Fluoride Action Network

Fluoride & Osteosarcoma: A Timeline

Fluoride Action Network | by Michael Connett | Updated April 2015

Several human epidemiological studies have found an association between fluoride in drinking water and the occurrence of osteosarcoma (bone cancer) in young males. (Bassin 2006; Cohn 1992; Hoover 1991). These studies are consistent with the National Toxicology Program’s (NTP) cancer bioassay which found that fluoride-treated male rats had an dose-dependent increase in osteosarcoma. (Bucher 1991). Although a number of studies have failed to detect an association between fluoride and osteosarcoma, none of these studies have measured the risk of fluoride at specific windows in time, which is the critical question with respect to fluoride and osteosarcoma.

As acknowledged by the NTP and most other observers, a fluoride/osteosarcoma connection is biologically plausible. The biological plausibility centers around three facts: 1) Bone is the principal site of fluoride accumulation, particularly during the growth spurts of childhood; 2) Fluoride is a mutagen when present at sufficient concentrations, and 3) Fluoride stimulates the proliferation of bone-forming cells (osteoblasts), which may “increase the risk for some of the dividing cells to become malignant.” (NRC 2006).

A number of studies have failed to detect an association between fluoride and osteosarcoma. None of these studies, however, have looked at the risk of fluoride during specific ages in life. This is important because, in 2001, an age-specific analysis of a national case-control study that previously reported no association between lifelong exposure to fluoridated water and osteosarcoma (Douglass 1995) found that boys consuming fluoridated water during their 6th, 7th, and 8th years of life (the mid childhood growth spurt) had a statistically significant, “remarkably robust,” risk of developing osteosarcoma during their teenage years. (Bassin 2001). Initially published as a PhD dissertation at Harvard, the study was later published in Cancer Causes & Control.

Although a study in 2011 purported to refute the findings that fluoride causes osteosarcoma (Kim 2011), the study’s methods — by the authors’ own admission — were incapable of assessing the age-specific risk during the critical window period (ages 6 to 8 ) that Bassin identified as the critical risk period from fluoride exposure.

Timeline of Fluoride/Osteosarcoma Research:


US Congress requests National Toxicology Program (NTP) to conduct animal study to determine if fluoride causes cancer. A report by the National Academy of Sciences (NAS), titled “Drinking Water and Health”, expresses concern about a possible connection between water fluoridation and osteosarcoma in young males:

“There was an observation in the Kingston-Newburgh (Ast et al, 1956) study that was considered spurious and has never been followed up. There was a 13.5% incidence of cortical defects in bone in the fluoridated community but only 7.5% in the non-fluoridated community… Caffey (1955) noted that the age, sex, and anatomical distribution of these bone defects are `strikingly’ similar to that of osteogenic sarcoma. While progression of cortical defects to malignancies has not been observed clinically, it would be important to have direct evidence that osteogenic sarcoma rates in males under 30 have not increased with fluoridation.” (NAS 1977)


In December of 1989, the Medical Tribune reveals that a preliminary review of NTP’s cancer bioassay indicates an increase in bone cancer among the fluoride-treated rats. According to John Sullivan, a representative for the American Water Works Association, “If fluoride turns out to be a carcinogen, it will be the environmental story of the century.”


A) NTP Study

NTP confirms dose-dependent increase in osteosarcoma among male rats. While controversy begins to surface about a “systematic downgrading” by NTP of non-bone tumors (liver, oral, and thyroid) found with increased incidence among the fluoride-treated animals, the NTP upholds and confirms the dose-dependent increase in osteosarcoma among male rats. However, the NTP rules that the statistically-significant, dose-dependent increase of cancer in the target organ (bone) for fluoride accumulation, only provides “equivocal evidence” – versus “some evidence” or “clear evidence” – that fluoride is a carcinogen. According to NTP, “Equivocal evidence of carcinogenic activity is demonstrated by studies that are interpreted as showing a marginal increase of neoplasms that may be chemically related.” The journal Chemical & Engineering News reports:

“A number of other government officials who asked not to be identified also have told C&EN that they have concerns about the conclusions of the NTP study. They, too, believe that fluoride should have been placed in the “some evidence” category, in part because osteosarcoma is a very rare form of cancer in rodents.”

