After, 13 years the American Dental Association (ADA) has finally updated its statement on Water fluoridation. Throughout its 100+ page document, Fluoridation Facts 2018 the ADA provides us with “opinions” dressed up as “Facts.” It’s central opinion – which it has held for over 60 years (the ADA endorsed fluoridation in 1950 and it has defended it ever since) -is that any study that finds any harm caused by fluoride, at doses close to those experienced by anyone drinking fluoridated water, cannot possibly be a sound study. This opinion is the “belief” that underpins its unscientific and unprofessional promotion of this reckless and unethical practice. From this belief it works backwards.
Here is an example on an issue that our readers know something about. The following paragraph is how the ADA handles the groundbreaking 12-year mother-child study funded by US-government agencies and carried out by some of the most prestigious researchers on cognitive function in the world.
In 2017 a study from Mexico City received some coverage in the popular press.
The authors concluded higher urinary fluoride levels- of pregnant women were associated with lower scores on tests of cognitive function in their children. This was an observational study that by definition could only show a possible association between fluoride exposure and IQ – not cause and effect. This small study did not adequately address a number of potential confounders that might explain the possible association such as breast feeding, maternal age, gestational age, birth weight and education as well as exposures to lead, mercury, arsenic and iodine that affect IQ and other measures of cognitive ability.Unlike conditions in the U.S., the pregnant women participating in the study were exposed to varied fluoride levels from naturally occurring fluoride in the water supply (in some cases at levels almost twice as high as the level recommended for community water fluoridation in the U.S.) and fluoridated salt.
Lets look at ADA claims about this study line by line. We’ll see that ADA is completely wrong or misleading in every statement they make about this study.
(ADA’s original words in red)
In 2017 a study from Mexico City received some coverage in the popular press.
This is misleading because it ignores the more important fact about the study, which is that it was published in the leading peer-reviewed journal of environmental health (Environmental Health Perspectives), which is published by the National Institutes of Environmental Health Sciences (NIEHS), a division of NIH. It was funded with $3 million in US government grants from the NIH, NIEHS and EPA. The authors were some of the world’s foremost experts in studying neurotoxicity from fetal and early childhood exposures.
The authors concluded higher urinary fluoride levels of pregnant women were associated with lower scores on tests of cognitive function in their children.
This is correct but incomplete and understates the strength of the findings. The results of the study showed that for every 1 mg/L increase in urine fluoride in the pregnant mothers, their children averaged 6 IQ points lower test scores, a large effect. This finding was highly statistically significant. A 1 mg/L difference in urine fluoride is roughly the difference one would expect between mothers in fluoridated parts of the USA and those in unfluoridated parts.
This was an observational study that by definition could only show a possible association between fluoride exposure and IQ – not cause and effect.
This is a red herring and is also misleading. On ethical grounds it is prohibited to conduct experimental studies on humans where an exposure is intentionally given in order to determine whether that exposure causes harm. Thus, virtually all studies of harmful outcomes of any chemical exposure are observational studies.
This particular study found a large dose-response effect that was statistically significant and they ruled out most possible alternative explanations. It can therefore be described as having found strong evidence that fluoride does cause a large loss of IQ at the exposure levels in the people studied. The ADA bringing in the philosophical question of what types of studies can prove cause and effect is an attempt to distract from the strength of this study.
This small study
This was one of the largest studies of fluoride and IQ to date. It included 300 mother-child pairs. This is as large as many of the studies that have been used to show that childhood lead exposure causes loss of IQ. Furthermore, this study has individual-level information on every mother and child. The mothers were followed throughout pregnancy and the children followed from birth through age 6-12 years. This was neither an ecological study design or a cross-sectional study design, it was the strongest study design possible.
…did not adequately address a number of potential confounders that might explain the possible association such as breast feeding, maternal age, gestational age, birth weight and education as well as exposures to lead, mercury, arsenic and iodine that affect IQ and other measures of cognitive ability.
This is the most serious criticism of the Bashash study, and it is also the falsest of the ADA’s cricitisms.
The ADA Fluoridation Facts 2018 lists nine potential confounding factors and says they were not adequately addressed by the study. This claim is simply not true. Maternal age, gestational age, birth weight, and mother’s education were all measured as continuous variables and controlled for fully. Lead and mercury were measured as continuous variables and controlled for fully. Arsenic was not measured but the authors did not find any evidence that arsenic was elevated in the Mexico City area of their study. More importantly, for confounding, arsenic would not only have to be elevated but it also would have had to be strongly positively correlated with fluoride exposure. But most fluoride exposure in the Mexico City women was from fluoridated salt, which would not have been contaminated with arsenic.
