Food and Water Watch sued the EPA in 2017 — after the agency denied its petition to end water fluoridation under the Toxic Substances Control Act (TSCA). This week’s trial is the first to challenge the dismissal of such a petition. Other plaintiffs include Fluoride Action Network (FAN), Moms Against Fluoridation and other advocacy groups and individuals.

Fluoride’s neurotoxic effects on children’s brain development were not in dispute during opening statements and in testimony delivered by the plaintiffs’ first expert witness, Dr. Howard Hu, an internist and preventive medicine specialist, with a doctoral degree in epidemiology.

Instead, attorneys for both sides faced off over the question of what level of fluoride in the water supply poses a risk to the developing brain of fetuses and children.

Levels of fluoride found in drinking water in the U.S. are typically 0.7 milligrams per liter (mg/L), which is lower than the 1.5 mg/L levels found to be neurotoxic by the key reports discussed in the trial.

Attorneys for FAN argued that according to the EPA’s own guidelines for chemical risk evaluation — which they allege the EPA is failing to implement — fluoridating water at a dose that is so close to a known hazard level is too risky, especially given that children are exposed to fluoride from other sources in their daily lives.

They also argued the EPA’s failure to follow its own guidelines is unprecedented. The agency bans other regulated toxic chemicals, such as methylene chloride or trichloroethylene at levels much lower than the known hazard level to ensure the chemicals won’t pose a risk to human health.

And, they said, water fluoridation is unnecessary because the benefits to dental health come from the topical application of fluoride, not from its ingestion.

The EPA argued there is no compelling evidence that fluoride is a neurotoxin at the current levels used for fluoridation in the U.S. and that therefore water fluoridation doesn’t pose a risk to children.

Over two hundred million Americans drink fluoridated water, a practice that has been backed by public health officials and dental associations for decades.

If Chen decides fluoride poses an unreasonable risk, the EPA will have to revisit its rules on water fluoridation.

Fluoride regulation ‘long overdue’

Wednesday’s trial was picked after a June 2020 ruling by Chen that placed the trial on hold pending the release of the National Toxicology Program’s (NTP) report on the link between fluoride exposure and neurodevelopment effects.

The report was released in draft form under court order in March 2023, after top public health officials at the U.S. Department of Health and Human Services (HHS) tried for almost a year to block its publication.

The NTP report concluded that fluoride exposure at levels equivalent to 1.5 mg/L is associated with lower IQ in children.

The second phase of the trial is scheduled to take place over nine days at the federal courthouse in San Francisco, with a Zoom feed available for up to 1,000 viewers to watch live.

FAN member Clint Griess told The Defender that fluoride regulation was long overdue, but he had confidence Chen was carefully considering the science. He said:

“This [phase of the trial] is long overdue. We won after the first trial in my opinion. The judge is being extremely cautious. He has recognized, in his own words, that ‘justice delayed is justice denied.’

“Here we are in 2024, and we are still delaying and denying justice to millions of Americans. I’m very glad we are finally here and our lawyers are doing a great job. And I have every confidence that we will be victorious.”

EPA must apply its own guidelines to fluoride

In his opening arguments, the plaintiffs’ attorney Michael Connett told the court it faced an issue of national importance, “whether the widespread addition of fluoride to water presents a risk of neurodevelopmental harm to children, including IQ loss.”

The EPA faced a similar question in the 1970s, he said, when it had to address the question of adding lead to gasoline.

The EPA was in a quandary, he said, because, at the time, there was no clear evidence that lead was damaging at the levels used. But the EPA decided the margin between the hazard level and the exposure level posed too great a risk — leading the agency to outlaw lead in gasoline.

Connett said that properly applying the EPA’s risk assessment framework for existing chemicals under TSCA is at the heart of the decision the court is facing regarding water fluoridation.

During the first part of the trial in 2020, the agency used the wrong standard to assess the evidence, he said, holding the plaintiffs to a burden of proof the EPA had never held anyone else.

