Large Swedish study finds no effect of fluoride on IQ


Fluoride levels in Swedish drinking water (mg/L). Variation between municipalities. Source: Aggeborn & Öhman (2016)

A significant new Swedish study shows fluoride in drinking water, at the concentrations used for community water fluoridation, has no effect on IQ or other measures of cognitive ability. Similarly, it has no effect on diagnosis or prescription of medicines for ADHD, depression, psychiatric illnesses, neurological illnesses or muscular or musculoskeletal diseases.

On the other hand, the study showed positive effects of fluoride on income and employment status – most probably because better dental health is beneficial in the labour market.

This work is reported in:

Aggeborn, L., & Öhman, M. (2016). The Effects of Fluoride In The Drinking Water.

The study covers most of the health effects that anti-fluoride campaigners complain about. It really should put all these complaints to rest as the quality of this new study is much better than anything the campaigners rely on for the following reasons:

  • It involved a much large sample. Over 700,000 individuals were involved. The numbers included in specific measurements varied but they were much greater than those used in the studies cited by anti-fluoride campaigners. For example, almost 82,000 were involved in the cognitive ability comparisons – compared with a few hundred at the most in the comparable studies cited by anti-fluoride campaigners.
  • Estimates of effects were much more precise (as expected with large numbers of subjects) than for previous studies. The effect of fluoride on cognitive ability was always close to zero and for practical purposes was zero.
  • Statistical analyses were based on continuously varying fluoride levels – a much better approach than the simple comparison of data for low and high fluoride villages used in the studies cited by anti-fluoride campaigners.

Sweden is an ideal country for studying effects of fluoride at these low concentrations. It does not have artificial water fluoridation but its drinking water contains naturally occurring fluoride. The fluoride concentration in drinking water depends on the geology of the region so different Swedish communities consume water with different fluoride concentrations.

This graphic from the paper shows the number of people drinking water with various concentrations of fluoride. Note – the steps are 0.1 mg/L and although concentrations above 2.0 mg/L occur they are relatively rare. Sweden makes no attempt to remove excess fluoride until the concentrations exceed 1.5 mg/L – the maximum recommended by the World Health Organisation. For comparison, the recommended optimum concentration in  New Zealand is 0.7 mg/L.


Histogram of numbers people drinking water containing naturally occurring fluoride at different concentrations. Source: Aggeborn & Öhman (2016)

Effects of fluoride on dental health

The Swedish data showed positive effects of fluoride on oral health. For example, the share of dentists visits “decreased by approximately 6.6 percentage points if fluoride is increased by 1 mg/l. This should be considered as a large effect.” Tooth repairs are closely related to fluoride. “If fluoride would increase with 1 mg/l, the share of 20-year-olds that had a tooth repaired would be decreased approximately 3.4 percentage points considering the 2013 sample. Again, this effect is large, especially for this cohort.”

Cognitive ability

Relevant data was used from national education tests and psychological tests during the years of the Swedish military conscription. The statistical analysis produced estimates which were all very small and often not statistically significant. The estimates were sometimes negative and sometimes positive. For example, an estimate including covariates showed that “cognitive ability is increased by 0.045 Stanine points [equivalent to about 0.3 IQ points] if fluoride is increased by 1 mg/l (a large increase in fluoride). This should be considered as a zero-effect on cognitive ability.”

Other possible health effects

The authors considered the effects of fluoride on the prescription of medicines for ADHD, depression, and psychoses. They also looked at psychiatric and neurological diagnoses from outpatient and inpatient registers, as well as diagnoses of muscular and skeletal diseases. Anti-fluoride campaigners often claim fluoride has a harmful effect on these health problems.

The was no effect of fluoride on the possibilities of being prescribed any of these medicines.  For example “the probability of receiving ADHD medicines is decreased by 0.2 percentage points if fluoride is increased by 1 mg/l. In economic terms, this effect is a zero-effect.”

It was the same for all the diagnoses considered –  “The estimated effects are small and often statistically insignificant.”

According to the authors:

“In conclusion, we do not find that fluoride has any effects on these health outcomes. This further strengthens our argument that fluoride does not have any negative effects for levels below 1.5 mg/l on human capital development or health outcomes related to human capital development. It is also interesting that we do not find any effects on diagnoses for muscular and skeleton diseases, which has been a question also discussed in connection to fluoride.”

