Miss a day, miss a lot. Subscribe to The Defender's Top News of the Day. It's free.
Editor’s note: The Defender is providing daily updates on the landmark trial pitting Fluoride Action Network against the U.S. Environmental Protection Agency. The trial started Feb. 1. To read previous coverage, click here. 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.
Over nearly three days of testimony, the U.S. Environmental Protection Agency’s (EPA) first key witness in the landmark fluoride trial, David Savitz, Ph.D., downplayed the link between fluoride and lower IQ in children.
However, the plaintiffs’ attorney and federal Judge Edward Chen pushed back on some of his conclusions.
Savitz, professor of epidemiology at Brown University School of Public Health, worked with the National Academies of Sciences, Engineering, and Medicines (NASEM), reviewing the National Toxicology Program’s (NTP) draft report linking fluoride to lower IQ in children.
The report is one of the key pieces of evidence in a lawsuit brought by Food & Water Watch, Fluoride Action Network (FAN), Moms Against Fluoridation and other advocacy groups and individuals, against the EPA in 2017, after the agency denied a petition to end water fluoridation under the Toxic Substances Control Act.
Savitz’s testimony supported the EPA’s three key arguments:
- Data on fluoride’s neurotoxic effects for children at current levels of water fluoridation is mixed or uncertain and therefore no action should be taken.
- There are limitations to the NTP’s conclusions, published in draft form last year, linking fluoride exposure and IQ loss in children at 1.5 milligrams per liter (mg/L).
- More recent studies not considered by the NTP cast doubt on the NTP’s findings.
Savitz’s testimony and cross-examination, which began last week and finished Monday, raised questions about the EPA’s key arguments.
EPA paid Savitz over $137,000 to testify
Testimony began with a long list of Savitz’s qualifications as an epidemiologist. Plaintiff’s attorney Michael Connett underscored in his cross-examination that Savitz is an expert in epidemiology, but also pointed out that he has no experience researching fluoride.
Savitz countered that his lack of experience researching fluoride was an asset because it allowed him to examine the evidence in an unbiased manner.
Savitz also was one of six expert consultants commissioned to advise Health Canada, the country’s public health agency, on water fluoridation. The panel reviewed data, including a new systematic review linking fluoride exposure at very low levels to lower IQ in children. Health Canada commissioned the review but didn’t publish it.
However, the panel wrote a report — admitted into evidence last week along with the systematic review — advising Health Canada to maintain its existing water fluoridation standards. This occurred during the same time the NTP’s report findings were being debated — and allegedly suppressed — by U.S. public health agencies.
Savitz testified the panel determined that the data linking IQ loss in children at existing water fluoridation levels contained too much “uncertainty” to set a hazard level for drinking water, so they advised Health Canada not to change its fluoridation levels.
Under cross-examination, Savitz told the court he sat on that panel at the same time that the EPA was paying him $500 per hour — totaling between $137,000 to $150,000 for 275-300 hours of work — as a litigation expert for the EPA in this trial examining that very question.
Chen asked Savitz if Health Canada knew he was serving as an expert witness in this case when they invited him to the panel. He said the agency did.
He also told Chen he was unaware that two other members of the panel, Steven Levy, DDS, and World Health Organization report author John Fawell were known for their activist work in support of water fluoridation.
Reviewers didn’t want NTP report findings to be ‘misused’
Savitz said NASEM determined the first draft of the NTP’s report — which classified fluoride as a neurotoxin — fell short of providing “a clear and convincing argument” that supported its assessment.
Savitz told the court he didn’t think NTP’s conclusions were “wrong” but that they were stated in a way that could be “misused” as a tool for setting or changing water policy on water fluoridation.
He testified that NASEM was aware of the sensitivity of the issue and that the report could be used in ways the authors “might not have intended.” The panel explicitly said the monograph could not be used to draw conclusions about low fluoride exposure concentrations.
Savitz said he thought that after the revisions, the communication was “tempered” and “more consistent” with what he thought they were trying to do.
As this review process was ongoing, former NTP director Brian Berridge, DVM, Ph.D., privately expressed frustration that political pressure was put on the NTP to change its evaluations. Berridge told The Defender he thought this raised issues for public health.
‘Inaction is an action’
Savitz testified that because two of the four major cohort studies discussed in the trial (MIREC and ELEMENT), found a statistically significant effect of fluoride on IQ at low levels, and two did not (Odense and INMA), there was too much uncertainty to definitively conclude that it posed a danger at current levels of water fluoridation.
Chen asked, “I take it the converse would also apply? Which is that given this mix [of results] you can’t foreclose that there is an effect at U.S. drinking levels?”
Savitz conceded this was true.
Chen asked, given Savitz’s response and the NTP’s findings, if it makes sense to assume that there is a concern about current drinking water levels.