Despite criticism, the NTP maintains their assessment of “equivocal evidence.” In 1991, NTP scientists publish a paper which concludes:

“The current findings are weakly supportive of an association between sodium fluoride administration and the occurrence of osteosarcomas in male rats, but are not conclusive… [I]n view of the widespread exposure of the population to fluorides from a variety of sources it would appear prudent to re-examine previous animal and human epidemiologic studies, and perform further studies as needed to evaluate more fully any possible association between exposure to fluorides and the occurrence of osteosarcomas of bone.” Bucher JR, et al. (1991). Results and conclusions of the National Toxicology Program’s rodent carcinogenicity studies with sodium fluoride. International Journal of Cancer 48(5):733-7. July 9.

B) NCI Study

In light of the concern raised by NTP’s animal study, the National Cancer Institute (NCI) looks through its National Cancer Registry data to determine if osteosarcoma rates are higher in fluoridated communities. In a report released by the Public Health Service in 1991, the NCI reveals that the incidence of osteosarcoma throughout the U.S. has increased at a greater rate among young males in fluoridated areas vs. unfluoridated areas. The NCI, however, dismisses this result because of an inability to demonstrate a linear-dose relationship between the duration of fluoridation and the increased osteosarcoma incidence in fluoridated areas. To quote:

“In summary, analysis of incidence data from the SEER program has revealed some age- and sex-specific increases over time for bone and joint cancers, and for osteosarcomas, which are more prominent in fluoridated than in non-fluoridated areas. However, on further analysis these increases are unrelated to the timing of fluoridation, and thus are not linked to the fluoridation of water supplies.” (Hoover 1991)

C) Proctor & Gamble Study

In the spring of 1990, with the NTP’s findings attracting media attention, Procter & Gamble, a manufacturer of fluoridated toothpaste, releases the findings of its own rat study of fluoride and cancer (which the company conducted between 1981-1983). While Procter and Gamble’s study finds several bone tumors in the fluoride-treated animals (versus none in the controls), the results do not achieve statistical significance and Proctor & Gamble’s scientists dismiss them as random. According to the published report:

“All bone neoplasms were considered to be incidental and spontaneous and not related to fluoride treatment, because of their low incidence and random distribution” (Maurer JK, et al. 1990. Two-year carcinogenicity study of sodium fluoride in rats. Journal of the National Cancer Institute 82(13):1118-26. July 4.)

Later, in 1991, the FDA publishes an independent review of Procter & Gamble’s rat study. In their analysis, the FDA identify 2 additional osteosarcomas in the fluoride-treated rats which were not identified in Procter & Gamble’s published report. According to the FDA:

“The adequacy of the gross examination at necropsy was questioned based upon the rat tumors that were not identified by the contract (Procter & Gamble) laboratory” (FDA 1991).

However, even with the newly identified osteosarcomas, the FDA notes that the incidence of bone tumors in the Procter & Gamble study still do not achieve statistical significance. The FDA thereby concurs with Procter & Gamble that the bone tumors are incidental.


In addition to reviewing Procter & Gamble’s rat study (see above), the FDA also reviews Procter & Gamble’s mouse study (Maurer 1993). The FDA notes that, among both sexes of the fluoride-treated mice, there is a very significant, dose-dependent increase in benign bone tumors (osteomas), although no occurrence of osteosarcomas. While the occurrence of the osteomas is believed to be related to the presence of a virus in the mice, the FDA makes it clear that if the virus played a role, it did so in conjunction with the fluoride. To quote:

“Active virus was found in the osteomas but not in animals that did not have osteomas. It is clear, nonetheless, that if [the virus] had a role it was only in the presence of fluoride.”

Nevertheless, the FDA rules that Procter & Gamble’s mouse study – in light of the absence of malignant tumors – “do not add to the concerns raised by the NTP rat study.”


