Prof. Newton's report to South Central SHA: A Critique
1.1 A report was written by the Regional Director of Public Health, Prof. John Newton, for the Board meeting of South Central Strategic Health Authority which was to decide whether or not to initiate a water fluoridation scheme in the Southampton area. It was one of several reports, but was a key paper from a key figure in determining "whether the health arguments in favour . . outweigh all arguments against water fluoridation", since it set out to "[summarise] the scientific evidence on water fluoridation and [address] issues highlighted in discussion during the consultation process".
1.2 The author of the present critique was responsible for the parliamentary questions which led to the 'York' systematic review of the evidence in 2000; was a member of the advisory board to that review, as an opponent of fluoridation; and has acted with some of the independent scientists from 'York' to correct publicly the misrepresentations of the review's findings.
Prof. Newton's report
2.1 The report to the SHA for its February 2009 meeting is a mixture of the scientific and unscientific. It is necessary to consider it carefully to assess what is reliable and what is not.
2.2.1 The two initial covering pages contain important statements, some of which raise issues which run through the report. Presumably the "independent critical appraisal" of new scientific material was by Bazian, in respect of the US National Research Council review which is addressed in 4.2 below; The Evidence Centre's report only summarised without appraising the input from the consultation. It is not clear how the bulk of the evidence from incoming submissions was handled and evaluated - who read them apart from the Evidence Centre, what proportion were fully studied, and in what ways the arguments raised were presented to the full board, if indeed they were. There have been suggestions that submissions that did not include the full scientific papers referred to may have been disregarded, which would be extraordinary. Prof. Newton's paper sheds no light into what is at first sight a major gap.
2.2.2 Prof. Newton is right to cite (later) the York report in 2000 as the most reliable source of evidence, and is correct here that (by implication) York suggested that fluoridation reduced tooth decay. Evidence for reduction in dental health inequalities, however, should not be put in the same bracket, as is done on his second page. There is a qualitative difference between an average study validity score of 5 out of 8 and one of 1.6. (See 3.5. below for more on this.) Since in Prof. Newton's words fluoridation "is intended to reduce inequalities in dental health", accuracy here is crucial. (See 3. below for the Equality Impact Assessment process referred to.) If York is to be relied on, it should have been made clear to decision-makers that the case for fluoridation rests on the suggestion of a benefit, combined with some "weak, contradictory and unreliable" evidence on reducing inequalities (see 3.6 below).
2.2.3 This is before safety comes into the equation. Prof. Newton says there is no reason to believe fluoridation is unsafe. This is a debatable statement. One needs to know whether the SHA as guardians of public health in this matter addressed the burden of proof. Is it enough that no definite harms have been proved? Or does a reasonable possibility of harm ring warning bells such as to invoke a precautionary approach (see 4.2.28 below)? The York reviewers restricted the standard for inclusion of effectiveness studies to levels A and B; but they admitted a lower level C for possible harms in accordance with usual precautionary practice. They also recommended good research into some possible serious harms, which has not been done. Did the SHA intend to adopt a more relaxed view on safety? No one should claim that harms have been scientifically proved; but nor should anyone claim as many have that there is "no evidence of harm", because this misuses language (see 4.2.25 below). This would be true if York had found all studies for harm suggesting no effect, but this was not the case since the studies on bone problems, cancer, etc., low-quality as they were, pointed both ways with roughly equal weight. Prof. Sheldon who chaired the York advisory board wrote on 3 January 2001: "The review did not show water fluoridation to be safe."
2.2.4 There is little merit in Prof. Newton's implication of fluoridation's safety on account of "widespread use for many decades". This is a populist argument, not a scientific one, as a moment's thought about smoking (since Elizabethan times) and lung cancer shows. Such associations do not spring out at you. Was the SHA's attention drawn to the lessons of the important article by Cheng et al in the British Medical Journal of 6 October 2007, which was largely addressed to people involved in consultations such as theirs? The distinguished authors, two of them from the York review, showed the difficulty of detecting even quite significant harms among populations by standard scientific methods. No such reliable methods have been used in safety studies in the fluoridation field.
