Expert witness testimony in fluoride litigation: strategic steps

The appellate court did not settle the scientific dispute. It faulted the way the district court handled the record.

That distinction matters to anyone studying expert witness testimony in fluoride litigation. A trial can turn on toxicology, epidemiology, exposure estimates, and dose-response analysis. It can also turn on what evidence the parties actually presented, what the judge was asked to decide, and whether the court stayed within those boundaries. The TSCA case offers a detailed procedural map. It does not offer a guaranteed formula for winning one.

How the TSCA case built its expert record

The dispute began with a petition, not a courtroom trial. In November 2016, petitioners asked the Environmental Protection Agency to issue a rule under Section 21 of TSCA. After EPA denied the petition, they filed suit in 2017 under the statute’s judicial-review provision.

The district court held a seven-day bench trial beginning June 8, 2020. The record included expert testimony on research concerning fluoride neurotoxicity. A second bench trial took place in early 2024. The later findings discuss testimony on epidemiology, fluoride exposure, cohort studies, and dose-response relationships. They also describe disagreement between experts over evidence at exposure levels below 1.5 mg/L.

That sequence is useful because it shows the scale and shape of the factual work. The case did not reduce to a single paper or a single witness. The parties presented competing interpretations of research and exposure, and the court had to evaluate how those pieces fit together. For litigators, the practical lesson is plain: an expert’s conclusion carries the weight of the chain supporting it. A dose estimate, a study population, the limits of a cohort, and the reasoning connecting exposure to an asserted health risk all belong in the record.

The 2024 district-court findings concluded that fluoride at 0.7 mg/L presented an unreasonable risk under TSCA and ordered EPA to manage the associated risks. That was a significant trial-court result. It is not the operative final ruling now. The Ninth Circuit vacated it in 2026.

The trial record mattered. So did the court’s authority to decide from that record.

This is where claims about the case often outrun the record. The litigation does not establish that expert testimony alone determines the outcome, that a particular type of witness will prevail, or that one exposure threshold resolves every fluoride dispute. It establishes that expert evidence formed part of a contested record in a specific TSCA proceeding.

Rule 702: the gate is about the opinion, not the résumé

Federal Rule of Evidence 702 governs expert testimony. Its 2023 amendment is reflected in the official House version of the rule. In practical terms, a court must decide whether proposed expert testimony satisfies the rule’s requirements before it becomes evidence for the factfinder. A respected title or a long publication list cannot do that work by itself.

For counsel qualifying scientific experts for water lawsuits, the core task is to connect a witness’s expertise to the opinion the witness will actually offer. A specialist in epidemiology may be qualified to interpret population studies, but that does not automatically qualify the same witness to calculate individual exposure or explain a toxicological mechanism. A toxicologist may speak to biological plausibility while lacking the foundation to resolve every epidemiological limitation. The line depends on the proposed testimony and the record, not on a broad label attached to the expert.

Preparation should make the boundaries visible:

• Identify the precise opinion the witness will give, and separate it from adjacent questions that require different expertise.
• Trace each conclusion to the data, methods, and assumptions on which it depends.
• Address study limitations directly, including population differences, exposure measurement, and uncertainty in dose-response interpretation.
• Keep the witness’s qualifications aligned with the subject of each opinion.
• Make clear which conclusions arise from the evidence and which remain uncertain.

These are not magic words that guarantee admissibility. The Ninth Circuit’s decision in the TSCA case did not set out a universal strategy for expert qualification, and the case-specific history cannot promise that a court will admit or credit any particular testimony.

A sound litigation record should also distinguish admissibility from weight. A court may admit an opinion and still find it unpersuasive after comparing it with competing evidence. Conversely, a forceful conclusion does not cure a weak foundation. Treating those questions as interchangeable invites avoidable trouble at trial.

The 2020 and 2024 trials: manage the record before the judge has to

Two bench trials created a procedural history with real consequences. The first began in June 2020 and lasted seven days. The second occurred in early 2024. The district court’s later findings addressed scientific subjects including epidemiology, exposure, cohort studies, and dose-response. But the appellate dispute focused on what the court could decide from the first-trial record and how it handled material developed later.

The Ninth Circuit held that the district court violated the party-presentation principle when it declined to decide the case on the first-trial record and insisted on a later National Toxicology Program monograph that the parties had agreed not to present at the first trial. The appellate court also concluded that both sides had introduced evidence during the first trial beyond the material in the 2016 EPA petition. In those circumstances, the district court did not err by considering that evidence.

