Fluoride in drinking water and risk of hip
fracture in the UK: a case-control study,
Hillier S, Cooper C, Kellingray S, et al., Lancet 22 January 2000; 355: 265-269.
A critique by John R. Lee, M.D.
In the continuing debate concerning water fluoridation, the question of its role in hip fracture is still unresolved. A number of studies found that hip fracture incidence rises with increased fluoride concentration in drinking water. A fewer number of studies found no detrimental bone effect of fluoridated water compared to unfluoridated water. In a previous report1, I have pointed out that the few studies that found no increase of hip fracture relative to fluoride concentration in drinking water are all faulty due to the short duration of exposure to fluoridated water (6 years or less), the small number of subjects, or the advanced age of the subjects.
In 1993, for example, the National Research Council reviewed 10 studies comparing fracture incidence with fluoridation status. Seven of these reported a positive correlation, while three did not. Of the later three, the elderly women in the Cauley et al study2 had drunk fluoridated water for a mean duration of only 6 years. This obviously can not compare to life long fluoridation. A 1993 study by Jacobsen et al3 of hip fracture and fluoridated water found remarkable variation in the year-to-year incidence of fracture and Jacobsen, himself, declared that the study could not be used to prove whether or not fluoridation had a deleterious effect on fracture incidence. This study, too, must be discarded. The third study, by Goggin et al4, looked at hip fracture incidence in women over 60 years of age for the 5 years before and after fluoridation of Elmira, NY, in 1960. The women in this study had all passed through menopause prior to fluoridation, and the number of women with hip fracture in the study was too small for the purpose of the study. It should be pointed out that, since bone loss in women of industrialized countries generally starts at age 35, it should be obvious that any study that does not compare the fluoridation status of women during the early bone-forming years of life is useless and must be discarded.
The seven studies5-11 that reported a positive correlation of fluoridated water with fracture incidence did not suffer from these same deficiencies. Since my 1993 report, a large 1995 study in France found the same positive correlation between water fluoride exposure and hip fracture incidence.12 The available evidence thus greatly supports the fluoridation/fracture hypothesis.
The authors of the present study are apparently aware of the deficiencies that invalidate the studies used to support the hypothesis that fluoridation is safe for bones. They argue, instead, that ecological studies that compare one population with another might fail to adjust for the effects of potential confounding variables such as physical activity, body build, cigarette smoking, dietary calcium intake, and reproductive variables between the two populations. It is of no small importance that the present study found no clear association between hip fracture incidence and the suspect variables such as age at menopause, alcohol consumption, cigarette smoking, or dietary calcium intake. Thus, criticism of previous ecological studies has no basis.
This present study of hip fractures in Cleveland County in northeast England included 169 cases from Hartlepool and 745 from elsewhere in the county. This is of interest since Hartlepool residents use naturally fluoridated water that is high in calcium. In this study, Hartlepool is the only source of hip fracture cases using water with a fluoride content over 1 ppm. The rest of the county uses water with low fluoride content. None of the subjects lived in artificially fluoridated communities. Technically, this is not a study of the bone effects of artificially fluoridated water.
There is a difference between natural and artificially fluoridated water. Natural fluoride in northeast England occurs as mineral/fluoride salts such as calcium fluoride, magnesium fluoride, etc., whereas artificial fluoridation uses industrial waste fluoride complexed with silica or sodium. Fluoride complexed with silica or sodium is readily ionized to free fluoride ions that are quickly absorbed in the gastrointestinal tract, whereas, when chemically bound to calcium, less of it ionizes and less is absorbed. Calcium inhibits fluoride absorption and is, in fact, the treatment of choice for fluoride ingestion overdoses. Note that fluoride concentrations in trabecular bone from Hartlepool patients were no different than that from "elsewhere" patients. (Hip and lumbar fractures are fractures of trabecular bone.) This is clear indication that fluoride absorption is essentially equal between the "high" fluoride Hartlepool water and the "low" fluoride water elsewhere in Cleveland County. Furthermore, industrial waste fluoride is not pure fluoride: the industrial waste "fluoride" includes several other toxic compounds that act synergistically with fluoride to increase overall toxicity.
More simply put, the Hartlepool water contains more beneficial minerals than surface water. These minerals not only inhibit fluoride absorption but also are themselves beneficial to bones and to general health. Adding industrial waste fluoride to mineral-poor surface water is not the same. Again, this study says nothing about the bone effects of artificially fluoridated water.
Fluoride concentration of drinking water does not necessarily indicate fluoride exposure: it depends on how much water is drunk. Manual work leads to more water drinking than sedentary occupations. Many people drink little water from the tap; they may use bottled water or watery beverages such as beer. Beer drinking is not unheard of in the pubs of England. Therefore, without knowing the amount of tap water consumed each day, little is known about water-borne fluoride exposure. Furthermore, fluoride absorption from fluoridated toothpaste can be a major factor of total fluoride intake. The study did not take this in to consideration.
