TSCA fluoride lawsuits: building the administrative record

In 2018, a federal judge rejected that limit: review under Section 21 of the Toxic Substances Control Act could proceed de novo, with evidence beyond the record the agency had assembled. Seven years later, a district court found that fluoridation at 0.7 mg/L presented an unreasonable risk of reduced IQ in children. Then, in May 2026, the Ninth Circuit vacated that ruling—not because it resolved the scientific dispute, but because the district court had taken control of the record.

That sequence is the legal terrain for anyone following a TSCA fluoride lawsuit against the EPA. The phrase “administrative record” sounds like a filing cabinet. In litigation, it can become a fight over who gets to put evidence before the judge, who selects it, and what the court is allowed to do with it. The latest appellate ruling makes one boundary especially clear: a court may allow parties to present evidence in a de novo case, but it cannot commandeer the record itself.

De novo review changed the evidence fight

A TSCA Section 21 case begins with a citizen petition to the EPA. Under 15 U.S.C. § 2620, petitioners ask the agency to initiate a rulemaking or take another action available under the statute. If the EPA denies the petition, the statute allows a challenge in federal court.

That path matters. A petitioner does not simply appeal an agency decision through the same process used for every form of judicial review. Section 21 provides for de novo review. In the fluoride litigation, the EPA argued that the court should limit its review to the administrative record. On February 7, 2018, Judge Edward M. Chen of the Northern District of California denied the agency’s motion.

The practical consequence is often misunderstood. De novo review does not mean the agency record disappears, or that a court must accept every study or claim a party offers. It means the court is not confined to asking whether the EPA’s decision can be defended from the materials the agency had before it. The parties can present evidence in court beyond the original administrative record.

That distinction gives plaintiffs room to build a litigation record. It also leaves them responsible for doing the work. A case still needs evidence tied to the statutory question, competent presentation, and a clear explanation of how the material supports the requested relief. “De novo” is a rule about the court’s review; it is not a magic phrase that makes weak evidence persuasive.

For readers tracking TSCA fluoride lawsuit EPA administrative record requirements, the first question should therefore be precise: which record is under discussion?

• The EPA’s administrative record contains material the agency considered in handling the petition.
• The litigation record contains evidence the parties present to the court.
• The trial record is shaped through the parties’ submissions and the court’s ordinary procedures.

Those categories overlap, but they are not interchangeable. Treating them as one undifferentiated pile invites procedural trouble.

De novo review opens the courthouse door to evidence beyond the agency file. It does not hand the judge a license to build the case.

What the 2024 ruling did, and what it did not do

On September 24, 2024, the district court ruled that water fluoridation at 0.7 mg/L presented an unreasonable risk of reduced IQ in children under TSCA Section 6(a). That was a consequential finding in the case. It was not the final word on federal water policy, and it should not be described today as an operative nationwide ban or mandate.

The distinction is no small technicality. A district-court ruling on statutory risk can shape the next phase of litigation and attract intense public attention. But its legal status depends on what happens next. On May 21, 2026, the Ninth Circuit vacated the 2024 order and sent the case back to the district court. The appellate court’s ruling addressed how the record had been handled. It did not settle the underlying scientific merits of fluoride neurotoxicity under Section 21.

That leaves two questions running on separate tracks. One concerns the evidence and the alleged risk. The other concerns the court’s authority to manage the case. A ruling on the second can erase the first court’s order without answering whether the scientific claim is right.

This is where public summaries often become sloppy. “The court ruled fluoridation unsafe” compresses a procedural history into a headline and can imply a present legal result that no longer exists. “The appeals court cleared fluoridation” makes the opposite mistake. The Ninth Circuit vacated the district court’s order over the party-presentation principle. The supplied case history does not establish that the appellate court resolved the substantive risk question in EPA’s favor.

For an EPA citizen petition vs. federal lawsuit, the handoff is the key. The petition asks the agency to act under TSCA. A lawsuit challenges the agency’s handling of that petition and invokes the statute’s judicial-review framework. The court case can involve evidence outside the agency’s file, but the parties—not the judge acting as an investigator—must present the case they want decided.

The Ninth Circuit drew a line around the judge’s role

The Ninth Circuit’s May 2026 decision turned on the party-presentation principle. In plain terms, courts decide disputes based on the case the parties put before them. The district court, the appellate court held, violated that principle by commandeering the record.

That holding sits beside the 2018 ruling, not in place of it. The 2018 decision rejected the EPA’s attempt to confine de novo review to the agency’s administrative record. The 2026 decision did not restore that rejected limit. Instead, it addressed who controls the evidence presented in court.