Annual income and employment status

The lack of any effect of fluoride on IQ and other psychological and non-psychological estimates suggest that fluoride would have no effect on long-term outcomes like income and employment status. However, the authors suggested that it could have a positive influence on these outcomes because of better dental health.

And this was the case. Estimates of the effect of fluoride on income were always positive and usually statistically significant. The authors estimated that “income increases by 4.2 percent if fluoride increases by 1 mg/l. This is not a negligible effect and the estimate should be considered as economically significant.”

Similarly for employment status. “If fluoride is increased by 1 mg/l, then the probability that the person is employed is increased by 2 percentage points. This result thus point in the same direction as the results for log income where both these results are significant in economic terms.”

Further analysis indicated “that when dental repairs increases by 1 percentage point, income decreases by 2 percent on the same aggregate level. This effect is clearly economically significant. This indicates that fluoride improves labor market outcomes through better dental health.”


This is an important study. It involved large numbers of people, estimated outcomes were far more precise than in previous studies, it used continuously varying concentrations of fluoride instead of simply comparing high fluoride and low fluoride villages, and it considered possible long-term outcomes like income and employment chances.

The advantages of this study compared with the generally poor quality studies cited by anti-fluoride campaigners should put to rest arguments used by those campaigners. In particular, it should make the current campaigns relying on to IQ and cognitive effects irrelevant.

The authors comment that their data shows there is no need to consider negative health effects on consideration of the cost-effectiveness of community water fluoridation. I wonder if, in fact, these results will encourage policy makers to consider the cost benefits of improved income and employment chances in future calculations of the cost-effectiveness of fluoridation programmes.

Similar articles


25 responses to “Large Swedish study finds no effect of fluoride on IQ

  1. So now we have positive proof or the advantages of “natural Fluoride’
    Forget that all fluoride is ‘natural’ The antis will push the its not “artificial” fluoride, that was in the paper.
    So now we will see some stupid scaremongering about that
    Clutching a big handfull of straws again
    No doubt the lawyer will have some strange reason why they should disregard it


  2. We now have the Swedes producing yet more evidence that changing fluoride levels in drinking water does not affect the IQ level of the population.

    Meanwhile, here in NZ, Canterbury demonstrates what can happen to the more impoverished portions of our population when fluoride is not present in the water supply…


  3. Are the authors of this study really suggesting that employment levels and incomes go up BECAUSE of fluoridation ?


  4. No, David, not at all. They do not have fluoridation in Sweden.

    However, the data shows that income levels and chances of employment actually increase with fluoride concentration in drinking water. They hypothesise that this is a result of improved dental health and that suggestion is supported by other data.

    These economic outcomes were investigated because of the idea that if fluoride causes a decrease in cognitive ability this would be reflected as lower incomes and poorer chances of employment.


  5. Sorry, I meant to say “because of higher fluoride ingestion”

    I am only a little ways into the paper but I am laughing about their methods for mapping who got how much. If you thought Malin’s research was suspect, perhaps you should apply the same scrutiny to this one. They both seem to have the same main problems. The measurement of total individual intakes is not available and the measurement of outcomes is suspect. The good new is that Malin’s new research, which should have been done in the 50’s, is going to be done.


  6. Interesting, though, that you didn’t say that.

    Malin and Till simply used the proportion of the state that was fluoridated. It did not look at individual levels at all. (Plus it ignored confounding factors which when considered showed the extent of fluoridation had absolutely no effect on prevalence of ADHD). Till’s new research may well suffer the same problems. I really must contact her to see if she is willing to discuss my analysis of the problems with her paper – now is the time to do this before he repeats her mistakes. Unfortunately the cost of publication in that journal is just too much for me.

    Agreed, total intakes are very rarely available for fluoride but it is well known that the levels of F in drinking water are a valid approximation because it is a major component of total intake. If that worries you then get onto Paul Connett and admonish him for relying on the poor quality Chinese studies which only consider village level – never individual intake.

    As for scrutiny of the Swedish paper – my assessment is that it is miles ahead of Malin and Till’s paper. The sample size is huge compared with previous studies. Covariates were considered – but as there was no effect of fluoride on the parameters considered the question of confounders really doesn’t come up.

