Health Stream Article - Issue 42 - June 2006

Fluoride And Bone Cancer

Researchers from the Harvard School of Dental Medicine have published the outcomes of a case-control study of osteosarcoma which found a significant association between exposure to fluoride and risks for this rare cancer (1). The newly published paper was not included in the recent review of evidence on the US EPA's fluoride standards conducted by the US National Research Council, although the review committee did assess the PhD thesis from which it was derived (2).

Osteosarcoma is a very uncommon primary malignant tumour of bone. In the US the incidence rate is about 3.3 cases per million children under the age of 15 years. There is evidence that suggests osteosarcoma is associated with skeletal growth, particularly for cases diagnosed during adolescence. Fluoride uptake into bone increases during rapid growth and a number of epidemiological studies have examined possible associations between fluoride exposure and osteosarcoma, with mixed results. Several animal studies have been conducted with one finding an apparent increase in risk only in males, while other studies were negative.

The new case-control study included histologically confirmed cases of osteosarcoma diagnosed between November 1989 and November 1992 at 11 teaching hospitals across the US. Individuals aged 40 years or older were excluded, as were those with a history of radiation therapy or renal dialysis. Matched controls were recruited from patients in the orthopaedic department in the same hospital as the case, who had been seen within 6 months (before or after) the case. Controls were matched by age (+/- 5 years), gender, and distance from hospital.

Information on usual drinking water sources, residential history, use of fluoride supplements and mouth rinses was collected by telephone interview. Where the case was incapacitated or deceased, information was collected from a relative. Information on fluoride levels in public water supplies was collected from water suppliers or relevant state or local authorities, and water samples from private wells were collected and analysed. The fluoride level in bottled water was assumed to be 0.1 ppm based on levels in leading brands. For those who reported bottled water as their drinking water source at home, it was assumed that some tap water would also be consumed in food and outside the home. Fluoride exposure levels in this group were assumed to be the mean of bottled and tap water values.

The data collected from case and controls did not include estimates of the volume of each type of water consumed. An adjustment for the relative exposure from all three drinking water sources in different geographic areas was made by standardising against CDC recommendations for optimum fluoride levels. For example in warmer climates the fluoride level was divided by 0.7 to correspond with the CDC recommended fluoride level of 0.7 ppm. In colder climates the fluoride level was divided by 1.2 to correspond with the CDC recommended level of 1.2 ppm. This procedure provided a weighting for the (presumed) increased consumption of water in warmer areas.

Only the results for cases under 20 year of age were presented in the paper. To examine the association between osteosarcoma and fluoride exposure at specific ages, separate conditional logistic regression models were fitted for each exposure age up to the age of diagnosis for each case and the same age for the matched controls. The genders were analysed separately.

Of 157 cases diagnosed before 20 years of age, 103 (60 male, 43 female) were included in the analysis together with 215 matched controls. The median age at diagnosis was 14 years. Median family income was significantly lower for cases than controls, and a significantly greater proportion of controls drank bottled water.

Exposure to fluoride at or above the target level among males was associated with an increased risk of developing osteosarcoma compared to the lowest fluoride exposure group. This association was most apparent between ages 4 and 12 with a peak at 6 to 8 years of age. The odds ratio was found to be 4.07 at 7 years for the high exposure group (95% CI 1.43-11.56). After adjusting for potential confounders a similar result was found with an adjusted odds ratio for males of 5.46 (95% CI 1.50-19.90) at age 7 years. In females, no association was found between fluoride in drinking water and osteosarcoma at any age of exposure.

Sensitivity analyses using different assumptions for the fluoride content of bottled water (0.1 to 0.5 ppm) or omitting the adjustment for climate-specific fluoride targets had little impact on the results. A subgroup analysis in males who reported never using fluoride supplements (n=33) produced similar results.

These results suggest that exposure to fluoride is associated with an increased risk of osteosarcoma in young males with the most sensitive exposure period being around 6 to 8 years of age. However there does not seem to be a significant increase in risk for females. In discussing the findings the authors note that odds ratio estimates for the intermediate exposure category (30-99% of target fluoride level) were generally similar to the highest exposure category (99% or more of target fluoride levels), rather than showing a clear dose-response effect. This may have been due to exposure misclassification as it has been reported that target levels for fluoridation are not maintained in many water supplies.

Fluoride exposure in this study was assessed in terms of a semi-quantitative estimate for drinking water sources plus ever/never use of fluoride supplements (rinses, tablets, drops, school programs). It appears that use of fluoride-containing toothpaste was not considered in the analysis. Estimates of average fluoride intakes in the US indicate that for a 6-12 year old child intake from fluoride supplements is 0.0250 mg/kg body weight /day, while intake from non-water sources (including toothpaste) is 0.0219 mg/kg/day (2). Intake from water (assuming a fluoride concentration of 1 ppm) is 0.0178 mg/kg/day. It is perhaps surprising therefore that the odds ratios in the subgroup with no use of fluoride supplements were similar to those for the total group given that their total fluoride intake would be lower than those who used fluoride supplements.

The case-control paper was accompanied by a letter to the journal editor advising caution when interpreting the findings of the paper. The letter from a senior member of the research group that conducted the study, points out that a second case-control study subsequently conducted in the same group of hospitals has failed to replicate the results. A second set of osteosarcoma cases was recruited between 1993 and 2000 from new incident cases, and similar exposure assessment methods were used. Preliminary finding from the overall analysis of the second set of cases do not appear to replicate the overall findings from the first part of the study. In addition, many of the cases and controls recruited in the second time period agreed to provide bone specimens. The preliminary analysis of the fluoride content of the bone specimens suggests that fluoride level within the bone is not associated with excess risk of osteosarcoma. It is recommended that conclusions should not be drawn until publication of the second study which will also examine whether age-specific exposure effects on risk are evident.

(1) Age-specific fluoride exposure in drinking water and osteosarcoma (United States). Bassin EB, Wypij D, Davis RB and Mittleman MA. (2006) Cancer Causes and Control, 17(421-8).
(2) The outcomes of the NRC review were reported in Health Stream Issue 41.
(3) Caution needed in fluoride and osteosarcoma study. Douglass CW and Joshipura K. (2006) Cancer Causes and Control, 17(481-482).