By Ken Perrott 12/10/2017


I think we do.

Something like the good faith scientific exchange I had with Paul Connett four years ago (see Connett & Perrott, 2014 – The Fluoride Debate).

After all, there have been a number of important scientific reports since then. They may have been thrashed out (and thrash is sometimes the operative word) in one of the “anti-fluoride” or “pro-fluoride” internet silos but there has yet to be a proper discussion.

I have been trying to get one going for a while. Paul Connett is no longer interested and everyone else on the “anti-fluoride” side seems unwilling. However, Bill Osmunson who recently replaced Paul Connett as director of the Fluoride Action Network has been contributing to the discussion on several of the posts here. He seems to be the obvious choice for a discussion partner and I  asked him if he is willing to participate in another scientific exchange of the sort I had with Connett.

So far he has not responded – but as he has made some relevant critiques of several recent scientific papers in these discussion contributions I think it is relevant to bring that discussion into the formal blog posts. Otherwise, some important points will just be lost because they are buried deep in the discussion threads.

Here I respond to criticisms Bill makes of two recent studies which looked for evidence of the influence of community water fluoridation (CWF) on IQ and cognitive deficits in general. I urge Bill Osmunson to respond to my points in a format which can be presented as a blog post here.

 

 

Community water fluoridation and IQ

The two studies were published after my exchange with Paul Connett and are:

Broadbent, J. M., Thomson, W. M., Ramrakha, S., Moffitt, T. E., Zeng, J., Foster Page, L. A., & Poulton, R. (2014). Community Water Fluoridation and Intelligence: Prospective Study in New Zealand. American Journal of Public Health, 105(1), 72–76.

And

Barberio, A. M., Quiñonez, C., Hosein, F. S., & McLaren, L. (2017). Fluoride exposure and reported learning disability diagnosis among Canadian children: Implications for community water fluoridation. Can J Public Health, 108(3), 229.

Broadbent et al., (2014)

This study used data from the Dunedin  Multidisciplinary Health and Development longitudinal study and found no difference in IQ of people in fluoridated and unfluoridated areas or any effect of fluoridated toothpaste or fluoride supplement use.

I hope I represent Bill correctly but his criticisms of this study are vague – I can’t help feeling he is succumbing to the general hostility anti-fluoride campaigners have had about this study.

Let’s deal with his last criticism:

” I have previously presented my reservations about the NZ study and Broadbent’s comparing fluoridation with fluoride supplements, which lacked power to evaluate IQ.”

It more or less encapsulates anti-fluoride criticisms of the study and does contain an element of validity in reference to the study’s “power.” However, Bill’s reference to “power” is far too vague. It needs to be quantified.

Is Bill claiming that there are declines in IQ caused by CWF but they are too small to be detected in a study like Broadbent et al., (2014)? Or was there something about that study which made it incapable of detecting a reasonable IQ decline? Or does it matter – after all someone who is ideologically committed to believing fluoride is bad for IQ can always fall back on this argument when experimental results don’t go their way. No study will realistically have the ability to detect an extremely small IQ change that they might argue for. And such a small change is more in the eye of the (biased) observer than a reality.

Fellow FAN members Hirzy et al., (2016) also argued that the “power” of the Broadbent et al.,  (2014) study was too low to detect their assumed change in IQ. They argued this case on the basis of total dietary intake of fluoride claiming that there was very little difference of total dietary intake between fluoridated and fluoridated areas.  Osmunson et al., (2016) made the same argument – appearing to give up completely on the contribution of CWF (as it “likely represents less than 50% of total fluoride intake”) and directing attention to total fluoride intake instead. However, their arguments are very subjective as they pull dietary data “out of a hat” and don’t deal with the real situation where the study occurred.

Osmunson mentioned the importance of fluoride supplements and fluoride toothpaste to fluoride intake but seemed to have missed the fact that Broadbent et al., (2014) had also included these as factors in their statistical analysis. Neither these factors nor CWF exhibited a statistically significant effect on IQ.

The apparent fallback position of Hirzy et al., (2016) and Osmunson et al., (2016) that the relatively small dietary F intake meant their assumed IQ differences were too small for the study to detect comes across as straw-clutching. Especially as oral health differences between fluoridated and unfluoridated areas were detectable See Evans et al., 1980 and Evans et al., 1984).

The “power” of a study

The “element of validity” I referred to in Bill’s complaint about the “power” of the experiment is one every practical researcher faces – especially when dealing with an existing programme rather than designing, from the ground up, a laboratory experiment. Numbers of participants, or samples, are always limited and researchers rarely have the luxury of the large number they would wish for to provide more “power.”

The “power” of a study is often represented by the  95% confidence interval (CI). This means that if the same population is sampled on numerous occasions and interval estimates are made on each occasion, the resulting intervals would bracket the true  population parameter in approximately 95 % of the cases.” Usually, more sample numbers mean a smaller CI and therefore more confidence in the value of the result.

Broadbent et al (2014) reported a 95%CI of -3.22 to 3.20 IQ points for the effect of community water fluoridation with children of 7 -13 years. (The equivalent CIs for the effects of fluoride toothpaste and fluoride tablets were -1.03 to 2.43 and -0.38 to 3.49 respectively). The observed effects were not statistically different to zero. Their study used just 990 children. If more participants had been available the 95%CI could have been reduced to less than the range of 6.4 IQ points actually found for the effect of CWF.