    The other advantage is that, except for the dental health data, the measured parameters were for individuals.


  7. Yes, they shifted the address when it became a working paper. Corrected.


  8. I haven’t read the paper yet but the “fnd precisely estimated” in the abstract has given me a good laugh to start the day. Looks like we are still looking at water fluoride instead of individual exposure. Do people in Sweden still drink the tap water or have they followed the North American trend?


  9. David, perhaps you should read the paper, then, before commenting.

    And, also, what about directing your concerns about use of drinking water concentrations to all those anti-fluoride people who have found such convincing evidence in studies made in areas of endemic fluorosis.

    Or, perhaps you could look at those many studies from health authorities which have looked at total dietary F intakes (there are plenty of them) and their conclusions about drinking water being the major source of dietary F in most countries like ours.

    Or, maybe you will shift your known commercial and financial interests away from marketing filters to remove fluoride from water to some sort of mechanism to remove it from other dietary inputs as well.

    Then again, perhaps you are just being hypocritical.


  10. I’m not retired yet Ken, so I had to get to the office this morning. I just thought it was funny that there was a typo in the abstract that no one caught and the phrase “precisely estimated” seemed like a funny contradiction suggesting overcompensation and bias by the researchers.

    I have seen the data showing the range of estimated doses in US men to be in the 1.6 to 6.6 mg/day range. I believe that came from the CDC in the 90’s so with the large increase in dental fluorosis between the data gathered in 1999 to 2004 and that collected in 2011-12, we can assume that the average dose has increased since the nineties, at least in the US. If there are now many people getting 7 or 8 mg/day, that is getting close to the 10 mg that is admitted to result in a high risk of harm. How do you propose that those at the higher levels of intake from other sources know how much they are getting, so they can avoid fluoridated tap water?

    What I haven’t seen is much research related to individual dose, especially early in life. That situation has been mentioned by the few researchers have completed that sort of study who also note that water level is not a great way to judge intake. I suggested that with fewer people drinking plain tap water, at least in North America, drinking water fluoride has become an even worse way to estimate total exposure. If you have anything, please post it. I would also be interested in seeing research on the elderly who have been exposed to 50 plus years of fluoridated water, compared to those who haven’t. Are they showing more arthritic symptoms? Diabetes? Heart Disease? Where is all the research that should have been done either before fluoridation was started, or when it was realized that dental fluorosis was occurring in more than the 10% or the population that was predicted by the promoters?


  11. David – why do you seel filters for removing fluoride from drinking water if you are conceding that fluoride in drinking water is not a problem?

    Are you just trying to make a quick buck from the gullible? 🙂

    As for the elderly – research definitely shows that they are living much longer since fluodiation was introduced. Sorry if that is a descent to the “logiuc: oif Declan Waugh.


  12. Conceding that fluoride in drinking water is not a problem?
    I don’t know where you got that from. I’m not sure what your reference to Declan Waugh means. I think you’ll agree that we could be living longer in spite of fluoridation, not because it is beneficial.


  13. David, you are the one saying we should look at other sources of fluoride. But, in an example of cognitive dissonance, you are also the one who attempts to raise fears about fluoride in drinking water so that you can sell more filters.

    Amazing how easy it was for you to pick up my joke – yes of course correlation is not causation. No one seriously suggests that fluoridation is the cause of our longer life.

    But you and your mates are the people who blindly accept Waugh’s citation trawling and arguments that fluoridation is responsible for any health problems he can find.

    Problem is, in contrast to me, Waugh is seriously attempting to misrepresent the situation. He is not joking.


  14. Sorry Ken, I don’t know how you translated my looking for research related to total dose and health effects into the idea that I am more concerned with what is in water than what is coming from other sources. I warn people about swallowing toothpaste, drinking tea and California wines, and mixing infant formula using fluoridated water; things that our public health people here should be very vocal about, but either don’t mention all or don’t say very forcefully. You may be joking about fluoridation causing longer life, but this isn’t the first time I have heard someone saying it. The other folks didn’t mention that they were joking. Thanks at least for confirming that you don’t have any research related to individual dose that looked at health effects and couldn’t find any.