The study population included only people aged 50 years or over. I have no difficulty with this since most non-traumatic hip fractures occur in people over age 50. The collection of hip fracture data involved only a 17-month period. Was there a reason for this? Why not a 24-month period? It would not be difficult to determine yearly incidence of hip fracture over a number of years to measure annual or seasonal variations that might be different in different parts of Cleveland County.
The study found no significant difference in hip fracture incidence between naturally fluoridated Hartlepool and unfluoridated communities elsewhere in Cleveland County. Given the considerations discussed above, this is not surprising. It found a positive correlation of low body-mass index and low physical activity with increased risk of hip fracture. This, too, is no surprise. However, the study also found no clear association with age at menopause, alcohol consumption, cigarette smoking, or dietary calcium intake.
This last factor should be of interest since conventional wisdom emphasizes calcium intake for the prevention of osteoporosis. The truth is that osteoporosis is not a disease of calcium deficiency; it is a metabolic disease in which the rate of new bone formation lags behind the rate of bone loss. In this matter, many factors are important, particularly the sex hormones. Estrogen inhibits bone loss while progesterone and testosterone both stimulate new bone formation. The study totally neglected these important bone factors. Diet, exercise, exposure to xenobiotics (pesticides of all sorts), and genetic factors all affect sex hormone production.
In testing sex hormone levels, it is important to keep in mind that circulating sex hormones exist in two different forms, protein-bound, relatively non-bioavailable form and the non-protein-bound, "free" form that is bio-available. Blood tests, as shown by Cummings et al,13 fail to distinguish between protein-bound and "free" hormones. In fact, their study found that the serum concentration of sex hormone binding globulin (SHBG), the binding protein of estradiol, is a far stronger correlate of hip fracture risk than estradiol serum concentrations are. The higher the SHBG level (i.e., less bioavailable estradiol), the greater is the rate of bone loss. Saliva hormone assay is far more relevant since it accurately reflects blood-borne "free" sex hormone.
Progesterone not only stimulates new bone formation but also increases the sensitivity of estrogen receptors. Thus, progesterone supplementation often prevents and/or reverses osteoporosis, thus preventing hip fractures without estrogen supplementation.
Finally, it is important to understand the bone effects of fluoride. It may increase bone quantity (osteofluorosis, osteosclerosis) but also decrease bone quality and bone strength. It is well known that pharmacological doses of fluoride increase the risk of torsion-type fractures (such as hip fractures) despite the appearance of greater bone density. Conventional medicine interprets the observed fluoride-induced increase of serum alkaline phosphatase concentration as a sign of osteoblast activity. Actually, it is a reflection of increased mortality of osteocytes within bone. Osteocytes are rich in alkaline phosphatase, which is released when the cells are killed by fluoride.14 It is unlikely, therefore, that a window of fluoride-induced bone benefit exists.
The editorial by Clifford J Rosen that accompanies the study reminds me of book reviews done by friends of the author. Entitled Fluoride and fractures: an ecological fallacy, it attempts to denigrate previous unfavorable (to him) ecological studies and praises the present study. Even though the study included no subjects using fluoridated drinking water, he claims that it presents "compelling evidence that lifelong exposure to fluoridated water does not increase the risk of hip fracture." Read it if you like but it offers no insight into the fluoridation/hip fracture problem.
Sensible conclusions from the Cleveland County, England, study:
Slow-Release Sodium Fluoride in the
Management of Postmenopausal Osteoporosis,
CYC Pak et al, Annals of Internal Medicine 120 625-632 1994)
by Dr. John Lee
Fluoride 1994 October, 27:4 227-228 (critical review).
Dr Pak's recent interim report in the Annals of Internal Medicine (1) is an interesting example of why science sometimes seems to be a bit out of touch with the real world. This randomized controlled trial of dosing postmenopausal women with slow-release sodium fluoride for the purpose of studying fluoride's effect on bones of postmenopausal women states that its objective is to test whether this treatment "inhibits vertebral fractures without causing fluoride complications." The bone effect of prolonged overdose of fluoride is, however, quite well known; it is a disease called osteofluorosis, also known as osteosclerosis, which is, according to Dorland's medical dictionary, "the hardening or abnormal denseness of bone." It is accompanied also by eventual calcifications in connective tissue such as ligaments, tendons, and peri-articular tissue. There are several points to be raised here.
The abnormal bone denseness resulting from fluoride is of poor quality and, while the increased density helps compression strength it generally leads to weakness of tensile strength. Thus, previous tests (2-4) of fluoride "treatment" for osteoporosis finds a decrease in vertebral (compression) fractures but an increased incidence of hip and long bone fracture, compared to control patients, after four years or so of the treatment. Other researchers have advised abandoning fluoride as a legitimate treatment of osteoporosis for that reason and fluoride's toxicity (5). Dr Pak et al are intent on detecting a difference in vertebral compression fractures, no matter how minor and insignificant that is in treating osteoporosis.
