A list of indicators of bad science – many of these are found in articles promoted by anti-fluoride activists.
Anti-fluoride activists have been pouring money into a scaremongering campaign warning pregnant women not to drink fluoridated water. They claim fluoride will lower the IQ of their future child.
Fluoride Free NZ (FFNZ) launched this campaign on the back of the recent publication of Canadian research on fluoride intake by pregnant women and child IQ (see Ground Breaking New Study – Top Medical Journal – Fluoridated Water Lowers Kids’ IQs). Now they are responding to criticisms of this paper by claiming it is supported by other research and claim a list of 6 papers support their claim that pregnant women drinking fluoridated water may be harming the IQ of their children.
Image used by FFNZ as part of their appeal for finance to support billboards and newspaper advertising promoting their false scaremongering claim.
None of these studies actually support the FFNZ claim. Let’s consider each of the six studies listed by the Fluoride Action Network (FAN) – but first, look at this data from the “New Study – Top Journal” mentioned by FFNZ:
This is from Table 1 in Green et al., (2019). There are no statistically significant differences in the IQ o children whose mothers lived in either fluoridated or nonfluoridated areas during their pregnancy. So the FFNZ claim is completely false.
The Fluoride Action Network (FAN) Bulletin Several In Utero Fluoride/IQ Studies Should Provoke A Nation-Wide Fluoridation Moratorium initially list 5 studies but expanded this to 6 studies in an article posted by Ellen Connett the same day – The 6 Mother-Offspring Fluoride Studies. Here is the list (descriptions are from the FAN articles):
- Green et al., (2019). Largest study with 512 mother-offspring. Lower IQ in children 3- 4 years of age.
- Bashash et al., (2017). Longest study. 299 mother-offspring pairs in Mexico. Lower IQ in children 4 and 6-12 years of age.
- Thomas et al., (2018). 401 mother-offspring pairs in Mexico. Lower IQ in children 1-3 years of age Only the abstract has been published.
- Valdez Jiménez (2017). Lower IQ between the ages of 3-15 months with 65 mother-offspring pairs in Mexico.
- Li et al., (2004). Significant differences in the neonatal behavioral neurological assessment score in 91 offspring aged 1-3 days old.
- Chang et al., (2017). Reported significant differences in the mental development index and psychomotor development index of the offspring at 3, 6, 9, and 12 months of age.
I will consider these in three groups and include some relevant theses.
The Chinese studies from areas of endemic fluorosis
These describe data from areas of endemic fluorosis in China. They are irrelevant to community water fluoridation as the drinking water fluoride concentrations are much higher and people in these areas suffer a range of health problems including severe skeletal and dental fluorosis. Remember, the optimum levels of fluoride recommended for community water fluoridation are generally below 1 mg/L and WHO recommends drinking water concentrations should be lower than 1.5 mg/L.
People in areas of endemic fluorosis suffer a range of health problems
Li, J., Yao, L., Shao, Q. L., & Wu, C. Y. (2008). Effects of high fluoride level on neonatal neurobehavioral development. Fluoride, 41(2), 165–170.
This is one of the generally poor quality studies FAN got translated to assist their campaign. They are generally published in Fluoride, an anti-fluoride journal.
It compared children from villages with drinking water concentrations of 1.7–6.0 mg/L with a control group from villages with drinking water concentrations of 0.5–1.0 mg/L.
Chang A, Shi Y, Sun H, Zhang L. (2017) Analysis on the Effect of Coal-Burning Fluorosis on the Physical Development and Intelligence Development of Newborns Delivered by Pregnant Women with Coal-Burning Fluorosis.
Another one of the poor quality papers FAN got translated but they have made only the abstract and a brief abstract available on their site. I cannot find a full test or even abstract anywhere else. It compares two groups:
“A total of 68 newborns delivered by pregnant women from coal-burning endemic fluorosis areas in this region were selected as an observation group, and 50 full-term newborns delivered by normal healthy pregnant women were selected as a control group. “
Both papers report statistically significant differences in some measurements between the two groups but that is to be expected for areas of endemic fluorosis and especially where coal-burning creates health problems. Of course, this is all irrelevant to community water fluoridation as only the control groups are likely to be drinking water with similar fluoride concentrations.
