Severe dental fluorosis the real cause of IQ deficits?

A new study finds cognitive function defects, like IQ, in children are not significantly related to fluoride in drinking water. But they are associated with medium and severe dental fluorosis.

This interests me for two reasons:

  1. The report is by Choi and Grandjean who had also authored the 2012 meta-review often used by anti-fluoride activists to claim that community water fluoridation causes a lowering of IQ (the authors subsequently pointed out the high fluoride concentrations in the papers they reviewed meant that conclusion is not valid)
  2. The data reported is consistent with my suggestion in Confirmation blindness on the fluoride-IQ issue that reported relationships between IQ and drinking water fluoride concentration could really indicate a relationship with severe dental fluorosis, and not drinking water fluoride itself.

The new report is:

Choi, A. L., Zhang, Y., Sun, G., Bellinger, D., Wang, K., Yang, X. J., … Grandjean, P. (2014). Association of lifetime exposure to fluoride and cognitive functions in Chinese children: A pilot study. Neurotoxicology and Teratology.

Firstly – this is only a pilot study and has several acknowledged weaknesses – the small number of children studied (51) being the most obvious. However, this is what was found:

“Sixty percent of the subjects examined had moderate or severe fluorosis. These children were exposed to elevated fluoride concentrations in drinking water. Children with normal or questionable Dean Index were all from households with a water fluoride concentration of 1 mg/L and had urinary fluoride excretion levels below 1 mg/L.”

The children were placed in 3 groups according to their degree of dental fluorosis:

  1. Normal/questionable (N=8)
  2. Very mild/mild (N=9)
  3. Moderate/severe (N=26)

The high proportion of children with moderate/sever dental fluorosis indicates the study involved an area of endemic fluorosis.

And the results of neuropsychological tests:

“Results of multiple regression models show that moderate and severe fluorosis was significantly associated with lower total and backward digit span scores when compared to the reference combined categories of normal and questionable fluorosis (Table 4). Although the associations between fluoride in urine and in drinking water with digit span were not significant, they were in the anticipated direction. Motor coordination and dexterity were not significantly associated with fluoride in drinking water and fluorosis although higher levels were associated with poorer scores as well. Other outcomes did not reveal any association with the fluoride exposure.”

The authors used a number of neuropsychological tests. The digit span test results suggest a “deficit in working memory” for the children with moderate and severe dental fluorosis. None of the other tests used show any signficant relationship with indices for fluoride exposure.

So, this pilot study did not show any association of neuropsychological tests with fluoride concentration in drinking water but it did find an association with medium and severe dental fluorosis. This is consistent with my speculation in Confirmation blindness on the fluoride-IQ issue that “a physical defect like dental and skeletal fluorosis could lead to decreasing IQ.”

I argued that:

“minor physical anomalies are known to be associated with learning difficulties and emotional illness in children (seeHilsheimer & Kurko 1979). It seems entirely reasonable that a physical anomaly like severe dental fluorosis could lead to learning difficulties in children which could be seen as lower IQ values.”

There are many problems with the studies anti-fluoride activists promote relating IQ to fluoride in drinking water. But it could be that any real effect seen with the higher fluoride concentrations could simply be explained by effects of the physical anomaly of medium and severe dental fluorosis common at these higher concentrations.

Unfortunately the authors of this study still do not consider this possibility. I guess it could be that someone with a hammer only sees nails, and chemical toxicologists are only capable of considering brain damage caused by toxic chemicals. The effects of physical anomalies on learning difficulties are probably quite outside their training and experience.

Their confirmation bias and mental blockage on this meant they were considering dental fluorosis as just another indicator of dietary fluoride intake. However, even that assumption has its problems because genetic differences are also known to be involved in dental fluorosis.

I think this must be why they ended with a conclusion that could well be quite unfounded:

“This pilot study in a community with stable lifetime fluoride exposures supports the notion that fluoride in drinking water may produce developmental neurotoxicity”

Dental fluorosis and community water fluoridation

Fluorosis is endemic in many parts of China and the high prevalence of medium/severe dental fluorosis (60%) among the children in the Choi et al (2014) pilot study shows their situation is not at all similar to that in areas of New Zealand and USA using community water fluoridation (CWF).