In clinical practice, the occurrence of atraumatic minor compression fractures of vertebra is common in postmenopausal osteoporotic women and is frequently asymptomatic, being found only by radiographs though the patient may have noted a slight decrease in her height over time. The more morbid consequence of osteoporosis is hip fracture which has the potential for seriously disabling patients. Thus, the goal of osteoporosis research should be directed to prevention of loss of tensile strength of bones, not merely compression strength. The protocol for Dr Pak's USPH funded and FDA approved investigational new drug application for Mission Pharmacal Company seems limited to demonstrating the obvious, i.e., that excessive fluoride causes osteofluorosis.
In his study, Dr Pak is administering about 25 mg of fluoride per day in a slow release form to postmenopausal women in order to raise their serum fluoride levels from 50 ng/ml to slightly over 100 ng/ml while avoiding fluoride's known gastric inflammatory effects such as mucosal erosions, ulcers, and bleeding which regularly accompany usual oral fluoride supplementation at this dosage. While the slow release form of administration will protect against these particular gastric symptoms of fluoride toxicity, it will likely have little or no effect on the later connective tissue abnormalities routinely resulting from fluoride intake at this level. In this report, the average duration of fluoride supplement exposure of the treated women is a little under two and one-half years. Since bone turnover time in postmenopausal women is relatively slow, one would not expect the decrease of bone tensile strength (and increase in incidence of hip fractures) to occur for several more years.
The one interesting finding in Dr Pak's interim report is the fact that fluoride supplementation did not cause any reduction in vertebral fractures in women on estrogen supplementation compared to controls. Among estrogen-treated women, the fracture-free rate of placebo (no fluoride) group compared to that of the fluoride group was 75.0% and 76.9% respectively, an inconsequential difference. Dr Pak et al. in commenting on this non-difference, rather obliquely state that "No statistical differences related to estrogen treatment were noted between the slow-release sodium fluoride and placebo groups in estrogen-treated patients, probably because of the small sample size." Unstated is corollary that, in women on estrogen replacement therapy, fluoride treatment offers no possible benefit, regardless of how minor it is or regardless of the potential toxicity of fluoride.
Further, if one looks at the reasoning underlying the context of fluoride treatment for osteoporosis in the light of multiple trials in the US and world-wide, the idea of a therapeutic window for fluoride dosage appears quite remote. As mentioned above, prior studies of "therapeutic" fluoride supplementation found an increase in hip fractures in the treated groups compared to controls, and a number of good studies (6-9) have shown hip fracture incidence is statistically significantly correlated with even the lower fluoride levels as found in fluoridated communities, compared to non-fluoridated communities. To believe that some dosage range between these two ranges of fluoride intake will be effective in preventing osteoporotic hip fractures is a true stretch of the imagination. Certainly no one believes that osteoporosis is a disease of fluoride deficiency.
Finally, regardless of the ultimate outcome of this study, it obviously has no bearing on the question of water fluoridation.
In summary: Dr Pak's trial of inducing what is essentially a controlled form of osteofluorosis (osteosclerosis) in postmenopausal women, and attempting to prove its effectiveness by measuring the incidence of radiographically-identifiable vertebral fractures is inherently flawed and has little relevance to effective treatment of osteoporosis.
Fluoridation and Osteoporosis 1992
by Dr. John R. Lee,
Fluoride,1992, 25:3,162-164 (Letter to the Editor)
The recent Journal of the American Medical Association report by Danielson et al (1) documents again the positive correlation between fluoride and increased risk of hip fracture. In typical knee-jerk fashion, some public health service spokesmen are quick to make unsubstantiated claims that the study may somehow be in error or that the picture is clouded by supposed conflicting studies. The purpose of this letter is to clarify the picture for those who might be confused by biased and misinformed rebuttals that have arisen with the publication of this latest study.
Fluoride "therapy" for osteoporosis
It has been proposed in the past that, since fluoride is "calcium seeking", it might be a good treatment for those with osteoporosis. This has resulted in a number of "therapeutic" fluoride trials. The results of these trials are now available and the conclusion is that fluoride has no place in the treatment of osteoporosis. These trials include those of Riggs (2), of Kleerekoper (3), and of Hedlund and Gallagher (4), all of which reported significantly increased hip fracture incidence as well as an unacceptable rate of gastrointestinal and osteoarticular side effects in the treated group compared to the controls. Professor Avioli of the Washington University School of Medicine concluded in 1987 that "sodium fluoride therapy is accompanied by so many medical complications and side effects that it is hardly worth exploring in depth as a therapeutic mode for postmenopausal osteoporosis" (5). Riggs, after several years of touting fluoride for osteoporosis, finally conceded in 1989 that fluoride had no place in osteoporosis treatment (6). In late 1989, the chairman of the FDA advisory committee reviewing fluoride"s effect on fracture incidence was quoted as saying the FDA "should quietly forget" about fluoride (5).