  15. David, I do not think, or say, you have any real concerns – except making money from the gullible by selling them filters to prevent harm for which there is no evidence.

    Interestingly, this particular study did not look at the length of life but it did look at the quality of life in terms of employment and income (and job opportunities) and found a positive relationship with fluoride concentration in drinking water. Probably not a surprise because poor oral health severely limits a person’s life opportunities – a fact I can personally confirm but there is research evidence for that.

    Your attempt to critique this study because it did not have data on total F intake is rather pathetic as you and your mates continually promote poor quality research from areas of endemic fluorosis where water F concentration is, except for a very small number of cases, the only parameter considered.

    In Sweden, they had a huge database of oral health and drinking water F concentration data. There were, obviously, no such detailed figures for other F intakes. But there is absolutely no evidence to suggest that the normal situation of drinking water being the major F dietary source is not true here. Nor is there any evidence of trends in other forms of dietary F intake. So, like all studies in the real world, one goes with the data one has – and there are real dangers in inventing data the way that Hirzy et al (2018) did.

    But, what about you reading the paper. I never feel confident in making any critique without actually reading the paper so can understand how worthless such uninformed comments can be.


  16. I don’t promote studies related to level in water. I promote the absence of studies related to individual dose, and the timing of that dose. I also promote the shadiness of the study promotions done by McLaren and Broadbent. If you complain about Hirzy, why don’t you complain about them? Studies that should have been done before fluoridation was ever started haven’t been. Surely 70 years is long enough that someone should have collected useful data. Your statement that you go with the data you have is an admission that good data isn’t available. I think that is because the promoters have never wanted to know, or am I the only one subject to confirmation bias? The good news is that my “mates” have been successful in raising enough questions that the people with the money are noticing and trying to answer.

    Did you read about the passing of Herbert Needleman?
    Interesting that he found an IQ drop of 5-6 points. The same range suggested for pre-natal fluoride exposure by Bashash.
    How do we know that fluoride isn’t the lead story all over again?
    I hope I am wrong.


  17. David – you are selling filters. the whole logic, and your story, is that F in water is harmful and you promote this to people receiving fluoridated water despite the lack of any evidence.

    The alternative to basing understanding on evidence is to rely on fairy tales.

    It is slander to describe honest research as Slander – the criticisms of Broadbent et al do not stand up (except the obvious and inevitable limitations imposed by sample numbers and I refer to that in my article here. Broadbent himself acknowledges the advantages of the Swedish study).

    I have written about the McLaren study and find the critiques of that to be faulty. Read Flaw and porkie in anti-fluoride report claiming a flaw in Canadian study.

    Yes, the Hirzy et al study is flawed and that is why they could not get it published anywhere but in the rag “Fluoride.” I have critiqued the study and my article is currently awaiting publication but you can read a draft here – Does drinking water fluoride influence IQ? A critique of Hirzy et al. (2016)

    You might be upset because I have exposed the flaws in Hirzy et al and exposed the flaws in the critiques of Broadbent et al and McLaren et al., but as a scientist, I go with the evidence, not ideological preferences and confirmation bias.

    You, of course, are welcome to critique my articles about those papers – I welcome all good faith scientific exchange.


  18. And of course you don’t suffer from confirmation bias like the rest of us.

    It is true that half of Broadbent’s non-fluoridated group was taking fluoride supplements, right?

    It is also true that McLaren had data that contradicted her conclusions, which she neglected to mention in the Discussion or to the press. That’s shady behavior and when combined with the weakness of her data, her study is of no value in demonstrating anything except we taxpayers were taken for a million dollars and the media doesn’t look to deeply before they report.

    You can speculate about my motivations for opposing fluoridation, but the other side of that coin is that many of our current customers are also opposed, judging by how many have signed our petition. If we are successful in ending fluoridation here, haven’t I removed a reason they have demonstrated for buying our water? My conflict is conflicted in the opposite direction.

    I don’t know where you got the idea that I am upset. I am quite relaxed about the whole issue since I haven’t been consuming fluoridated water for the past 35 years and my wife and kids haven’t been exposed since they have been with me-about 23 years. No fillings in any of their heads and I still have the same number I had when I was in my teens -3. Of course if any of my kids had dental fluorosis, i would be extremely upset and suing everyone in sight for failing to warn me that fluorosis might be a marker for harm to other parts of the body.


