Prenatal Fluoride Exposure and IQ in Kids: Is There a Link?

Tara Haelle

August 19, 2019

Pregnant women's fluoride intake was associated with a reduction in their children's IQ at ages 3 to 5 years, in an observational study published online today in JAMA Pediatrics.

Various outside experts have attested to the quality of the study's methodology and statistical analysis — which included sensitivity analyses and controlled for a multitude of potential confounders — but critics and other experts still point to important limitations that encourage caution in how the results are interpreted until additional research replicates the findings.

"Given there was limited evidence to support or refute fluoride safety, we weren't really sure what to expect in the Canadian population, so we let the data tell the story," lead author Christine Till, PhD, CPsych, an associate professor of psychology at York University in Toronto, Ontario, Canada, told Medscape Medical News.

"We're certainly concerned about the size of the effect we're seeing," she said, but she acknowledged research in this field is in its early stages. "In terms of where we go from here, it will really depend on policymakers and public health scientists to weigh the benefits and risks of fluoridation, and our hope is that our results combined with other recent studies are going to inform this."

For the prospective study, Till and colleagues collected maternal urine samples to measure fluoride concentration during each trimester and self-reported data on maternal intake of tap water, including drinks such as tea and coffee, once in the first trimester and once in the third trimester. The team enrolled 601 women who gave birth between 2008-2012 in six cities across Canada: Vancouver, Halifax, Hamilton, Kingston, Montreal, Toronto, and Vancouver. Their children's IQ was tested using the Wechsler Primary and Preschool Scale of Intelligence-III when they were 3-4 years old.

Among 512 mother–child pairs with all three trimesters of urinary exposure and covariate data, median urine concentration was 0.41 mg/L, though it was unsurprisingly higher for the 40% of women who lived in cities with fluoridated water (0.69 mg/L) than those in cities without (0.40 mg/L).

For these children, researchers found each additional 1 mg/L of maternal urinary fluoride was associated with 4.49 fewer IQ points in boys (P = .02) but there was no significant association in girls. (In fact, the data show a slight nonsignificant increase in IQ with increasing fluoride exposure in girls.)

However, self-reported data on water intake showed no such sex differences. Among 400 women with complete data on self-reported intake — excluding those who didn't drink tap water or live outside water treatment zones — each 1 mg increase in daily fluoride intake was linked to a 3.66 decrease in IQ score for boys and girls (P = .04). Overall median estimated daily intake among these women was 0.39 mg; it was 0.93 mg in fluoridated and 0.30 mg in nonfluoridated areas.

Experts Point Out Study Rigor and Caveats

Till said the research team did "everything we possibly could" to account for limitations of observational studies, including consideration of over 30 covariates and running multiple sensitivity analyses. Sensitivity analyses, for example, showed no change in the result after consideration of other neurotoxic contaminants and creatinine (a biomarker of tobacco smoking).

The covariates included child sex, gestational age, and age at testing; home secondhand smoke exposure; quality of the child's home environment; city; household income; and paternal age, education, employment status, smoking status, and race/ethnicity. Maternal factors included those associated with fluoride metabolism, as well as age, parity, race/ethnicity (white or nonwhite), education, smoking, pre-pregnancy body mass index, marriage status, birth country, smoking status, pregnancy alcohol consumption, breastfeeding duration, and chronic conditions during pregnancy, including gestational diabetes and hypertension.

F. Perry Wilson, MD, MSCE, an associate professor of medicine and interim director of the Yale University School of Medicine's Program of Applied Translational Research in New Haven, Connecticut, told Medscape Medical News that the biostatistical analysis looks solid. However, unmeasured confounders, such as maternal IQ, and other limitations remain, including the use of urine concentration instead of blood concentration, the latter being more precise for assessing potential fetal exposure, he said.

"In the end, it's an interesting observational study which adds importantly to the existing literature in this space by examining the lower end of fluoride exposure," said Wilson, who is writing a separate commentary about the study for Medscape Medical News. Past studies have mostly examined populations with extremely high fluoride exposure, he said, "but the subtle overall effect, isolated only to boys, and the presence of unmeasured confounding means we should be quite cautious about jumping to conclusions."

But Scott Tomar, DMD, MPH, DrPH, a professor of dental public health at the University of Florida College of Medicine in Gainesville, was more skeptical of the findings. Tomar, a consultant to the US Food and Drug Administration (FDA) and American Dental Association, questioned why the only association seen in the urinary concentration analysis is in boys and how much the data supported the conclusions given that most participants consumed less than 1 mg/L of fluoride daily.

