A case-control study of topical and supplemental fluoride use and osteosarcoma risk

Abstract

Background.

A relationship between fluoride and osteosarcoma has been hypothesized but not validated. To the authors’ knowledge, there are no published studies examining topical fluoride or dietary fluoride supplements and osteosarcoma risk. The purpose of this study was to examine the association between ever or never use of topical and dietary fluoride supplements and osteosarcoma.

Methods.

The authors performed a secondary data analysis on data from 2 separate but linked studies. Patients for Phase 1 and Phase 2 were selected from US hospitals using a hospital-based matched case-control study design. Case patients were those who had received diagnoses of osteosarcoma, and control patients were those who had received diagnoses of other bone tumors or nonneoplastic conditions. In Phase 1, case patients (N = 209) and control patients (N = 440) were those seeking treatment at orthopedic departments from 1989 through 1993. In Phase 2, incident case patients (N = 108) and control patients (N = 296) were identified and treated by physicians from 1994 through 2000. This analysis included all patients who met eligibility criteria and on whom the authors had complete data on exposure, outcome, and covariates. The authors used conditional logistic regression to estimate odds ratios and 95% confidence intervals (CIs) for the association of topical fluoride use and supplemental fluoride use with osteosarcoma.

Results.

The adjusted odds ratios were 0.94 (95% CI, 0.60 to 1.46) and 0.78 (95% CI, 0.46 to 1.33) for topical fluoride and supplemental fluoride, respectively.

Conclusions.

Neither topical nor dietary fluoride supplements are associated with an increased risk of developing osteosarcoma.

Practical Implications.

Supplemental and topical fluorides used in the dental office and in over-the-counter products are not related to an increased risk of developing osteosarcoma.

Fluoride occurs naturally in water and soil. The effectiveness of controlled amounts of fluoride in preventing caries has been well documented.^1-5 Fluoride tablet administration may have both preeruptive and posteruptive influences on teeth, each through different mechanisms.^6,7 Fluoride has an affinity for calcified tissues such as bone and developing teeth, and 99% of fluoride in the body is contained in the skeleton. This uptake increases the rate of osteoblast proliferation and bone formation.⁸ Topical fluorides are available in many forms, primarily mouthrinses, toothpastes, or professional application such as fluoride varnish. Most commercial toothpastes contain stannous fluoride or sodium fluoride. Topical fluorides enhance the remineralization process.

Dietary fluoride supplements are available via prescription from physicians and dentists as drops, lozenges, or tablets. Most supplements contain sodium fluoride as the active ingredient.^6,7 The recommendations for use of fluoride supplements vary across the world.⁷ In the United States, fluoride supplements are prescribed on the basis of age and level of fluoride in the drinking water.⁷ Topical fluorides, in the form of mouthrinses, are available readily in pharmacies and most supermarkets and convenience stores across the United States.

Fluoride ingestion has been suggested as a possible risk factor for osteosarcoma in a National Toxicology Program animal study in which 4 of 130 male rats given very high doses of sodium fluoride (100 parts per million and 175 parts per million) for 2 years developed osteosarcoma.⁹ In the same National Toxicology Program study, female rats and male and female mice with similar levels of exposure did not develop osteosarcoma,⁹ and 2 follow-up animal studies did not find an association between fluoride and osteosarcoma.^10,11

Numerous human studies examining this potential association also have been conducted that found no association between fluoridation and osteosarcoma occurrence.^12-27 Of these, time trend studies of cancer in communities before and after introduction of fluoridation and all-cause cancer mortality also have reported no difference in osteosarcoma incidence rates between the fluoridated and nonfluoridated areas.^12-15 A number of ecological studies comparing the rates of cancer mortality and incidence in areas with nonfluoridated and fluoridated drinking water supplies and time trends of all-cause mortality also have failed to find any association or evidence of increased mortality after the introduction of fluoride into the water supply.^13,16-19 Additional ecological studies have found that water fluoridation was not associated with increased incidence rates of osteosarcoma during childhood or adolescence.^20,21 A 2014 ecological study from Great Britain also did not find an association between naturally fluoridated areas or those with adjusted fluoride levels and osteosarcoma.²² Five case-control studies evaluating the fluoride-osteosarcoma association have been published.^23-27 Three studies did not find any significant association between fluoridated drinking water and risk of developing osteosarcoma.^23-25 One study found that total lifetime fluoride exposure was not associated significantly with osteosarcoma.²⁶ In an exploratory analysis, Phase 1 of a linked case-control study, an increased risk was reported among males in 1 age group who were exposed to fluoride in drinking water during childhood.²⁸ However, the fluoride levels in the bone samples obtained in Phase 2 of this linked case-control study did not show an association with osteosarcoma at any age.²⁵ A 2020 analysis of community water fluoridation and osteosarcoma also found that none of the results for age-specific multivariable analyses showed a significant association.²⁷