Fluoridation and osteoporosis prevention
The hypothesis that fluoridation might somehow help prevent osteoporosis and bone fracture received its impetus in 1966 when Bernstein et al reported a difference in osteoporosis prevalence between high- and low-fluoride areas in North Dakota (8). What appears to have been forgotten is that Bernstein went on to study the metabolic effects of fluoride on bone and concluded that fluoride was toxic to bone and not likely to be of any bone benefit (9). Similar studies abroad (10-13) reached the same conclusion: fluoride is toxic to bones.
In a 1986 retrospective study (14) and a prospective study concluded in 1991 (15), Sowers et al reported a definite correlation of fluoridation status and an increased susceptibility to fracture when comparing communities with fluoride at 1 ppm and 4 ppm. Kleerekoper, in his JAMA editorial concerning the recent Danielson study, mistakenly suggests that Sowers ignored other differences in the water, such as calcium content. In fact, Sowers deliberately chose communities with differing levels of calcium in the drinking water so that this factor could also be tested and it was found not to be of any importance to fracture incidence.
In 1990, Jacobsen et al reported on an assortment of ecologic variants relative to over 500.000 hip fractures in white women in all counties of the 48 contiguous US states; he found a definite positive correlation between fluoridation status and hip fracture incidence. After correcting for the other possible variants, this positive correlation became even stronger (16).
In 1991, Jacobsen joined with Cooper of Great Britain to correct an error in applying statistical methods to Cooper's earlier paper concerning hip fracture incidence and fluoridation status in 38 districts in England. When properly addressed, the correct results showed a positive correlation of hip fracture with fluoridation status even though the difference in water fluoride levels ranged only from <0. 1 to 0.9 ppm (17).
Fluoride promoters, however, claim that there exists one study that found fewer hip fractures in a fluoridated community compared to an unfluoridated one. This is the Finnish study by Simonen and Laitinen (18), a two page report that appeared in Lancet in 1985. Apparently, the authors had collected a computer-generated readout of hip fracture diagnoses from Kuopio and another community that had recently started using computer data in their respective hospitals. They did not notice that their Kuopio data produced results indicating approximately equal numbers of hip fractures for both men and women. Since the common ratio of hip fractures between men and women is 1:4, this result means that their Kuopio case finding was faulty; it was missing 3/4 of the female hip fractures. The control community data exhibited the usual sex ratio of fractures and thus was more likely to be correct. The authors also were unaware that researchers at the Kuopio University had already published several papers (10-13) on the bone damaging effect of fluoride. A further defect of the Simonen and Laitinen study is that the residence of the fracture patients was never established; it is likely that a university hospital will attract patients from outside the city itself and thus from a population not drinking the fluoridated city water.
Though the defects in the study were obvious to all knowledgeable researchers, Lancet never published an admission of this error. The city fathers, however, were less obtuse; in comparing the work of their own university researchers with the obviously faulty Lancet article, they realized that continuing fluoridation was a mistake and shortly thereafter, voted to discontinue the practice. Thus, the only community in Finland that was fluoridating their water in those years, is now fluoride free. One wonders at the mental competence of those that still insist this one report, faulty as it is, is sufficient reason to continue to place the rest of us at the greater risk of fracture induced by fluoridation.
The Danielson JAMA report (1)
Simply put, Danielson found that the risk of hip fracture was approximately 30% higher for women and 40% higher for men exposed to fluoridated (1 ppm) drinking water when compared to those with unfluoridated water. The effect was particularly strong in those women who were exposed to fluoridated water during the time of their menopause, a time of active bone remodelling. In older women, when bone remodelling is less and the incorporation of fluoride in bone is less, the effect was less strong. Further, the confounding factors of smoking and alcohol use which might dilute the fluoride significance of other studies, are not present in the populations of this study; thus the effect of fluoride is all the more clear to see.
Fluoride is toxic to bones and increases risk of fracture at all levels of exposure including fluoridation at 1 ppm. Regardless of any other consideration, this is reason enough to discontinue fluoridation immediately.
FLUORIDATION AND HIP FRACTURE
National Research Council Report: "Health Effects of Ingested Fluoride"
by Dr. John R. Lee, Fluoride 1993 October, 26:4 274-277, (Review)
Fluoride's bone effect is considered by most authorities world-wide to be toxic: bone density may increase but histomorphologic studies reveal areas of bone destruction, and fracture rates, especially of the hip, increase. The NRC reviewed three US studies (1-3) and one European study (4) of osteoporosis "therapy" using sodium fluoride (NaF) in doses of 50-80 mg (20-32 mg of fluoride) daily, equivalent to 5-10 times the daily fluoride intake of persons living in fluoridated communities. They report that in all three US studies hip fracture rates increased significantly compared to controls and that, in all the studies, periarticular joint pain and gastrointestinal side effects afflicted the fluoride-treated groups. The NRC stated no conclusion from these findings. However, in an independent review of "therapeutic" fluoride treatments, Dr Avioli, Shoenberg Professor of Medicine and Director, Division of Endocrinology and Mineral Metabolism, at the Washington University School of Medicine, concluded that "sodium fluoride is accompanied by so many medical complications and side effects that it is hardly worth exploring in depth as a therapeutic mode for post-menopausal osteoporosis" (5). Not mentioned by the NRC was the Hedlund and Gallagher study (6) which likewise found a fluoride-induced increase in hip fracture.
