Expert reaction to study looking at maternal exposure to fluoride and IQ in children
Research, published in JAMA Pediatrics, reports that maternal exposure to fluoride in pregnancy can lead to lower IQ scores in young children.
Prof Thom Baguley, Professor of Experimental Psychology, Nottingham Trent University, said:
“First, the claim that maternal fluoride exposure is associated with a decrease in IQ of children is false. This finding was non-significant (but not reported in the abstract). They did observe a decrease for male children and a slight increase in IQ (but non-significant) for girls. This is an example of subgroup analysis – which is frowned upon in these kinds of studies because it is nearly always possible to identify some subgroup which shows an effect if the data are noisy. Here the data are very noisy. A further issue is that the estimate of the decrease in IQ for male offspring is unfeasibility large – at 4-5 IQ points. This level of average deficit would be readily detectable in previous studies and is likely a reflection of bias or very noisy data (the interval estimate here is very wide). As high fluoride areas are not randomly assigned there are also countless uncontrolled confounders. While they did correct for a limited set of covariates, the overall effect was non-significant with and without covariates. In summary it is not correct to imply that the data here show evidence of a link between maternal fluoride exposure and IQ. The average change in IQ is not statistically significant.”
Dr Oliver Jones, Associate Professor of Analytical Chemistry, RMIT University, said:
“While this paper’s title (which hints that high fluoride may affect your children’s intelligence) might at first sound scary I think there are several factors that should be kept in mind when reading it.
“The key words in the paper are “higher levels”. The authors state that an increase of 1 milligram per liter (1 mg/L) increase in fluoride was associated with a 4.49 point lower IQ score but fluoride intake appears to have been below 1 mg/L for most people in the study, even for those with fluoridated water, and nearly everyone (bar a few outliers) had a fluoride intake of less than 2 mg/L (which multiple previous studies have shown is safe) . There is also a Lot of variation in the data – which makes drawing firm conclusions/ predictions from it difficult.
“There are also a number of potential confounding factors including the fact that the water intake was self-reported and, as the authors admit, some of the methods used are not validated. One observed association also only seems to occur in boys and not girls but no explanation is given as to why this should be the case?
“In conclusion, while the work in interesting and the authors have tried to be thorough, in the end they only claim to have shown an association between two factors – not prove they are linked; the “possibility” of an effect justifying future research is not the same as concluding that there is an effect and; nearly everyone in this study had low levels of fluoride intake anyway. So, although the work is interesting, I don’t think there is a need to worry.”
Dr Joy Leahy, Statistical Ambassador, Royal Statistical Society, said:
“This study aimed to specifically assess the impact of maternal fluoride exposure during pregnancy and IQ scores in offspring. However, we should expect that fluoride exposure during pregnancy will be strongly associated with fluoride exposure in children after birth. For example, if a woman is living in an area with fluoridated water during pregnancy, then her child is likely to grow up drinking this same fluoridated water. Therefore, it is difficult to say whether any association found is driven by the fluoride consumption in pregnancy, or an assumed fluoride consumption in the infant after birth.”
Prof Alastair Hay, Professor (Emeritus) of Environmental Toxicology, University of Leeds, said:
“The Green et al. study looks at the association between maternal exposure to fluoride during pregnancy and IQ at ages 3 and 4 in the children. The authors find an association between maternal intake of fluoride and reduced IQ in boys, but not girls.
“I have a range of concerns about this paper.
“Firstly, as the authors acknowledge maternal intake of fluoride has not been validated. I see this as a crucial failure. For a substance with a short half-life, such as fluoride, urine concentrations vary hugely and are really only representative of the last drink. Validation of intake is something you must do before looking at associations.
“The authors look at mid-stream urine concentrations in pregnancy and base findings largely on these and the association with IQ at ages 3 to 4. Whilst the authors correct for changes in urine concentration (early morning urine is more concentrated than later samples in the day, for example) the half-life of fluorine is going to have a much greater impact on concentrations of the chemical than any adjustment to take account of urine dilution.
“Although the authors adjust for a range of substances (including lead in maternal blood and say this has no effect on findings, for me, the major serious gap is the range of exposure to multifarious substances, including lead, that the children would have had between birth and IQ assessment at ages 3 or 4. We know that lead exposure has devastating effects on IQ in children and this study takes no account of postnatal lead exposure.
“A curious finding is that the link between maternal urine fluoride and IQ decrements is only seen in boys and not girls. And the IQ decrement is not present for verbal IQ in boys. Whilst the authors are just reporting what they found I find these sex differences difficult to explain. With a neurotoxicant you might expect both sexes to be affected.
“There are many limitations which could affect the findings reported in this paper and until these are addressed, I regard the results as interesting, but not something to alter fluid or food intake by mothers.”
