Advances in Dental Research

Advanceshttp://adr.sagepub.com/ in Dental Research

Non-Dental Tissue Effects of Fluoride M. Kleerekoper ADR 1994 8: 32 DOI: 10.1177/08959374940080010801 The online version of this article can be found at: http://adr.sagepub.com/content/8/1/32

Published by: http://www.sagepublications.com

On behalf of: International and American Associations for Dental Research

Additional services and information for Advances in Dental Research can be found at: Email Alerts: http://adr.sagepub.com/cgi/alerts Subscriptions: http://adr.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav

Downloaded from adr.sagepub.com by guest on July 13, 2011 For personal use only. No other uses without permission.

NON-DENTAL TISSUE EFFECTS OF FLUORIDE

M

M. KLEEREKOPER

Division of Endocrinology Wayne State University School of Medicine 4201 St. Antoine, UHC-4H Detroit, MI 48201 Adv Dent Res 8(l):32-38, June, 1994

Abstract—The anti-caries effects of water fluoridation are well-established. The non-dental tissue effects of fluoride in drinking water, either naturally occurring or as an additive, have been too poorly studied to permit definitive conclusions to be drawn. Claims have been made that fluoride results in an increased occurrence of malignancies, particularly osteogenic sarcoma. Experimental rat data have not resolved this issue, and epidemiologic studies are equally unclear. Initial claims that fluoride offers protection against atherosclerosis remain viable, but here too, much more directed research is needed. Early studies suggested that a water fluoride content greater than 1 ppm resulted in a lower prevalence of osteoporotic fractures. Recent epidemiologic data seriously question this conclusion and raise the possibility that even this relatively low level may increase the prevalence of osteoporotic hip fractures. Other elements, including calcium and magnesium, also vary in amount as water fluoride content varies, and it has proved difficult to distinguish the independent effects of the various nutrients in water from each other. Therapeutic use of fluoride has been largely restricted to studies of its effect on the osteoporotic vertebral fracture rate. After more than 30 years of detailed study, this important issue remains unresolved. This review provides an overview of these issues, focusing on the uncertainties alluded to, and attempting to develop strategies for future research.

This manuscript was presented during a Workshop on Methods for Assessing Fluoride Accumulation and Effects in the Body, sponsored by the National Institute of Dental Research (Bethesda, MD), January 13-15, 1993.

any claims, particularly of deleterious effects, were loudly voiced when water fluoridation programs were initiated as an anti-caries public health measure. Few have been subject to scientific inquiry, and none of them has demonstrated validity. For example, the belief that water fluoridation would lead to an increase in birth defects, Le., that fluoride is a mutagen, has not been systematically studied. There have been some attempts to link fluoridation to carcinogenesis, and, on the benefit side, there are suggestions that fluoride offers protection against atherosclerosis. With respect to the skeleton, there are conflicting reports of fluoride either protecting against the development of osteoporosis or hastening its development. In contrast to these limited studies of non-dental tissue effect of fluoride in drinking water, fluoride as therapy for osteoporosis has been the subject of intense study for more than three decades. In this review, I have attempted to summarize the effects of water fluoride on carcinogenesis, atherogenesis, and osteogenesis and to focus the controversy surrounding therapeutic use of fluoride in osteoporosis.

FLUORIDE AND CANCER Claims that water fluoridation programs resulted in increased cancer mortality began to surface in the late 1970's (Yiamouyiannis, 1977). Apparently prompted by these reports, the National Cancer Institute evaluated the cancer mortality and cancer incidence for the years 1973-1987 through the Surveillance, Epidemiology, and End Results (SEER) program. Particular focus was placed on the relatively uncommon cancer, osteosarcoma. The National Toxicology Program (NTP) of the United States Public Health Service has demonstrated "equivocal evidence" [one of five standard terms used by the NTP to describe the strength of evidence of carcinogenicity of individual experiments (Younge, 1991)] that long-term exposure to sodium fluoride results in a small number of osteosarcomas in male rats. This was not seen with similar exposure in female rats or in mice of either sex (Department of Health and Human Services, 1991a). A two-year carcinogenicity study by Maurer et al. (1990) also failed to find a link between fluoride and cancer. The SEER registry demonstrated an increase in osteosarcoma in young males (younger than 20) from 3.6 to 5.5 cases per 10^ population during the years 1973 to 1987. During the same period, rates in females were essentially stable, at 3.8 to 3.7 cases per 10^ population. Although the rate of osteosarcoma in males was higher in the area where the water was fluoridated, there was no demonstrable relationship between these incidence rates and the introduction of fluoride into the water supply (Hoover et al, 1990). Epidemiologic studies in the state of New York (Mahoney et al, 1991) and England (Chilvers and

32 Downloaded from adr.sagepub.com by guest on July 13, 2011 For personal use only. No other uses without permission.

VOL8(I)

