Does community water fluoridation reduce diabetes prevalence?


Maybe community water fluoridation reduces the prevalence of diabetes?

You will be seeing headlines claiming a link between community water fluoridation (CWF) and diabetes. Or even that fluoridation can predict an increase in the prevalence of diabetes. But they are misleading

These articles report results from a recently published study indicating that in the majority of situations CWF is linked to a decrease in diabetes prevalence. But many of these articles, and especially those from anti-fluoride activists are making opposite claims.

Why the confusion? Well, the study used modelling to relate a number of factors to the prevalence of diabetes. According to the model’s prediction CWF using fluorosilicic acid and sodium fluorosilicate is related to a decrease in diabetes prevalence. However, the saving clause for anti-fluoride activists is that the model predicts an increase in diabetes prevalence when the least common fluoridation chemical, sodium fluoride, is used.

A 1992 survey found that only 9% of the US population received water fluoridated with sodium fluoride – compared with 63% for fluorosilicic acid and 28% for sodium fluorosilicate. I got the latest figures from a fluoridation engineer at the US Center for Disease Control. The current figures are 75% for fluorosilicic acid, 13% for sodium fluorosilicate and 7% for sodium fluoride.

In New Zealand only on water treatment plant for a small community uses sodium fluoride.

So this subheading by the Fluoride Action Network (FAN) is completely  misleading – “Regression analyses suggest association between increases in consumption of fluoridated water and type 2 diabetes.” The only way anti-fluoride propagandists can make mileage out of this study is by deliberately ignoring the results indicated for over 90% of the population!

Perhaps supporters of CWF should be the ones reporting and promoting this study – arguing that CWF could reduce diabetes prevalence! However, I would not push that idea on the basis of a single report. This study has a number of deficiencies – and recommendations should not be based on individual cherry-picked studies anyway.

This is the paper reporting the study:

Fluegge, K. (2016). Community water fluoridation predicts increase in age-adjusted incidence and prevalence of diabetes in 22 states from 2005 and 2010. Journal of Water and Health.

Here are some of the problems I see with it.

Insufficient consideration of confounders

It is a modelling study looking for correlations between selected parameters. Such studies often suffer from little or no consideration of important confounders. Statistically significant correlations can disappear when such confounders are later included. For example, consider my criticism of the Malin and Till (2015) ADHD study – see ADHD linked to elevation not fluoridation and ADHD link to fluoridation claim undermined again.

Fluegge included obesity prevalence and leisure time physical inactivity as confounders but more could have been considered.  One that sticks out like a sore thumb to me is the community size. It could be that the sodium fluoride data he used could be acting as a proxy for community size as these days sodium fluoride is usually only considered for small water treatment plants.

Adjustment of fluoride exposure data

Fluegge compared his model prediction for diabetes prevalence using two different measurements of fluoride exposure – drinking water fluoride concentration (ppm) and an adjusted estimate of fluoride intake (mg/day). His estimation was made from per capita domestic water deliveries per county. I find this questionable as the proportion of water consumed will vary by location where there are different requirements for things like lawn and garden watering, car washing, swimming pools etc.

Whereas the drinking water fluoride concentration showed a negative correlation with diabetes prevalence (the prevalence decreased with increasing fluoride concentration), the adjusted exposure values showed a positive correlation (the prevalence increased with increasing fluoride concentration). He declared the second correlation more “robust” but his reasons seem more related to confirmation bias than any proper analysis.

Confused discussion

Fluegge seems completely unaware that sodium fluoride is now only rarely used as a fluoridating chemical. He even suggests a possible policy outcome of his research could be switching from sodium fluoride to fluorosilicic acid!

He refers to Hirzy et al. (2013) claiming it showed cost savings from using sodium fluoride but critiques Hirzy for not including consideration of effects on diabetes prevalence. He seems completely unaware that Hirzy’s paper was discredited and he had to withdraw its claims about cost savings.

This suggests to me that Fluegge is not familiar with fluoridation research. In fact, his very brief publication history indicates his interest is more associated with cherry-picking various health measures to find fault with by using statistics and modelling.

How reliable is the modelling?

I have drawn attention to possible problems with poor selection of confounders and lack of familiarity with the fluoridation literature. But there may also be problems with the modelling methods used.