The figures below give some context.

Here are examples of the different degrees of dental fluorosis.

The graph below shows the situation reported for New Zealand in the 2009 New Zealand Oral health Survey (see Our Oral Health). No severe and only 2% moderate dental fluorosis reported.

This figure (taken from Fluorosis Facts: A Guide for Health Professionals) shows the amount  of moderate and severe dental fluorosis in the US is also very small.

Perhaps we can now contrast the situation here, in areas where CWF is common, with the situation in China in areas with endemic fluorosis where these studies were undertaken. The figure below is a slide from a presentation by Xiang (2014) to Paul Connett’s recent anti-fluoride “get-together” (Xiang 2014). This is not the very mild dental fluorosis attributed to CWF.

(Anti-fluoride people also often single out the study of Xiang, et al (2003). Effect of fluoride in drinking water on children’s intelligence. Fluoride, 36(2), 84–94, because unlike the others it is more detailed.  Xiang’s team has studied areas where fluorosis is endemic.)


Anti-fluoride activists often promote the meta-review of Choi et al (2012) in their arguments against CWF. However, there are many problems with these studies including the fact reported IQ effects were associated with much higher drinking water fluoride concentrations than occurs with CWF.

The new study of Choi et al (2014) did not confirm any association of neuropsychiatric measurements with drinking water fluoride concentration. However, it did show association of negative neuropsychological effects with medium/severe dental fluorosis.

This is consistent with the physical anomaly of severe dental fluorosis being the real cause of IQ effects and not any direct chemical toxic effect.

Similar articles

16 responses to “Severe dental fluorosis the real cause of IQ deficits?

  1. Whenever genetic differences are associated with an affliction it is interesting to consider the basis and whether the individuals should be penalised, in this instance by the cost of avoiding fluoride burden that others may cope with.

    Genes exist in pairs. Sometimes one of the pair is wrong but the result may be of benefit.

    The classic example of genetic defect as a benefit is that of sickle cell anaemia carrier with one copy of the disease providing a survival benefit in a malaria-afflicted society.

    A Tay Sachs carrier may be more able to survive the diarrhoea of cholera.

    It needs to be considered what a carrier of genetic propensity to fluorosis may also be carrying in a beneficial capacity. Do they have better teeth if they do not get fluorosis? Note that administering fluoride through the water is only claimed to offer some improvement in tooth health in a population.

    Ken, is it your hypothesis that the same fault which may cause fluorosis may also cause IQ deficit? I seems to me a bit like saying that in a city where salt is put on the road to disperse salt, and where cars get weak from rust, that some cars may be more rust-prone which is the real cause of the weakness.


  2. No it is not my “hypothesis that the same fault which may cause fluorosis may also cause IQ deficit?” Not at all. Quite the opposite in fact.

    It is known that a physical disfigurement like severe tooth decay or severe dental fluorosis can cause learning difficulties. These could be reflected in IQ measurements and similar.

    Please note this research has nothing to do with “administering fluoride through the water.” It is a study of an area where endemic fluorosis occurs/ There is no addition of F to the water supply – the F in the drinking water is natural.

    My suggestion is that the results they found are consistent with problems caused by physical anomalies, not by F in the drinking water. They found an association with severe dental fluorosis, not with F in the drinking water or urine.


  3. Sorry salt to disperse ice.

    And I correct cystic fibrosis carrier may be the cholera-resistant one, with Tay Sachs carrier having some immunity to TB.

    I think there is need for more understanding about balanced polymorphism by persons controlling our food supplies.

    Another example:
    “The heterozygote advantage of the Chuvash polycythemia VHLR200W mutation may be protection against anemia”

    I think understanding needs to go beyond what is related to what and further to what deeper mechanism organises this benefit to heterozygotes. I may touch a sore point when I say I believe genetic engineers of many leanings do not understand such matters and what effects they may be having.


  4. Ken your chart about NZ fluorosis. The table you have taken it from appears to wrongly average out for severe fluorosis for fluoridated areas a range of from 0 to 1.5 as 0. Likewise for non-fluoridated a range of from 0 to 1.8 as 0. And I am not sure why they include “Questionable fluorosis” where they are not sure if it exists. Leaving out that questionable column, fluorosis from very mild to severe is always greater in the non-fluoridated areas. At mild level it is over double in the non-fluoridated areas to what it is in the fluoridated. I think that data needs some comment before it may be used as evidence.


