Fluoridated Water

Grand Rapids, Michigan. January 1945. City officials begin receiving complaints about sore gums and damaged teeth. Residents blame the fluoride being added to the local water supply.

It turns out their water wasn't fluoridated yet. The local press had made a mistake, announcing that fluoridation would begin several weeks before it actually did.

Why would anyone complain about the side effects of a change that hadn't occurred yet?

Perhaps chemophobia (an irrational fear of chemicals), perhaps something else. One thing for sure is that since the 1940s, we've experienced quite a bit of irrational opposition to fluoridated water (including the extreme view, lampooned in the 1964 film Dr. Strangelove, that it's a communist plot).

In recent months, the opposition seems less irrational than it used to be:

—On August 21, the National Toxicology Program published a review linking fluoride exposure to lower IQ among children.

—On September 24, a federal judge ordered the EPA to increase regulation of fluoridated water.

—Last Friday, a new review showed that fluoridated water benefits dental health less than previously found.

The latter prompted some mainstream news outlets, like CNN, to suggest that the cost-benefit ratio of fluoridating water has changed. The practice may even do more harm than good, according to some journalists. Is that a reasonable inference?

Why this is important

Fluoride is a naturally-occurring mineral found in water as well as a variety of foods ranging from shellfish to black tea. According to the CDC, fluoride is also added to the public water supply for 72.3% of Americans. This isn't a federal requirement, although the EPA does set maximum exposure limits. It's up to state and local governments whether or not to fluoridate the public water supply.

Since 2023, more than 170 communities (including Portland, OR) have banned or refused fluoridation, and others are currently debating the issue, including three states (Georgia, Kentucky, and Nebraska) considering bills to reverse state-wide fluoridation mandates. Unfortunately, the debates are rife with propaganda, misinformation, and conspiracy thinking.

We need to understand the data on fluoridated water. Is fluoridation beneficial, and if so, what levels are safe? Do federal recommendations and standards protect us and our children? For some communities, the practical question is how much fluoride can be safely added to the water; for others, the question is how much to filter out.

The benefits of fluoridated water

There's no question that fluoride helps prevent tooth decay by strengthening tooth enamel and reducing bacteria. (Dr. Andrea Love provides a deep, accessible dive into the chemistry here.)

"Colorado brown stain"

In 1901, a young dentist named Fredrick McKay moved to Colorado Springs to open a new practice and discovered that most of the residents had brown, mottled teeth. Local practitioners called it "Colorado Brown Stain." As McKay and a prominent colleague tried to figure out what caused the disorder, they discovered that the brown teeth were surprisingly resistant to decay.

Within a few decades, scientists figured out that high levels of naturally-occurring fluoride in the water of Colorado Springs (and other places around the world) were responsible for the brown stains as well as more decay-resistant enamel. On January 25, 1945, Grand Rapids became the first city to add fluoride to its public water supply. Since then, seven decades of research have linked fluoridated water to reductions in tooth decay, particularly among children.

Is fluoridated water still beneficial?

Last Friday, a review led by Dr. Zipporah Iheozor-Ejiofor (University of Central Lancashire) appeared in the influential Cochrane Database of Systematic Reviews. Iheozor-Ejiofor's review was an update of a 2015 Cochrane paper.

Iheozor-Ejiofor and colleagues looked at 21 studies published between 1959 and 2022 on the effects of adding fluoride to a water supply that contains little or none of the mineral. Each of these studies included a "control group" or comparable water supply where no fluoride was added. The studies tracked dental outcomes for anywhere from 3 to 12 years. The main outcome variables were measurements of decayed, missing, or filled teeth (DMFT). (The results were essentially the same for baby teeth and permanent teeth, so I'll just refer below to "teeth".) Here are some key findings:

(a) In areas where fluoride had been added to the water, children showed significantly greater mean reductions in DMFT.

(b) In areas where fluoride had been added to the water, significantly higher percentages of children exhibited no tooth decay.

(c) The benefits of fluoridating the water were smaller in recent studies than in pre-1975 studies.

Although not part of the review, the authors also cited independent data suggesting that adults also experience less tooth decay when fluoride is introduced to their water.

In sum, fluoridated water reduces tooth decay among children, but not to the extent that it used to. Presumably this is due to the increasing prevalence of fluoridated toothpaste and other products, though the data doesn't directly demonstrate this.