2.2.5 Moreover, detection of harm presupposes some suspicion in the first place; without this, why would any studies be worth doing? History shows that there is no index of suspicion among those who believe in fluoridation and, most significantly, that fluoridation is in Prof. Newton's words the "professional orthodoxy" among UK health workers, and has been for decades. Given this cultural mindset there is a bias against looking for harm, which can account for the paucity of evidence which surprised the York reviewers. Many would think that to argue from there that there is no cause for concern, as Prof. Newton does, is to take risks with public health. More will be said about this in 2.5 below.
2.2.6 There is no strong basis for Prof. Newton's claim that the scientific evidence supports the Primary Care Trust's choice of fluoridation against other options, least of all for the claimed reduction in inequalities, since all the science is weak (Prof. Newton's reference 19 shows this weakness among the options). What is clearly known is that topical fluorides are effective means of reducing caries, at concentrations much higher than 1 part per million, although again reliable safety studies are lacking.
2.2.7 It is worrying but indicative that the US National Research Council's 2006 report on possible harms from fluoridated water had not been "adequately considered" earlier by the SHA. This was a major review, publicised at the time, cited by the 2007 Australian systematic review, and of interest to anyone involved in the fluoridation debate worldwide. It did not feature in the SHA's consultation document, which claimed to give "all the facts", so the public were not told of it, and it seems that only the consultation process "identified" it. It will be dealt with below (4.2.1-13), together with the question of fluoridation and IQ which formed part of it.
2.2.8 Lastly and importantly, it must be pointed out that "all arguments against water fluoridation" which are referred to in these introductory pages include the ethical argument which does not form part of Prof. Newton's report. It impacts on the choice of options in 2.2.6 above.
2.3.1 The 2007 Nuffield Council on Bioethics report is cited in the 'Introduction' section. This was a report from an expert cross-disciplinary working party that took an interventionist line in certain health areas that surprised some critics, and that, in addition to the view cited by Prof. Newton under 1, advocated "Democratic decision-making procedures" for fluoridation, while confirming the uncertainties of the evidence as shown by York. Unfortunately we do not have such democratic procedures; it seems that it is possible for an unelected SHA to act regardless of local opinion, which cannot have been what Nuffield or indeed many in Parliament had in mind. A government minister in 2005 stated that "fluoridation schemes would only be introduced where the local population were in favour".
2.3.2 The dental health of local childen is considered in 5 and 6, where Prof. Newton builds a case for intervention. Two questions arise. The first is why figures for 5-year-olds are taken as the basis for the consultation; it is 12-year-olds who provide the usual benchmark. This has not been answered, and the suspicion must remain that since the statistics for 12s are much better than for 5s in Southampton the case for fluoridation, the "professional orthodoxy", would be weaker.
2.3.4 The second has been noted by researchers with a broader focus on health trends, which is that there has been a uniform decline in caries in first-world countries over the last 30 years, regardless of fluoridation status (see Cheng et al's article), that Southampton children's teeth have been part of this trend and are not bad (especially for 12-year-olds) by historical standards, and that what has been seen recently may be no more than a blip to which fluoridation of 200,000 people is an excessive reaction. (Caries incidence in 5-year-olds over the last two measured periods has risen from 1.35 to 1.99 dmft, then fallen back to 1.76 against a national average of 1.47; publication of the most recent figures is awaited.) It would have been helpful to explain why caries is thought to have increased locally, since this is important information in deciding what to do next. This information has been lacking throughout.
2.4.1 The section on evidence illustrates a common dichotomy in argument over fluoridation. You cannot both adhere to the York review and cite with approval the statement by the US Centers for Disease Control in Prof. Newton's 11 and the views of the WHO and others in 12. They are mutually exclusive: you have to decide which horse you will ride. It should not need restating that York is not only the most recent of these reports but in scientific terms the best - indeed, because of its systematic nature and transparency, the only truly reliable assessment of the evidence. Anything that came before it has been superseded, unless it can be persuasively shown why York was in error. (See 2.4.5 below for the 2007 Australian review.)