Accordingly, the hypothesized relationship between fluoride use and osteosarcoma has not been validated. To our knowledge, there are no studies reported in the literature that specifically address the use of topical fluoride or fluoride supplements and osteosarcoma. Hence, the purpose of our study was to assess topical fluoride and dietary fluoride supplement use and risk of developing osteosarcoma.

METHODS

The data we report in this article were collected from 2 separate but linked studies, hereafter referred to as Phase 1 and Phase 2. Phase 1 was carried out from 1989 through 1993, and Phase 2 was carried out from 1994 through 2000. The 2 phases were conducted under similar but not identical protocols in most of the same hospitals. Phase 1 was approved by the institutional review boards of the respective hospitals and Harvard Medical School. Phase 2 was approved by the institutional review boards of the respective hospitals, Harvard Medical School, and the Medical College of Georgia.

Study population

Study patients for Phase 1 and Phase 2 were selected from 12 US research hospitals using a hospital-based case-control study design (eTable 1, available online at the end of this article), of which 8 hospitals participated in both phases. All patients with osteosarcoma were eligible. Control patients were identified on the basis of their age, sex, and geographic distance from these hospitals that were similar to case patients. In both phases, control patients were matched to case patients according to sex, age within plus or minus 5 years, and distance from the hospital. An inclusion-exclusion flowchart describing the patients included in the analysis can be found in the Figure.

Figure.

Patient recruitment.

Case patient definition

In Phase 1, all patients diagnosed histologically with osteosarcoma by the participating orthopedic departments from 1989 through 1993 were identified from the patient record files and recruited for telephone interviews. In Phase 2, incident case patients with osteosarcoma were identified by physicians in the orthopedic departments of participating institutions, and all patients who received a diagnosis of and were treated for primary osteosarcoma from 1994 through 2000 were considered eligible. Diagnosis of osteosarcoma was confirmed from surgical pathology reports.

Control patient definition

In Phase 1, control patients were those of record from 1989 through 1993 and chosen using the matching criteria from within the same orthopedic surgery department as the case patients. The control patients were those with other bone tumors or nonneoplastic conditions. In Phase 2, control patients were also recruited from the same orthopedic departments as the case patients from 1994 through 2000. The control patients included those with newly diagnosed malignant bone tumor other than osteosarcoma, such as Ewing sarcoma, chondrosarcoma, and malignant fibrous histiocytoma (tumor control patients); orthopedic and surgical patients with benign tumors and nonneoplastic conditions (inflammatory diseases, trauma, sports injuries), including day surgical orthopedic patients (orthopedic control patients).

Exclusion criteria for study entry

Patients older than 40 years at diagnosis or who reported prior radiation treatment for cancer or a history of kidney dialysis were excluded in Phase 1 but not in Phase 2. In Phase 2, no age restrictions or other exclusion criteria were placed on patients entering the study, and thus Phase 2 attempted to enroll all case patients. In addition, in Phase 2, patients were eligible if they were residents of the United States or were US citizens temporarily living outside the United States, for example students, members of the military, or people with work assignments abroad. However, foreign nationals who were in the United States solely for treatment were not eligible.

