I blame my dentists. Not for poor dental care—Barbara and Gordon do great work. I blame them for sending me into a vortex of dento-epistemological anxiety.

On a tooth-cleaning visit not long ago, Barbara told me that in the late 1970s, when she attended dental school, her professors expected that most middle-class patients would lose a lot of their teeth and need dentures by the time they were in their 60s. Today, she said, most middle-class people keep their teeth until they are 80. The main reason for this, Barbara explained, was fluoridation—the practice of putting fluoride compounds in community drinking water to combat tooth decay.

To hear more feature stories, get the Audm iPhone app.

For reasons I can’t now recall, I mentioned this remark on social media. The inevitable but somehow surprising response: People I did not know troubled themselves to tell me that I was an idiot, and that fluoridation was terrible. Their skepticism made an impression. I found myself staring suspiciously, as I brushed, at my Colgate toothpaste. strengthens teeth with active fluoride, the label promised. A thought popped into my head: I am now rubbing fluoride directly onto my teeth. So why is my town also dumping it into my drinking water?

Surely applying Colgate’s meticulously packaged fluoride paste directly onto my teeth, where it bonds with the surface to create a protective layer, was better than the more indirect method of pouring fluoride into reservoirs so that people drinking the water can absorb the fluoride, some of which then makes its way into their saliva.

Then I wondered: How much fluoride is in my water, and how did public-health officials set the dose? Fluoride in large quantities is bad news. Potential side effects, I quickly discovered, include joint pain, bone fractures, sperm decline, dementia, premature puberty, gastrointestinal distress, immune-system dysfunction, (possibly) cancer, and (also possibly) lower IQ in children. Children have smaller bodies than adults and thus are at risk of relatively greater exposure when they drink. In calculating the dose, I thought, the authorities must have taken into account the weird thirsty kid who guzzles water by the quart. But if they lower the dose to avoid harming that child, where would that leave my mother-in-law, who for some reason has decided she no longer wants to drink much water at all? Is she getting shortchanged?

Fluoridation of public water supplies is backed by every mainstream dental organization in the nation and opposed by a lot of people who spend too much time on YouTube. The most-watched anti-fluoridation video in my YouTube search results—from the series Stuff They Don’t Want You to Know—hauls out the specter of Nazi Germany before the one-minute mark. Another video, from the series Brainwash Update, states categorically that “fluoride is poison.” It has the high production value one associates with its sponsor, Russia Today. When I was growing up, anti-fluoridation campaigns were the province of the John Birch Society and other right-wing cranks. Now I myself seemed to have become a candidate for the tinfoil-hat brigade.

Yet the more I looked, the more I realized that fluoridation encapsulates several recurring medical dilemmas. How much trust should we give to expert judgment? How much potential harm can we expose one group to in the course of helping another? And how much evidence should be required before we allow governments to force people to do something for their own good?

photo of model of teeth and toothbrush in jar
Sergiy Barchuk

Modern dentistry is a formidable example of human progress. Our grandparents’ jaws used to hurt all the time. Tooth decay plagued everyone—rich and poor, famous and obscure. George Washington, an affluent planter, had lost all but one of his teeth by age 57, when he was first sworn in as president. His quest to fill his mouth led him to wear sets of dentures made from his own pulled teeth, from animal teeth (donkey and horse up top, cow on the bottom), and from other people’s teeth, possibly including those of his slaves.

Washington was not alone. People on both sides of the Atlantic participated in a lively black market in cadavers’ teeth. Fortunately for denture customers, Europe had a ready supply. Scavengers followed wartime armies, according to the medical historian Lindsey Fitzharris. After the shooting stopped at the battle of Waterloo, many of the dead were toothless within hours.

The widespread introduction of sugar worsened society’s dental difficulties. In the first decades of the 20th century, American dentists regularly made full sets of dentures for teenagers so that they would look presentable at graduation. American soldiers were required to have a minimum number of opposing teeth: six on the top, six on the bottom. Thousands of would-be doughboys and GIs were barred from service in the First and Second World Wars for failing to meet this standard.

