A Memphis Team Linked Tap-Water Fluoride to Kids' Testosterone, Sodium, and Uric Acid

I always filed fluoride under dentistry. Cavities, the little foam tray, the hygienist telling me to swish and spit and not swallow. In my head it was a mouth story, sealed off from the rest of the body, and I had never asked the obvious follow-up: after you swallow the mouthful you did not spit, where does the rest of it actually go?

A study published September 26 in the journal Environmental Health asked exactly that, in 4,470 American children and teenagers, and the answer is not “the teeth.” Researchers in the Division of Epidemiology, Biostatistics, and Environmental Health at the University of Memphis pulled national survey data from 2013 to 2016 and linked household tap-water fluoride to predicted plasma fluoride, and then to places fluoride is not supposed to be part of the conversation: blood pressure, serum sodium, uric acid, and testosterone.

What makes this one hard to wave away is the design. Most fluoride studies just correlate a kid’s measured plasma fluoride with their lab values, and that move is easy to dismiss, because sicker or poorer or differently fed children differ in a thousand ways at once. This team used the fluoride concentration in each household’s tap water as an instrument instead, a stand-in set by geography, water source, and what the municipal plant decides to do, not by the child. Where you live decides how much fluoride pours into your glass. An instrument like that is stronger than an ordinary correlation, but it leans on its own assumptions, and it is still not the randomized trial nobody will ever run on children.

And here is what it found, expressed per one micromole-per-liter increase in predicted plasma fluoride. Testosterone came in at a geometric mean ratio of 0.49, roughly half, with a confidence interval running from 0.30 to 0.80. Serum sodium rose by 3.41 millimoles per liter. Uric acid climbed 0.92 milligrams per deciliter. Systolic blood pressure dropped 7.91 millimeters of mercury. That testosterone figure is not a claim that every child on fluoridated water carries half the hormone. It is what the model predicts for each additional unit of fluoride in the blood, and the exposure gap between a low-fluoride and a high-fluoride town can be wide.

Wait. Why would the mineral in your tap water reach down into a hormone built in the testes and ovaries? That was the question I could not let go of, and the biology has a fairly direct answer. In lab dishes, sodium fluoride disrupts testosterone synthesis in Leydig cells, the exact cells whose whole job is assembling the hormone from cholesterol, by jamming the steroidogenic line partway through. So fluoride touching the gonads is not magic, it is plausible: a small, reactive ion that circulates and settles on enzymes, and steroid-making is a chain of enzymes. Whether ordinary tap-water levels hit those enzymes hard enough in a living child to explain these numbers is the one thing a cell dish cannot tell you. What it does is make the association biologically believable instead of a fluke. And it is not the first hint. An earlier analysis of the same national survey found male adolescents in the top third of plasma fluoride had about 21 percent less testosterone. The new study, with its cleaner design, points the same way, harder.

What convinces me most is the sodium and uric acid moving together, because it points at one organ. Your kidney clears fluoride from the blood, and your kidney also sets serum sodium and pulls urate out of circulation. Lean on the tubules that handle the first job and you would expect the other two numbers to drift, which is the pattern here. Animal work backs the plausibility: fluoride at high doses inflames the kidney through the NF-κB pathway, the body’s master switch for inflammatory signaling. High doses in mice are not tap water in kids, so read the pattern as kidney-consistent, not kidney-proven: a mechanism that could exist, not one this study caught in the act.

I want to be square about the wobbly one. The blood pressure result is the softest of the four. Its confidence interval nearly touches zero at the top (the drop could be as small as 0.68 millimeters), and a lower pressure is an odd direction to build a worry around. The demographic breakdowns were noisy too, significant differences by sex, age, and race but no clean story about who gets hit hardest. This is observational, cross-sectional data. It shows associations, not a stopwatch-clean cause, and the authors say so plainly. It cannot tell you whether a halved teenage testosterone level rebounds in adulthood or leaves a mark. Hold that uncertainty honestly. Then hold it next to the last two years.

In September 2024 a federal judge ruled that fluoride at 0.7 milligrams per liter, the exact level the government endorses for community water, poses an unreasonable risk to children’s IQ and ordered the EPA to act. It was the first time a citizen petition under the toxics law reached trial and won. In April 2025 the HHS secretary moved to tell the CDC to stop recommending fluoridation outright. The dental establishment’s defense for eighty years has rested on one claim: the benefit is topical, it works on the surface of the tooth, so swallowing it is incidental and harmless. This study, and a 2026 paper reporting fluoride’s systemic footprint in children’s urinary proteins, keep saying the same inconvenient thing. The swallowed fraction does not vanish. It circulates, and it lands on organs.

The line I keep circling back to is the funding statement. The authors declared no external funding. No agency, no manufacturer, no advocacy group is named as having paid for it. An independent look at whether the fluoride added to most of the country’s tap water does anything to kids’ hormones, with no sponsor to keep happy, and the numbers came back loud.

So this is my conclusion. I am not going to pretend one clever survey analysis settles a fight the whole field is still having. But I have a countertop fluoride filter, it costs less than a month of streaming, and my kids drink from it. I did not need a guideline change to make that call, and I am not waiting for one. When the dental associations and public-health agencies that spent decades reassuring you keep insisting the swallowed part stays in your mouth, and the unfunded data keeps finding it in the blood, I would rather filter now and argue later.

Sources

1. Environmental Health, plasma fluoride and systemic biomarkers in U.S. children (2026)
2. Environmental Research, fluoride and sex steroid hormones in U.S. adolescents (2020)
3. Science of the Total Environment, sodium fluoride disrupts testosterone biosynthesis in Leydig cells
4. Scientific Reports, sodium fluoride induces renal inflammation via NF-κB signaling in mice
5. CBS News, federal court rules fluoride in drinking water poses unreasonable risk to children (2024)
6. STAT, RFK Jr. plans to tell CDC to stop recommending fluoride in drinking water (2025)
7. Ecotoxicology and Environmental Safety, urinary proteomic biomarkers of systemic fluoride exposure in children (2026)