Following NCI’s analysis of national data, the New Jersey Department of Health conducts a study of osteosarcoma occurrence in Central New Jersey (Cohn PD. 1992. An Epidemiologic Report on Drinking Water and Fluoridation). The study finds a statistically significant relationship between fluoridation and osteosarcoma among males less than 20 years old:

“Recently, a national study of drinking water fluoridation at the country level found a significant association with osteosarcoma incidence among males under 20 years of age (Hoover et al., 1991). However, the meaning of the association was questioned by the authors because of the absence of a linear trend of association with the duration of time for which the water supplies were fluoridated… As a follow-up to the study by Hoover et al., a small study of similar design was initiated by the New Jersey Department of Health to compare drinking water fluoridatiuon at the municipal level with the municpal residence of osteosarcoma cases at the time of diagnosis… The study observed an association between fluoridation of water and osteosarcomas among males under 20 years of age in seven Central New Jersey counties.”


Dr. John Yiamouyiannis, the biochemist whose research on fluoridation/cancer in the 1970s led to Congress’ request for animal studies on fluoride/cancer, analyzes the National Cancer Institute’s data in addition to two other databases containing fluoride exposure/osteosarcoma information. Like NCI’s investigators (Hoover 1991), Yiamouyiannis finds osteosarcoma rates to be higher among young males under 20 in fluoridated versus unfluoridated areas. To quote:

“Recent studies showing substantial increases in the incidence of bone cancer and osteosarcoma in males (but not females) exposed to fluoride gave us the unique opportunity of using females as a control group to determine whether there is a link between fluoridation and bone cancer in males. Using three different data bases, we found that 1) the bone cancer incidence rate was as much as 0.95 cases a year per 100,000 population higher in males under age 20 living in fluoridated areas; 2) the osteosarcoma incidence rate was 0.85 new cases a year per 100,000 population higher in males under age 20 living in fluoridated areas; and 3) for males of all ages, the bone cancer death rate and bone cancer incidence rate was as much as 0.23 and 0.44 cases higher per 100,000 population, respectively, in fluoridated areas. These findings indicate that fluoridation is linked to an increase in bone cancer and deaths from bone cancer in human populations among males under age 20 and that this increase in bone cancer is probably all due to an increase in osteosarcoma caused by fluoride.”


A) New York Study (Gelberg)

A study on a population in New York State reports that fluoride exposure is not correlated with osteosarcoma. Although published in the American Journal of Public Health, the study contains a number of surprisingly glaring errors, uses a questionable method for calculating fluoride exposure from toothpaste, and fails to address the fact that its own data, when limited to fluoride exposure from water, shows an elevated risk of osteosarcoma when comparing those with some exposure to fluoridated water versus those with no exposure. Although this risk is not duration-dependent, this doesn’t necessarily mean that the effect is not real.

B) Harvard Study (Douglass)

A team of Harvard scientists, led by Dr. Chester Douglass, publish the preliminary findings of a large case-control analysis of fluoride and osteosarcoma (McGuire et al 1995). In the preliminary analysis the authors report no association between fluoride and osteosarcoma. In a Progress Report to the NIH, however, Chester Douglass notes that “all” of his analyses which assumed bottled water contains no fluoride found that fluoridated drinking water (>0.7 ppm) is associated with elevated, but not statistically significant, rates of osteosarcoma. In subsequent correspondence, Douglass expresses concern about the ramifications to water fluoridation from reporting that fluoridation is associated with an elevated, even if not statistically significant, rate of bone cancer:

“Because of the importance of the question at hand, we think the policy implications of reporting that the relative risk maybe higher than 1.5 would have consequences for fluoridation health policies.”

Consistent with Douglass’s concern, he never publishes any of his analyses that found elevated rates of osteosarcoma from fluoride exposure. Indeed, in all three of Douglass’s public summaries of his data (in 19951998 & 2002), he states that the study shows fluoridation has either no effect, or a slightly protective effect, on osteosarcoma rates.


A team of Japanese researchers publish results showing that fluoride is genotoxic to rat bone. (Mihashi 1996). The authors note that the fluoride-induced genotoxicity in bone reinforce the biologic plausibility of a fluoride-osteosarcoma connectioin. (Many compounds that cause genotoxicity also cause cancer). The authors used the same type of rat (F344/N) used in NTP’s cancer bioassay.

