0.22-0.49 µg/L (fluoride in drinking water) |
Serum fluoride concentrations were 0.018 ± 0.012 µg/mL in mothers, 0.030 ± 0.015 µg/mL for umbilical cord samples, and 0.038 ± 0.016 µg/mL for neonates. Statistically significant differences were found between maternal and cord serum fluoride (P ≤ 0.05) and between neonatal and cord serum fluoride (P ≤ 0.05). No statistical difference between maternal and neonatal serum fluoride. No correlation in fluoride concentrations between maternal and neonatal sera, between maternal and cord sera, or between neonatal and cord sera. |
Shimonovitz et al. 1995 |
Fetuses (14-36 weeks of intrauterine life; n = 64) |
Drinking water |
0.2 mg/L (fluoride concentration in drinking water) |
Higher contents of Ca, Mg, and P were disclosed in the diaphyseal part of the bones. Higher concentrations of fluoride were recorded in the metaphysis than in the shaft. Statistically significant correlations between fetal age and content of calcium and phosphorus in the bones; the fluoride contents in the shaft and in the metaphyseal part. No influence of fluoride on calcification of fetal bony tissue. |
Mokrzynski and Machoy 1994 |
Children from India (ages 5-12; n = 50) with dental and/or skeletal fluorosis |
Drinking water |
≤ 1.5 (control), 4.5, and 8.5 mg/L (fluoride) |
A total of 22 (44%) of the 50 children in the study group, and 6 (12%) of the children in the control group revealed spina bifida occulta in the lumbosacral region. Proportion of children with spina bifida occulta in fluoride-rich areas was 44%. |
Gupta et al. 1995 |
Data for mothers under age 30, Down’s syndrome birth rates in five counties of metropolitan Atlanta, Georgia (reanalysis of Erickson 1976) |
Drinking water |
Not specified; comparison of fluoridated and nonfluoridated communities; authors selected 0.1-0.3 mg/L as a reasonable range assumption for nonfluoridated areas |
Highly significant association between fluoridated water and Down’s syndrome births (P < 0.005) in a selected subset of previously published data. |
Takahashi 1998 |
Subjects |
Exposure Route, Duration |
Concentration/Dose |
Results |
Reference |
Data from literature search on studies of Down’s syndrome and exposure to fluoride |
Drinking water |
Range from all studies was 0-2.8 mg/L |
Six ecological studies were included in the evaluation. Crude relative risk ranged from 0.84 to 3.0. Four studies found no significant association between Down’s syndrome and water fluoride concentration. Two studies showed increased incidence of Down’s syndrome with increased water fluoride concentrations (P < 0.05). All the studies scored poorly on the validity assessment. Only two studies controlled for confounding factors, only one of which presented summary outcome measures. |
Whiting et al. 2001 |
Data from literature search on SIDS mortality rate for 1980-1984 in New Zealand |
Drinking water |
Median fluoridation was ≤ 1 g/m3 |
Strong negative correlation between SIDS and mean daily temperature of -0.83 (P = 0.0001). Nonsignificant correlation between SIDS and average fluoridation (P = 0.24). Mean daily temperature was significant while average fluoridation was not. Daily temperature was a significant predictor of SIDS after removing average fluoridation from the model. |
Mitchell et al. 1991 |
485 postneonatal deaths attributed to SIDS; 1,800 control infants |
Drinking water |
0.7-1.0 mg/L (artificial) 0.1-0.3 mg/L (natural) |
Exposed infants to fluoridated water in utero were not at increased risk for SIDS, adjusted odds ratio 1.19. Fluoridated water was not associated with increased risk for SIDS among breastfed infants. Fluoridated formula feeding, compared with unfluoridated formula, showed no increase of SIDS. No interaction between fluoridation and infant feeding. |
Dick et al. 1999 |
ABBREVIATIONS: SIDS, sudden infant death syndrome. |