The difference is narrow enough to be missed and important enough to drive a remand. Plaintiffs may introduce new evidence in a de novo Section 21 case. The judge must remain a neutral arbiter, rather than expanding the trial record independently beyond what the parties choose to present. A court that takes over that task risks crossing from judging the dispute into assembling it.

That is a serious constraint for both sides. Plaintiffs cannot expect the court to find missing support or repair an incomplete presentation. The EPA cannot rely on the administrative file as an automatic wall against all outside evidence. Each party has to identify the material it wants considered, explain its relevance, and meet the applicable procedural requirements.

The appellate ruling also changes how the 2024 order should be discussed. It has been vacated. The case has been remanded. The final outcome of the remanded proceedings, limited to the record the parties present, is not established in the available facts. Nor is there a settled answer here on whether the Ninth Circuit or the Supreme Court will ultimately rule on the substantive merits of fluoride neurotoxicity under Section 21.

That uncertainty is not a loophole for confident predictions. It is the present posture of the case.

Building evidence without confusing the records

The phrase “administrative record preparation for fluoride litigation” can send petitioners down the wrong route if they assume that collecting a thick agency file will settle the court case. The agency record matters because it shows what the EPA considered and how it handled the petition. But de novo review means litigation evidence can extend beyond that record.

A disciplined evidence plan should separate the work into stages.

1. Preserve the petition trail. Keep the citizen petition, the EPA’s response, supporting submissions, and correspondence in an organized file. These documents show what was requested and how the agency acted on it. They also help identify what belongs to the agency’s decision-making history.

2. Identify the claim each item supports. A scientific study, exposure estimate, or water-system document should not sit in the file as decoration. The party presenting it needs to explain its connection to the statutory issue and the relief sought. A record full of material without a clear evidentiary purpose is still a poor presentation.

3. Distinguish agency materials from litigation evidence. Labeling everything “the record” obscures the procedural question. Keep a clear account of what the EPA had, what the parties later offer, and what the court actually receives through the litigation process.

4. Present the case through the parties. The Ninth Circuit’s ruling makes this point hard to dodge. Parties should identify and submit the evidence they want the court to consider. They should not build a strategy around expecting the judge to expand the trial record independently.

5. Track the order’s procedural status. A district-court finding, an appellate vacatur, and a remand are different legal events. Any public account should state which order remains in force and what issues the appellate court actually decided.

These steps are not a substitute for legal counsel or court rules. They are a way to avoid a basic category error: treating the agency’s file, the parties’ evidence, and the court’s record as one thing.

What the 0.7 mg/L figure can and cannot tell you

The 2024 district-court ruling concerned fluoridation at 0.7 mg/L, the recommended U.S. drinking-water concentration noted in the case materials. That figure anchors the dispute. It does not, by itself, answer every question about exposure, risk, statutory standards, or the scope of any remedy.

A concentration is a measurement. A legal claim requires a chain of reasoning connecting the evidence to the statutory standard. The court must assess the case presented by the parties; the number alone cannot carry that burden. Nor does the existence of a district-court finding mean that the same legal conclusion now governs after the Ninth Circuit vacated the order.

This is where documenting water toxicity for legal challenges demands more care than collecting alarming excerpts. Evidence has to be organized around the actual claim and presented in a form the court can evaluate. The supplied case history does not provide a complete evidentiary inventory or settle which materials will control on remand. It would be reckless to pretend otherwise.

For communities following the litigation, the useful questions are procedural as much as scientific:

• Has a petition been filed under TSCA Section 21, and what did the EPA decide?
• Is the dispute at the petition stage, in district court, on appeal, or back on remand?
• Is a cited finding still operative, or has a later court vacated it?
• Does the claim concern the EPA’s administrative process, the evidence of risk, or both?
• Which party put the evidence before the court?

The answers map the route through a case more reliably than a headline announcing a win or a defeat.

The next case will turn on discipline

The fluoride litigation now carries a paired lesson. Section 21 de novo review can permit evidence beyond the EPA’s administrative record. The Ninth Circuit has also warned that the district court must remain a neutral decision-maker and cannot take over the parties’ role in building the trial record.

That is a more demanding framework than either side’s cleanest slogan. Plaintiffs cannot say the agency record is irrelevant. The EPA cannot say it is the only evidence that matters. And the court cannot fill gaps by assembling a case of its own.

The immediate legal route is a remand, with the final outcome still unresolved in the available record. Anyone watching for the next ruling should keep the procedural posture in view and separate it from the underlying scientific dispute. That separation is tedious. It is also where liability arguments, agency defenses, and public claims stop masquerading as the same thing.

The courthouse has not issued a final answer on the merits. It has issued a warning about who must do the presenting. In litigation, even the strongest evidence can lose its force when the wrong actor puts it on the record.