Data collection of covariates

In Phase 1, a telephone interview lasting approximately 15 through 20 minutes was conducted with patients 18 through 40 years old at their homes or with a parent if patients were younger than 18 years. If the patient was deceased at the time of the telephone interview, data collection was ascertained from a family proxy of the deceased patient. The questionnaire included demographic characteristics (age, sex, race or ethnicity, education), information on residences, and ever or never use of topical fluoride (in the form of mouthrinses) and supplemental fluoride (drops, tablets). Income was determined using a surrogate measure based on the ZIP Code at the time of diagnosis according to data from the US Census Bureau.

In Phase 2, all eligible patients providing consent to participate were interviewed in-person by trained interviewers in English during hospitalization or pre- or postadmission. If the patient was younger than 21 years, a parental or guardian consent was required for participation in the interview as a corespondent. The questionnaire included information on demographic characteristics (age, sex, race or ethnicity, education, income), residence, topical and supplemental fluoride use, occupational history, medical history, family medical history, and growth and development.

Fluoride exposure

Fluoride exposure from sources such as toothpaste, fluoride supplements, and mouthrinses was ascertained in the questionnaire. In telephone interviews in Phase 1 and in in-person interviews in Phase 2, patients were asked about which brand or brands of toothpaste they usually buy, use of mouthrinses (current use, ever routinely used, brand, frequency, number of years used), participation in a school mouthrinse program (number of years of participation, frequency, school, city or town and state), and use of fluoride tablets, pills, or drops (number of years used). The history of living in a fluoridated or nonfluoridated community was collected using each patient’s residential history. The date when each town or city was fluoridated is a publicly recorded health policy decision.

Statistical methods

We summarized and compared baseline characteristics between case and control patients using χ² tests (Tables 1 and 2). We used conditional logistic regression models to account for matching. We initially performed analyses on data from Phase 1 and Phase 2 study groups independently to estimate the association of topical fluoride use or supplemental fluoride use with osteosarcoma in each phase. We considered age and distance from the hospital for inclusion in the models to adjust for possible residual confounding. Subsequently, we omitted age because it did not have a significant effect in univariable or multivariable models. We included the matching factor, distance in miles from the hospital (0-100 miles, > 100 miles), in the final model because there was a significant association shown in the multivariable analysis. There was no significant result when distance was considered as a continuous variable. A sensitivity analysis showed there was no effect on the model with regard to the independent variables of interest; for topical fluoride use, odds ratio (OR) was 0.99 (95% confidence interval [CI], 0.64 to 1.53), and for supplemental fluoride use, OR was 0.78 (95% CI, 0.46 to 1.33) when dichotomous distance was dropped from the analysis. The final models also included the following adjustment variables: race, ethnicity, income, urban-living status, bottled water use, and history of living in a fluoridated area. We assessed access to fluoridated water by first determining the residential history, which we then linked with public information on community water fluoridation. We included this information to account for other sources of fluoride. The ubiquitous use of fluoride toothpaste led to dropping that variable from the final model.

Table 1.

Univariate associations comparing matched case patients with control patients, Phases 1 and 2 topical fluoride use, all matched patients.^*