2.4.2 Paragraph 14 provides clear evidence that the Regional Director of Public Health is having it both ways, where he says that "the body of evidence in support of fluoridation is based on direct observation that there is less dental disease in people who drink fluoridated water . ." It was largely to get away from reliance on observation, which is open to bias, and to bring good science into the debate, that the Government set up the York review. From reactions since it was conducted there is small doubt that York has been endorsed through gritted teeth by most adherents to the professional orthodoxy who had expected a verdict that fluoridation was indeed very effective and safe. Such a verdict could only be reached by methods of inferior quality and a selective approach to the evidence.
2.4.3 "Much of this effect may be topical rather than systemic." It is now thought that you do not get protection from decay principally by swallowing fluoride, as was once believed, but by its contact with surfaces of teeth. This is a significant enough change in dental thinking to warrant more than one sentence and a phrase "However, regardless of how it acts . ." Was the SHA's attention drawn to the implication of this change, which is that one of the planks supporting a policy of population fluoridation has been removed?
2.4.4 In the following sections Prof. Newton gives a fair account of York's findings on effectiveness, although the reviewers did not dwell on the lack of randomisation as stated so much as on the other methodological weaknesses in the evidence base, including some confounding variables that he does not mention (see also 4.2.24 below).
2.4.5 The one comparable exercise to the York review was the thorough systematic review published in 2007 by the Australian National Health and Medical Research Council. This took York as its basis over all relevant questions, examined the relatively few studies of acceptable quality since then, and found no reason to alter York's conclusions. In doing so, however, it overstated York's finding on effectiveness, cited by Prof. Newton in his paragraph 22. York did not "strongly" suggest that fluoridation was beneficial; and to say that York "showed" that it "significantly increased the proportion of caries-free children" is also to quote York's assessment as firmer than it was. Prof. Newton may not have spotted and did not address the conflict with 17 in which York was accurately quoted. The SHA may not have noted it either.
2.4.6 The thrust of paragraphs 23 and 24 "To illustrate the nature of the available evidence" is puzzling. Prof. Newton gives details of a particularly weak study suggesting benefits from fluoridation, with a graph to highlight the message. This is the joint-lowest ranked study in the York review (Riley et al has a validity score of 0.8 out of 8) from a section that thought 'level C' studies inappropriate for the purpose but could find few that were better. Prof. Newton goes on to state that this was "a study of low quality according to the York review" and explains why. The impression on this author is of a comparison to show York as over-scrupulous in the face of a study suggesting "not only that dental health was generally better in fluoridated areas but also" that it reduced inequalities. If this was not the purpose, the charge becomes explicit when Prof. Newton deals with criticisms by opponents of fluoridation later in his report (see 4.2.22 below).
2.4.7 More weak evidence follows. Direct observation of patterns of dental health and child surveys (25, 26) may suggest associations between fluoridation and dental caries, but are little use in confirming suggestions of causation from existing studies, for which more rigorously conducted studies are required as the York reviewers emphasised.
2.4.8 Prof. Newton cites a 2007 systematic review in 27 (Griffin et al) for the suggestion that fluoridation benefits adults. This review applied a lower quality threshold than York, found only 9 studies (all pre-York) of which 1 only had been deemed acceptable by York, did not conduct a comprehensive literature search and used some questionable statistical methods. The authors made some acknowledgement of the paucity of evidence, and as yet little weight can be put on this for adult benefits. Griffin published after York, but all her studies were from the last century.
2.4.9 Prof. Newton follows a line that has become familiar in contrasting York's findings on unsightly dental fluorosis to those of the MRC working party that reported in 2002. The latter did not conduct a systematic review, contained a disproportionate membership from the dental profession, and the evidence cited in 28 pre-dated York and so had been taken into account by it; yet the MRC report is preferred by the public health community, and it seems here, because it is closer to the existing professional orthodoxy in playing down disadvantages of fluoridation. No good scientific reason has been given for departing from the best available evidence, which was defended by the senior York reviewers in the British Medical Journal of 16th June 2001, and which suggests that 1 fluoridated child in 8 is likely to be concerned about the appearance of its teeth. As Prof. Newton says, however, the majority of fluorosis cases are mild.