So dire was the state of U.S. dentition that in 1901, Frederick McKay’s discovery that many of his patients’ teeth were mottled with ugly brown stains generated little notice. McKay was a dentist in Colorado Springs. Intrigued, he and two colleagues examined 2,945 schoolchildren for what they called “Colorado stain.” To their shock, 87.5 percent had stained teeth.

McKay contacted a famous Chicago dentist (famous in dental circles, anyway) and got him to describe the syndrome to the Colorado state dental association. Hardly anyone paid attention. Trying again, McKay and the Chicago dentist evaluated students at Colorado College, in Colorado Springs. They found that students raised in Colorado Springs had discolored teeth, whereas students from other areas had normal teeth. Hardly anyone paid attention. The two researchers then published an article, “An Investigation of Mottled Teeth: An Endemic Developmental Imperfection of the Enamel of the Teeth Heretofore Unknown in the Literature of Dentistry.” Unknown in the Literature of Dentistry! Still, hardly anyone paid attention.

In the 1930s, McKay and others identified the staining agent: naturally occurring fluoride compounds in water supplies. (This kind of staining, along with the other negative effects of fluorine absorption by bones and ligaments, is now called fluorosis.) The researchers also discovered something else: Although the staining looked terrible, people with fluoride stains had fewer decayed and missing teeth. A small group of dentists began agitating to add low levels of fluoride to drinking water—low enough to avoid staining and also low enough to be safe.

Those dentists would soon get corporate reinforcement. Fluorine, a chemical element, is lethal in small doses and extremely reactive. Fluorides—compounds of fluorine—can be nearly as toxic but are much more stable. They are a common waste product of the fertilizer, pesticide, refrigeration, glass, steel, and aluminum industries. In the ’30s, many of these industries were facing protests and lawsuits for poisoning workers, polluting the soil, and contaminating water supplies. Understandably, executives were thrilled to discover that the chemicals they had to get rid of because they could seep into city water systems might be gotten rid of by being jettisoned into city water systems. Less understandably, some later anti-fluoridation activists described the corporate embrace of fluoridation as evidence of a Communist plot.

It was more like a capitalist plot. From 1921 to 1932, the secretary of the Treasury was Andrew W. Mellon, a founder of the Aluminum Company of America, better known as Alcoa. The U.S. Public Health Service was then under the jurisdiction of the Treasury Department. In January 1931, Alcoa chemists discovered high levels of fluoride in the water in and around Bauxite, Arkansas, an Alcoa company town. By May, at Mellon’s urging, a Public Health Service dentist had been assigned to examine the link between fluoride and reduced cavities. Eight years later, a biochemist at the Mellon Institute, in Pittsburgh, became the first researcher to call for the widespread fluoridation of water.

Additional impetus came during the Second World War. The Manhattan Project—the crash effort to develop the atomic bomb—processed uranium by combining it with huge amounts of fluorine to form uranium hexafluoride. Large quantities of other fluoride compounds, including the DuPont refrigerant Freon, were needed. Accidents exposed employees to these little-understood substances, killing some and sickening others. Fearing litigation, the Manhattan Project created a “medical section” to study fluorides. Together with industry, it pushed for clinical trials of fluoride’s effects. Under the guise of protecting teeth, the Manhattan Project set about obtaining data on long-term fluoride exposure.

Starting in 1945, tests were conducted in Grand Rapids, Michigan, and Newburgh, New York. Both cities added fluoride to their water. In both cases, the control was a nearby city that did not add fluoride. The experiments were supposed to continue for at least a decade, with dentists in each city examining their patients to evaluate long-term effects. As it happened, one of the control cities fluoridated its water within seven years because its citizens had heard rumors about the benefits.