The NRC also reviewed 10 studies comparing fracture incidence relative to fluoridation status. Seven (8,10-13,16,17) found a positive correlation (ie., increased hip fracture incidence) of fluoridation and hip fracture while three (7, 14, 15) reportedly did not. None reported any fluoride bone benefit. Two of the seven positive studies were presentations made at the Workshop on Drinking Water Fluoride Influence on Hip Fractures and Bone Health (April 10 1991, Bethesda MD) and all of the remaining five were published in major peer-reviewed journals. In spite of this array of evidence linking fluoridation to hip fracture, the NRC concluded that the three negative studies were of sufficient relevance to out-weigh the findings of the seven positive studies. Let us look at the three studies that reported no fluoride effect.
The first, by Cauley et al referenced only by an abstract in the American Journal of Epidemiology, looked at 1, 878 white women aged 65-93 years (mean age: 70.9) only 73% of whom had exposure to public drinking water, with a mean exposure time of only 6.0 years. Since bone turnover (remodelling) rate is relatively rapid before menopause and slow after menopause, fluoride's major effect on bone is most likely to occur during the years before menopause (i.e., before age 45-50), as was clearly shown in Danielson's study, Thus, none of the women in this study were exposed to fluoride during the time in their life when fluoride would be expected to affect bone. The study, therefore, is useless and should be discarded.
The second, the last of the Jacobsen studies reviewed by the NRC, and reported as a Public Health Brief in the American Journal of Public Health, suffers the same problem, i.e., few if any of the women studied were exposed to fluoridated water prior to menopause. For reason unknown, he reviewed data only through 1969, the ten years after fluoridation of Rochester MN, the site of the Mayo Clinic. Where were the data through 1979 or 1989? The Mayo Clinic is good at keeping such data. By using those, he could have observed the hip fracture incidence of women who had been premenopausal prior to being fluoridated. Even worse is the fact that the population studied (<50,000) was just too small. During the 20 years covered by the study, there were only 383 hip fractures in women and 268 in men; i.e., less than 20 per year among women and only about 1 per month among men. The annual variation in the incidence of hip fracture prior to fluoridation was over 100% and the range of standard error (95% confidence interval) was approximately 300%. If one were looking for, say a 40% fluoride-induced increase in fracture incidence, it could not have been found. The authors' own conclusions state that fluoridation in these age groups was "not associated with an immediate increase in rates of hip fractures" and that "further studies .... are clearly required before public policy decisions can be made." It has been reported that, when Jacobsen himself was asked if his study could be used to prove whether or not fluoride had a deleterious effect on fracture rates, his answer was "No". This study, too, must be discarded.
The third, a 28-year-old study by Goggin et al, looked at the hip fracture rates of women over 60 years of age for the 5 years before and after fluoridation of Elmira NY, in 1960 when the population was only 46,517 (similar to Rochester MN). Again we see that 1) the women had all passed through menopause prior to fluoridation, and 2) the number of women in the study is too small for the purpose of the study. Therefore, this study, too, is useless and must be discarded.
Interestingly, the Lancet of July 24 1993, in their section Facts, Figures, and Fallacies, carries a timely article entitled "Is the study worth doing? ". It points out the elementary statistical fact that the degree of uncertainty increases with decreasing sample size and states: "studies that are too small may fail to detect medically important effects or produce estimates too imprecise to be useful". It is apparent that NRC failed to employ the services of any competent statistician in even including these three useless studies for their consideration.
With seven studies showing a positive correlation of fluoridation with increased hip fracture incidence and not one acceptable study showing the contrary, on what basis does the NRC conclude that the fluoride MCL (maximum contaminant level) be left at 4 mg/L? They argue, without the least evidence, that there exists a "publication bias" in favor of ecologic studies showing positive results. Even if true, what does it matter? A study can either stand on its merits or it can't. They do not mention that public health reports are routinely published as a "public service" by legitimate medical journals without any independent peer review. Their second argument is that, because of un-named "limitations of ecologic analyses" and "the potential in all studies for confounding factors", the positive studies "offer only limited support for a hypothesis of a weak association between fluoridated water and hip fracture." They do admit, however, that "when results from a number of such studies converge to indicate an exposure-disease relation, confidence in collected findings is bolstered." In a remarkable feat of denial, they then conclude "there is no basis at this time to recommend that EPA lower the current MCL of fluoride of 4 mg/L."