I do not have the modelling skills or time to delve into his model in any depth but note there has been some controversy about another modelling paper he was involved in.

He co-authored a paper with his brother claiming a link between glyphosate and ADHD. This created some controversy because it was rejected by the journal and then published by mistake. So the journal had to retract the paper. You can read about it at Retraction Watch – A mess: PLOS mistakenly publishes rejected ADHD-herbicide paper, retracts it.

The paper was rejected because it did not satisfy the standards of experimental and statistical analysis required, or describe these in enough detail. Also because the conclusions were not presented in an proper way or supported by the data.

OK, we should not discredit future work because an earlier paper was rejected, even for the given reasons. Authors can learn from their mistakes. But it does ring warning bells. With this history, I would prefer a deeper critique of the methods used and the reliability of his conclusions.

These questions just underline my warning that one should never base policies, or final interpretations, on single papers – especially cherry-picked ones. Conclusions should be based on a more complete reading of the scientific literature.


So, always take headlines with a grain of salt. In this case they will be completely misleading – especially if promoted by anti-fluoride activists.

Similar articles



11 responses to “Does community water fluoridation reduce diabetes prevalence?

  1. All very interesting if you were thinking of trying a dose of fluoride . . . but totally irrelevant to the issue of artificial fluoridation of public water supplies. No one has the right to medicate people against their wishes . . . end of argument!


  2. Ken you ought to take another look at the Malin and Till study.
    I do not think it to be correct to use just CWF figures, but better to use actual fluoride levels in drinking water.

    Fluoride levels were allowed to be up to 4 mg/litre and to combine such a location with the non-CWF States would be wrong.

    As for diabetes the measured symptom of that is blood sugar. Silica may reduce blood sugar and some even warn about that, that it may reduce it too much if certain herbs are taken.

    If sodium fluoride be used to fluoridate water that would not add the silica which fluorosilicic acid would. So it is not really valid to ascribe any reduced diabetes to just the fluoride component of fluoridation.


  3. Brian, why don’t you write to Malin and Till and suggest this. It is not my paper, I have no responsibility to correct it.

    You might also bring their attention to my critiques of their paper at the same time.


  4. If this correlation was indicative of causation (and the two are not necessarily related), then I’d expect to see a much higher rate of diabetes in those parts of the world where the natural levels of F- are well in excess of that produced by CWF. Parts of India & China, for example. Surely a thorough review of the data would have included that as a comparator?


  5. David Fierstien

    Wonder Mouse: ” No one has the right to medicate people against their wishes . . . end of argument!”

    Medicate? By your logic you are medicating yourself when you brush your teeth.

    The fact is you use the word “medicate” in an attempt to drum up paranoia. Why is that? What exactly is your agenda? Do you sell bottled water? . . water filters? . . or some kind of homeopathic “medicine?” I have found that fanatics like you usually have some financial agenda driving them.

    But let’s say – for the sake of argument – that optimally fluoridated water is medicine. You are literally arguing for the freedom to have the choice to have poorer oral health. That makes you irrational. If you’re in it for the money, you would be completely rational.

    So, which is it? Are you a shill who tries to sell more paranoia to the world? . . . or are you irrational?


  6. Thanks Alison for stating the obvious (which I missed) but the result is also is also surprising given the 50 years of successful us of CWF with no obvious problems. Given the scrutiny any problems would have shown clearly by now. My thought is: most likely a spurious result, wait for confirmation before even beginning to change recommendations.


  7. [I believe the herbicide is known as glyphosate not glycophosphate]


  8. Thanks Richard – I’ll correct that. I knew I should have checked.


  9. The problem is that this paper is somewhat of a dog’s breakfast – probably because of the author’s motives.

    His model actually produced a negative correlation of fluoride with diabetes prevalence – indicating a reduction of prevalence with an increase ion fluoride. This was true for both “natural” fluoride and added fluoride when using the fluoride conbentration. When he used the “adjusted” fluoride intake values the correlation with natural fluoride was still negative but was positive for added fluoride.

    So he would probably say that in regions of endemic fluorosis because of high fluoride intake the diabetes prevalence would be lowered!

    The other strange thing is that his adjusted model suggests added fluoride increases prevalence – but when fluorosilicic acid is used it decreases it. He targets in on sodium fluoride as the “bad” fluoride and yet only 7% of the people receiving fluoridated water get it from plants using NaF!