  5. Check the table (Table 92 on page 172). The numbers you quote are the 95% CIs – not ranges.

    “Questionable” is one of the 6 grades (1 in 0-5) in the Dean Index.

    The publication makes the following comment on Table:

    “Table 92 presents the prevalence of fluorosis (by the six categories of Dean’s Index of Fluorosis), among dentate children and adults aged 8–30 years, overall and by fluoridated and non-fluoridated areas. Overall, the prevalence of moderate and severe fluorosis was very low in the population, with 2.0% of people aged 8–30 years with moderate fluorosis and virtually no people with severe fluorosis (0.0%).

    These results suggest there was no significant difference in the prevalence of fluorosis between people living in fluoridated and non-fluoridated areas. “



  6. Some other relevant comments from the survey (page 171):

    “Many studies on fluorosis confirm that, in optimally fluoridated areas, dental fluorosis is usually only mild or very mild. A recent review concluded that mild fluorosis was not a concern for people, and that mild fluorosis was sometimes found to be associated with improved oral health-related quality of life. Severe fluorosis was consistently reported to have negative effects on oral health-related quality of life (Chankanka et al 2010).

    This section presents the prevalence of fluorosis among dentate children and adults aged 8–30 years living in fluoridated areas and those living in non-fluoridated areas. Unadjusted results are presented, because the numbers of respondents with moderate or severe fluorosis were too low to adjust for other variables.”



  7. Ken wrote:
    “Check the table (Table 92 on page 172). The numbers you quote are the 95% CIs – not ranges.”

    They are the intervals between which the values are expected to fall with 95% confidence.

    They examined 1324 people aged 5-34 so somewhere in the vicinity of 1000 for their fluorosis study ages 8-30. I can’t see their sample size for fluorosis.

    For severe fluorosis their confidence level is 0.0-1.5 fluoridated and 0.0-1.8 non-fluoridated. Wouldn’t that mean they are 95% confident the number is about 0-15 people (F) and 0-18 (NF)? What justification is there for calling that 0.0?

    For mild fluorosis they give 0.8-7.6 (F) and 4.3-12.7 (NF)
    So they must be 95% confident that at least 8 people from the F areas and 43 from the NF areas have mild fluorosis, if the sample size were 1000.

    The significance of this is that if your hypothesis is correct then the NF areas should have lower IQ.

    >“Questionable” is one of the 6 grades (1 in 0-5) in the Dean Index.

    Judging from your photos I feel there is a wide margin for error in distinguishing between between normal and questionable.

    >The publication makes the following comment on Table:

    “Table 92 presents the prevalence of fluorosis (by the six categories of Dean’s Index of Fluorosis), among dentate children and adults aged 8–30 years, overall and by fluoridated and non-fluoridated areas. Overall, the prevalence of moderate and severe fluorosis was very low in the population, with 2.0% of people aged 8–30 years with moderate fluorosis and virtually no people with severe fluorosis (0.0%).”

    Which I have addressed above.

    ” These results suggest there was no significant difference in the prevalence of fluorosis between people living in fluoridated and non-fluoridated areas. “

    I don’t think so.


  8. Sound hill, perhaps you should read up a statistical text on the meaning of statistical significance.

    My “hypothesis” does not say “the NF areas should have lower IQ” at all. I suggest you read my article more carefully. It is not relating IQ to fluoridated or non-fluoridated areas. It is specifically about a study of an area in China suffering endemic fluorosis – something we don’t have in NZ.

    My graphs for NZ and the USA were included purely to show the completely different situation in our countries where severe and medium dental fluorosis is very rare whereas 60% of the children in the Choi et al study suffered from it.



  9. A suggestion is to test IQ before many teeth have erupted.

    “Significant figures”
    0.0 giving a co-ordinate of a point on a graph would indicate 1 significant figure accuracy or between -0.95 and +0.04 usually. That is 1 significant figure accuracy.
    Since we can’t have a negative here it would mean up to 0.04, not 1.5 or 1.8.