Further, they call for additional studies that "should use information from individuals rather than population groups," stating that "it is important that individual information be collected about fluoride intake from drinking water and all other sources" plus hormone status, dietary factors, and other factors that "might influence the risk." If the NRC truly believed that, they should be calling for the throwing out of all the previous dental caries studies as well since the same argument applies.
This obvious selective reasoning does not speak well of the NRC report. Such deviousness is employed not for truth finding but for political ends and, as such, is a disservice not only to the cause of science but to the people of the US. Given the data available, it is impossible to conclude that fluoride is safe. When the NRC concludes otherwise, one must question whether they are competent to evaluate scientific studies, or whether they are guilty of a deliberate lie. If merely the data of this one chapter of the report (chapter 3) is properly analyzed, the only responsible conclusion is that 1) total daily fluoride exposure should be reduced, 2) the MCL for fluoride should be lowered, and 3) the addition of fluoride to public drinking water should be ended.References
DISCUSSIONSubject: The Case for Eliminating The Use of Dietary Fluoride Supplements Among Young Children, by Brian Burt, Fluoride, 1994, 27(2), Abstract presented at the Dietary Fluoride Supplement Conference, American Dental Association, Chicago, Jan 31-Feb 1, 1994. [Click here for abstract.]
Discussion published as Letters to the Editor in: Fluoride, 27:3, July 1994, pp 180-182.
For well over ten years it has been apparent that fluoride supplements are unwise for young children. Indeed, a Rand Corporation literature review in 1981 found that supplements are "simply not warranted" by the results (RAND report N 1732-RWJF Dec 1981). Furthermore, their use clearly increases the risk of dental fluorosis, which is now at an all time high in the US. It is equally clear that there is little or no evidence of any dental benefits from pre-eruptive supplementation. For these reasons, fluoride supplements are banned or strictly regulated in Africa, Europe, Canada, Japan, and India.
It is heartening, therefore, to read that a long time fluoride advocate agrees that "fluoride supplements should no longer be used for young children in North America" (Abstract in Fluoride 27 (2) 1994). The question that Dr. Burt should now address is: Given the uselessness and risk of fluoride supplements, why should the practice of giving young children the same fluoride dissolved in water (fluoridation) be continued?
John R Lee MD
9620 Bodega Hwy, Sebastopol CA 95521, USA
Professor Burt responds:
Dr. Lee has quoted accurately from the abstract, but it seems that he has not read the full paper. (Along with other papers from the workshop, this is to be published in the Journal of Public Health Dentistry in due course). My conclusion, correctly quoted by Dr. Lee, was that the risks of using fluoride supplements in young children outweigh the benefits. "Risk" was described as the likely develop development of the mildest forms of fluorosis from regular use of supplements by infants and young children, while "benefits" were the minor (at best) cariostatic effects likely to result from this use. Dr. Lee then goes on to extrapolate that line of thinking to the use of water fluoridation, but this does not follow because with water fluoridation the benefits outweigh the risks.
Dr. Lee seems to be suggesting that the relative lack of cariostatic effect of systemic fluoride is a new finding, but this is not true. The first review of literature that I know of to point this out was in 1976 (Levine RS. The action of fluoride in caries prevention: a review of current concepts. British Dental Journal 140 9-14 1976). The evidence for a primarily topical cariostatic effect of fluoride has grown and has been cited many times since then, culminating with the 1989 Georgia conference (proceedings in Journal of Dental Research 69 special issue 1990).
There is considerable evidence that vehicles like fluoridated water and table salt have powerful topical cariostatic effects. These methods are highly effective public health approaches to caries control, though they are accompanied by about 12% prevalence of the mildest forms of fluorosis. I consider the benefits of water fluoridation exceed the risks associated with its use, and my argument about eliminating supplements for young children in fact depends partly upon the effectiveness of fluoride in water and toothpaste. Supplements are unlikely to add much to this existing effectiveness, but do increase the risk of fluorosis.
Dr. Lee may not have read the following part of my discussion, which I think helps put the issue in perspective:
The exposure to fluoride from multiple sources, a fact of life in the United States today, is a prime reason why dental caries experience has been reduced to its current low levels. The caries decline is a major public health achievement which must be preserved in those who have benefited from if, and extended to those remaining segments of society which need if most.
I would like to finish by summarizing my philosophy on fluoride use, which I believe is well-based on published evidence. I hope that this will counter any wrong impression that Dr. Lee's inferences may have produced.
Dr Lee replies:
Common sense would dictate that the amount of fluoride touching the teeth during the act of swallowing fluoridated water is extremely small. In fact, I doubt that it would even be measurable. It is certainly uncommon for children to swish their drinking water back and forth through their teeth while imbibing a drink. One would think that brushing with fluoridated toothpaste would be more effective in bringing fluoride into contact with the teeth.