As for the extent of benefit, whether it's water or toothpaste or anything else, there's not much truth value (should that be tooth value?) to any of the specific estimates. The research mostly documents average effects, and there's a lot of variability from one study to the next in findings. All we can be sure of is that fluoridated water reduces decay; we can't make blanket statements about how much, since people differ in dental health, in the amount of fluoride they're exposed to from various sources, and so on.

The risks of fluoridating water

Some activists label fluoride a toxic chemical. This is not wrong but meaningless.

Fluoride is indeed harmful if you're exposed to enough of it, but one could say that about any substance. Even water can kill you if you drink too much too quickly (a tragedy that occurs on occasion among endurance athletes and military trainees). As the ancient maxim goes, the dose makes the poison. In the case of fluoride, ingesting too much can cause dental fluorosis (the technical term for "Colorado Brown Stain"), skeletal fluorosis (a painful condition affecting bones and joints), and other health problems.

In short, it's meaningless to label fluoride as either toxic or safe. Some amount is unsafe, and we need to know how much. Can we fluoridate water enough to prevent tooth decay without endangering anyone's health? This is a scientific question that can't be answered by obnoxious billboards or fearful parents screaming at city council members.

New data: The NTP report

This August 21, the Department of Health and Human Service's National Toxicology Program (NTP) published a review linking fluoride exposure to lower IQ among children.

The NTP reviewers identified 19 high-quality studies on the relationship between fluoride exposure and children's IQ. 18 of these studies showed an inverse relationship: The more fluoride children were exposed to, the lower their IQs.

At first glance this is stunning. 18 out of 19 studies show that fluoride, a mineral most of us are exposed to in our water and toothpaste, undermines children's intelligence. But the devil is in the details. If we stop here, without looking at least a few of those details, we will deeply misunderstand what the data says.

The NTP reviewers focused on "high" levels of fluoride exposure, which they defined as anything exceeding the World Health Organization's safe water guideline of 1.5 milligrams of fluoride per liter of drinking water (1.5 mg/L). That's 1.5 parts per million, or roughly one teaspoon of fluoride mixed into 10 ordinary bathtubs full of water.

The reviewers concluded that once drinking water exceeds 1.5 mg/L of fluoride, the higher the levels of fluoride, the lower children's IQs will tend be.

The reviewers aren't claiming that something special happens at precisely 1.5 mg/L. The original studies were very different from each other in methods and findings, and only a few of them showed IQ effects marked off at or near the 1.5 mg/L value. Rather, the NTP reviewers looked at the totality of the evidence before deciding, with moderate confidence, that fluoride exposure above 1.5 mg/L is linked to lower IQ. As you can imagine, the American Dental Association and other stakeholders are upset by this conclusion and have pushed back.

Can we trust the data?

I wouldn't call the NTP review flawed or meaningless. However, the evidence does seem weaker than the review suggests, and if there's cause for concern, we should be only be worried about a small minority of children.

1. The strength of the evidence is overstated.

The NTP team took great pains to identify potential bias in the original 19 studies and in the way they gleaned information from them. All the same, the review feels biased. Specifically, biased toward the view that fluoride diminishes IQ. For instance:

(a) In one study, the fluoride-IQ association was actually nonsignificant. The reviewers shouldn't have counted this study among those showing an association, because the significance game isn't meant to be played this way. You have to either follow the rules or use different rules, and statisticians do provide many options.

(b) In one study, the researchers administered 15 IQ measures to a total of 51 children, finding significant effects for just two of those measures. The researchers didn't control for Type I error. Informally speaking, this means that by performing many analyses, the researchers ran the risk of some turning up significant just by chance, but they didn't control for that risk. I don't think the reviewers should have trusted this study.

(c) In one study, the researchers used the wrong statistical test to compare groups. (A paired-samples t-test was used to compare a total of 93 children from villages with high vs. low levels of fluoride in their water. These are independent samples.) On general principle, we shouldn't trust a study where the main statistical comparison fails to reflect a simple distinction taught to undergraduates in their first statistics class. (I'm not trying to be snarky. Children's health is at stake here.)

(d) In one study, the fluoride-IQ association isn't linear. If fluoride is damaging, you'd expect more exposure to cause greater damage. That's not what the study shows. Instead we see, for instance, an association in the range from 1.60 mg/L to 2.50 mg/L, but not at a lower range (0.01 mg/L to 1.60 mg/L) or at a higher range (2.50 mg/L to 5.54 mg/L). How could a substance be increasingly toxic within a medium range of exposure but not within a higher range? The researchers speculate vaguely about saturation, but this doesn't make sense. The saturation point here would be brain damage to the point that one couldn't take an IQ test. I'm not sure how much to trust this study.