Fluoridation took off. So did the anti-fluoride movement, a loose coalition of Christian Scientists, Boston society ladies, chiropractors, biochemists, homeopaths, anti-Semites, and E. H. Bronner, the spiritualist soap-maker. A woman named Golda Franzen, from San Francisco, testified before Congress in the early 1950s that fluoridation was a Communist plot to turn Americans into a race of “moronic, atheistic slaves.” Franzen was later convicted of violating state health laws for peddling a “cancer cure” machine consisting of a speakerless tape recorder that vibrated as it played “Smoke Gets in Your Eyes.”

The opposition mostly failed. At an annual cost of about $325 million, more than 70 percent of Americans now have fluoridated water. Still more Americans get fluoride from soft drinks, most of which are made with fluoridated water. Some bottled water is fluoridated too. In 2007, Grand Rapids, celebrating its historic role, erected a 33-foot-high powder-blue sculptural monument to fluoridation.

The fluoride revolution was not restricted to the United States. The Organization for Economic Cooperation and Development regularly surveys the progress of its 36 member nations. One variable it tracked until recently was the number of decayed, missing, or filled adult teeth in 12-year-olds, a measure of overall dental health. The top graph below depicts the results—uniformly positive—for six nations that have widely adopted fluoridation.

Graph: Tooth Decay in Countries With Fluoridated Water
Adapted from Harvard Public Health magazine

Graphs like this help explain why the Centers for Disease Control and Prevention in 1999 called fluoridation one of the top 10 public-health advances of the 20th century. Curiously, they also help explain why fluoridation is opposed by the surprisingly durable cohort of activists who barraged me on social media. The bottom graph, based on the same OECD surveys, tracks the number of decayed, missing, or filled adult teeth in 12-year-olds from countries that have not embraced fluoridation in a significant way or at all.

Graph: Tooth Decay in Countries Without Fluoridated Water
Adapted from Harvard Public Health magazine

The differences between the two graphs don’t leap out at the viewer. Nonfluoridated nations such as Belgium, Luxembourg, and Denmark actually have better dental health by this measure than the United States, one of the world’s fluoridation champions. Finland, Germany, Japan, the Netherlands, Sweden, and Switzerland tried fluoridation, abandoned it years later—and saw no rise in tooth decay. What’s going on?

One of the lesser-known advantages of government-run health-care systems, such as Britain’s National Health Service, is the fact that because taxpayers are funding everything, the government occasionally tries to determine whether the money is being spent usefully. In 1999, the government asked the NHS to “carry out an up-to-date expert scientific review of fluoride and health.” A research team based at the University of York evaluated every study of fluoridation it could find—about 3,200 of them. The team’s conclusion was, it said, “surprising.” Despite the long fight over fluoridation, few of the thousands of studies counted as “high-quality research.” The implication was that Britain had been tinkering with its water supply with little empirical support. Trevor Sheldon, the head of the York review’s advisory board, was blunt: “There’s really hardly any evidence” that fluoridation works, he told Newsweek. “And if anything there may be some evidence the other way.” These findings were respectfully ignored.

In 2015, the Cochrane organization waded into the debate. Founded in 1993, Cochrane is a London-based global network of about 30,000 medical researchers in multiple countries that provides systematic analyses of medical issues. The goal is to produce painstaking, rigorous assessments of what research has—and hasn’t—established about a given subject. Cochrane has a fiercely guarded reputation for impartiality and thoroughness. Its verdicts have global impact. Which may be why the pushback on its fluoridation work was so strong.

To evaluate the efficacy of water fluoridation, the Cochrane researchers wanted to select properly conducted scientific research, discarding studies that were badly designed (too few participants to produce sound data, for example) or incompetently executed (for instance, the researchers didn’t follow their own protocols). To evaluate the studies, the team used two simple but strict criteria: They needed to have two large groups of subjects, one with fluoride (the intervention group) and one without (the control group), and each group had to be examined at least two times. Moreover, the studies needed to be prospective (meaning the scientists announced beforehand what they were looking for, then measured it) as opposed to retrospective (meaning the scientists sifted through historical data looking for patterns). Scrutinizing medical databases, the Cochrane team found 4,677 fluoridation studies. All but 155 of them—20 that focused on tooth decay, and 135 that focused on dental fluorosis—failed to meet the two criteria. Worse, all of the tooth-decay studies and all but a handful of the fluorosis studies were, in the jargon, “at high risk of bias”—for example, variables such as age and income hadn’t been properly taken into account.