Mexican study from an area of endemic fluorosis
The paper is:
Valdez Jiménez, L., López Guzmán, O. D., Cervantes Flores, M., Costilla-Salazar, R., Calderón Hernández, J., Alcaraz Contreras, Y., & Rocha-Amador, D. O. (2017). In utero exposure to fluoride and cognitive development delay in infants . Neurotoxicology
This study found a relationship of child cognitive deficits with maternal prenatal urinary fluoride but, again, it is not relevant to community water fluoridation. The authors acknowledge that the study was done in an area of endemic fluorosis. About 90% of the drinking water samples in the study contained fluoride above the World Health recommended maximum of 1.5 mg/l and the Fluoride in the mothers’ urine was also high – with the mean concentration for all the mothers of 1.9 mg/l for the 1st trimester, 2.0 mg/l for the 2nd and 2.7 mg/l for the 3rd trimester. Urinary fluoride concentrations as high as 8.2 mg/l were found. This compares with a mean value for urinary F of 0.65 mg/L for pregnant women residents in areas with low levels of F in drinking water (0.4 to 0.8 mg/l – similar to that recommended in community water fluoridation).
I have written about this study in my post Premature births a factor in cognitive deficits observed in areas of endemic fluorosis?‘ There I speculated that the effect of fluoride on cognitive deficits may be indirect because of the observede higher incidence of prematurity and low birth rate.
The Bashash study
I have separated these from the Green et al., (2019) study although they both report relationships between maternal prenatal urinary fluoride and the IQ of offspring and many of the authors are common to both studies Bashash is the senior author on the paper reporting data from Mexico city and Green on the paper reporting data for Canada.
There are really three citations for this study. The main paper:
Bashash, M., Thomas, D., Hu, H., Martinez-Mier, E. A., Sanchez, B. N., Basu, N., … Hernández-Avila, M. (2017). Prenatal fluoride exposure and cognitive outcomes in children at 4 and 6–12 years of age in Mexico. Environmental Health Perspectives, 125(9), 8–10.
A conference poster:
Thomas, D., Sanchez, B., Peterson, K., Basu, N., Angeles Martinez-Mier, E., Mercado-Garcia, A., … Tellez-Rojo, M. M. (2018). OP V – 2 Prenatal fluoride exposure and neurobehavior among children 1–3 years of age in Mexico. Environmental Contaminants and Children’s Health, 75(Suppl 1), A10.1-A10. https://doi.org/10.1136/oemed-2018-ISEEabstracts.23
And Deena Thomas’s Ph. D. thesis which also reported data from the study:
Thomas, D. B. (2014). Fluoride exposure during pregnancy and its effects on childhood neurobehavior: a study among mother-child pairs from Mexico City, Mexico. University of Michigan.
The anti-fluoride activists have waxed lyrical about the reported negative relationship of child IQ with maternal prenatal urinary F concentrations but they are clutching at statistical straws as, in fact, the relationship is very weak – explaining only a few percent of the IQ variance. I explained this in my post Fluoride, pregnancy and the IQ of offspring, and described several other problems (correlation is not evidence of causation, information about the mothers is scant with no indication if they lived in areas of endemic fluorosis, possible important risk-modifying factors were not considered as confounders, urinary fluoride is not a good indicator of fluoride intake by individuals, and there was no association of child IQ to child urinary fluoride).
When data has this much scatter the marginal statistical significance of relationship are easily altered by tweaking the data. (Fig 3a from Bashash et al., 2017).
It is easy to be misled by marginal statistically significant relationships when considering data with such a high scatter. This is illustrated by the conclusions of one of the authors, Deena Thomas, in her Ph. D. thesis that:
“Neither maternal urinary or plasma fluoride was associated with offspring MDI scores” [page 37);
“This analysis suggests that maternal intake of fluoride during pregnancy does not have a strong impact on offspring cognitive development in the first three years of life.” [page 38];
“Maternal intake of fluoride during pregnancy does not have any measurable effects on cognition in early life.” [page 48].
Yet, in the conference poster based on her thesis she concluded:
“Our findings add to our team’s recently published report on prenatal fluoride and cognition at ages 4 and 6–12 years by suggesting that higher in utero exposure to F has an adverse impact on offspring cognitive development that can be detected earlier, in the first three years of life.”
Her conclusions reported in her thesis are exactly the opposite of the conclusions reported in her conference poster!
I suggest in my article A conference paper on the maternal prenatal urinary fluoride/child IQ study has problems that the different conclusions in the poster resulted from the fact that at least 30 mother-child pairs were removed from the data set used in her thesis (the thesis consider 431 mother-child pairs but the poster considered only 401 pairs). Perhaps even some data pairs were added (the maximum urinary F value is higher in the smaller data set used for the poster).