Prof Rick Cooper, Professor of Cognitive Science, Birkbeck, University of London, said:
“Green et al.’s headline claim that “maternal exposure to higher levels of fluoride during pregnancy [is] associated with lower IQ scores in children aged 3 to 4 years” is not supported by the data. In fact, when using the more reliable measure of fluoride exposure, a significant decrease in IQ was found only in boys – girls showed a non-significant increase in IQ. The negative effect was driven by a small number of boys whose mothers had extreme levels of fluoride exposure, but even these children had IQ in the normal range.
“There are numerous other complicating factors that limit the interpretability of the results. For example:
“While attempts were made to control for various chemical toxins, no attempt was made to take account of socio-economic status.
“There was no attempt to assess whether the fluoride estimate measure is accurate (yet there is opportunity to do so). Relative to the fluoride urine test, this estimate produces lower levels of fluoride in non-fluoride water areas, but higher levels of fluoride in fluoride water areas, so the accuracy of this estimate is questionable.
“The male sample includes a couple of extremely low IQ scores (below 70 would be special needs, but they have two boys with IQ in the 50s). The results would be more convincing if it were shown that they did not depend on these children.
“The fluoride distribution is not homogenous. There are very few children/observations at high levels and most at low levels, so the assumptions of the statistical test are not met, and any conclusions must be interpreted with caution.
“There is no consideration of whether an effect, if present, might be non-linear. For example, it appears that there is no effect at low concentrations, with the effect only arising at extreme levels. The chosen statistical approach does not allow consideration of this possibility.
“Lastly, one cannot tell if any effect (if present) is due to prenatal exposure or postnatal exposure.”
Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:
“This question would need to be researched in a lot more detail, in my view, before any clear conclusions can be drawn. I think that there are far too many questions to be answered before pregnant women should feel they must reduce their fluoride intake, or public health authorities should change their advice of policies. The reduction in average IQ scores in young children, associated with higher fluoride intake during their mothers’ pregnancies, is really fairly small, and there are reasons to doubt whether the reduction is entirely caused by fluoride intake anyway.
“The researchers used two different measures of the amount of fluoride the mothers consumed during pregnancy. One was based on the fluoride in the mothers’ urine. The researchers point out that such readings can change quite rapidly over time and be affected by things like drinking fluoride-free bottled water shortly before the urine samples were taken, or swallowing fluoridated toothpaste. So, their measure (based on only three urine samples across the whole pregnancy) may not be a very accurate measure of the exposure of the unborn baby to fluoride. The other method used an estimate of daily fluoride intake in the mothers. This measurement method was developed by this research team and they make it clear that it has not been validated. It is based on what the mothers recalled and told the researchers about their consumption of water and of tea (which contains some fluoride), on two occasions during pregnancy, but couldn’t take into account the actual brand of tea (because the fluoride content varies), and didn’t take into account any fluoride from food (which depends on exactly what was eaten) or from swallowing some toothpaste after teeth-cleaning. So, the accuracy is not known and may or may not be low.
“This is an observational study – the women were not allocated to particular levels of fluoride intake by the researchers, but just did what they would have done anyway. Therefore, there are likely to be several differences between mothers with high and low fluoride intake, apart from the levels of fluoride. They will tend to live in different places, if nothing else. Maybe these other differences (called confounders) are the real cause for any differences in the children’s IQ levels, and not their mothers’ fluoride consumption at all. It’s possible to make statistical adjustments to allow for the effect of these differences, though these adjustments aren’t perfect. The researchers did make some such adjustments, for things like the town the mother lived in, the mother’s age and number of previous pregnancies, a very simple measure of the mother’s educational level (was it above or below degree level), and a standard measure of the home environment that is often used in studies of learning and child development (but does not really relate to the physical environment of the home). But adjustments cannot be made for possible confounders on which there is no data. The researchers point out that they could not adjust for the mother’s IQ, and there may well be other important confounders that weren’t dealt with. In particular, apart from the home environment measure, nothing was taken into account that occurred between birth and the age of 3 or 4 when the child’s IQ was measured. A child’s IQ certainly isn’t completely determined before birth, and any coincidental difference in early childhood between children of mothers with different measured fluoride intakes could explain some or all of the observed IQ differences.