Iodine was not measured, but there are strong grounds for expecting that if iodine confounded the relationship between fluoride and IQ, that it would attenuate the effect not exaggerate the effect. That is because in Mexico City the main source of fluoride is from fluoridated salt and that salt is also iodized. Therefore, mothers with higher fluoridated salt intake would also have higher iodine intake. Iodine is protective against loss of IQ, so it would counteract the fluoride in the salt.
The only potential confounder suggested by ADA that was not considered in the study was breast feeding. However, for a factor to actually confound the relationship between fluoride exposure and loss of IQ, the factor must be correlated with both fluoride exposure and the outcome, IQ. The main source of fluoride exposure was fluoridated salt. There is unlikely to be a correlation between breast feeding a child and fluoridated salt consumption in pregnant women. Furthermore, even if there was a very close correlation between breast-feeding and fluoride exposure, with the women with the lowest urine fluoride levels during pregnancy being much more likely to breast feed their child than the women with the highest urine fluoride levels, this would be unlikely to explain more than about half of the decrease in IQ. That’s because meta-analyses of the benefit of breast feeding on IQ have found it only increases IQ by about 3 points [Anderson et al 1999], whereas the Bashash study found a 6 point difference in IQ between those with high and low fluoride exposure.
So, in direct contradiction to the ADA claims that the study did not adequately account for potential confounders, in fact, one of the most important strengths of the study was how comprehensively and carefully it did control for potential confounding.
Unlike conditions in the U.S., the pregnant women participating in the study were exposed to varied fluoride levels from naturally occurring fluoride in the water supply (in some cases at levels almost twice as high as the level recommended for community water fluoridation in the U.S.) and fluoridated salt.
This is a very misleading statement on several grounds. Most importantly, it ignores the fact that it is the total fluoride exposure, not exposure from a single source, that is relevant for comparing the Bashash study to exposures in the U.S. Mother’s urine fluoride level is a good reflection of total fluoride intake. It is also likely to be strongly correlated with the fetal fluoride exposure.
Therefore, the best way of assessing whether the Bashash study exposure levels are relevant to the U.S. is to compare the urine fluoride levels in Bashash to those in the U.S. No studies of the urine fluoride levels in pregnant women in the U.S. are known, however studies of U.S. adults have found a similar range of urine F as found in the Bashash study. Also, national surveys of urine F in Canada and the United Kingdom have also found a similar range as the Bashash study.
There has been one study of urine F in pregnant women in a city with artificial fluoridation in New Zealand. This study, by Brough et al (2015) found an average urine F and range of urine F that were almost exactly the same as found in the Bashash study. The existing evidence is therefore strong that pregnant women in the U.S. will have a similar range of exposures as was found in the Bashash study. Whether their fluoride exposure comes from artificially fluoridated water, water with a natural high fluoride level, fluoridated salt, or eating a lot of canned sardines (which are high in fluoride) is immaterial.
The ADA’s implication that the Bashash study is not relevant to the U.S. because the main source of fluoride was not artificially fluoridated water is the same fallacy used by Broadbent et al. (2015) in the study of fluoride and IQ in New Zealand that was not able to find an effect. Broadbent only considered fluoridated water as a source of fluoride and did not properly account for fluoride supplements. Fluoride supplements were given to those without fluoridated water so that those children had total fluoride intake almost as high as the children with fluoridated water. There was little difference in total fluoride exposure so it is not surprising Broadbent found little difference in IQ.
There is another misleading, if not outright false “fact” in this ADA statement about fluoride exposures:
…naturally occurring fluoride in the water supply (in some cases at levels almost twice as high as the level recommended for community water fluoridation in the U.S.)
This uses the same false logic as the ADA and other fluoridation defenders have used when criticizing many previous studies of fluoride and IQ. They frequently cite only the highest level of exposure that showed an effect, not the average level or lowest levels. In toxicology and when setting public policy, it is the lowest levels of exposure that cause harm that are relevant, not the highest levels. This should be self-evident.
So, for ADA to state “in some cases” the natural water fluoride level was “almost twice as high” as the 0.7 mg/L level currently recommended for water fluoridation in the U.S. is simply an attempt to distract from the relevant exposures, which are the lowest exposures that caused harm.
Furthermore, it ignores the evidence from a separate study by Martinez-Meir et al (2005), one of the co-authors of the Bashash study, that found the average water fluoride level in Mexico City was only 0.3 mg/L, which is less than half of the recommended fluoridation level in the U.S.
In summary, every one of ADA’s “Fluoridation Facts” about the Bashash study are false, misleading, or deceiving because of omission.
The ADA is attempting to deny the science. It plays “fast and lose” with the “facts” to support its agenda to defend fluoridation at all costs. And the costs in this case is a serious risk that fluoride exposure -at levels experienced in artificially fluoridated communities – threatens our children’s brains.
Paul Connett, PhD.