Chen also asked Savitz if he took issue with NTP’s conclusion that there is an association between fluoride exposure and lowered IQ at 1.5 mg/L — just over two times current fluoridation levels.
Savitz said he had no reason to challenge it, but he hadn’t corroborated it.
Savitz said another flaw was that the NTP used high-quality ecological studies — studies of endemic fluoride in other countries — as some evidence to show the effects of fluoride and that those could be confounded by other variables.
Chen pointed out that the studies would have controlled for that issue. Savitz conceded they did.
On cross-examination, Connett also pointed out that in Savitz’s own work on arsenic in China, his team studied endemic arsenic at high concentrations to show evidence for arsenic’s toxic effects.
They also used that data to consider toxic exposure levels in the U.S., using the same methods NTP scientists and other researchers were using endemic fluoride data, which Savitz criticized.
Connett also asked Savitz if he believed his own statements on uncertainty by quoting from Savitz’s textbook, “Interpreting Epidemiological Evidence: Connecting Research to Applications.” Savitz wrote in the book that “to claim we have insufficient evidence does not resolve the problem for those who make public health decisions, because inaction is an action.”
Does gender affect brain’s response to fluoride?
Chen also asked Savitz if the gender-based differences in response to fluoride, which were identified in the studies, were meaningful, regardless of statistical significance. Savitz said they were simply “flukes” because there is no reason to think gender would affect outcomes.
He said he wouldn’t expect to see sex differences and he didn’t know anyone who could explain them, so he thought it made more sense to focus on the aggregate data.
On cross-examination, plaintiffs’ attorney Connett challenged Savitz’s conclusions, introducing seminal work on sex differences in response to neurotoxins, explaining how common that is.
Connett also introduced the NTP’s report on fluoride and neurotoxicity in animals, which identified sex-differentiated responses and called for more research into the matter.
Savitz responded that he was an epidemiologist and therefore not familiar with the toxicology literature on sex differentiation.
Was Savitz cherry-picking the data?
Savitz identified several recent “high-quality” studies, published after the NTP completed its review, that he said found no statistically significant relationship between fluoride and IQ loss in children: the Mexican Cantoral 2021 study, the Canadian Dewey 2023 study and the Australian Do 2023 study.
Savitz said these three studies contributed to the uncertainty about water fluoridation’s risks.
On cross-examination, Connett asked Savitz why, in his expert report to the court, he never disclosed that some of the findings in both the Cantoral and the Dewey study did show statistically significant associations between fluoride and cognitive ability in children.
Savitz said those pieces of data were only part of the findings, but not what he considered important as part of the overall evidence from the paper.
When the trial resumed Monday, Connett walked through each paper, highlighting the statistically significant findings in the papers that Savitz had not reported.
For example, Dewey found significant adverse associations between maternal fluoride exposure and executive function in girls, and the authors reported this as a key finding.
Savitz said he thought the authors were simply highlighting the most “interesting” findings, but that it was bad practice.
Chen asked whether he found a problem in the findings he did not report. He said he didn’t, but he didn’t place a premium on them.
Connett then moved to the Do study, which Savitz correctly reported found no relationship between fluoride and IQ loss. He asked Savitz if he was aware the study was published in a dental journal — not a neurotoxicology journal —- by a dentist with no prior experience studying neurotoxicology using a methodology his co-author said had validity problems.
Savitz said he didn’t.
On redirect, EPA lawyers again showed the data presented in the studies Savitz summarized, and Savitz explained he mentioned only the results that showed no link between fluoride and neurotoxic effects because the preponderance of evidence in those studies made that finding.
The EPA also presented the conclusion from the recent Canadian systematic review, which also stated effects below 1.5 mg/L remained “uncertain,” which Savitz said was consistent with his conclusions.
Throughout his testimony, Savitz maintained there was no strong evidence for the neurotoxic effects of fluoride exposure at “low levels,” which extended up to 2 mg/L.
On cross-examination, Connett presented him with data from the NTP report and also from at least one key study showing this link.
Savitz conceded he hadn’t read those studies. In fact, in addition to the NTP report, he said he had read only about 10 studies on fluoride and neurotoxicity.
After Savitz, the EPA’s risk analyst Stanley Barone, Ph.D., took the stand again. Plaintiffs called him earlier in the trial to comment on the EPA’s risk analysis methodology, although he is a witness for the EPA.
The EPA called him back to testify as to the quality of the evidence on fluoride and IQ for a hazard assessment. His testimony is scheduled to end Tuesday, ending testimony for the trial.
Chen will review the recorded deposition of Jesús Ibarluzea, Ph.D., the plaintiff’s final witness, and review the evidence so he can pose any final questions to attorneys, who will present their closing arguments on Feb. 20 via Zoom.