"I really don't think this should have an effect on the policy of recommending water fluoridation in the United States or Canada," Tomar told Medscape Medical News. "There's nothing remotely resembling a linear relationship in the range at which the overwhelming majority of these subjects fell."

The authors made no recommendations related to water fluoridation policies but did suggest that women should reduce their fluoride intake during pregnancy. Still, Tomar said he does not see a real association "between exposure to fluoride in the concentrations typically found in water fluoridation and IQ" in the study.

"If this effect were real," he continued, "we would have seen widespread declines in IQ from the 1940s through the early 21st century in the United States, as exposure to fluoridated water increased from about 10% of the population to around 80%. We saw just the opposite."

Tomar also pointed out that other differences in the cities may account for the IQ differences — though the authors included city as a covariate — and suggested the regression analysis relied too much on outliers with higher fluoride exposure levels. Till, however, said they found no effect difference in the model when they excluded outliers from the analysis.

Other experts praised the study's methodology.

"I think this is one of the most rigorous studies published in this whole field because they are using biological markers," a substantial improvement over previous epidemiological studies that used only geographic or other indicators to determine fluoride exposure levels, said Howard Hu, MD, ScD, an affiliate professor in environmental and occupational health science at the University of Washington in Seattle. Hu was a coauthor, along with some authors from the current study, on the only previous study to use a biomarker in pregnant women, also urinary concentration, in Mexico City. He said that "urinary concentration tends to be more stable than blood concentration."

"It would be hard to criticize this study as being subject to confounding or bias. The sensitivity analyses that were done strengthen the ability to conclude that the relationships they found were quite strong," Hu said. However, he stopped short of making any recommendations based on the findings. "I'm committed to doing the science. What I can say is this study is really rigorous, but I'm going to leave the policy pronouncements to others."

In an accompanying editorial, David C. Bellinger, PhD, a professor of neurology and psychology at Harvard Medical School, Boston, Massachusetts, points out that past studies associating fluoride with reduced cognition — most often ecological ones — were "limited by their generally poor quality." He notes that the authors of the current study "considered numerous potential threats to the validity of the findings," resulting in apparently robust findings.

"The hypothesis that fluoride is a neurodevelopmental toxicant must now be given serious consideration," Bellinger writes, noting how long it took researchers to conclude no safe lower threshold of lead exposure exists. "Research on fluoride as a potential neurodevelopmental toxicant is still at an early stage, and a compelling weight of evidence from high-quality epidemiological studies has not yet developed."

But he also noted the many questions the study brings up, such as replicability and generalizability, possible effects on other neurodevelopmental domains, the most critical exposure windows, and risk from postnatal exposure.

Questions Remain Over Sex Differences and Effect Size

Several experts questioned why boys might show an effect when girls showed almost the opposite effect in the urinary analysis and why no sex differences showed up in the self-reported intake analysis. For the former, Till pointed out that the developmental neurotoxicity literature already shows males tend to be more sensitive to exposures, including lead and endocrine disruptors.

Even if the findings are taken at face value, there is also disagreement on the magnitude of the effect.

"At the population level, we're talking about a very large effect," especially when there is no benefit of fluoride exposure to the fetus, Till said. Her former research colleague Hu agreed. Because epidemiological studies focus on average effects in a total population, some kids lose more points and some lose fewer, he said. "If you shift the entire [IQ] curve over 3-4 points, some kids can't really afford to lose IQ points."

Bellinger also writes that the "total cognitive loss at the population level...could be substantial" at the effect sizes reported, though he adds that early studies of subsequently affirmed associations "tend to report larger effect sizes than do later studies."

But biostatistician Regina Nuzzo, PhD, a senior advisor for Statistics Communication and Media Innovation at the American Statistical Association, Washington, DC, points out that IQ study measures already include a degree of uncertainty, often with a margin of error of up to 5 points in either direction.

"So it's very possible that the uncertainty in any precise IQ score is much higher than the effect of any kind of fluoridation in the water," she told Medscape Medical News. "Overall, this looks like a classic case of statistical significance vs practical significance, plus a lot of uncertainty that might dwarf any actual effects."

She also pointed out that urine collection only occurred three times during pregnancy, so otherwise mundane actions such as accidentally swallowing toothpaste that morning or having more black tea than usual on those days could affect the readings.

Hu also noted the possibility of measurement errors but suggested they would lead the other direction. Using just two or three measures of fluoride to estimate cumulative exposure during pregnancy may attenuate the effect, he said, so the "the real impact on IQ might be somewhere higher."

Still, given the potential bias in self-reported intake, such as recall bias and social desirability bias, Nuzzo said a potentially substantial amount of uncertainty and error in the intake estimate could translate into "highly uncertain numbers coming out of the model."