VARIABLE	TOTAL, NO. (N = 755)	CASE PATIENTS, NO. (%) (N = 261)	CONTROL PATIENTS, NO. (%) (N = 494)	ODDS RATIO (95% CONFIDENCE INTERVAL)	P VALUE
Age, y
1-10	85	26 (10.0)	59 (11.9)	[Reference]	NA†
11-20	437	164 (62.8)	273 (55.3)	1.36 (0.83 to 2.25)	.225
2-30	170	48 (18.4)	122 (24.7)	0.89 (0.51 to 1.58)	.697
31-40	63	23 (8.8)	40 (8.1)	1.31 (0.66 to 2.60)	.450
Sex
Female	304	107 (41.0)	197 (39.8)	[Reference]	NA
Male	451	154 (59.0)	297 (60.1)	0.96 (0.70 to 1.30)	.766
Race
Non-White	115	47 (18.0)	68 (13.8)	[Reference]	NA
White	640	214 (82.0)	426 (86.2)	0.73 (0.48 to 1.09)	.124
Ethnicity
Non-Hispanic	709	245 (93.9)	464 (93.9)	[Reference]	NA
Hispanic	46	16 (6.1)	30 (6.1)	1.01 (0.54 to 1.89)	.975
Income, $
< 40,000	366	139 (53.3)	227 (46.0)	[Reference]	NA
40,000-60,000	233	78 (29.9)	155 (31.4)	0.82 (0.58 to 1.16)	.264
> 60,000	156	44 (16.9)	112 (22.7)	0.64 (0.43 to 0.96)	.033
Ever Lived in an Urban Area
No	211	76 (29.1)	135 (27.3)	[Reference]	NA
Yes	544	185 (70.9)	359 (72.7)	0.92 (0.66 to 1.28)	.602
Distance From Hospital, Miles
0-100	458	149 (57.1)	309 (62.6)	[Reference]	NA
> 100	297	112 (42.9)	185 (37.5)	1.26 (0.93 to 1.70)	.144
Patient Education
< High school	366	142 (54.4)	224 (45.3)	[Reference]	NA
High school-some college	277	95 (36.4)	182 (36.8)	0.82 (0.60 to 1.14)	.242
≥ College degree	109	23 (8.8)	86 (17.4)	0.42 (0.25 to 0.70)	.001
Missing	3	1 (0.4)	2 (0.4)	NA	NA
Ever Drank Well Water
No	514	174 (66.7)	340 (68.8)	[Reference]	NA
Yes	241	87 (33.3)	154 (31.2)	1.10 (0.80 to 1.52)	.545
Ever Drank Bottled Water
No	629	235 (90.0)	394 (79.8)	[Reference]	NA
Yes	126	26 (10.0)	100 (20.2)	0.44 (0.28 to 0.69)	.001
Ever Used Topical Fluoride
No	625	214 (82.0)	411 (83.2)	[Reference]	NA
Yes	130	47 (18.0)	83 (16.8)	1.09 (0.73 to 1.61)	.676
Ever Lived in a Fluoridated Area
No	169	66 (25.3)	103 (20.9)	[Reference]	NA
Yes	586	195 (74.7)	391 (79.2)	0.78 (0.55 to 1.11)	.165
Ever Used Supplemental Fluoride
No	481	173 (66.3)	308 (62.4)	[Reference]	NA
Yes	120	36 (13.8)	84 (17.0)	0.76 (0.50 to 1.18)	.220
Missing	154	52 (19.9)	102 (20.6)	NA	NA
Fluoride Toothpaste Use
No	1	0 (0.0)	1 (0.2)	NA	NA
Yes	752	260 (99.6%)	492 (99.6%)	NA‡	NA‡
Missing	2	1 (0.4)	1 (0.2)	NA	NA

^*Percentages may not add up to 100 due to rounding.

^†NA: Not applicable.

^‡All but 1 patient (control) used fluoride toothpaste. Fluoride toothpaste use was unknown for 2 patients. The odds ratio was not calculated. That variable was not included in the multivariable logistic regression model.

Table 2.

Univariate associations comparing case patients with control patients, univariate analyses Phases 1 and 2 supplement fluoride use, all matched patients.^*