Connett said:

“What you see in this trial is the clash of fundamentally different paradigms. On one hand, you have the sort of 70-year-old longstanding approach by the CDC [Centers for Disease Control and Prevention] and dental interests where basically it’s not a risk until you prove beyond a reasonable doubt that 0.7 [mg/L] fluoride water is causing harm, and that’s been their approach.

“But that’s not how the EPA does business. They use risk assessment. And we are in a position where the plaintiffs are the ones explaining how the EPA is supposed to do risk assessment.”

The EPA’s risk assessment framework, he said, begins with determining whether and at what level a chemical poses a hazard through a dose-response analysis. Then it assesses community exposure. The third piece, he said, is that the EPA looks at the margin between the hazard level and exposure level.

Connett said there are two types of risk. The first is when human exposure exceeds the hazard risk, but that is very rare. For example, the EPA didn’t have that type of data when it decided to ban lead in gasoline.

Then, he said, there is inferred risk, where exposure is lower than the hazard level. This scenario focuses on whether that margin between hazard and exposure may put some people at risk. TSCA mandates the EPA protect the most susceptible people from risk, he said.

The EPA typically requires a margin of 30-fold to determine whether something has a risk. However, some are much higher — for example, tetrachloroethylene is banned at levels 89 times lower than the hazard level, and methylene chloride exposure is not allowed at levels 27 times lower.

In this case, he said, rather than inferring risk as it ought to, the EPA is requiring a risk hazard at the exposure level, which for fluoride is 0.7 mg/L.

Connett outlined the evidence the plaintiffs will present. It includes undisputed evidence that fluoride passes through the placenta and gets into the fetal brain. FAN also will present data from animal studies and human studies, including the NTP report at the center of the trial.

“The NTP found that a large number of studies have been published on fluoride and human IQ. In total, they identified 72 human studies, of which 64 found a connection between fluoride and IQ deficiency. Of the 19 highest quality studies, 18 found lowered IQ, a 95% consistency,” he said.

Connett introduced the first witnesses, Hu and Dr. Bruce Lanphear, professor of health sciences at Canada’s Simon Fraser University.

Connett also previewed evidence the EPA would introduce to attempt to show fluoride is not neurotoxic at low levels, namely a study conducted in coastal Spain by Jesus Ibarluzea, Ph.D., and published in 2022 after the NTP finished its systematic review.

That study did not find evidence that fluoride is neurotoxic at low levels. Instead, it found fluoride increased IQ for boys by 15 points — a finding Connett called “implausible.”

Connett told The Defender, “The EPA has never applied the principles of risk assessment to fluoridation and this case is finally getting them to confront the principles on this issue.”

Chen pushed back on EPA during opening comments

In its opening statement, the EPA argued that anything can be toxic at high levels. The agency’s attorney laid out the EPA’s core argument that there is not enough data showing fluoride’s neurotoxicity at low levels present in drinking water and the law requires a “preponderance of evidence” of risk.

He highlighted a line in the NTP report indicating that more studies at lower exposure levels were needed to fully understand the potential associations with neurotoxicity.

Chen paused the remarks to ask the EPA to confirm the NTP report did establish that with moderate confidence that fluoride caused neurotoxicity at 1.5 mg/L, a relatively low level, which the EPA attorney confirmed.

“Do you disagree with the NTP’s use of 1.5 [mg/L as a hazard level]?” Chen asked. The EPA’s lawyer said they did not.

The EPA also argued that TSCA says “must be a preponderance of the evidence that the chemical substance presents an unreasonable risk.”

According to the EPA, studies of fluoride’s neurotoxicity at low levels have mixed findings — some show there are statistically significant adverse effects at low levels and others found there are not.

Given that, EPA’s attorneys argued the data is “too inconsistent” to conclude that low-level fluoride exposure presents an unreasonable risk.