  10. Sorry not -0.95, rather -0.04, not that it makes any difference here.

    And I notice from the abstract, “Dental fluorosis score was the exposure indicator that had the strongest association with the outcome deficits,…”
    So thinking in terms of my rusty cars and salt on the road, rust-proneness and salt on the road both cause rusty cars. There may be a stronger association to one than the other. That does not mean no effect of the other.


  11. I don’t know what is worrying you about this soundhill. The authors found a significant association only with dental fluorosis grades. This of course does not mean that fluoride intake has no association – especially as we know that the severe dental fluorosis results from excess F intake.

    The real question is whether one should make an assumption of chemical neurotoxicity when there is no evidence for this at low F intake.

    It seems to me at least as credible, if not more so given the associations which have been reported between physically anomalies and learning difficulties, that the results have nothing to do with neurotoxicity. It could be that the association with F intake is indirect.


  12. Ken you acknowledged the small sample:
    Normal/questionable (N=8)
    Very mild/mild (N=9)
    Moderate/severe (N=26)
    What comment did they give to the urine test? No fluoridated water the night before, so presumably higher urine fluoride would be looking at a person who was slower to excrete.
    They said to that and the water level that the result was trending though not significant?

    It would be interesting if the multiple regression could extract effects of fluoride on in-utero or infancy of brain development. Those with fluorosis will have had more or been less able to throw it off.

    I suggested looking at IQ between NF and F areas in NZ. Can we get from the NZ data where those severe fluorosis people were? Was there a DHB pushing fluoride tabs & toothpaste &c. Then look at IQ there, on top of natural fluoride in water which may vary a bit with previous volcanic activity.


  13. “associations which have been reported between physically anomalies and learning difficulties,”

    Physical anomalies in a family, if inherited may show up as a fistula in one, and maybe in another something that cannot be seen, maybe even in the brain.

    We have two kidneys, and may live with one, but our backup is then gone, if the second one is hurt. It is my hypothesis that the brain can rewire itself up to a point, with the same sort of backup, but then under further stress it would not cope where a brain which has not already used its backup would cope.

    We have talked about tests to forewarn if a particular individual may not cope with a drug or vaccination. Iatrogenic disease can be very severe, as can effects of recreational drugs on some persons.

    Maybe as the world advances we shall research genetic or other predisposition to susceptibilty to fluoride, and so avoid the imposition on some individuals. A well-functioning brain may help EQ as well as IQ.

    Just watched the Nutters Club on Maori TV.


  14. Ken,

    Have I got this right?

    The authors have already said that a conclusion that CWF lowers IQ is not valid.

    They say that severe dental fluorosis may be a marker for lowered IQ – but small study, etc.

    We already know that children with physical anomalies tend to have learning difficulties.

    You suggest that the physical anomaly of severe dental fluorosis may be contributory to learning difficulties, and have contributed to the study findings.

    I think that’s a good hypothesis, easily testable, and a good example of keeping your mind open to the interpretation of data. (Unfortunately, it’s not testable in NZ, because of the almost nonexistent incidence of severe dental fluorosis)


  15. They previously said in a press release commenting on their 2012 meta review that the results were not applicable to the situation with CWF in the US. However, they bet both ways by saying that the results don’t indicate CWF lowers IQ, but then again they don’t indicate that it doesn’t.

    The only significant association they found in their pilot study is with severe/medium dental fluorosis. While no significant association was seen with F in urine or drinking water one would expect that a larger study might show such an association.

    However, if the results, in the end, are due to learning difficulties resulting from the physical anomaly presented by severe dental fluorosis (which could be a reasonable explanation) one should expect that a significant correlation with drinking water F could be found – if only because drinking water F would be one of the major contributors to severe dental fluorosis.

    The annoying thing is they seem to be blind to another possible mechanism like this. They interpret the result as evidence that F is a neurotoxin. As I said, someone with a hammer only sees nails. A chemical toxicologist only see toxic chemicals.

    In the absence of any data specifically supporting a neurotoxin mechanism these authors should have included other mechanisms in their speculation.

    I have found some of the literature relating learning difficulties to physical anomalies but really should search out more. I am sure it is there – although it may consider the anomaly of severe tooth decay rather than dental fluorosis when considering oral health.



  16. Ken,



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 )

Connecting to %s