VARIABLE	TOTAL, NO. (N = 609)	CASE PATIENTS, NO. (%) (N = 212)	CONTROL PATIENTS, NO. (%) (N = 397)	ODDS RATIO (95% CONFIDENCE INTERVAL)	P VALUE
Age, y
1-10	73	23 (10.9)	50 (12.6)	[Reference]	NA†
11-20	356	133 (62.7)	223 (56.2)	1.30 (0.76 to 2.22)	.345
21-30	132	38 (17.9)	94 (23.7)	0.88 (0.47 to 1.64)	.684
31-40	48	18 (8.5)	30 (7.6)	1.30 (0.61 to 2.80)	.496
Sex
Female	233	88 (41.5)	145 (36.5)	[Reference]	NA
Male	376	124 (58.5)	252 (63.5)	0.81 (0.58 to 1.14)	.228
Race
Non-White	85	36 (17.0)	49 (12.3)	[Reference]	NA
White	524	176 (83.0)	348 (87.7)	0.69 (0.43 to 1.10)	.117
Ethnicity
Non-Hispanic	575	201 (94.8)	374 (94.2)	[Reference]	NA
Hispanic	34	11 (5.2)	23 (5.8)	0.89 (0.43 to 1.86)	.757
Income, $
< 40,000	303	110 (51.9)	193 (48.6)	[Reference]	NA
40,000-60,000	200	67 (31.6)	133 (33.5)	0.88 (0.61 to 1.29)	.519
> 60,000	106	35 (16.5)	71 (17.9)	0.87 (0.54 to 1.38)	.543
Ever Lived in an Urban Area
No	179	65 (30.7)	114 (28.7)	[Reference]	NA
Yes	430	147 (69.3)	283 (71.3)	0.91 (0.63 to 1.31)	.616
Distance From Hospital, Miles
0-100	356	120 (56.6)	236 (59.5)	[Reference]	NA
> 100	253	92 (43.4)	161 (40.6)	1.12 (0.80 to 1.58)	.498
Patient Education
< High school	293	114 (53.8)	179 (45.1)	[Reference]	NA
High school-some college	224	78 (36.8)	146 (36.8)	0.84 (0.58 to 1.20)	.341
≥ College degree	89	19 (9.0)	70 (17.6)	0.43 (0.24 to 0.75)	.003
Missing	3	1 (0.5)	2 (0.5)	NA	NA
Ever Drank Well Water
No	417	144 (67.9)	273 (68.8)	[Reference]	NA
Yes	192	68 (32.1)	124 (31.2)	1.04 (0.73 to 1.49)	.831
Ever Drank Bottled Water
No	503	190 (89.6)	313 (78.8)	[Reference]	NA
Yes	106	22 (10.4)	84 (21.2)	0.43 (0.26 to 0.71)	.001
Ever Used Supplemental Fluoride
No	488	175 (82.6)	313 (78.8)	[Reference]	NA
Yes	121	37 (17.4)	84 (21.2)	0.79 (0.51 to 1.21)	.276
Ever Lived in a Fluoridated Area
No	148	55 (25.9)	93 (23.4)	[Reference]	NA
Yes	461	157 (74.1)	304 (76.6)	0.87 (0.59 to 1.28)	.490
Ever Used Topical Fluoride
No	499	171 (80.7)	328 (82.6)	[Reference]	NA
Yes	102	38 (17.9)	64 (16.1)	1.14 (0.73 to 1.77)	.564
Missing	8	3 (1.4)	5 (1.3)	NA	NA
Fluoride Toothpaste Use
No	1	0 (0.0)	1 (0.3)	NA	NA
Yes	606	211 (99.5)	395 (99.5)	NA‡	NA‡
Missing	2	1 (0.5)	1 (0.3)	NA	NA

^*Percentages may not add up to 100 due to rounding.

^†NA: Not applicable.

^‡All but 1 patient (control) used fluoride toothpaste. Fluoride toothpaste use was unknown for 2 patients. The odds ratio was not calculated. That variable was not included in the multivariable logistic regression model.

The matching criterion for age was plus or minus 5 years, which created a large age interval when considering the patient’s education level (less than high school, high school-some college, ≥ college degree) in our models. Because osteosarcoma mostly affects young people, case patients may not have had a chance to reach higher levels of education, as shown by our findings in which the control patients in both the topical and the supplemental fluoride analytic groups were more likely to have a college degree or higher than the case patients (17.4% versus 8.8% for topical, respectively, and 17.6% versus 9.0% for supplemental, respectively). For these reasons, we did not consider patient education as a covariate in the multivariable analysis. Mother’s education, a known influential variable in children’s health outcomes, was collected only in Phase 2.²⁹ A multivariable analysis using only Phase 2 data showed that mother’s education had no influence on the models with regard to topical and supplemental fluoride use and osteosarcoma. The adjusted OR for topical fluoride and osteosarcoma in Phase 2 was 0.1.31 (95% CI, 0.65 to 2.65) and for supplemental fluoride, 0.88 (95% CI, 0.21 to 3.57) when mother’s education was included as an adjustment variable in the model.

