Quality and selection counts in fluoride research

By Ken Perrott 16/02/2014

I think the paper most quoted by anti-fluoridation activists must be Choi et al (2012), Developmental fluoride neurotoxicity: A systematic review and meta-analysis. I say quoted but, I suspect, not read. It is always coming up in articles on natural health web sites and continually thrown into blog and Facebook discussions. Often as a link without explanation.

It is also heavily promoted on Twitter. The same tweet is often sent from the same account daily, or more often. Here are examples.

Harvard Study: #Fluoride Lowers Children’s Intelligence By Seven IQ Points
Harvard Study Confirms Fluoride Reduces Children’s IQ huff.to/111fe5o via @HealthyLiving
Harvard Study Confirms Fluoride Reduces Children’s IQ: articles.mercola.com/sites/articles…

The paper also gets a lot of mention when anti-fluoride activists petition local bodies to prevent or stop fluoridation. It was one of the papers that had a big effect on the Hamilton City Council during the hearings they held last year (see When politicians and bureaucrats decide the science).

I recently reread the paper – this time paying special attention to the selection and quality of the papers reviewed. I think both of these are important to anyone attempting to understand the significance of this review.


Choi et al (2012) selected 27 papers for their review. Their selection was clearly not random – 25 of these are Chinese studies with 2 Iranian. Very few of these papers are directly available to western readers. One of the Iranian papers has an English abstract and 3 of the Chinese studies were published in English (guess which journal – you are right – Fluoride). 

The authors do not give any details of translation of the papers but 8 of them seem to have been translated under the auspices of Paul Connett’s Fluoride Alert (FAN) activist group (FAN describes them as the “FAN English translation”). Copyrights for these English translations are held by the International Society for Fluoride Research (ISFR) and included in their journal Fluoride. They are also available on FAN. (It is sort of difficult to locate the boundaries between FAN, ISFR and Fluoride. And did you know the ISFR has charity status in New Zealand. Yes, as taxpayers we are subsidsing them through their tax exemption!).

The authors acknowledge the reviewed studies were selected and give several reasons:

1: Studies from rural China had not been included in earlier reviews:

“We specifically targeted studies carried out in rural China that have not been widely disseminated, thus complementing the studies that have been included in previous reviews and risk assessment reports.”

2: High fluoride concentrations in drinking water are not common in the west:

“Opportunities for epidemiological studies depend on the existence of comparable population groups exposed to different levels of fluoride from drinking water. Such circumstances are difficult to find in many industrialized countries, because fluoride concentrations in community water are usually no higher than 1 mg/L, even when fluoride is added to water supplies as a public health measure to reduce tooth decay.”

Well, I can understand the logic behind that selection – provided readers don’t think they are seeing a balanced, representative review of all the existing literature. And the reasons given for this selection makes nonsense of Paul Connett’s charge that the lack of material in the industrialised countries indicates at least an unwillingness to research problems or at worst a conspiracy not to do the research and/or hide the results.


Here I will just take the size of the reviewed reports as a possible indicator of their quality. Not that I am against short papers, far from it. But in this cases most of the papers were very short – basically because they reported only a simple relationship found between IQ and fluoride in drinking water. Very few papers consider confounding factors like family education, schooling, breast-feeding, etc. Factors known to influence IQ. Nor did they discuss their results in any depth.

The authors acknowledge the brevity of the reports reviewed:

“In regard to developmental neurotoxicity, much information has in fact been published, although mainly as short reports in Chinese that have not been available to most expert committees.”


“Although most reports were fairly brief and complete information on covariates was not available, the results tended to support the potential for fluoride-mediated developmental neurotoxicity at relatively high levels of exposure in some studies.”

The histogram below gives an idea of the size of these reports – obviously some were extremely short –  19 of the 26 considered were of 3 pages or less!


Another reason for brevity is that most papers give hardly any discussion, let alone critical assessment, of the reported results. I wonder if this is because of the well-known problem of excessive levels of fluoride in many Chinese well waters and the associated incidence of dental and bone fluorosis. Perhaps this encourages researchers to simply consider fluoride as a factor in other problems when we might think it more rational to look at factors traditionally related to IQ. Like education and breastfeeding.

Even Paul Connett has conceded the poor quality of many of the studies considered in this review (while of course still scare-mongering that fluoridation is somehow going to make us all dumb). The authors of the study itself warned their paper was not relevant to the fluoridation issue (see Harvard scientists: Data on fluoride, IQ not applicable in U.S):

“Two of the scientists who compiled the Harvard study on fluoride said it really doesn’t address the safety of fluoridation levels typical of American drinking water.

“These results do not allow us to make any judgment regarding possible levels of risk at levels of exposure typical for water fluoridation in the U.S.,” the researchers said in an e-mail response to questions from The Eagle. “On the other hand, neither can it be concluded that no risk is present.”

The researchers noted that the fluoride levels they studied were much higher than what is found in fluoridated water in the United States and recommended “further research to clarify what role fluoride exposure levels may play in possible adverse effects on brain development, so that future risk assessments can properly take into regard this possible hazard.””

I guess this paper, and its continual promotion, must impress many anti-fluoridation activists and even some local body councillers and staff. But scientists and other experts familair with the subject are not so impressed.  The NZ National Fluoridation Information service last year reviewed  literature on possible effects of fluoridation on IQ (see A review of recent literature on potential effects of CWF programmes on Neurological ):

“The available evidence raises the possibility that high levels of fluoride in drinking water may have subtle effects on children’s IQ. However all of these studies have limitations in design and analysis, a clear dose-response relationship between DWFCs and assessed IQ are often not evident. The study authors are frequently very cautious in their comments, and several noted that any indicated negative effect applied only to high DWFCs. An hypothesis of fluoride neurotoxicity would also be supported by some experimental animal studies, however the great majority of these have only considered high fluoride intakes.
However collectively the data described are not robust enough to draw a firm conclusion that high fluoride levels in drinking water supplies contribute to retarded development of children’s brains. Also there is no clear evidence to suggest an adverse effect on IQ at lower fluoride intakes such as that likely to occur in New Zealand, where fluoridated water supplies contain fluoride in the 0.7 to 1.0 mg/L range.”

Since then a local New Zealand study has failed to find any relationship between fluoridation and IQ (see Dunedin fluoride-IQ study finds no ill-effect). The study did find a positive influence of education and breast feeding on IQ though – just as we would expect.

In a typical sour grapes comment Paul Connett, who was told of this research by a reporter, quipped “rather convenient.”

Confirmation bias in action! I guess he won’t be promoting the New Zealand research in the way he does the poor quality research in the Choi et al (2012) review.

Similar articles

0 Responses to “Quality and selection counts in fluoride research”

  • Lack of sufficient fluoride in the drinking water of anti-flouradation campaigners results in mental illness due to traumatic decay of tooth enamel, other wise known as “bear with a sore tooth syndrome”. Note: extensive and reputable research provides evidence of clinically significant reduction in brain weight due to trauma and chronic pain. (see the Oxford Pain Group)