NON-DENTAL

33

TISSUE EFFECTS OF FLUORIDE

Conway, 1985) also failed to demonstrate any evidence of increased cancer mortality related to water fluoridation. The current opinion of the Public Health Service is that fluoride in drinking water is not a carcinogen (Younge, 1991). It is worth noting that the therapeutic use of fluoride in much larger doses in older osteoporotic subjects has never been linked to an increase in occurrence of cancer, even after 10 or more years of therapy. ATHEROSCLEROSIS In 1966, Bernstein et al. reviewed the results of lateral lumbar spine radiographs in 380 persons over age 45 (of whom 166 were male) living in a community in North Dakota where the fluoride content of water was 4-5.8 ppm. Results were compared with those of 715 persons (of whom 312 were male) in a similar community where the fluoride level was from 0.15 to 0.30 ppm. The communities were similar in most habitual respects. In particular, more than 50% of those in each community had lived in the area all their lives, so that a comparison could be made concerning lifetime exposure to different fluoride levels. In all age categories, the percentage of radiologically detectable calcification of the abdominal aorta was less in males and females residing in the high-fluoride area (Fig. 1). These differences reached statistical significance in males at all ages and in females in the 55-64-year age group. In discussing this observation, they noted that animals fed a magnesiumdeficient diet had accelerated atherosclerosis and soft-tissue calcification (Vitale, 1957; Hellerstein, 1960) and that this could be prevented by concomitant feeding with large doses of fluoride (Chiemchaisri and Phillips, 1963). Luoma et al (1976) confirmed these observations between magnesiumdeficiency-induced calcification of the aorta and reversal by fluoride in a rather complicated dietary manipulation study in experimental rats. This experiment was supported by preliminary observations in men in four different rural communities in Finland (Luoma et al., 1973). These 300 men lived in communities with a 50-fold difference in water fluoride content (from 0.05 to 2.57 ppm) and a similar pattern (with less variance) in water magnesium content (from 5.8 to 11.0 ppm). The overall prevalence of cardiovascular disease was four-fold higher in the men living in the area with the lowest water fluoride and magnesium concentrations compared with those in the region with the highest concentrations of these elements. The prevalence of heart disease was eight-fold higher in the low-fluoride/lowmagnesium area. In a follow-up reported 10 years later, this same group found a relative risk for hospitalization for acute myocardial infarction of 3.0 to 4.4 for men in communities where the fluoride was < 1 ppm. The relative risk for low magnesium (< 1.2 ppm) was 2.4 to 4.7 (Luoma et al., 1983). In contrast, Miller (1980) was not able to demonstrate any protective effect of fluoride on mortality due to ischemic heart disease. While it is tempting to link the recent overall decline in cardiovascular disease in the United States to the water

AGE 45-54

55-64

65+

45-54

55-64

65+

Fig. 1—Percentage of radiologically detectable calcification of the abdominal aorta. Reproduced with permission from Bernstein et al. (1966). fluoridation programs, such a link would be unfounded speculation. Regrettably, very little attention has been paid to this potentially very important public health aspect of water fluoridation. One major problem with epidemiologic studies of a condition as prevalent as atherosclerotic cardiovascular disease is that the pathogenesis is multi-factorial, and it will probably not be possible to unravel the potential permutations, combinations, and confounding variables no matter how large the sample size of the study.

WATER FLUORIDE AND THE PREVALENCE OF OSTEOPOROSIS Leone et al (1955, 1960) noted that radiographic evidence of osteoporosis was substantially less in people living in Bartlett County, Texas (water fluoride content, 8 ppm), than in Framingham, Massachusetts (water fluoride content, 0.09 ppm). These observations led to the North Dakota study (Bernstein et aL, 1966) previously referred to. (In fact, observing the effect of fluoride on calcification of the abdominal aorta was a fortuitous by-product of what was initially an osteoporosis study.) They noted less radiographic evidence of decreased bone density in 300 subjects living in the high-fluoride area. They also noted a significant reduction in osteoporotic fractures in the lumbar vertebrae of the women in this community. There was no difference in the prevalence of these fractures in the men in the two communities studied. A curious and, to my knowledge, stillunexplained observation in this report was that the men had substantially more fractures than the women in both communities. This is so contrary to the universal observation that the prevalence of spinal osteoporosis is greater in women over 50 than in men that it lessens the credence of these radiographic studies. Bernstein et al. did recognize this phenomenon but were unable to offer a satisfactory explanation. Nevertheless, the prevailing opinion for the last


34

KLEEREKOPER

1.0-

:0.9^ Males 5 0.8 \s

1

••

T

XN >

s

Females

••

0.6-

25

35

45

55

65

75 Age

Fig. 2—Average bone densities in a high-fluoride area {broken lines) and low-fluoride area (solid lines). Reproduced with permission from Groot (1977).