Two early papers (Rapaport 1956, 1963) reported an association between elevated rates of Down’s syndrome and high water fluoride concentrations. Rapaport also was the first to suggest that maternal age might be an important consideration, with the association between drinking water fluoride concentrations and elevated rates of Down’s syndrome particularly pronounced among young mothers. However, the impact of Rapaport’s observations is limited by some significant methodological concerns, including the use of crude rates as opposed to maternal age-specific rates, limited case ascertainment, and the presentation of crude rates per 100,000 population as opposed to per live births. Several subsequent reports (Berry 1958; Needleman et al. 1974; Erickson et al. 1976; Erickson 1980) studied the association of Down’s syndrome with fluoride or water fluoridation. Berry (1958) found little difference in rates of Down’s syndrome between communities with relatively high and low water fluoride concentrations; however, the populations evaluated were small, and maternal age was not considered in the analysis. Needleman et al. (1974) found a positive association between water fluoride concentration and Down’s syndrome incidence when crude incidence rates were compared; however, this apparent association was largely lost when the comparison was limited to before and after fluoridation for a subset of towns that introduced water fluoridation, an attempt to partially control for maternal age. Erickson et al. (1976) used data from two sources, the Metropolitan Atlanta Congenital Malformations Surveillance Program and the National Cleft Lip and Palate Intelligence Service. The metropolitan Atlanta database is particularly robust, with detailed retrospective ascertainment. Erickson et al. (1976) found no overall association between the crude incidence rates of Down’s syndrome and water fluoridation; however, their data suggested a possible increased rate of Down’s syndrome among births to mothers below age 30. Takahashi (1998) grouped Erickson’s metropolitan Atlanta data for mothers under 30 and calculated a highly significant association (P < 0.005) between fluoridated water and Down’s syndrome births to young mothers. A recent review (Whiting et al. 2001) has evaluated the quality of the literature and concluded that an association between water fluoride concentration and Down’s syndrome incidence is inconclusive. While the committee agrees with this overall characterization, the review by Whiting et al. was problematic. For example, it described all six studies as ecological and all but one (Rapaport 1956) as having found the majority of cases. However, some studies were partially ecologic, assigning exposure at the group level but categorizing case status and limited covariates (age, race) at the individual level. Erickson (1980) ascertained cases via birth certificates and explicitly acknowledged problems with this approach.
Overall, the available studies of fluoride effects on human development
are few and have some significant shortcomings in design and power, limiting their impact.
FINDINGS
A large number of reproductive and developmental studies in animals have been conducted and published since 1990, and the overall quality of the database has improved significantly. High-quality studies in laboratory animals over a range of fluoride concentrations (0-250 mg/L in drinking water) indicate that adverse reproductive and developmental outcomes occur only at very high concentrations. A few studies of human populations have suggested that fluoride might be associated with alterations in reproductive hormones, fertility, and Down’s syndrome, but their design limitations make them of little value for risk evaluation.
RECOMMENDATIONS
-
Studies in occupational settings are often useful in identifying target organs that might be susceptible to disruption and in need of further evaluation at the lower concentrations of exposure experienced by the general population. Therefore, carefully controlled studies of occupational exposure to fluoride and reproductive parameters are needed to further evaluate the possible association between fluoride and alterations in reproductive hormones reported by Ortiz-Perez et al. (2003).
-
Freni (1994) found an association between high fluoride concentrations (3 mg/L or more) in drinking water and decreased total fertility rate. The overall study approach used by Freni has merit and could yield valuable new information if more attention is given to controlling for reproductive variables at the individual and group levels. Because that study had design limitations, additional research is needed to substantiate whether an association exists.
-
A reanalysis of data on Down’s syndrome and fluoride by Takahashi (1998) suggested a possible association in children born to young mothers. A case-control study of the incidence of Down’s syndrome in young women and fluoride exposure would be useful for addressing that issue. However, it may be particularly difficult to study the incidence of Down’s syndrome today given increased fetal genetic testing and concerns with confidentiality.