All the recent research anti-fluoride campaigners promote as “evidence” of harm from community water fluoridation amount to cherry-picking a very few statistically significant results from a large number of non-significant results. The whole exercise is a bit like the “Mouse that Roared.” Credit: The Mouse that Roared – TMTR Intro animation
Another one of those papers with weak or vague relationships with fluoride intake has just been released and is already being promoted by anti-fluoride activists. These activists probably don’t even read these papers but because they confirm their biases they will promote them anyway.
The evidence in this paper is very poor. A large number of sex steroid hormone measures were investigated. None were significantly related to fluoride in water and very few related to blood plasma F levels.
So this new paper is no better than those I reviewed in What are the recent fluoride-IQ studies really saying about community water fluoridation? (Incidentally, I have updated that article to include the recent paper I had discussed in Sleep disorders and fluoride: dredging data to confirm a bias and provided all the results from that study).
The new paper is:
Bai, R., Huang, Y., Wang, F., & Guo, J. (2020). Associations of fluoride exposure with sex steroid hormones among U.S. children and adolescents, NHANES 2013–2016. Environmental Pollution, 114003.
It uses data from a US database so is relevant to community water fluoridation (unlike most studies promoted by anti-fluoride campaigners which are from areas of endemic fluorosis). Basically, it looks for significant relationships of drinking water fluoride or blood plasma fluoride with three sex steroid hormones (testosterone, Estradiol and sex hormone-binding globulin -SHBG) in male children, male adolescents, female children and female adolescents. So a lot of measures and relationships from which to find significant ones.
The results they found are shown in the three figures below -which are figures 1, 2 and 3 from the paper. The methodology used appears quite confused. Instead of simply reporting regression results for the correlation of the hormone levels with water and plasma F they appear to have divided the fluoride data into tertiles. That is the set of one third lowest values, the set of middle third of the values and the set of one third highest values.
They then report differences between the mean hormone levels in the 2nd and first tertile and differences between the mean hormone levels in the third and first tertile. The p-values in the second to the last column indicates the statistical significance of these differences.
The last column reports a p-value for the “trend” – but this appears to be simply a correlation of the geometric mean hormone level against the median water-F or blood plasma-F for each tertile. That is only 3 sample pairs for each set considered. It is a mystery to me that they didn’t simply report linear regressions of all the values in the sample sets. Perhaps this is the only way they could find anything significant.
It would have been more helpful to present all the data graphically so readers could see how scattered it is. (The confidence intervals shown in the graphics below indicate a large amount of scatter).
There are very few statistically significant results – those shown in red. These were for male adolescents and the total sample where there was a decline in levels of testosterone and estradiol with increaser of blood plasma-F. The result foir the total sample probably reflects the male adolescent result as there was no significant difference between either the second or third tertile and the reference first tertile, or any significant decline, for any of the other groups (male children, female children and female adolescents). There were no significant effects with water-F for these hormones with any of the groups.
Ther only significant effects seen with SHBG was a decrease in the SHBG level for male adolescents when comparing the third tertile of water-F with the first tertile. And for female children a decrease in SHBG levels when comparing the values for the second tertile of blood plasma-F against the first tertile.
So all this is a bit like the mouse that roared. Despite claims in the conclusion that the public health applications of their finding “are substantial” I do not think there is anything here for public health experts to get concerned about.
My conclusions parallel those expressed for the cognitive studies in What are the recent fluoride-IQ studies really saying about community water fluoridation?.
This is the sort of picture one might expect from exploratory studies using several hormones for different population groups and several measures of fluoride exposure. While these results may be useful in suggesting possible hypotheses to check in future better-designed experiments they are not sufficiently coherent to inform social health policy.
My conclusions about the recent fluoride studies for levels relevant to community water fluoridation applies equally to this study:
“These recent studies do not provide sufficient evidence for revision of community water fluoridation policies because of possible effects on cognitive abilities. Anti-fluoride activists have only been able to use these studies in their scaremongering propaganda by cherry-picking results and ignoring the weakness of the relationships they cite.”