2.5.1 Safety is crucial, especially in a health issue affecting hundreds of thousands of people whose medical histories are not known . Prof. Newton gives an accurate account of York's findings in paragraphs 33 - 38. The question is what conclusion to draw from a finding of "no clear association" with cancer, bone problems and other harms which have had even less research. As mentioned above there was evidence of harm, though it was no stronger than the evidence of no harm. It is up to individuals what degree of risk they find acceptable. But individuals cannot control their risk here because mains water is unavoidable, and in this case a "stewardship" model such as Nuffield's might suggest a margin of caution, not least for the "vulnerable" (such as the old and infirm) of whom play is commonly made when promoting fluoridation's benefits. Prof. Newton does not address the significance of coercion in his preferred anti-caries measure. (This is sometimes commended as "minimising problems of compliance" by those of an Orwellian turn of phrase.)
2.5.2 Many, and not just the lay public, would argue that "a good level of assurance" (40) is not what the evidence provides, and might take a dim view of a fluoridation lobby that, as they saw it, had not bothered to do proper safety studies and were using this very lack of evidence (for harm, as for some benefits) as justification for imposing yet more schemes. York wrote of the possibility that adverse effects might take a long time to show themselves and so might not have been captured by the studies that were reviewed (cf. Newton 41 - 43). Cheng et al showed that a 20% increase in bladder cancer "would be difficult to detect", stating that "lifetime exposure of the whole population may have large population effects". As with some of Nuffield's conclusions, Prof. Newton gives an accurate account here of the best evidence but draws an unwarranted lesson from it.
2.5.3 Paragraphs 40 - 44 are therefore open to criticism because assurances about safety cannot be given. Appeals to fluoridation's extent over decades and populations are spurious because of the difficulty in detecting harms among populations, as well as the absence in official quarters and among the research community of any index of suspicion over fluoride's potential for harm. The author learnt this at first hand when tabling a parliamentary question in 1999 about an increase in perinatal mortality in areas of fluoridated Birmingham: the Government's reply was that fluoridation had not been considered since only "known risk factors" were being looked into - a telling sidelight on Prof. Newton's claim in 40 that routine monitoring "has not revealed any unusual health problems associated with water fluoridation". If you do not look you will not see.
2.5.4 Those who have staked their professional judgement on fluoridation since its beginnings 60 years ago have had little motive to do the kind of solid work which might turn up problems. The "theoretical risk" in paragraph 44 may involve undetected adverse effects that are neither "small" nor "rare". No one within the fluoridation culture and not many outside it have pointed up the high toxicity of the fluorides used, which can be found in any standard textbook and which should surely be fed into the debate about the plausibility of harm. Prof. Newton's "certainty of adverse outcomes if [fluoridation] is withheld" (44) is a clear overstatement of York's findings.
2.5.5 There is an error in paragraph 37. It is stated that "the same result [of increased risk of osteosarcoma] is not seen in the bigger study". This was not referenced: no bigger study has appeared. A letter to a journal in 2006 warned against drawing conclusions from Bassin's results because a larger study shortly to be published by the author would provide reassurance about osteosarcoma. We are still waiting for this study, and until it appears, which seems increasingly unlikely, it is wrong to cite it in opposition to Bassin's preliminary findings about cancer in young males.
Equality Impact Assessment
3.1 A report was produced by the Public Health Resource Unit to assess any inadvertent discrimination that might result from water fluoridation in relation to ethnicity, gender, disability, age, sexual orientation or religion. Prof. Newton cites this as generally reassuring (46) in predicting definite or probable reductions in inequality over some of these. It is clear that access to dental prophylaxis, of an uncertain kind, will be facilitated for many of the disadvantaged if fluoridation happens, as it will to everyone served by the water pipes.
3.2 PHRU rightly takes York (principally) and the Australian systematic review as its basis for the scientific evidence on fluoridation's effects, while recognising that there is overlap though not congruence between York's findings and the specific concerns of the EqIA. While much of what the unit wrote was faithful to York, there are errors in the PHRU's account which may affect judgements about fluoridation's likely effect on the above groups.