That ruling puts a spotlight on record management. In litigation under TSCA Section 21, lawyers must keep track of what the petition contains, what each side introduces, what the parties stipulate, and what the court is being asked to decide. A trial record is not a general library of everything relevant to fluoride. It is a defined body of evidence assembled by the parties under the rules of the case.

For expert witness deposition preparation, that means rehearsing the reasoning, not just the headline conclusion. Counsel should test whether the witness can explain the source of each key figure, describe the limits of the underlying studies, and answer questions about alternative interpretations without sliding into advocacy. Depositions expose gaps. They also establish what an expert can support under questioning. A polished report that cannot survive a careful walk through its assumptions is a liability, not a shield.

Presenting infant health data in litigation calls for the same discipline. The available case facts establish that experts addressed neurotoxicity research and related scientific questions; they do not establish a universal evidentiary protocol for infant-specific claims. Any such claim must be tied to the actual studies, populations, exposure measures, and conclusions in the record. If the evidence does not support a particular inference, counsel cannot repair that weakness by making the language more emphatic.

The Ninth Circuit’s ruling: a procedural reversal, not a scientific verdict

The appellate court’s May 21, 2026 decision vacated the district court’s ruling and remanded the case. Its focus was procedural: party presentation, the record, and the district court’s approach to deciding the case. It did not resolve the scientific merits of fluoride exposure. It did not establish a nationwide fluoridation mandate. It did not declare the district court’s 2024 risk finding to be the final word.

The court also left an important statutory question open. It declined to decide whether Section 21 permits evidence beyond the original EPA petition in circumstances different from those before it. The Ninth Circuit found that both parties had introduced evidence beyond the petition in this case, and it addressed the district court’s consideration of that evidence in that setting. That is narrower than a general rule that extra-petition evidence is always allowed, or always barred.

For lawyers building a federal record, the distinction is not academic. It affects how they frame the scope of the dispute, preserve objections, and explain why particular expert evidence belongs in the case. A party should not treat the appellate ruling as permission to expand the record without limit. Nor should it assume that the petition alone necessarily defines every evidentiary boundary. The broader issue remains unresolved on the facts supplied here.

There is another practical point: procedural victories can erase a headline result without answering the underlying scientific question. The district court’s 0.7 mg/L finding drew attention because it was concrete. After vacatur, it cannot be presented as the final operative determination. Anyone describing the case should state both parts: what the district court found in 2024 and what the Ninth Circuit did in 2026.

Dose-response disputes need careful translation

The district-court findings discussed disagreement over evidence below 1.5 mg/L. That figure belongs to the court’s account of the experts’ dispute. It should not be repackaged as a universal threshold, a safe level, or a finding by the Ninth Circuit. The appellate court did not decide the scientific merits.

This is the terrain where toxicology evidence in federal court can become either legible or distorted. Expert witnesses may work with different kinds of evidence: population studies, exposure estimates, biological mechanisms, or dose-response analyses. Each answers a different question. An epidemiological association does not by itself establish a mechanism. A plausible mechanism does not by itself quantify risk for a particular population. Exposure estimates depend on their inputs. The court must assess the opinions in the context of the record, not collapse them into a single dramatic number.

The admissibility of fluoride research in court therefore depends on the actual opinion and its foundation. Counsel should be ready to explain why a study is relevant, what it can and cannot show, and how the expert moves from the study to the conclusion offered. Where opposing experts disagree, the disagreement should be made precise: which data, assumption, method, or inference is contested? General claims that the science is settled, or that it is too uncertain to consider, do little work unless the record supports them.

TSCA adds a procedural layer. Section 21 separately provides that a court may award reasonable expert-witness fees when it finds such an award appropriate. That provision is not an assurance that fees will be awarded, and it does not determine whether an expert opinion is admissible. It is one part of the statute’s litigation framework, not a substitute for building a defensible evidentiary record.

What a careful litigation strategy can take from the case

The case offers lessons in discipline, not a recipe. Expert evidence must be tied to the claims and to the record. The witness must be prepared to defend the steps between data and conclusion. The parties must preserve the boundaries of what they ask the court to decide. And any account of the result must track the procedural posture accurately.

The district court’s 2024 findings remain part of the case history. The Ninth Circuit’s 2026 vacatur changes their status. The scientific questions that drove the dispute remain contested in the record described here, while the appellate ruling leaves them unresolved. That is the map: a detailed trial record, a vacated decision, and a remand shaped by procedural limits.

In fluoride litigation, expert testimony can carry substantial weight. It cannot carry the case past the record the parties built, the rule the court applies, or the question the appellate court actually decided. Companies and advocates alike may prefer a cleaner story. The docket is less accommodating.