“Because the origin of osteosarcoma is considered to be osteoblastic/osteogenic cells, the ability of sodium fluoride to induce chromosome aberrations in these cells provides a mechanistic basis for the occurrence of osteosarcomas observed in sodium fluoride treated animals in the NTP study. Ingested fluoride is accumulated in bone, suggesting that osteoblastic/osteogenic cells in the bone microenvironment can be exposed to high levels of fluoride during bone formation. Our data and the NTP findings provide evidence that bone can be an organ for NaF carcinogenesis.” (Mihashi M, Tsutsui T. (1996). Clastogenic activity of sodium fluoride to rat vertebral body-derived cells in culture. Mutation Research 368(1):7-13. May.)


A doctoral student, Elise Bassin, working with Dr. Chester Douglass receives her doctorate from the Harvard School of Dental Medicine for a PhD dissertation that analyzes Douglass’s fluoride/osteosarcoma data (a large, NIH-funded case control study). Bassin finds a very strong, statistically-significant relationship (OR = 7.2) between consumption of fluoridated water during the mid-childhood growth spurt (ages 6-8) and osteosarcoma among boys less than 20 years old:

“Among males, exposure to fluoride at or above the target level was associated with an increased risk of developing osteosarcoma. The association was most apparent between ages 5-10 with a peak at six to eight years of age.. [T]he results continue to demonstrate an effect after adjusting by zipcode, county population, ever use of bottled or well water, age, and any use of self-administered fluoride products. For males, the odds ratio for the high exposure group was 7.20 at 7 years of age with a 95 percent confidence interval of 1.73 to 30.01… All of our models are remarkably robust in showing this effect during the mid-childhood growth spurt, which, for boys, occurs at ages seven and eight years. For females, no clear association between fluoride in drinking water during growth and osteosarcoma is shown in this study… Our results are consistent with findings from the National Toxicology Program animal study which found ‘equivocal evidence’ for an association between fluoride and osteosarcoma for male, but not female, rats and from two ecological studies that found an association for males less than twenty years old.” (Bassin EB. 2001. Association Between Fluoride in Drinking Water During Growth and Development and the Incidence of Osteosarcoma for Children and Adolescents. Thesis presented Doctor of Medical Sciences in the subject of Epidemiology. Harvard School of Dental Medicine, Boston, Mass.)


The National Research Council (NRC) begins a review of the safety of currently allowable levels of fluoride in drinking water.  Chester Douglass submits a summary of his fluoride/osteosarcoma study to the NRC. In his summary,  Douglass states that his study found no significant association between fluoridation and osteosarcoma. Although he cites Bassin’s study as one of 2 supporting references for this summary, Douglass makes no mention of the fact that Bassin found a statistically significant, 5-to-7-fold risk of osteosarcoma among boys drinking fluoridated water.


The Fluoride Action Network (FAN) obtains a copy of Elise Bassin’s PhD thesis and Douglass’s submission to the NRC. FAN provides these documents to the Environmental Working Group (EWG) which files a complaint of scientific misconduct with the National Institute of Health (NIH). In response to EWG’s complaint, the NIH launches an investigation. According to the Washington Post (7/13/05):

“Federal investigators and Harvard University officials are probing whether a Harvard professor buried research suggesting a link between fluoridated tap water and bone cancer in adolescent boys. The National Institute of Environmental Health Sciences (NIEHS), which funded Chester Douglass’s $1.3 million study, and the university are investigating why the Harvard School of Dental Medicine epidemiologist told federal officials he found no significant correlation between fluoridated water and osteosarcoma, a rare form of bone cancer. Douglass, who serves as editor in chief for the industry-funded Colgate Oral Care Report, supervised research for a 2001 doctoral thesis that concluded boys exposed to fluoridated water at a young age were more likely to get the cancer. The Environmental Working Group, an advocacy organization, urged federal officials late last month to explore whether Douglass had skewed his 2004 report to the institute to play down possible risks associated with fluoridation.”

The investigation is managed and run by Harvard University. No one from EWG is contacted during the course of the investigation, and Chester Douglass is given a draft copy of the report’s findings several months prior to Harvard issuing a short, one page press release announcing that Douglass did not “intentionally misrepresent” the research.