3.3 PHRU overstates York's findings on benefit. In most places the York reviewers were careful to say that the evidence, which they found surprisingly weak, "suggests" that it prevents caries in children. PHRU uses the stronger "concluded". Care with words is as important as care with figures where public health is at issue. York made no such suggestion about adult teeth since it found one admissible study only; PHRU, while stating that "Further robust reviews" would be "beneficial", could have stressed that Griffin's cited 2007 review on adults (EqIA, 188.8.131.52) drew on a handful of studies which York had found unacceptably weak (see 2.4.8 above). Griffin's abstract says "These findings suggest . ."; PHRU however states "the reviewers concluded . ." Griffin's "27%" reduction in caries was also on a different scale of measurement ("prevented fraction", producing a higher figure) from York's. Evidence for adult benefit is at present minimal.
3.4 On possible harms PHRU gives a fair summary of York, with three exceptions. Under cancers (7.1.4) it is wrong to cite "authors of the primary research" as failing to replicate tentative findings of bone cancer in young males, since as mentioned in 2.5.5 above the fuller paper on this that was foreshadowed has not been published and the question-mark against osteosarcoma remains. And while York could indeed come to no conclusions about other possible harms because of the weakness of the evidence, it did show enough concern to recommend further research in some areas, a call which has been echoed and indeed extended by the major 'weight of evidence' review by the US National Research Council in 2006 (see 4.2 below).
3.5 PHRU makes a common error over York's findings of unsightly dental fluorosis. Under 8.3.1 in box E7 it states that there "could [i.e. might] be no risk at all" of 1 additional child in 22 suffering from this if the fluoride concentration is increased to 1 ppm. What this quotation omits to state is that York's calculations related to an increase from a notional 0.4 ppm (York 7.2.1 Table 7.8), i.e. a 2½-fold increase. This will not equate to the situation in Southampton where the present baseline is 0.08 ppm, producing a 12½-fold increase. From this much lower concentration it would be likely not only that the difference in the proportion of sufferers would be statistically significant, but that the figures would be considerably higher than 1 extra child in 22. The later statement on this topic under 8.3.5 A4 is inaccurately presented. References may be unreliable in this whole section (e.g. York cited inappropriately under 8.3.3 G1), suggesting that the PHRU may have been working under time pressure.
3.6 PHRU goes wide of the mark over evening out dental inequalities between social classes. This is important because there is political pressure to combat inequalities in society, and it can be tempting for public-spirited (and career-minded) health professionals to see evidence for this when it is not there. The Chief Dental Officer led the way in his February 2008 letter urging SHAs and PCTs to fluoridate, referred to by Prof. Newton in 3 and criticised in a paper by this author which was endorsed by two of the independent York scientists (see Reports). York found some weak evidence for dmft in 5-year-olds, but none for proportions of caries-free children of any age. PHRU should also have stressed that this section of the report contained the weakest evidence in the review, with an average study validity score of 1.6 out of 8, leading the reviewers to state that "the ability to answer this question is low". The four leading independent York scientists later wrote that the evidence here was "weak, contradictory and unreliable". To claim as PHRU does that the evidence "showed" that disadvantaged 5-year-olds benefited more in fluoridated areas, and to summarise that York's studies "provide evidence . . that the impact [of fluoridation] is strongest for more disadvantaged groups", is misleading. This unwarranted conclusion is repeated in the later summary.
3.7 Neither PHRU nor any other body has pointed out that measures to protect the "vulnerable", which is often claimed for fluoridation, are a two-edged weapon. What helps may also harm. While the question of adverse effects is still unresolved, it is the very old as well as the very young (see 4.2.9 & 16 below for the latter) who may plausibly suffer consequences from fluoridation that may not be manifest in robust adults in good health. It would have been responsible for some body to bring this to the SHA's attention, and the EqIA was the ideal vehicle since its brief included "age". The chronically sick are another category who might have featured in an equality impact exercise. The contrast to the public caution over folic acid supplementation is striking.
3.8 For these reasons the PHRU summary tables on likely impacts in the EqIA can be misleading. PHRU is appropriately sparing in assigning "definite" to proportionate benefits and disadvantages (1.2); but even the "probable" benefits are questionable where dental inequalities are concerned, and "speculative" would apply more accurately to these and to benefits to adults. The weakness of the evidence found by a high-class systematic review in distinguishing degrees of possible benefit between social classes and geographical areas does not suggest that the specific subdivisions targeted in the EqIA will be supported by anything better, an impression that the references provided do little to dispel. The conclusions of the NRC review which Bazian (see 4.2.4 below) largely rejected have at least as good a claim to be taken seriously on evidence grounds as do those of the PHRU. Failure to discriminate over quality of evidence, to draw the right conclusions from it, and to admit the possibility of harms other than fluorosis, has been a besetting sin in the fluoridation field for too long. And it must be stressed that the only reliably and systematically assessed evidence applies solely to children in the present state of the science, and even then (in York's words) with no "clear confidence".