Concern About Effect of Findings

The variety in perspectives about the study's findings and potential bias reflect the stakes of research into water fluoridation. Within environmental epidemiology, "the intensity of the debate inspired by the fluoridation of municipal water supplies to reduce dental caries is perhaps unrivaled," Bellinger wrote in his editorial.

The study authors did not call for any changes to water fluoridation policies, but some worry about what might result from these findings given the history of anti-fluoridation advocacy — despite broad public, health, and medical organizations' support for water fluoridation, which the US Centers for Disease Control and Prevention has called one of the greatest "achievements in public health" of the 20th century.

"It's unfortunate because I think this is going to unnecessarily raise lots of concerns," Tomar said. He suggested that confirmation bias may play a role in the findings as the same group of researchers have published several other studies on fluoride exposure and neurological concerns in children, such as increased incidence of attention deficit hyperactivity disorder.

"This is a group that has conducted a series of studies [on fluoride], and it's hard to conclude anything else other than they seem to want to find adverse effects of fluoride on various health outcomes," he said.

Other population-based studies that looked at fluoride exposure in children after birth, such as one in New Zealand, have not found an association with intelligence. Jonathan M. Broadbent, PhD, University of Otago Faculty of Dentistry in Dunedin, New Zealand, who led that study, said the topic certainly warrants further investigation given these findings.

JAMA journal editors are cognizant of the potential reverberations of the study's findings among the general public. In fact, they devoted a podcast to it, and JAMA Pediatrics editor in chief Dimitri A. Christakis, MD, MPH, University of Washington, Seattle, wrote in an Editor's Note that deciding to publish the study "was not easy" given its potential implications.

He said in the podcast that the journal editors "try not to accept observational studies where an experimental design could be done, and that clearly could not be done here ever, so we're stuck with observational studies." The idea that fluoride as a neurotoxin was "totally news to me," something he'd previously have dismissed as "junk science," Christakis said.

"One of the things we struggled with as an editorial staff was we certainly would not want to be the publishers of the Wakefield study and be the article that pulls fluoride out of water and led to decreased dentition," he said, referring to the fraudulent study published in the Lancet that unnecessarily spurred a generation of fears related to vaccines. "We are left with the challenge of how to advise pregnant women," he said.

But he noted that science is an imperfect, iterative process and implied the journal had a responsibility to publish the study after subjecting it to "additional scrutiny for its methods and the presentation of its findings."

"It is rare that a single study provides definitive evidence," he wrote in the editor’s note. "This study is neither the first, nor will it be the last, to test the association between prenatal fluoride exposure and cognitive development. We hope that purveyors and consumers of these findings are mindful of that as the implications of this study are debated in the public arena."

US federal agencies have not ignored public health research into water fluoridation. The Environmental Protection Agency (EPA) re-evaluated the science on fluoride in 2015, when the existing maximum fluoride level in water was 4 mg/L with a non-enforceable (albeit recommended) secondary standard of 2 mg/L. That amount is aimed at protecting against adverse effects of fluoride, while Health & Human Services recommends 0.7 mg/L as the optimal fluoride level for balancing the benefits of tooth decay reduction and potential harm of dental fluorosis. Following the EPA's 2015 re-evaluation, they advised a reference dose of 0.8 mg/kg/day as the "estimate of the daily exposure that is likely to be without harmful effect during a lifetime." FDA limits on fluoride in bottled water range from 0.8 to 2.4 mg/L in the United States.

The World Health Organization's quality guideline value for fluoride in drinking water is 1.5 mg/L but emphasizes the need for local public health officials to consider existing geographic variation in climate, water intake, and other sources of fluoride, including geologic/geographic differences.

"At present, there is no reference value for urinary fluoride concentration in pregnancy," Bellinger writes, "and guidance on fluoride use to prevent caries does not include separate recommendations for pregnant women."

The research was funded the National Institute of Environmental Health Science, Chemicals Management Plan at Health Canada, Ontario Ministry of the Environment, and Canadian Institutes for Health Research. Coauthor Lanphear is serving pro bono as an expert witness in upcoming litigation involving the EPA and water fluoridation. The other authors have reported no relevant financial relationships. Bellinger, Nuzzo, and Wilson have reported no relevant financial relationships. Tomar consults pro bono for the American Dental Association and FDA. Hu has worked with some of the authors on a previous fluoridation study and is currently working with Till on other research.

JAMA Pediatr. Published online August 19, 2019. Study, Editorial, Editor's Note

Follow Medscape on Facebook, Twitter, Instagram, and YouTube.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.