‘It’s just one study’: Why experts say it’s not time to give up on fluoridation
By Genna Buck
There are good reasons that a scientific paper — even a very good one — shouldn’t form the basis of policy
York University psychology professor Christine Till, the lead author of a Canadian study that made waves internationally because it found a link between exposure to fluoridated drinking water during pregnancy and lower IQs in children, is “shocked” at the way it has been characterized.
The study measured the levels of fluoride in about 500 pregnant women’s urine three times during pregnancy, and also collected self-reported information about their consumption of tap water and black tea, which is naturally rich in fluoride. Their children’s IQs were tested just once, at three or four years of age. Self-reported fluoride was associated with lower IQs in both girls and boys. Urinary fluoride concentration was associated with lower IQ only in boys. For every 1 milligram-per-litre increase in urine fluoride, a corresponding drop of 4.5 IQ points was found.
The study controlled for a large variety other factors that could account for IQ differences, including economic status, parents’ education and exposure to other environmental chemicals.
This makes it the most largest and most rigorous study of this kind ever done, and the only one in the Canadian context, Till said. She said it’s possible some “mystery variable” is related to both fluoride and IQ, but they controlled for everything they could think of.
The research, being controversial, was subject to “review after review after review” before it could be funded, Till said, and then extensive scrutiny during peer review. The authors hired a third-party statistician to look over the numbers to make sure everything checked out.
“Four and a half IQ points is not negligible. People have been calling this a small effect. I’m shocked. (The effect on IQ) of lead exposure is of this magnitude. It’s a big deal. The impact is going to be felt — it means fewer people in the gifted (IQ) range and more people falling in the range of intellectual disability,” Till said.
University of Calgary community health sciences professor Lindsay McLaren, who has studied fluoridation, said the size of the sample was reasonable and the study is “methodologically sound” and makes an important contribution.
“Public policy is ideally informed not by any one study, but by the best available evidence as a whole,” she said.
“At the end of the day, it is still one study –
Adding fluoride to drinking water, routine since the mid-20th century in some cities across Canada, helps strengthen tooth enamel and reduce cavities. Studies show people with the poorest oral health and access to dentistry tend to benefit the most. Good oral health is linked to a host of good health outcomes, and cavities and poor dental hygiene to a motherlode of other health issues.
Ray Copes, chief of environmental and occupational health at Public Health Ontario, said the link between lead exposure and IQ is not comparable to fluoride.
“What we know about the effects of lead on IQ doesn’t come from one study. It is … based on a far larger set of research findings that extend over many more populations and many more years,” he said.
He added the sex differences observed in the study were very “puzzling.” He said it’s an indication that the question should be tested using different research methods, and the research replicated by other researchers, before we jump to telling pregnant women to limit fluoride. Till believes there’s enough evidence to tell pregnant women to limit fluoride now.
If the study looked at a chemical that was thought to disrupt sex hormones, or a disorder that is more common in boys, like ADHD, you would expect to see sex differences, Copes said. But there’s no reason to expect them in measures of IQ.
The fact that the more robust measure of fluoride intake — fluoride in the mothers’ urine — was less convincingly associated with IQ suggests other factors are at play, he added.
He said there might be differences in support and programming available to children in different cities, and that this wasn’t captured by the variables the study controlled for. Early educational interventions can have a strong effect on children’s IQ.
“Public policy is ideally informed not by any one study, but by the best available evidence as a whole –
Although a difference of 4.5 IQ points should not be dismissed, it’s within the ballpark of “test-retest reliability,” Copes added. That means it’s not unheard of for the same child to take the same IQ test on different days and receive a score that is four points off.
“At the end of the day, it is still one study. Although I like a lot of what was done in this particular paper, at the end of the day, you’re left scratching your head — why aren’t the results internally consistent?” Copes said.
He said it’s something he sees all the time as a researcher: Academics who are doing great scientific work — really innovative, interesting, beautiful work — and want to see it translated into policy right away.
“I’ve had that discussion with many, many colleagues, in many different research areas, over many years. And my advice to them is: I’m with you on the research. It’s really neat stuff … but the policymakers have to consider far more than just one paper. And some of my research colleagues get that more than others. For some, it’s a really tough sell.”