Scientific issues raised in opposition
4.1 Prof. Newton devotes a section of his report to examining the main objections to fluoridation raised at meetings during the consultation. He is unwilling to concede any merit in them.
4.2.1 He begins with the question of the effect of fluoridation on IQ, but this can usefully be dealt with under his next heading of the US National Research Council report 'Fluoride in Drinking Water' of 2006. This expert review, which was unusual in drawing members from opponents as well as supporters of fluoridation, worked for three years to look at the toxicological, clinical and other data on fluoride in water to assess for the Environmental Protection Agency whether the current maximum contaminant level goal of 4 parts fluoride per million was adequate "for protecting children and others from adverse health effects". They concluded that it was not. An assessment of the value of this review and the validity of criticisms of it is given below.
4.2.2 IQ was one of the areas examined. Any claims that fluoridation has been "shown" to reduce IQ can be dismissed because of the weakness of the studies. The NRC looked at more studies than York had done because its brief was not confined to York's criteria of human studies of a certain standard. It trawled much more widely, with a consequent broadening of the perspective on fluoridation but a diminution in the reliability and applicability of evidence considered. What both review bodies noted was the consistency of the direction of findings, and this moved them to suggest that the area warranted further research. Prof. Newton is right that this did not "constitute convincing evidence of harm"; but some evidence was present. How one deals with this situation in a public health context is the question.
4.2.3 The NRC reviewers were most concerned about risks of severe dental fluorosis, which they considered "a toxic effect". They also felt the present 4 ppm level was "not likely to be protective against bone fractures." They wished to see research into skeletal fluorosis, and had concerns about brain chemistry and function and endocrine effects. They considered the links between fluoride and cancer to be "tentative and mixed".
4.2.4 The NRC report was heavily criticised by Bazian, a respected body commissioned by the SHA to assess its findings. Prof. Newton lists its main charges. Bazian's 50 pages show no concession to anything of worth or real relevance to Southampton's situation, except where it suggests reassurance (it also wrote a separate report on the IQ question). It gives no credit to the depth and breadth of the inquiry, the inclusion of scientists from different perspectives and disciplines which will have reduced the chances of bias in study selection and interpretation, or the thorough peer-review procedures according to the NRC's usual practices under the National Academy of Sciences. The importance of including people with opposing attitudes to fluoridation, as York also did, in a dominant fluoridation culture in a fluoridating country, should not be underestimated as a check against the biases endemic through half a century of fluoridation. This author saw at first hand the usefulness of having people of different persuasions (and none) on the York advisory board.
But there was substance in what Bazian said.
4.2.5 Bazian's first point highlighted by Prof. Newton is that the NRC did not conduct a systematic review, the "gold standard". It was in fact a "weight-of-evidence" review with experts drawing on material of many types, synthesising and weighing its value in reaching conclusions. Given this approach it could not be as reliable as York over water fluoridation (which addressed human studies and fluoridation only) or its Australian counterpart, though it went wider in its scope; it is also uncertain how systematic it was in searching for the whole of the relevant evidence on a given topic. Bazian was unstintingly critical of every section of the NRC report for the quality of its evidence and approach.
4.2.6 It was also critical of the lack of a clear "audit trail", meaning that one could not follow the methods used to gather and assess the evidence. If this is correct it must detract from the impact of the report. It is good practice to be transparent so that others can follow step by step and if they wish replicate what has been done. Few may follow best practice, but the criticism is just if in fact no trail exists.