In the wake of allegations that Chester Douglass was suppressing data suggesting a link between fluoride and osteosarcoma, Dr. Elise Bassin and several other Harvard scientists publish the findings of Bassin’s PhD dissertation. As with Bassin’s dissertation (which had been completed in 2001), the published study reports that boys drinking fluoridated water during the ages of 6 to 8 have a five-fold increased risk of developing osteosarcoma during their teenage years:

“We observed that for males diagnosed before the age of 20 years, fluoride level in drinking water during growth was associated with an increased risk of osteosarcoma, demonstrating a peak in the odds ratios from 6 to 8 years of age. All of our models were remarkably robust in showing this effect, which coincides with the mid-childhood growth spurt. For females, no clear association between fluoride in drinking water during growth and osteosarcoma emerged.” (Bassin EB, et al. 2006. Age-specific fluoride exposure in drinking water and osteosarcoma (United States). Cancer Causes & Control 17(4):421-8. May.)

In the same issue that Bassin publishes this study, Chester Douglass publishes a letter in which he publicly discloses for the first time that he had found some associations between fluoride exposure and osteosarcoma in the (retrospective) dataset that Bassin analysed. Douglass states that he was unable to replicate these findings in a new (prospective) dataset, and thus cautions readers from making any conclusions based on Bassin’s findings. Douglass notes, however, that he has yet to conduct an age-specific analysis on the prospective data. He notes though that he is planning on doing so. To quote:

 “A parallel analysis of age-specific exposure to fluoride, especially during growth periods, is also being pursued by our study team in the second set of cases of our study. Accordingly, readers are cautioned not to generalize and over-interpret the results of the Bassin et al. paper and to await the publications from the full study, before making conclusions, and especially before influencing any related policy decisions.” (Douglass CW, and Joshipura K. 2006. Caution needed in fluoride and osteosarcoma study. Cancer Causes & Control 17(4):481-82. May.)


Nineteen years after receiving his grant from the NIH to study fluoride/osteosarcoma, Chester Douglass publishes his first paper on his fluoride findings. Although Douglass indicated in his 2006 letter that he would be conducting an age-specific analysis, the published paper does not do so. Instead, the paper focuses exclusively on the relationship between bone fluoride content and osteosarcoma — an analysis incapable of determining age-specific risk. Although the paper concludes that there is no relationship between bone fluoride content and osteosarcoma, there are several notable shortcomings in the study’s methods. These shortcomings include:

  • Douglass’s study was much smaller and weaker than Bassin’s: It had only 20 control subjects under age 30, a fifth the size of Bassin’s sample. For this key age group, Douglass’ study was so small it could provide no reliable conclusions. Even Douglass admitted this serious limitation.
  • The controls were much older than the cases: Controls were much older (average = 41 years) than the cases (average = 18 yrs). Since fluoride builds up in bone with age, it is obvious that a 41-year-old will have higher bone-fluoride levels than an 18-year-old. The fact therefore that Douglass did not find higher bone fluoride levels in 18-year-old osteosarcoma patients than 41-year-old non-osteosarcoma patients can hardly be deemed surprising or significant. It is unlikely that Douglass could have adequately compensated for this gross mismatch in age.

FAN is currently working on a more thorough analysis of Douglass’s study which we will be releasing soon.


Canadian researchers publish an analysis of water fluoridation and osteosarcoma rates in children in the United States. (Levy & Leclerc, 2012). The study concludes that water fluoridation in the U.S. is not associated with an increased risk of osteosarcoma. The authors use a notably crude measurement for determining fluoride exposure. Specifically, the authors used the National Cancer Institute’s SEER data, and limited their analysis to the average fluoridation rate of the child’s STATE of residence at the time of diagnosis. By contrast, when the NCI conducted its analysis of the SEER data in 1990 (in which NCI found elevated rates of osteosarcoma among young males in fluoridated areas), the NCI considered the fluoridation status on the COUNTY level — a smaller unit which is less prone to classification error. Also, by focusing on the child’s residence at the time of diagnosis, the study is incapable of determining the effect of fluoride intake at the relevant moment in time: namely, the age at which the osteosarcoma was induced. The study thus sheds little light on fluoride’s possible relationship to osteosarcoma.

References – Timeline:

References: Studies Reporting No Relationship between Fluoride & Osteosarcoma:

For a review and critique of the following studies, click here