The Grand Rapids study is an example of these problems. Not only was it cut short when the control city, Muskegon, started fluoridating its water, but the experimenters had not established whether the two populations had similar incomes or ethnic backgrounds. Nor did the researchers evaluate people’s teeth blindly, by taking X-rays to be examined by technicians who did not know which group a patient belonged to. Instead the study dentists simply looked into patients’ mouths and subjectively reported what they saw—a recipe for what is called “confirmation bias,” in which people tend to interpret what they see in ways that reinforce their prior beliefs.

The Grand Rapids researchers cannot be much faulted for these lapses, according to the Cochrane spokesperson Anne-Marie Glenny, a researcher at the University of Manchester School of Dentistry. In the late ’40s and early ’50s, the proper procedures for clinical trials were just being established. Few scientists understood how small imbalances between the intervention and control groups could compromise an entire trial. And the researchers definitely cannot be blamed for the unhappy fact that their experiment—indeed, all of the original fluoride research—occurred before the introduction of Crest, the first fluoride toothpaste, in 1956. Today, given that almost all toothpaste contains fluoride, and that most people brush their teeth, assessing the impact of fluoridated water remains highly problematic.

“It’s a really difficult area to evaluate,” Glenny told me. “You can’t really do the ideal experimental study,” because it is next to impossible to assemble two large, similar groups of people, one of which is not drinking fluoridated water or brushing their teeth. On top of that, “measuring the confounders—sugar consumption, socioeconomic status, and so on—is really tricky.” How much, I asked, of the improved dental health of the ’60s and ’70s was due to water fluoridation? How much was due to the soaring popularity of fluoride toothpaste and mouthwash? And how much was due to rising affluence, which generally translates into more visits to the dentist? “I’m not sure you can answer that question,” Glenny said.

Sergiy Barchuk

The Cochrane group reported its work carefully. The evidence, it said, is poor and sparse, but what little there is “indicates” that the fluoridation of water reduces cavities in children. But, the group said, “these results are based predominantly on old studies”—from before 1975—“and may not be applicable today.” For adults, there is “insufficient evidence,” old or new, to determine whether fluoridation is effective. The report did not support or attack fluoridation; it only asked for more research.

Nonetheless, it set off an uproar. A blog post on the Cochrane website attracted so many vitriolic comments from anti-fluoridation zealots that the organization eventually removed it. When a writer for Harvard Public Health magazine used the Cochrane report to ask “Is Fluoridated Drinking Water Safe?,” the heads of the American Dental Association, the American Public Health Association, the American Dental Education Association, the American Association of Public Health Dentistry, the American Association for Dental Research, and the Harvard School of Dental Medicine demanded that the article be amended or taken down. (The story included earlier versions of the two charts in this article.) Fluoride, Glenny told me, is “the only topic that I’ve been involved in that has created so much angst and controversy.” The responses also critiqued the Cochrane report itself. The president of the American Dental Association said that it was “shaped by its unusually narrow inclusion criteria, excluding 97 percent of the more than 4,000 relevant studies that it identified.” In a joint letter, the president of the American Dental Education Association and the executive directors of the American Dental Association and the American Association for Dental Research concurred, scoffing at Cochrane’s “rigid inclusion criteria.” But the inclusion criteria were not “unusually narrow” or “rigid”—they were based on those in a standard textbook, now in its fourth edition. The implication of the dental experts’ critique seemed to be that if only statistical analysts would lower their standards, everything would look good.