To determine whether associations with fluoride use varied by study phase, we added a phase indicator and phase by fluoride use interaction term to the multivariable logistic regression model run on the combined data from both study phases. Neither term was statistically significant. The nonsignificant interaction term indicates that the association between fluoride exposures and osteosarcoma was not significantly different between the 2 phases, allowing the 2 phases to be combined for analysis. Hence, all results reported in our article are for the combined phases study groups in both the topical and the supplemental fluoride models. Results are reported as ORs and 95% CIs. We performed all statistical analyses using SAS software Version 9.4 (SAS Institute), and we considered P values below .05 as significant.

RESULTS

Table 1 (ever or never use of topical fluoride) and Table 2 (ever or never use of supplemental fluoride) show the univariable comparisons for the matching factors used in the study design for each fluoride type: age (1-10, 11-20, 21-30, 31-40 years), sex (male, female), and distance from hospital (0-100 miles, > 100 miles). None of the matching factors were significantly different for case and control patients. The prevalence of topical fluoride use was 18.0% among case patients and 16.8% among control patients. The prevalence of supplemental fluoride use was 17.5% among case patients and 21.2% among control patients.

Annual income (< $40,000, $40,000-$60,000, ≥ $60,000), did show a significant univariable association with osteosarcoma in the topical fluoride analysis group (Table 1). The highest income group showed a protective effect in its association with osteosarcoma compared with the reference group in the topical fluoride analysis (P = .033). This protective effect was not present in the supplemental fluoride results (Table 2). Drinking well water showed no association with osteosarcoma in the topical fluoride analysis (OR, 1.10; 95% CI, 0.80 to 1.52) or the supplemental fluoride analysis (OR, 1.04; 95% CI, 0.73 to 1.49). The univariable comparison showed a significant protective association of drinking bottled water with osteosarcoma in the topical fluoride analysis (OR, 0.44; 95% CI, 0.28 to 0.69) and the supplemental fluoride analysis (OR, 0.43; 95% CI, 0.26 to 0.71). The results for the univariable comparisons by study group for Phase 1 and Phase 2 are similar to the combined analysis and may be found in eTables 2-9, available online at the end of this article.

The ORs from the multivariable conditional logistic regression that included only living in a fluoridated area as the covariate for the combined Phase 1 and Phase 2 analysis group are shown in Table 3 (topical fluoride) and Table 4 (supplemental fluoride). The adjusted ORs for living in a fluoridated area and osteosarcoma were 0.98 (95% CI, 0.65 to 1.49) for topical fluoride and 0.71 (95% CI, 0.43 to 1.19) for supplemental fluoride.

Table 3.

Conditional logistic regression: odds ratios and 95% confidence intervals for ever used topical fluoride and risk of developing osteosarcoma for all matched patients (N = 755, N_cases = 261).

MODEL	ODDS RATIO	95% CONFIDENCE INTERVAL	P VALUE
Adjusted for Fluoridation Status Conditional Logistic Regression	0.98	0.65 to 1.49	.936
Adjusted Conditional Logistic Regression *	0.94	0.60 to 1.46	.773

^*Adjusted for race, ethnicity, income, residing in urban setting, distance from hospital, ever drank bottled water, and ever lived in a fluoridated area.

Table 4.

Conditional logistic regression: odds ratios and 95% confidence intervals for ever used supplemental fluoride and risk of developing osteosarcoma for all matched patients (N = 609, N_cases = 212).

MODEL	ODDS RATIO	95% CONFIDENCE INTERVAL	P VALUE
Adjusted for Fluoridation Status Conditional Logistic Regression	0.71	0.43 to 1.19	.197
Adjusted Conditional Logistic Regression *	0.78	0.46 to 1.33	.364

^*Adjusted for race, ethnicity, income, residing in urban setting, distance from hospital, ever drank bottled water, and ever lived in a fluoridated area.

Bottled water analysis

We included bottled water use as an adjustment variable in the multivariable analyses because the univariable analysis showed a protective association for bottled water use and disease state. The rate of consuming bottled water was low in our study (16.7% for the topical fluoride use analysis, 17.4% for the supplemental fluoride use analysis), likely attributable to the period of the data collection (1994-2000).

