EXPERT REACTION: Fluoride exposure during pregnancy may be linked to neurobehavioural problems in kids

The Malin et al 2024 paper does not add to the body of knowledge due to major shortcomings with its exposure and outcome measurements, and small sample size. Those shortcomings render the study’s findings unreliable.

Exposure to fluoride in this study was measured by spot urinary fluoride analysis, which is highly variable within a person even during a day. This major shortcoming was even acknowledged in a paper, of which Malin was a co-author. Even 24-hour urinary fluoride measurement (a much more reliable exposure measurement) cannot measure chronic fluoride exposure in individuals (Rugg-Gunn et al. 2011). An inexplainable observation when those with the highest income or education had markedly high mean urinary fluoride levels,  pointing to the unreliability of this measure in this study.

The outcome measure has high measurement error, requiring relatively high literacy, and close attention to the child. Those with the highest income or education (also with the highest mean urinary fluoride in this study) could observe and report more problems about their child than others, leading to assessment bias.

The sample was very small (only 229), and not adequate to balance expected measurement errors. Even a few individuals with biased measurements could completely alter the overall findings.

The paper did not cite all the literature that contradicted their findings. An Australian nationwide study with a large population-based sample (2,682 children) did not find any effect of early life exposure to fluoride on child neurobehavioural development and executive functioning. The exposure measurement of fluoride exposure in this study was objective and fixed. It used two sophisticated instruments of child neurobehavioural development and executive functioning.

I think the results of this paper are interesting but not something to be alarmed about. Several limitations in this paper mean the results should be treated very carefully.

Firstly, the paper notes an ‘association’ between the fluoride concentration of a mother’s urine and the behaviour of her child at three years of age. An association between two factors does not mean the two factors are linked or that another factor (that was not tested for) is the cause of said association. The previous Canadian and Mexican studies that the authors refer to also had limitations that restrict the conclusions drawn from them.

Secondly, according to the paper, the participants in this study were mostly, from one specific group - 'Hispanic women of low socioeconomic status living in urban Los Angeles'. This means the results can’t be extended to other populations. Fluoride also has a short half-life in the human body and urine concentrations vary hugely so taking a single measurement is not accurate yet this paper also used the 'spot' method of urine collection, rather than the more accurate 24-hour method, so one can’t be sure of the actual fluoride levels the children were exposed to.

Thirdly, if fluoride really did cause problems with brain development in children, we would expect to see this effect consistently worldwide, but we don’t. In fact, at least two recent studies found that maternal fluoride levels were actually associated with better cognitive scores in children.

I could go on, but to summarise; Fluoride is a naturally occurring substance found in water and some foods. It is not toxic at the levels we are exposed to and there is no robust evidence of any relationship between fluoride exposure and IQ level or behaviour issues. This study does not change that.

This new study purports to show an association between prenatal fluoride consumption and child neurobehaviour at age 36 months. The study is based on a small sample of parents (229 mother-child pairs), with only 32 children (14%) reaching the threshold for total problems on the Preschool Child Behaviour Checklist. However, this number is inflated by including children who scored in the borderline clinical range with those who scored over the clinical threshold – and the paper does not report the numbers in each group. This means it is possible that the associations found are for children who have not reached the clinical threshold of the neurobehaviour measure.

There are also associations between a number of demographic factors (such as parental income and maternal education) and fluoride exposure, but there is no data presented on whether there is also an association between these factors and neurobehaviour. Blood lead levels are also reported to be associated with increasing the magnitude of the effect of urinary fluoride on CBCL scores, but it is not clear if there is also a direct association between blood lead levels and neurobehaviour.

Importantly, as the authors note, fluoride exposure is a single measure at one point in time during the third trimester – even allowing for the fasting (and it is not clear from the methodology whether the sample was taken during the fasting period or not) – it cannot be considered representative of the total prenatal fluoride exposure.

The study 'Maternal Urinary Fluoride and Child Neurobehavior at Age 36 Months' explores the association between prenatal fluoride exposure and child neurobehavioral outcomes in a US-based cohort. This prospective cohort study included 263 mother-child pairs from the MADRES cohort, focusing on third-trimester maternal urinary fluoride (MUF) levels. Neurobehavioral outcomes were assessed using the Preschool Child Behavior Checklist (CBCL).

The study found that higher prenatal fluoride exposure was significantly associated with increased odds of borderline or clinical range neurobehavioral problems, particularly internalising problems like emotional reactivity, anxiety, and somatic complaints. The findings suggest a need for recommendations to limit fluoride exposure during pregnancy to protect children’s neurodevelopment.
 
It is a prospective study, which adds weight to its findings. Also, the level of maternal fluoride exposure was reliably assessed and took into account all sources of fluoride. The findings are consistent with previous studies, which demonstrated that increased fluoride exposure during pregnancy may be associated with several adverse childhood outcomes, including symptoms of executive dysfunction, poorer inhibitory control, inattention, ADHD risk, and lower IQ levels. One must agree with the authors who state that the weight of scientific literature supports an association between increased fluoride exposure in pregnancy and adverse child cognitive and neurobehavioral developmental outcomes.
 
In Australia, the primary source of dietary fluoride is fluoridated drinking water. Community water fluoridation is widely practised in Australia to help prevent dental caries. Additionally, fluoride can be found in other sources, such as tea, certain fish and seafood, and foods and beverages prepared with fluoridated water. While not a dietary source, fluoride from toothpaste and mouth rinses can be ingested. Overall, the fluoridation of community water supplies remains the most significant contributor to fluoride intake in the Australian population.
 
There are no current recommendations to limit fluoride intake in pregnancy, but it may be prudent to do so in light of these research findings. Fluoridated tap water can be avoided by drinking bottled water and using it for cooking. Some, but not all, water filters remove fluoride from the drinking water and can be used to limit fluoride intake. Reducing the amount of tea one drinks and using dental products that do not contain fluoride while pregnant may also limit fluoride intake.