January 6, 2025

expert reaction to systematic review and meta-analysis of fluoride exposure and children’s IQ scores

A study published in JAMA Pediatrics looks at the link between fluoride exposure and children’s IQ scores.

Prof Kevin McConway, Emeritus Professor of Applied Statistics, Open University, said:

“This is a complicated piece of research, and the interpretation of its findings needs some care. I’ll make some points relevant to my expertise as a statistician. There are many aspects of what’s in the research paper that are not within my own expertise.

“Overall, the conclusions from the research are moderate and measured. The researchers conclude that their findings may help assessments of the risks and benefits of fluoride, for example of adding fluoride to water supplies with the aim of reducing tooth decay. They also call for more and better research on the association between relatively low levels of fluoride and children’s average IQ, such as those that would be produced by fluoridation of water in countries like the UK or USA. They report that there isn’t much existing data on that specific question, and that there is considerable uncertainty (for statistical and other reasons).

“First, it’s worth saying that, arguably, much of the new research paper isn’t really news at all. As the paper itself points out, a considerable proportion of the work was carried out within a much larger systematic review* by the National Toxicology Program (NTP) in the USA, with a long report published in August 2024. That report indeed mentions the new paper several times (as a paper “in press”), and includes a summary of its results. But the NTP report goes beyond the new paper in some directions, including findings on other neurodevelopmental or cognitive results in children, cognitive findings in adults, and results on possible ways that fluoride might lead to effects on the nervous system (from humans, from animal studies and from other study types).

“That previous report includes the following statement in bold type:

“This Monograph and Addendum do not address whether the sole exposure to fluoride added to drinking water in some countries (i.e., fluoridation, at 0.7 mg/L in the United States and Canada) is associated with a measurable effect on IQ.”

“In the UK, most of the population does not live in areas where fluoride is added to drinking water supplies. Where it is added, as a measure to prevent tooth decay, the added fluoride can bring up the level to 1.0 mg/l (milligrams per litre). In some parts of the UK, there is a natural level of fluoride in the water supply. Legally, the maximum allowed in drinking water supplies is 1.5 mg/l.

“The majority of the studies that are systematically reviewed in the new paper, 45 out of 72, were carried out in China, often in areas where the level of fluoride in the water is well above the level permitted in UK drinking water. No studies in the UK (or the USA) were included – not because they were deliberately omitted, but because no studies of the appropriate types had been performed in the UK or USA.

“The fact that the data that are analysed in the new study doesn’t mostly deal with anything like the conditions relating to fluoridation of drinking water in the UK, the USA, or similar countries, is the reason why the disclaimer about adding fluoride appears in the NTP report.

“Though the new research uses data from some extra, recently published, studies that couldn’t be included in the earlier NTP report, the overall finding that, on average, IQ scores in children are rather lower in areas where people’s exposure to fluoride is higher is the same in the new research as in the previous report. (This relates, in different studies that were reviewed, either to fluoride levels in the drinking water supply, or to measures of fluoride in the participants’ urine.)

“Compared to the NTP report, the new research paper includes more detail on some aspects, and also analyses the amount by which average IQ scores differ in places where the fluoride levels differ by certain amounts. (I’ll return to that.)

“I should mention that the studies that are reviewed in the new research paper are all observational – the researchers in those studies did not make the participants consume different amounts of fluoride, they just observed their exposure to fluoride and carried out IQ tests.

“As always in observational studies, there remains a possibility that the correlation between higher fluoride exposure and lower average IQ isn’t caused by the fluoride at all, but by some confounding factors related separately to fluoride and IQ. That’s why it is essentially never possible to use results from one observational study to conclude anything about cause and effect. Finding consistent results from several observational studies and from different types of statistical analysis, as is the case in this new research, does tend to increase confidence that the fluoride is somehow causing the differences in average IQ, but really, evidence of different types would need to be considered too, before one could really be confident about what is causing what.

“The researchers in the new systematic review rated each of the studies that they included, in a systematic way, to decide which were at high risk of bias. That could be bias because they did not make appropriate statistical adjustments for possible confounding factors, or for several other reasons. In fact, of the 74 studies reviewed in the new research paper, 52 were rated as being at high risk of bias.

“The association between higher fluoride exposure and lower average IQ was still there in the studies rated as being at lower risk of bias, but showed much smaller differences in IQ between high and low fluoride levels than in the studies with higher risk of bias. I’d strongly recommend concentrating only on the results from studies rated as being at low risk of bias, and, even then, not forgetting that some doubt about cause and effect remains.

“In the low-risk-of-bias studies using measures of fluoride in urine (as a measure of exposure to fluoride from all sources), the researchers concluded that an increase of 1 mg/l in fluoride would correspond to a decrease in average IQ of just over 1 point. There’s considerable statistical uncertainty about that figure – it could plausibly be somewhere else between about 0.6 points and 1.7 points. But in any case that’s not a huge decrease, even if it turned out that it is all caused by the difference in fluoride exposure (and we can’t really be sure of that yet, in my view).

“On an individual basis, you can’t know people’s IQ exactly to the nearest point anyway – the tests give a score to the nearest point, but there’s a statistical margin of error around that score anyway, for example because if a person does IQ tests a few days apart, the resulting scores may well not be exactly the same.

“All this uncertainty does mean that comparing the potential benefit of adding fluoride to water for people’s dental health with the potential disbenefit of possible small changes in average child IQ is not straightforward in a country like the UK. This research perhaps helps a bit, but as it points out, it certainly doesn’t answer everything. Lack of data on low fluoride levels is a real issue.

 (Reference 12 in the research paper – available at https://ntp.niehs.nih.gov/publications/monographs/mgraph08)

‘Fluoride Exposure and Children’s IQ Scores A Systematic Review and Meta-Analysis’ Kyla W. Taylor et al. was published in JAMA Pediatrics at 16:00 UK time on Monday 6^th January.

DOI: doi:10.1001/jamapediatrics.2024.5542

Declared interests

Prof Kevin McConway: Previously a Trustee of the SMC and a member of its Advisory Committee.