Topical and supplemental fluoride analysis

Also shown in Tables 3 and 4 are the results of the full multivariable conditional logistic regression that included race (non-White, White), ethnicity (non-Hispanic, Hispanic), categorized annual income, urban versus rural living status, distance from hospital, ever drank bottled water, and history of living in a fluoridated area as adjusting factors. Adjusting for these factors, the OR of topical fluoride use was 0.94 (95% CI, 0.60 to 1.46), and the OR for supplemental fluoride was 0.78 (95% CI, 0.46 to 1.33). Drinking bottled water is significant in the multivariable conditional logistic regressions performed that included the bottled water indicator and other adjustment variables: for topical fluoride analysis, OR was 0.41 (95% CI, 0.24 to 0.70), and for supplemental fluoride analysis, OR was 0.37 (95% CI, 0.19 to 0.72).

The multivariable results were similar for Phase 1 and Phase 2 when the phases were run separately (eTables 2-9, available online at the end of this article).

DISCUSSION

The results of our hospital-based case-control study provide no evidence of an association between osteosarcoma and topical fluoride use or systemic fluoride intake in the form of dietary fluoride supplements. If the hypothesis that topical or supplemental fluoride use is associated with osteosarcoma were valid, it would be expected that fluoride used topically or as supplements would show a relationship with osteosarcoma given that the proportion of the population exposed to fluoride topically or systemically has grown dramatically since 1945 when the first community was fluoridated in the United States. If such a widespread exposure truly were associated with osteosarcoma, we would expect to see an increase in cases over time. Osteosarcoma has remained a rare disease with a relatively steady prevalence in the United States. The lack of any such relationship is consistent with our previous report showing that there was no difference in bone fluoride levels among case patients or control patients in that study.²⁵ It is unlikely that such a common exposure as fluoride, which is ubiquitous in the environment, toothpastes, and many beverages, would be related to an increased risk of developing osteosarcoma.

A finding of our analysis, which belies explanation, is the consistently shown protective effect of the highest income level, which in this study was defined as greater than $60,000 per year in the 1990s. Given that supplements must be prescribed by a dentist or physician, those receiving prescriptions are using the health care system and thus may be practicing healthier behaviors than nonusers. The protective effect of higher income may be a spurious finding; however, it deserves greater exploration in studies specifically designed to understand the economic and social determinants of health. Similarly, we found that bottled water use was protective for osteosarcoma. This was not a hypothesis of our study; thus, the study was not designed to detect an association with bottled water use, and detailed information of this variable was not collected. Given the variability of fluoride concentration in various brands of bottled water and the low prevalence of bottled water use in this study population, a firm conclusion cannot be drawn. Future studies could evaluate this association as the primary hypothesis.

Strengths of this study include the multicenter design ensuring a broad cross section of communities nationwide, confirmation of diagnosis, and comparable selection and enrollment procedures for case and control patients, as well as the use of trained interviewers with a standardized intake procedure.

Potential limitations of this study include the case-control study design, which may be prone to selection and recall bias. However, it seems unlikely that selection bias played a role in these findings given that case and control patients were from the same hospitals and clinics. Recall bias is possible as with any case-control study.

Any study with fluoride as the exposure is fraught with the challenge of discerning the amount of fluoride exposure. In our study, the exposures of topical fluorides or dietary supplements was ever or never, given that the quantification of fluoride exposure was challenging.

In our study, we report a secondary data analysis. The original study was not powered to examine topical fluoride or dietary fluoride supplements and osteosarcoma risk.

An additional limitation is whether these findings are generalizable.

CONCLUSIONS

There is no evidence that the use of dietary fluoride supplements or topical fluoride is associated with an increased risk of developing osteosarcoma.

Appendix

The National Osteosarcoma Etiology Group is represented by MA Simon (University of Chicago, Chicago, IL), MC Gebhardt (Massachusetts General Hospital, Boston, MA), MT Scarborough (Shands Medical Center, University of Florida, Gainesville, FL), S Gitelis (Rush Presbyterian and St. Luke’s Medical Center, Chicago IL), JJ Eckardt (UCLA School of Medicine, Los Angeles, CA), JR Neff (Nebraska Health System, Omaha, NE), MJ Joyce (Cleveland Clinic Foundation, Cleveland, OH), M McGuire (Creighton University, Omaha, NE), and HC Anderson (University of Kansas Medical Center, Kansas City, KS).