In her thesis, Deena Thomas also reported: “concurrent urinary fluoride exposure has a strong positive impact on cognitive development among males aged 6-15 years.” [page 54]. The relationship was not significant for females. But the actual paper, Bashash et al., (2017), concluded “there was not a
clear, statistically significant association between contemporaneous children’s urinary fluoride . . . and IQ. “
I have discussed the Mexican maternal prenatal urinary fluoride- IQ study in more detail in the following articles:
- Fluoride, pregnancy and the IQ of offspring,
- Maternal urinary fluoride/IQ study – an update,
- Paul Connett’s misrepresentation of maternal F exposure study debunked,
- Mary Byrne’s criticism is misplaced and avoids the real issues, and
- A conference paper on the maternal prenatal urinary fluoride/child IQ study has problems
A draft of my article critiquing the Bashash et al., (2017) paper, “Predictive accuracy of a model for child IQ based on maternal prenatal urinary fluoride concentration” is also available online. I have also discussed another paper from this study (Bashash et al., 2018) which reported a weak relationship of ADHD prevalence with maternal urinary fluoride in my article Fluoridation and ADHD: A new round of statistical straw clutching.
Green et al., paper/thesis
This study is actually the only one that included people exposed to community water fluoridation – hence the relevance of the data I presented in the introductory table which showed no effect. But the study it is basically the same as that of Bashash et al (2017) except it involved Canadian mother-child pairs and most of the criticism of Bashash et al., (2018) are relevant to the Green et al., (2019) study which has been reported in the following forms:
Green, R., Lanphear, B., Hornung, R., Flora, D., Martinez-Mier, E. A., Neufeld, R., … Till, C. (2019). Association Between Maternal Fluoride Exposure During Pregnancy and IQ Scores in Offspring in Canada. JAMA Pediatrics, 1–9.
Green, R. (2018). Prenatal Fluoride Exposure and Neurodevelopmental Outcomes in a National Birth Cohort (MSc thesis, Graduate Program in Psychology York University Toronto, Ontario).
My original critique included a conclusion that the reported negative relationship of child IQ with maternal prenatal urinary F concentration was extremely weak. I found that it explained only 1.3% of the child IQ variance using data extracted from the figures. Subsequently Rivka Green claimed an R-squared value of 4.7% which is still very low (we can reject her claim that it was “quite high” as simple promotion of her work).
For further discussion of the Green et al (2019) study see my articles:
- If at first you don’t succeed . . . statistical manipulation might help
- Politics of science – making a silk purse out of a sow’s ear
- More expert comments on the Canadian fluoride-IQ paper
- An evidence-based discussion of the Canadian fluoride/IQ study
- Fluoridation – A new fight against scientific misinformation
- Biostatistical problems with the Canadian fluoride/IQ study
A problem of self-promotion and confirmation bias
Science and the scientific literature are, of course, not immune to self-promotion and confirmation bais and I think the maternal urinary fluoride-child IQ studies show this. I discussed this in If at first you don’t succeed . . . statistical manipulation might help as well as pointing out that these scientific politics are amplified by activist propaganda.
This is a pity because such confirmation bias and self-promotion may result in important information being overlooked. I discussed this in my article A more convincing take on prenatal maternal dietary effects on child IQ which referred to another paper from the Mexican maternal urinary fluoride study which showed a relationship of child IQ with maternal nutrition:
Malin, A. J., Busgang, S. A., Cantoral, A. J., Svensson, K., Orjuela, M. A., Pantic, I., … Gennings, C. (2018). Quality of Prenatal and Childhood Diet Predicts Neurodevelopmental Outcomes among Children in Mexico City. Nutrients, 10(8), 1093.
Again the relationships reported were weak, but the negative relationship of child IQ with poor prenatal maternal nutrition explains 11.2% of the variance in child IQ – much better than the data for prenatal maternal urinary fluoride (which explained only 3% of the variance).
So what do these mother-child studies say about community water fluoridation and IQ?
Well, nothing really – except that the only study which compared fluoridated and non-fluoridated areas showed absolutely no effect.
But this will not stop activists (and unfortunately self-promoting scientists and their institutions) from making unwarranted claims. Their propaganda relies on unsupported “authority” opinion and misrepresentation. This violates many of the rules in my first image above.
It tries to present correlation as proof of causation, misinterprets results, promotes unsupported conclusions, selectively reports the data and findings, and expands these unsupported conclusions way beyond the small sample sizes used.
This is just bad science!