Room for cherry-picking to confirm a bias. Separate Swedish studies report that fluoride can either prevent or promote the risk of hip fracture in the elderly. Image credit: Are hip fracture patients treated quickly enough?
Anti-fluoride activists are promoting a recent study linking fluoride intake and bone fractures. No surprise there. But they are cherry-picking a single study to support their agenda and scientifically literate people should see the wider picture and not ignore other studies which, on the whole, convey a different story. This issue illustrates problems with epidemiological studies producing variable results and shows why people should avoid cherry-picking and look at the full range of studies in a field.
Here I consider just two studies on fluoride intake and bone fracture which produced different conclusions. Both studies involved people from Sweden where the natural fluoride levels in drinking water vary across the country.
Drinking water fluoride may protect against hip fractures
First a study from 2013:
Näsman, P., Ekstrand, J., Granath, F., Ekbom, A., & Fored, C. M. (2013). Estimated drinking water fluoride exposure and risk of hip fracture: A cohort study. Journal of Dental Research, 92(11), 1029–1034.
The main findings are illustrated in the figure showing the calculated Hazard Ratios for people of different ages living in areas of Sweden with “very low” (less than 0.3 mg/L), “low” (0.3 – 0.69 mg/L), “medium” (0.7 -1.49 mg/L) or “High” (greater than 1.5 mg/L) fluoride levels in the drinking water. The Hazard Ratio in the figure below is a measure of the number of hip fractures at these levels compared with the number of hip fractures at “Very low” fluoride concentration. The bars represent the 95% confidence intervals. The Hazard Ratios for the “very low” group are 1.0 and Hazard Ratios statistically significantly different to 1 (no effect) are coloured red.

Considering all people there is no statistically significant increase in the number of hip fractures for any level of water fluoride concentration compared with the “very low” levels. The number of hip fractures experienced by people in the two lower age groups (less than 70 years and 70 – 80 years) was significantly lower at higher water fluoride concentrations than at the “very low” concentrations. The authors say:
this “suggests a protective effect of fluoride among the younger (age younger than 80 years): however, the majority of fractures occurred above the age of 80 years (median age at time of fracture, 82.0).”
So a study suggested that the fluoride in Swedish drinking water does not encourage bone fractures and may actually protect against them in the lower age groups.
Fluoride may encourage hip fractures
Now a study from 2021 – the one anti-fluoride activists are promoting (for obvious reasons):
Helte, E., Vargas, C. D., Kippler, M., Wolk, A., Michaëlsson, K., & Åkesson, A. (2021). Fluoride in Drinking Water , Diet , and Urine in Relation to Bone Mineral Density and Fracture Incidence in Postmenopausal Women. Environmental Health Perspectives, 129(April).
Unlike Näsman et al (2013) which used drinking water fluoride concentrations as a measure of fluoride exposure, Helte et al (2021) used urinary fluoride and estimated dietary fluoride intake as measures of fluoride exposure. The Hazard Ratios were calculated from the number of hip fractures in the Tertile 2 groups (0.88 – 1.30 mg/g urinary fluoride or 1.74 – 2.41 mg/day dietary fluoride intake) and Tertile 3 groups (1.30 – 116.51 mg/g urinary fluoride or 2.41 – 11.16 mg/day dietary fluoride intake) compared with hip fractures in the tertile 1 groups (0.14 – 0.88 mg/g urinary fluoride or 0.26 – 1.74 mg/day dietary fluoride intake).
Note: The urinary fluoride units of mg/g represent mg of urinary F/g urinary creatinine. Creatinine levels were used to correct the spot values for dilution.
The Hazard Ratios that statistically significantly different to 1 (no effect) are coloured red in the figure below.

A bit complicated I know, but what the figure shows is no statistically significant increase in hip fracture numbers for the tertile 2 groups compared with the lower F intake tertile one group. But a significant increase in fracture numbers for the tertile 3 groups except for the women exposed to constant water fluoride concentrations since 1982 in the dietary F group.
Hertle et al (2021) also considered other types of fracture. There were no statistically significant increases in fractures in the upper tertiles for either the “all fractures” or “major osteoporotic fractures” classes.
So, a bit of a mixed bag but this paper is currently being promoted by anti-fluoride activists as evidence of a harmful result from community water fluoridation (CWF).
Critically assessing the evidence for bone fractures
It is easy to see why supporters of CWF may cite Näsman et al (2013) as evidence for lack of harm and opponents may cite Helte et al (2021) as evidence of harm from CWF. But neither approach is really scientific. The methodological differences and choice of factors considered can easily explain variable results. One should critically and rationally assess both of these papers, together with the many other papers reporting similar studies, before coming to any conclusion.
On balance, the published studies probably support the findings of Näsman et al (2013) and not Helte et al (2021). In fact, a systematic review and meta-analysis published in 2015 concluded that chronic exposure to fluoride in drinking water was not associated with a significant increase in hip fracture risk. The citation for this review is:
Yin, X.-H., Huang, G.-L., Lin, D.-R., Wan, C.-C., Wang, Y.-D., Song, J.-K., & Xu, P. (2015). Exposure to Fluoride in Drinking Water and Hip Fracture Risk: A Meta-Analysis of Observational Studies. PLOS ONE, 10(5), e0126488.
It’s worth reproducing one of the figures from that review because it illustrates how epidemiological studies may, individually, support a claim of harm but when considered as a whole these studies do not support the claim. The figure below shows the range of Hazard Ratios obtained by a number of studies.
The lesson here is to be very careful of claims made on the basis of single cherry-picked studies. Especially when those making the claim have a bias they wish to confirm. Every claim should be critically and rationally considered using all the available studies.