The dental establishment’s argument for fluoridating water in a society where a majority of people use fluoridated toothpaste and go to the dentist boils down to a contention that fluoridation will likely help people who are unable to afford good dental care. The idea is that poor children don’t brush their teeth, and fluoridation will fill the gap—a notion, incidentally, that the Cochrane team found no good evidence to support. (Last year, JAMA Pediatrics published a large, careful study that suggested fluoridation gave extra benefit to poor children and adolescents, but it, too, had limitations—the authors could not establish whether the different families in the study ate similar amounts of sugar, for instance.) Still, the argument runs, it is ethically acceptable to force a majority to do something potentially useless if it might benefit a minority. Unless, of course, fluoridation at current levels is unsafe in some way, and the many are harmed in pursuit of a potential benefit for the few.

Is it safe? Some fluoride perils are well documented. Over the long run, the body incorporates fluoride into bone, making it more prone to fracture, and into ligaments and joints, making them less flexible and sometimes making movement very painful. Severe cases of fluorosis are crippling; most victims are elderly. As a result, fluoridation advocates and people in government must thread a needle: enough fluoride to protect against tooth decay in children, but not enough to cause problems in the long term.

Alas, epidemiologists have been complaining about the safety studies for decades, according to Sander Greenland, an emeritus professor of epidemiology and statistics at UCLA. Greenland, who is a co-author of the standard textbook Modern Epidemiology, began his own fluoridation work in the ’70s by examining a “typical crap ecological study” supposedly showing that fluoride caused cancer. “But then I got into the literature, just because I wanted to do a thorough job, and I noticed there was really no safety information. They didn’t have any good rationale for the dose.” The current U.S. recommendation is 0.7 milligrams per liter.

Greenland went on: “Since they didn’t have any good long-term data, the precautionary approach would be ‘What’s the smallest amount we can put in [so that] we get most of the benefit and minimize the likelihood of long-term harm?’ Instead, that mentality was totally absent from the literature.” Moreover, a seemingly prudent level doesn’t account for the possibility that certain people may be extra-sensitive to fluoride’s negative effects, because they are very young or very old, or are unlucky genetically, or have nutritional deficiencies. Nor, Greenland said, would it “take into account the errors you always expect in a large-scale system, where there are accidents that put in too much, and the monitoring is not that good.”

Howard Pollick, an ADA spokesperson and a dental scientist at the UC San Francisco School of Dentistry, defended water fluoridation in a recent interview: “The water systems are operated by professionals. With the new equipment, they can control the fluoride level within a very narrow range.” As for general safety, he noted, “there’s a 2015 review by the U.S. Public Health Service that looked at this. I’m comfortable with it.”

Matters get more complex for less well-documented risks. In October, a research team published the results of a long-term study in Canada that correlated concentrations of fluoride in the urine of pregnant women with the IQ scores, three to four years later, of their children. The IQs of the boys (but not the girls) in fluoridated communities were, roughly speaking, one to three points lower than those of boys in nonfluoridated communities. Another long-term study, published in 2017, had found a similar effect in Mexico (where the fluoride exposure was higher than in Canada). An analysis in 2012 of 27 fluoride-IQ studies from China had also found effects on cognition (these were retrospective studies, though).

Fluoridation advocates rightly point out that the IQ studies have limitations. However, their position necessarily involves making the gymnastic argument that you should put fluoride in water because its positive effects have been shown in a bunch of mostly retrospective studies, but you should ignore the risk to IQ because the negative effects have been shown only in a bunch of mostly retrospective studies. How should one weigh the potential small harm to a broad population against the potential broad benefit to a small population? What if neither the harm nor the benefit is well established? What if constraints (moral, financial, logistical) on our ability to experiment with human beings mean that these questions can never be answered definitively?

I asked Anne-Marie Glenny whether there were other ways of reaching poor children who can’t go to dentists—training them to brush their teeth in school, for instance. Or providing free dental care in impoverished communities. She said she was unaware of any research that compared the outcomes of fluoridation with these alternatives.

Given all the uncertainties, I asked, can we really say that fluoridation works? “There’s no argument that fluoridation doesn’t work,” Glenny said. “The question is whether it is still the right way forward.”

This article appears in the April 2020 print edition with the headline “Something in the Water.”