several decades has been that, at least up to the levels that are effective in protecting against dental caries without causing fluorotic pitting and mottling, the higher the fluoride content of the drinking water, the lower the incidence of osteoporosis (Fig. 2) (Leone et al, 1955). Higher levels, > 10 ppm, are associated with a crippling bone and joint disease, endemic fluorosis (Groot, 1977; Christie, 1980). Several recent reports have brought this thinking into question. Now we must consider the very real possibility that fluoride (even at levels felt to be safe and optimal for the prevention of dental caries) may in fact aggravate the risk of developing osteoporosis. These reports formed the basis of an NIH/NIDR Workshop in 1991, the results of which have recently been published (Gordon and Corbin, 1992). Most of the relevant publications are referenced in that report, and only a few will be summarized here. In a county-by-county survey of hip fractures recorded in the United States in persons aged 65 or older, Jacobsen et al (1990) reported a significant relationship between osteoporotic hip fractures and the water fluoride content. The strength of this relationship was quite small and required accounting for other (stronger) factors to reach significance. Shortly thereafter, Cooper et al (1991), in a re-analysis of an earlier report (Cooper et al, 1990), reported a similar pattern in 39 counties in England (Fig. 3). More recently, Danielson et al (1992) reported an increased relative risk in hip fractures (RR, 1.27 for women, 1.41 for men) for persons in Brigham City, Utah, a community whose water had been fluoridated to 1 ppm since 1966, compared with persons in two control communities in Utah that had not been exposed to water fluoridation and where the water fluoride level was 0.3 ppm. Again, the strength of this association is not great, given that the investigators had to find two control groups in order to achieve statistical significance. Nonetheless, this is a non-

ADV DENT RES JUNE 1994

trivial increased risk of osteoporotic hip fracture, given the goal of the Public Health Service (Department of Health and Human Services, 1991b) to reduce the incidence of such fractures by 20% by the year 2000. In a prospective study conducted in three rural communities in Iowa, Sowers et al (1991) reported a faster rate of bone loss from the radius in women in the community with the highest fluoride content. The relative risk of (presumed) osteoporotic wrist, spine, and hip fractures in these women was 2.2 (95% confidence interval 1.1 to 4.7) when the community with 15 ppm of calcium and 4 ppm of fluoride was compared with the control community containing 67 ppm of calcium and 1 ppm of fluoride. As a prospective study, this report should perhaps carry more weight than the epidemiologic studies. However, careful review of the data highlights the complexities in this field. Water that is naturally fluoridated at 4 ppm is clearly quite different from water that is artificially fluoridated to 1 ppm. Perhaps more importantly, potential confounding calcium-fluoride interactions are evident by the Ca/F ratios being 15/4 and 67/1. It is clearly impossible to lay all the blame for any observed effects simply on the different absolute amounts of either fluoride or calcium. An additional confounding variable for both studies is that local community drinking water is not the only source of dietary fluoride (the so-called "halo" effect). For example, 80% of the subjects exposed to 1 ppm of fluoride used fluoride-containing toothpaste compared with only 39% in those exposed to 4 ppm. How much this and other non-local sources of fluoride contributed to total dietary fluoride is difficult to estimate. In fairness, Sowers et al. (1991) appeared to have used all available questionnaire techniques in an attempt to address this issue and clearly acknowledged the limitations. Not all of these studies have linked water fluoridation to an increase in incidence of osteoporotic fracture. Simonen and Laitinen (1985) in fact reported lower incidence of hip fractures in a Finnish city with its water supply artificially fluoridated compared with one with low fluoride levels. In a recently published study, Jacobsen et al. (1993) were able to demonstrate no difference between the incidence of hip fracture in Rochester, Minnesota, for the 10 years before compared with the 10 years after the introduction of water fluoridation in 1960. The issue of a potential cause-and-effect relationship between the water fluoride content and the presence of osteoporotic hip fractures looms as large a public health issue as is the protection against dental cavities offered by the same fluoride. As is the case with the possible relationship between water fluoride levels and cardiovascular disease, the studies are complicated by the multi-factorial pathogenesis of osteoporotic fractures and may prove impossible to answer by current epidemiologic/ecologic methods. There is no reliable means for monitoring the total body burden of fluoride, particularly the total skeletal burden. Measurement of plasma or urine fluoride is not well-correlated with skeletal fluoride content. One possible approach would be to measure the bone fluoride content of a large number of femoral heads surgically removed from elderly subjects because of


VOL.8(1)

osteoporotic hip fracture, osteoarthritis, or avascular necrosis. Concomitant measurement of plasma and urine fluoride as well as estimates of fluoride intake would also be helpful to such a study. As previously suggested by Kleerekoper (1992), all of the prospective osteoporotic fracture epidemiologic studies sponsored by the NIH should monitor water fluoride intake and plasma and urine fluoride in their study cohorts. Even these studies, no matter how carefully conducted, are not guaranteed to provide definitive answers to this crucial question.

NON-DENTAL TISSUE EFFECTS OF FLUORIDE

35

Population Size • ° •

19 000-44 000 44 000-75 000 75 000-33 000

Weighted (r=.41, P