Chen interrupted the opening comments again to ask whether, as the plaintiffs argued, that uncertainty is precisely what should inform the discussion of risk. “If the outcome wasn’t lowered IQ but cancer or death,” he asked, “would that change things?”

The EPA closed by telling the judge that what matters for TSCA is whether 0.7 mg/L presents an unreasonable risk. Chen pushed back again, “Shouldn’t we consider that in context,” he asked, because fluoride exposure occurs through sources other than water?

The EPA named the expert witnesses it will call later in the case, including David Savitz, Ph.D., and the EPA’s Stan Barone.

‘The evidence is quite persuasive’

The first witness, Dr. Howard Hu, an environmental epidemiologist and chair of the Department of Population and Public Health Sciences, Keck School of Medicine, University of Southern California took the stand yesterday to begin the trial’s deep dive into the science.

Hu has authored more than 300 papers in peer-reviewed journals and published several studies on fluoride. He also advises the EPA and collaborates with its scientists on issues related to lead exposure.

In 1993, Hu co-founded the ELEMENT research project, a pregnancy and birth cohort funded by the EPA and the National Institutes of Health and used to study how prenatal exposure to environmental toxins, including lead, mercury and fluoride affects children’s neurodevelopment.

In such cohorts, researchers collect epidemiological data during pregnancy and then from children over their lifetimes to study a variety of health outcomes tied to environmental exposures.

More recently at San Diego, Hu analyzed data on fluoride and neurotoxicity from the MADRES cohort, comprised of Los Angeles County residents, largely Latino. That research is not yet published.

Hu testified about his research, which consistently finds a link between fluoride and lowered IQ in children.

One of his fluoride studies examined the ELEMENT cohort and found that prenatal levels of fluoride that appeared in maternal urine predicted offspring intelligence scores at ages 4 and 12, with IQ levels lower with incremental increases in maternal fluoride levels.

A second paper expanded the analysis of the 2017 paper and made similar findings. Hu said the neurotoxic effects of fluoride were the strongest in the nonverbal domains, which he said is similar to lead.

Hu also addressed other cohort studies that have different findings, such as the MIREC study in Canada or the Danish study referred to as Odense where the research was conducted, which Hu also used in some of his research.

For example, the MIREC study found sex-specific findings whereas the ELEMENT study did not. The Danish cohort study did not find statistically significant toxic effects.

Hu told the court that different sexes and demographics can have different life experiences that can account for different outcomes.

Overall, he said, his research supports the idea that fluoride at current exposure levels in drinking water is toxic.

Hu also discussed his concerns about the Spanish study the EPA is using as a basis to argue fluoride is not toxic at low levels. He testified it did not control for seafood consumption, which creates high levels of fluoride exposure. He testified it did not control for seafood consumption by pregnant mothers, which creates high levels of fluoride exposure and also has been shown to confer IQ benefits, so it could be a confounding factor in an analysis.

He also criticized the EPA’s opening statements. He said the EPA was presenting data as black and white. Epidemiology, he said, is moving away from characterizing things in that way. Even when a study, like the Danish Odense study, is “negative,” as the EPA put it, the data in the study can indicate a more nuanced reality.

On cross-examination, the EPA asked Hu to concede that the Spanish study was well done. Hu agreed but said he had serious reservations about it, which he had previously discussed.

The EPA also challenged the work he did with Grandjean reporting the Danish study. The results of the Danish study, which did not identify neurotoxic effects, were only published in 2023 as part of a “pooled” study where he and his colleagues used the Danish, Mexican and Canadian data to characterize the dose-effect of fluoride exposure, which the EPA’s lawyer implied was a form of selectively reporting results.

Hu told the court combining the studies increased the power of the analysis and the ability of the research to address questions of public health.

After his testimony, Hu told journalist Derrick Broze, “The evidence is quite persuasive that there is a negative impact of fluoride exposure on the neurodevelopment of children.”