(e) In one study, effects on IQ were observed among boys but not girls. If fluoride is harmful, you'd expect damage to occur regardless of biological sex. The researchers couldn't account for the sex difference, much less the fact that exposure to fluoridation seems to have very slightly improved girls' IQ. (Since boys are neurologically more vulnerable than girls in some respects, and since the literature hasn't explored sex differences systematically, I'm on the fence about this study.)

(f) Most of the other studies were cross-sectional and/or relied on on single samples of urine, a method that many experts consider an unreliable indicator of long-term fluoride exposure.

2. The evidence concerns high levels of exposure.

Ignore everything I just wrote and assume for a moment we can trust the data. The reviewers note that they found no clear evidence of a fluoride-IQ relationship at exposure levels below 1.5 mg/L.

The CDC estimates that only about half of one percent of Americans (0.59%) use water that contains more than 1.5 mg/L. This suggests that the vast majority of American children are not exposed to worrisome levels of fluoride in their drinking water.

A residual concern

My takeaway is that the NTP review mainly raises concerns about that tiny percentage of children exposed to fluoride levels above 1.5 mg/L. These include children whose families use private sources such as well water, particularly in regions in the southwest and elsewhere where fluoride levels are naturally high.

The reviewers found no clear evidence that fluoride undermines intelligence at lower exposure levels. That's good news but, to tweak an old saying, absence of clear evidence doesn't equal clear evidence of absence. Is there any chance that lower levels could be harmful?

I reached out to Dr. David Eaton, a professor emeritus at University of Washington. Dr. Eaton, a leading toxicologist, chaired the original committee of experts tasked with evaluating the NTP's response to external feedback on an earlier version of the document.

Dr. Eaton addressed an important issue I haven't touched on here: The difficulty of measuring IQ precisely, particularly when studying children.

"One of the biggest challenges in assessing impacts at lower doses is that assessment of the adverse outcome (generally IQ measurements) is not very robust, and there are many other variables that affect children’s IQ and must be controlled. Even the tool(s) used to measure IQ have a lot of noise, making it virtually impossible to accurately determine a change in IQ of 1-2 points."

(I suspect the NTP authors would respond that the consistency of the fluoride-IQ relationship across studies outweighs any imprecision in measuring IQ, but, as I've noted, one could question the validity of many of those studies. IQ measurement issues join a long list of reasons one could question the findings.)

Dr. Eaton reiterated that we don't know whether doses of fluoride lower than 1.5 mg/L are harmful, and thus the data doesn't support the notion that there's no safe level of exposure. With respect to lower levels, he used the U.S. Public Health Service's 0.7 mg/L recommendation as a reference point:

"[T]he impacts, if any, at the current recommended level of 0.7 mg/L are very likely to be insignificant, if not zero. What is lacking thus far is a really balanced assessment of risks and benefits at lower levels (e.g., < ~1 ppm) for infants and children..."

In other words, even if lower levels of exposure introduce small risks, we would need to balance those risks against potential benefits, just as we do with any other health practice. Vaccines, for instance, have side effects, but most health professionals believe the benefits vastly outweigh the risks.

Conclusions – and what you can do

The CDC has listed the fluoridation of drinking water among the ten great public health achievements of the 20th century, and I'm inclined to agree.

Fluoridated water prevents tooth decay. The new Cochrane review suggests small benefits over and above exposure to fluoride from other sources. Children benefit most; adults seem to benefit as well. We can't treat fluoridation as a substitute for good dental hygiene, but it's clearly helpful.

At sufficiently high, sustained doses, fluoridated water may affect children's intelligence, but if we can trust the somewhat questionable findings of the NTP review, effects are only observed at fluoride levels that exceed what only about half of one percent of American children are exposed to.

Personally, I'm not concerned about the water that my grand-babies are drinking. I moved to Rockville, Maryland recently. According to the city's website, the goal is for our water to contain the equivalent of 0.4 mg/L fluoride; the actual range has been 0.41 to 0.64. I anticipate benefits for dental health but no particular health risks.

If you have concerns, you can check the CDC database on fluoridated water systems, reach out to your local water provider for details, or procure a test for use at home (the tests are cheap and, in some places, free). Neither boiling nor conventional water filters remove fluoride, but you can purchase specialized filters that do the job. Most bottled water products contain no fluoride, but I wouldn't recommend buying more plastic than you need to. Plastic is...well...a topic for another day.

Thanks for reading!