  19. The Broadbent study used the highly respected Dunedin long-term study. It included confounders like the use of fluoride tablets and breastfeeding. As I have explained the precision or confidence interval of the results are of course influenced by sample numbers – and that is expressed in the results of the statistical analysis. Use of breastfeeding or fluoride tablets does not negate the finding, although it may influence the confidence intervals. This study found no statistically significant effect on IQ.

    Of course, the current Swedish study has much larger sample numbers and the confidence intervals are much smaller as a result. Read the study – it confirmed no significant negative effect on IQ.

    Could you please provide the actual evidence for your claim “McLaren had data that contradicted her conclusions, which she neglected to mention?” I know FAN makes this sort of claim but they have been shown wrong. They have confused the different data measurements from the different cities and ignored their coverage in another paper by McLaren. I went into this in my article.

    You have simply demonstrated your ideological convictions, use of blinkers in your reading, reliance on biased sources and uncritical adoption of their mythology to confirm your own bias.

    And, yes you are upset. Why else would you comment here – especially as you have not read the paper my article describes.


  20. Thanks for telling me how I feel. You have confirmed your bias in my mind by first not acknowledging that it exists, then pretending to not be aware of the problems with McLaren’s public relations exercise sold as useful research. You also seem to have avoided answering my question about the fluoride free group taking fluoride supplements in Broadbent’s paper. I think I”ll keep looking for better data, before I make any sort of judgement about what harms there may or may not be from the doses of fluoride people are now consuming. Even if no issues are found currently, we still won’t know what the long term will hold for the unfortunate kids who are showing more severe Dental Fluorosis than any generation before them. If you want to see anger, wait to see what happens if those kids find out their health has been harmed.


  21. David, you are simply repeating FAN dogma – not engaging with the discussion.

    You seem unable to even describe your claimed “problems with McLaren’s research” or respond to my article discussing it.

    You seem unable to understand how the Broadbent study included confounders like breastfeeding and use of F tablets in its multiple regression – a completely normal scientific procedure.

    In fact – you make your judgment on ideological and commercial grounds. After all, your scaremongering helps you to sell more fluoride filters, doesn’t it?

    You are not “looking for better data” – you avoid good data and good studies like the plague – they may inhibit your sales.


  22. Now you are telling me what I am not looking for. For a research scientist, you have spent a great number of words talking about my bias, even after I have pointed out that my bias may be in the opposite direction, You didn’t answer my question about the fluoride supplements, except indirectly by claiming it doesn’t matter, so I guess that is a yes. Since I know you are aware of most of the major weaknesses with McLaren’s effort, I’ll only mention the one that I haven’t heard anyone else talk about before: How can Calgary and Edmonton be considered comparable cities when they were starting from such different baselines?

    Since you have no problems affixing motivations and feelings to me, let me tell you about my opinion of your “Open Parachute”. It is not as open as you like to pretend. To paraphrase that line about racism, “a biased person sees bias in everyone but themselves”.

    Here’s funny story about our local MOHs response to the Bashash paper. She cited a paper that McLaren and another researcher published that used a yes/no self report, by the subjects or their parents, to judge cognitive effects. I call that damage control. Why don’t you ever talk about public health dogma?

    I am afraid I have to sign off and give you the last word. Let me know if you find any research that connected or attempted to connect dose to health effects. We had another one come out here in Canada that found a negative effect, but it was still looking at water fluoride..


  23. Raymond R. White, Ph.D.

    Since higher fluoride exposure correlates significantly with higher income and therefore higher class, the study compares lower class populations with higher class populations in an uncontrolled way. Fluoride certainly does not raise incomes (what a ridiculous and biased idea) but lowered IQs may be masked by higher class populations being those exposed to higher fluoride levels. The compared populations should be MATCHED by income levels.


  24. Raymond, you say “Fluoride certainly does not raise incomes (what a ridiculous and biased idea).”

    But why ridiculous and biased? Surely it is reasonable to postulate that better child dental health could be a factor in a person reacher a higher income and better job. It seems obvious to me that poor dental health as a child leaves a person with horrible psychological consequences that inhibit their development and reaching their potential.


Leave a Reply: please be polite to other commenters & no ad hominems.

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s