In a very large Swedish study, Aggeborn & Öhman (2016) included between 20,000 and 80,000 participants and estimated a confidence interval of -0.23 to 0.89 IQ units when fluoride is increased by 1 mg/L. (They were able to consider a continuous measure of fluoride and not simply fluoridated or unfluoridated treatments). This study has far more “power” than that of Broadbent et al., (2014), and therefore a smaller CI value. But the conclusion was the same – fluoride at these concentrations had “a zero-effect on cognitive ability.”

Barberio et al., (2017)

This is a Canadian study with a large representative sample and individual estimates of fluoride exposure and reported learning disability diagnosis. Overall it concluded there was no “robust association between fluoride exposure and reported learning disability diagnosis.”

Bill Osmunson argues that this study “has limitations” and that the “conclusions overstate their data.”

I agree with Bill that diagnosis of learning disability based on a household questionnaire is not the same as a proper professional diagnosis, although presumably the question aimed at finding out if a professional diagnosis had been made – and what it was in some cases. The authors acknowledge that weakness but argue that more objective assessments are probably only feasible in small-scale studies.

Interestingly Bill and his fellow anti-fluoride campaigners did not raise this problem of reliance on parental answers to a questionnaire when they considered and argued strongly for, the Malin and Till (2015) ADHD study. (See  Perrott 2017 – Fluoridation and attention deficit hyperactivity disorder – a critique of Malin and Till (2015)for more details of this study and its problems.

Of course, these are the real-world problems faced by researchers attempting to extract useful data from large-scale surveys. One of the reasons why readers should not consider single studies as definitive and should consider each one critically and sensibly.

However, I think Bill is straw-clutching when he quotes the authors:

“When Cycles 2 and 3 were combined, a small but statistically significant effect was observed such that children with higher urinary fluoride had higher odds of having a reported learning disability in the adjusted model (p = 0.03).” [Cycles 1 and 2 are two separate parts – 2009-20011 and 2012-2013 respectively – of the Canadian Health Measures Survey]

And then argues:

“Barberio could have concluded they found harm. Instead, they focused on data which did not show harm.”

Bill is aware that a statistically significant effect of fluoride exposure was observed in only a limited case – when data from two cycles were combined and the urinary fluoride data had not been corrected by using either creatine concentration or specific gravity. This correction is necessary as an attempt to overcome the shortcomings of single spot-samples of urine. As the authors point out “spot urine samples used to measure fluoride are vulnerable to fluctuations.” And :

“creatinine-adjusted urinary fluoride or specific gravity-adjusted urinary fluoride . . .  are thought to be more accurate because they help to correct for the effect of urinary dilution, which can vary between individuals and different points in time. Accordingly, these adjusted measures help to offset some of the limitations associated with spot urine samples. The finding that the effect was reduced to non-significance when creatinine-adjusted and specific gravity-adjusted urinary fluoride were used, suggests that the association between urinary fluoride and reported learning disability diagnosis may not be robust.”

So Bill would prefer that the authors had based their conclusions on uncorrected urinary fluoride data and not the more reliable corrected figures? And why? Because that would have confirmed his bias. That is an unfortunate personal foible – our biases often encourage us to go with unreliable conclusions and not allow them to be challenged by the more reliable data.

Conclusions

Here I have simply considered the Broadbent et al., (2014) and Barberio et al.,. (2017) papers because these are the ones Bill Osmunson has responded to. I urge him, to also consider the Aggeborn and Öhman (2016) paper.

I hope Bill Osmunson will respond to this post with his refutations of my points or further arguments about these and other papers. I hope also that he takes up my offer of space here for an in-depth exchange of the sort I had with Paul Connett four years ago.

References

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

Barberio, A. M., Quiñonez, C., Hosein, F. S., & McLaren, L. (2017). Fluoride exposure and reported learning disability diagnosis among Canadian children: Implications for community water fluoridation. Can J Public Health, 108(3), 229.

Broadbent, J. M., Thomson, W. M., Ramrakha, S., Moffitt, T. E., Zeng, J., Foster Page, L. A., & Poulton, R. (2014). Community Water Fluoridation and Intelligence: Prospective Study in New Zealand. American Journal of Public Health, 105(1), 72–76.

Evans, R. W., Beck, D. J., & Brown, R. H. (1980). Dental health of 5-year-old children: a report from the Dunedin Multidisciplinary Child Development Study. The New Zealand Dental Journal, 76(346), 179–86.

Evans, R. W., Beck, D. J., Brown, R. H., & Silva, P. A. (1984). Relationship between fluoridation and socioeconomic status on dental caries experience in 5-year-old New Zealand children. Community Dentistry and Oral Epidemiology, 12(1), 5–9.

Hirzy, J. W., Connett, P., Xiang, Q., Spittle, B. J., & Kennedy, D. C. (2016). Developmental neurotoxicity of fluoride: a quantitative risk analysis towards establishing a safe daily dose of fluoride for children. Fluoride, 49(December), 379–400.

Malin, A. J., & Till, C. (2015). Exposure to fluoridated water and attention deficit hyperactivity disorder prevalence among children and adolescents in the United States: an ecological association. Environmental Health, 14.

 

Osmunson, B., Limeback, H., & Neurath, C. (2016). Study incapable of detecting IQ loss from fluoride. American Journal of Public Health, 106(2), 212–2013.

Perrott, K. W. (20217). Fluoridation and attention deficit hyperactivity disorder – a critique of Malin and Till (2015)).  British Dental Journal, In press.

Similar articles


Site Meter