    As I said – a dog’s breakfast. But he will move on to the next health “scare” – he has no real interest in fluoridation.

    I really don’t place much value on his model at all.


  10. Fluegge’s paper seems to provide some of the more illogical conclusions of recent studies that purport to find harmful effects of community water fluoridation. The findings are so incongruous that the Case Western Reserve University apparently couldn’t even produce an accurate news release.

    I read the news release before reading Fluegge’s paper and came to the conclusion that Ken had misinterpreted the study because he listed both fluorosilicic acid and sodium fluorsilicate as additives that were protective against diabetes.

    After all, the CWRU news release reported that sodium fluorosilicate along with sodium fluoride were both additives that increased the risk of diabetes. The CWRU news release title and sub title also stated that fluoride consumption was linked to an increase in type 2 diabetes with no qualifications as to the type of fluoride additive.

    However, as far as I could determine after reading Fluegge’s paper, sodium fluorosilicate was never linked to an increase in diabetes as reported by the CWRU news release.

    Unfortunately, every report of the study in the news articles I read has quoted the CWRU news release, used the misleading title, stated that both sodium fluoride and sodium fluorosilicate were linked to an increase in the incidence of diabetes and often ignored the conclusion that “fluorosilicic acid was significantly and robustly associated with decreases in incidence and prevalence of diabetes”.

    And, of course, fluoridation opponents have completely misinterpreted whatever conclusions there were to try and “prove” yet again the dangers of community water fluoridation.
    Details from: the Case Western Reserve University news release:
    Title: Fluoride consumption linked to diabetes using mathematical models
    Subtitle: Regression analyses suggest association between increases in consumption of fluoridated water and type 2 diabetes

    Quote: “Digging deeper revealed differences between the types of fluoride additives used by each region. The additives linked to diabetes in the analyses included sodium fluoride and sodium fluorosilicate. Fluorosilicic acid seemed to have an opposing effect and was associated with decreases in diabetes incidence and prevalence. Counties that relied on naturally occurring fluoride in their water and did not supplement with fluoride additives also had lower diabetes rates.”
    Details from: Fluegge’s paper:
    From the abstract: “The findings suggest that a 1 mg increase in the county mean added fluoride [no mention of which additives] significantly positively predicts a 0.23 per 1,000 person increase in age-adjusted diabetes incidence (P<0.001) and a 0.17% increase in age-adjusted diabetes prevalence percent (P<0.001), while natural fluoride concentration is significantly protective. For counties using fluorosilicic acid as the chemical additive, both outcomes were lower: by 0.45 per 1,000 persons (P<0.001) and 0.33% (P<0.001), respectively.”

    From the discussion: “Among the three fluoridation chemicals used in this data set (sodium fluoride, fluorosilicic acid, or sodium fluorosilicate), only fluorosilicic acid was significantly and robustly associated with decreases in incidence and prevalence of diabetes.”

    From the results: “Sodium fluoride produced significantly positive associations with incidence (β= 0.93, P< 0.001) and prevalence (β= 0.76, P< 0.001), whereas fluorosilicic acid and sodium fluorosilicate produced significantly negative associations respectively (fluorosilicic acid: β= –0.72, P< 0.001 and β= –0.54, P= 0.002; sodium fluorosilicate: β= – 0.55, P= 0.05 and β= –0.49, P= 0.02). Thus the comparisons are all relative. The protective effects of fluorosilicic acid and/or sodium fluorosilicate in the multivariable GEE models are, alternatively stated, a deleterious consequence of sodium fluoride use.”

    Liked by 1 person

  11. Thank you Ken for another insightful dissection of the latest piece being used around the world to convince citizens and leaders to discontinue or avoid fluoridation.

    For me the two most important criticisms are:

    1. The title misleads as to the actual findings. Notwithstanding methodological criticisms, Fluegge found that about 90% of the US fluoridation programs are associated with fewer cases of diabetes. The title misleadingly implies the opposite.

    2. Because of complete hydrolysis/dissociation there is no conceivable biological mechanism for water additives to have differential effect at consumers’s taps. Any findings must therefore be due to confounders, errors in methodology, calculation or non-representative data selection.


Leave a Reply: please be polite to other commenters & no ad hominems.

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s