The argument that fluoride in public drinking water is responsible for the observed decline in children's dental caries is contrary to numerous studies in the US and world-wide which find that the same decline occurred also in unfluoridated communities. (See references 1 - 11 in the list that follows below.) In fact, I have continuously challenged any dental authority to provide one valid study of the past two decades justifying the presumption of fluoridation's dental benefits and none has been forthcoming.
Dr. Burt's statement that fluoridation results in only "12 % prevalence of the mildest forms of fluorosis" is contradicted by the US Public Health Service (the Hoover. report) indicating a fluorosis prevalence of 22.3% in fluoridated communities and by other authoritative reports in which dental fluorosis prevalence was variously found to be 30-60 % In communities with supposedly "optimal" fluoridation. (See references 4 and 12-15 in list below.)
Dr. Burt is seemingly unaware that dental fluorosis connotes fluoride toxicity far more important than mere dental disfigurement. Dental. fluorosis is a visible indicator in developing teeth of generalized fluoride toxicity throughout the body, including damage to connective tissue, bone tissue, immune functions, and enzyme functions. As such, any rise in the prevalence of dental fluorosis is cause for concern.
I shall look forward to continuing this discussion when Dr. Burt's full paper is published (in due course) in the Journal of Public Health Dentistry.
John R Lee MD
[Editor: Fluoride 23 (3) 1990 page 106 listed 7 studies which reported dental fluorosis prevalences in 12 fluoridated communities. The average prevalence was 29%. One of the studies, co-authored by Dr Burt, reported dental fluorosis prevalences in 3 Michigan fluoridated communities of 32%, 49% and 51% (Journal of Dental Research 67 802-806 1988).]
Fluoride Exposure and Childhood
Osteosarcoma A case-control study
A report by Gelberg KH et al., of the N.Y. State Department of Health,
in The American Journal of Public Health, 1995, 85, 1678-1683)
A Critique, by John R. Lee, M.D.
Fluoride Exposure and Childhood Osteosarcoma A case-control study (A report by Gelberg KH et al., of the N.Y. State Department of Health, in The American Journal of Public Health, 1995, 85, 1678-1683), Fluoride,1996, 29:4, 237-240
Does fluoride increase the risk of osteosarcoma in young men? This case control study by the New York Department of Health tested this hypothesis by comparing the estimated fluoride intake of 130 osteosarcoma victims with that of an equal number of presumed healthy sex- and age-matched surrogates. The authors report finding little difference, and concluded that fluoride does not increase the risk, and may even be protective. Differing conclusions are not uncommon in science, and especially in medical science since underlying causes are often exceedingly complex, subtle and heterogeneous. It is important, therefore, to examine this report's results, its test design, and the assumptions on which the test and the conclusions are based.
Several lines of investigation suggest that fluoride intake increases the risk of cancer in general and, in particular, the incidence of osteosarcoma in males. As the authors admit, in vivo studies show fluoride to be mutagenic, inducing chromosome aberrations, sister chromatid exchanges, cytotoxicity, and neoplastic transformation in cultured mammalian cells. The authors also agree that fluoride accumulates primarily in bones; and that children, who are actively forming bone, have a higher uptake of fluoride into bone than adults. Further, bone in knees, ankles, shoulders, and wrists, where childhood osteosarcoma most often occurs, shows a high response to fluoride.1
In 1990, a two-year carcinogenicity study by the National Toxicology Program (NTP) found a statistically significant, dose-related increase of osteosarcoma rates in male rats, but not in mice.2 That the so-called peer review members at the time quixotically chose to call this fluoride/osteosarcoma correlation "equivocal" (as reported by the authors of this present study) does not change the facts. This same study revealed a strong correlation of fluoride intake with nasal and oral cancer and precancerous lesions in test rats and mice. A coincidental Proctor and Gamble study reported an increased incidence of cancer in rats but this was discounted later on the basis of a concomitant viral contamination in the test rodents.3 Time trends for bone and joint cancer and osteosarcoma derived from the Surveillance, Epidemiology and End Results (SEER) data of the National Cancer Institute (NTP) revealed a positive association of osteosarcoma incidence and water fluoridation among males under 20 years of age.4 In 1993, an ecological study performed by the New Jersey Department of Health found a strong statistical association between fluoridation and osteosarcoma among young men.5 It would appear that the fluoride/osteosarcoma hypothesis is credible and convincing, if not yet "conclusive" to fluoridation proponents.