“The researchers reported the size of the associations between fluoride in the mothers’ urine, or their estimate of fluoride intake, and the children’s IQ in several different ways, but arguably none of the associations is actually very sizeable. Perhaps the most useful measure is the difference in average IQ between children whose mother’s fluoride intake or urine concentration was just in the lowest quarter of fluoride levels (the lower quartile), and children whose mother’s fluoride level was just in the top quarter (the upper quartile). For boys, in terms of the mother’s urine fluoride, the IQ for those at this lower quartile fluoride level had an IQ on average 1.5 points above those whose mothers were at the upper quartile level. Now 1.5 IQ points isn’t much, given the variation that exists between people in IQ anyway. Imagine we could get 100 pairs of boys, matched up in terms of the quantities used for statistical adjustment in this research, but with, in each pair, one boy whose mother was at the lower quartile of urine fluoride and one whose mother was at the upper quartile. Then in about 53 of those pairs, the boy whose mother had the lower fluoride level would have the higher IQ, but in the other 47 pairs, it would be the other way round and the boy whose mother had the lower fluoride level would have the lower IQ. Even if the difference was entirely caused by the fluoride – and we definitely don’t know that it is – it’s certainly not the case that children’s lives are being blighted on a large scale. In girls, those whose mothers had lower levels of fluoride in their urine actually had lower average IQs in this study, not higher, though the difference was not statistically significant and so might well have been due entirely to chance variation. Looking at the measure of maternal fluoride intake, there were slightly larger effects (averaged across both boys and girls), but again not huge. Again making the comparison on the basis of the quartiles (of the mothers’ fluoride intake measure), the difference was on average about 2.3 IQ points, but again looking at 100 pairs of children where one child’s mother had the lower fluoride intake and the other had the higher, in about 54 pairs the child whose mother had a lower fluoride intake would have the higher IQ, but in the other 46 pairs it would be the other way round.
“I think that differences of this sort of size could well be largely, or even entirely, due to residual confounding. That doesn’t mean that fluoride during pregnancy definitely has no effect on IQ – just that it’s far from clear that it’s even the main explanation for these relatively small IQ differences.”
Dr Stuart Ritchie, Lecturer, Institute of Psychiatry Psychology & Neuroscience, King’s College London (IoPPN), said:
“This study comprises two analyses – the first one is about the maternal urinary fluoride content, and the second is about the self-report of how much fluoride the mother had ingested. Some points to highlight about the results are:
- In the first analysis, there’s only a statistically significant result if they split the sample up into boys and girls: the effect only exists in boys. In the second, there’s an overall effect, but it’s no stronger in boys than girls. So those two results are inconsistent.
- Not only that, but as far as I can see there’s no pre-registration of the study, so we can’t know whether the decision to split it by sex was done post hoc. Doing further analyses on the same data makes it more likely a false positive result will be found. No theoretical reason for there being an effect in boys but not girls is given, which is further evidence that the split wasn’t done for a reason set out at the start of the study, before the data was analysed.
- For the second analysis, where there’s an overall effect, the p-value is .04 – that is, it’s JUST below the standard threshold used for declaring something to be significant (0.05). Given that they ran lots of other hypothesis tests in the paper, and didn’t correct for how many times they did so, I wouldn’t have much confidence in this finding being robust or replicable.
- The same point can be made for the sex difference in the first analysis – the p-value there (for the interaction, which is the critical test of whether the effect is bigger in one sex than another) is only .02, which I’d suggest is nothing to write home about either.
“So overall, I think the findings here are pretty weak and borderline. They might be interesting as part of a larger set of studies on this question, but alone they shouldn’t move the needle much at all on the question of the safety of fluoride.”
Prof Grainne McAlonan, Professor of Translational Neuroscience, Sackler Centre for Translational Neurodevelopment, King’s College London, said:
“I would be very cautious about over interpreting this data. Statistical significance does not equal ‘importance’.
“This is a large sample and although a statistically significant relationship is reported, any shift in IQ said to be elicited by exposure to more fluoride is not very dramatic. The authors describe the relationship between urinary fluoride and IQ in terms of a 1mg/L increase in urinary fluoride levels in pregnancy being associated with an almost 4.5 point reduction in IQ, but in reality the average difference in fluoride levels between the majority living in low and high fluoride areas is no-where near 1mg/L. The average urinary fluoride levels in non-fluoridated areas is 0.4mg/L while in higher areas it is almost 0.7mg/L. A difference of only 0.3mg/L.
“Also, if you look at average IQ in the children from fluoridated and non-fluoridated groups these are virtually the same: 108.07 vs. 108.21 respectively. I was therefore surprised that the study went on to look for a relationship between fluoridation and IQ, given these figures.
“If you look at the figures it’s only the individuals which have the very highest levels of fluoride whose children have any IQ lowering and it’s a pretty small drop in IQ score with a range of error that includes the average of the non-fluoridated children’s’ IQ.
“IQ, like many other measurements, contains a degree of error. Taking a test at different times or in different environments, with different administrators all might give a score difference. A few points difference in IQ may not necessarily be meaningful within the average range – that’s why we tend to talk about IQ in terms of ranges instead – low, average, high.”
‘Association Between Maternal Fluoride Exposure During Pregnancy and IQ Scores in Offspring in Canada’ by Rivka Green et al. was published in JAMA Pediatrics at 16:00 UK time on Monday 19 August.
Dr Oliver Jones: I have no conflicts of interest to declare.
Prof Kevin McConway: Prof McConway is a member of the SMC Advisory Committee, but his quote above is in his capacity as a professional statistician.
Prof Alastair Hay: I have no conflicts of interest.
Dr Stuart Ritchie: No conflicts of interest
Prof Grainne McAlonan: No relevant conflicts of interest
None others received.