Testing for the fluoride/osteosarcoma link is a daunting prospect. Osteosarcoma is quite rare, the incidence being only 2.9 cases per million people in the US. Test design always follows from assumptions made. The susceptibility of a cell to be cancer-prone may stem from a variety of subtle influences. In the case of xenobiotics (petrochemical compounds such as pesticides and various plastics), it is now known that exposure during embryonic tissue differentiation is far more toxic than later in life; yet the effects show up much later in life as an increased susceptibility to cancer of urogenital tissues such as the vagina, cervix, ovary, or testes.6
Fluoride readily crosses the placenta. Maternal fluoride intake determines whether baby's bones are fully fluoridated or not. It is likely that fluoride intake later in life will be more toxic to the child whose bones are already fully fluoridated than to one whose bones are not so fluoridated. Thus, one might well assume that exposure to fluoride during embryo life could be a factor in developing bone cancer later in life, regardless of whatever other factors may also play a role. Oncology researchers often make the distinction between cancer initiators and cancer promoters. Whether fluoride is considered a cancer initiator or a promoter, one's test design must include the fluoride intake by the mother prior and during the time of her pregnancy. This factor is missing in this present study.
In testing the fluoride/osteosarcoma link, one must be able to calculate total fluoride intake at various stages of life preceding the onset of the cancer. This is more difficult than it might at first seem. In calculating total fluoride intake, the study included fluoride tablets used, mouth rinses, toothpaste used, dental treatments, and water fluoridation levels. Missing from this list are calculations of differences in water actually consumed based on differences in ambient temperature, individual work or athletic exercise that greatly increases water consumption, and dietary habits such as processed beverages versus "plain" water. It is not difficult to understand that commercially processed beverages made from fluoridated water are sold in unfluoridated communities. Likewise, it is not difficult to understand that some children drink more processed beverages than water from the tap. Thus, knowing the fluoride concentration of the tap water is not the same as knowing the fluoride intake from one's drinking of fluids.
Further, it is well established that much of our US diet choices are canned or processed foods rather than fresh, unprocessed foods. Community water fluoridation adds fluoride not only to one's drinking water but to foods processed with the fluoridated water. It is for this reason that processed foods of different brands can differ greatly in their fluoride content, and this difference is not recognized when making food purchase choices. It is likely that one family will routinely choose one brand while the next family always uses another brand. Estimating averages does not help since the "average" does not exist; one brand will be fluoridated and the other is not. Fluoride from processed foods comprise a major portion of one's total fluoride intake, often equalling or exceeding that obtained from tap water.' This calculation, too, is missing from the fluoride exposure variables listed in this study.
In the present study, something is odd about the case subjects. While it is routinely found that osteosarcoma is more common in young men than in young women, this study's list of 130 cases included only 42 males, or 32% of the total. Thus, the osteosarcoma cases used were not typical of the disease in question. Did the males go elsewhere for treatment? Did some male cases of osteosarcoma slip through undetected in the study's case selection method? Did the young women with osteosarcoma drink more fluoridated beverages and less unfluoridated water than the young men? From the information given, no clue is found. The authors seem unconcerned over this discrepancy.
Finally, one must question the case-control method of the study. In the case control method, patients with the disease in question are compared to similar appearing, same-age people without the disease. In effect, patients susceptible to osteosarcoma were selected controls, i.e. those without evident osteosarcoma. Given the rarity of osteosarcoma, and the fact that the sources of fluoride exposure are so ubiquitous, it would be no problem to find an equal group of healthy people living in the same communities and using the same toothpaste as those with osteosarcoma.
The fact that the two groups' drinking water and toothpaste choices are the same does not invalidate the conclusion that fluoride was a factor in the development of osteosarcoma. The study's authors apparently assume that osteosarcoma victims require higher fluoride exposure than those without the disease. An equally plausible assumption is that variable individual susceptibility exists such that equal fluoride exposure will affect only those with the requisite susceptibility. Given the rarity of the disease, this seems more probable. The susceptibility for osteosarcoma may stem from early prenatal fluoride exposure or from factors not yet known. The later occurrence of the cancer may require only the level of fluoride exposure common to fluoridated communities. If this assumption was correct, as case-control study such as this comparing only post-natal fluoride exposure between osteosarcoma victims and controls would find no difference.
When polio "epidemics" were common, it was clear that only a small percentage of children in any given community developed clinically apparent poliomyelitis while well over 90% of the children showed an equal rise in polio antibodies. That is, despite equal exposure, only a few children were sufficiently susceptible to be stricken with polio. A similar scenario might well apply to the osteosarcoma problem. Since we often do not know all the factors that "cause" or "promote" a given cancer, we do not know what factors are important in selecting comparison groups. Case-control study designs are not appropriate for all illnesses and this, one might suspect, is one of them.
This present study, while being used to cast doubt on the relationship of fluoride to osteosarcoma, is flawed by: 1) disregard of prenatal fluoride exposure; 2) inadequate calculation of postnatal total fluoride intake; and 3) inappropriate choice of study design. Thus, the study carries little weight in negating the fluoride/osteosarcoma connection or in any consideration of continuing fluoridation as a public policy.
Published by the International Society for Fluoride Research