4.2.7 These are substantial criticisms. But Bazian goes further in its reluctance to acknowledge virtually anything good in the NRC report. In its frequent use of phrases such as "no convincing evidence" of harm, "no conclusive proof", "not necessarily indicative of what would happen in humans", it only states what is (or should be) generally acknowledged. The issue where public safety is concerned is how cautious one should be where there are some indications of possible risk. York, which was the best review of its kind, admitted 'level C' studies for harms but not for benefits on the basis that one needs to flag up potential dangers. Bazian states rightly that in vitro and animal studies can only generate hypotheses. But one needs these hypotheses, which the NRC's approach was capable of generating, and where there is consistency in preliminary studies, or harms are plausible, it is reasonable to expect such hypotheses to be taken further in the interest of public safety. The serious charge against the fluoridation community is that red flags have not been attended to and adequate safety studies have still not been done after 60 years.
4.2.8 As an illustration, in the 1990s US dentists commissioned a toxicologist, Dr. Mullenix, to look into the possible neurotoxicity of fluoride. She worked in the accepted manner, starting with rat studies at a high dose where she found adverse effects. When she reported these her research was closed down. It could therefore be claimed that no convincing evidence of harm to humans had been found. But few researchers would maintain that it was right professional practice to stop at the first hurdle: if you find a red flag, as Dr. Mullenix protested, it is mandatory to lower the dose and look further. It seems that the professional orthodoxy preferred the weak evidence to remain as it was - a 'see-no-evil' approach.
4.2.9 Bazian states that the great majority of studies cited by the NRC are not directly applicable to Southampton because of the differences in fluoride concentration - the second criticism highlighted by Prof. Newton but exaggerated in his statement "Their conclusion was that the report is not relevant" to Southampton. This approach however lends a spurious precision to arguments about 1 as against 4 parts per million, since there is no such clear cut-off in human terms. People drink widely different amounts of water. This author has spoken to a laboratory worker analysing blood samples from people claiming to be sensitive to fluoride in water whose daily reported intakes vary from half a litre to ten litres. People have different personal susceptibilities, especially when ill (in one place Bazian dismisses health concerns because only osteoporosis sufferers were shown to be involved, but this is hard on sufferers in fluoridated areas). The very old and very young are at special risk, and the NRC stated that infants and young children have 3 - 4 times the exposure per body-weight of adults, which means that any harms affecting adults at the NRC's 4 ppm will affect many toddlers at Southampton's 1 ppm. Water is only one source of fluoride exposure, and the intake by target populations of fluoride from food and other sources remains a major unknown. Prof. Newton himself writes of "dose" instead of "concentration", and it becomes easy to think of distinct categories instead of the overlap in population responses to different levels of fluoride. Bazian nowhere addresses this "indirect" applicability, which may affect a significant number of Southampton residents. It found "direct" relevance in one or two cases of reassurance on safety. A clear distinction between direct and indirect is not meaningful.
4.2.10 Toxicologists, who were represented on the NRC committee, ideally like to have a ten-fold margin of safety for unforeseen harms and another ten for individual variation. The small margin between 1 ppm and levels at which harms appear likely is the reason why pharmacologist Arvid Carlsson, Nobel laureate in 2000 in Medicine or Physiology, considers fluoridation unsafe and persuaded his native Sweden to follow his advice. Toxicology seldom features in fluoridation debates. Neither Bazian nor Prof. Newton addresses the significance of narrow therapeutic margins.
4.2.11 The American Dental Association (see also 4.2.16 below) gave an award in 2006 to one of its members in part because he "served nearly three and a half years on [the NRC fluoride review], which will have long lasting effects on the health of the public... He worked tirelessly to bring an objective, evidence-based approach to the Subcommittee's deliberations and recently released final report, which will have direct impact on the Nation's policies regarding fluoride in drinking water." One of the leading promoters of fluoridation thus recognises a "direct" relevance in the report criticised by Bazian.
4.2.12 The third line of criticism is that the NRC's purpose was to look at environmental contamination and "not to assess the health effects of fluoridation schemes". This is disingenuous. While the report was to the Environmental Protection Agency, the preface and summary make plain that establishing the maximum contaminant level goal is about "adequacy for protecting children and others from adverse health effects". The American Dental Association, represented on the NRC committee, clearly thought that the review had implications for children's health, dental or other. "EPA's guidelines are maximum allowable concentrations in drinking water intended to prevent toxic or other adverse effects that could result from exposure to fluoride" (NRC report). The NRC was looking at the health of people, not plants.
4.2.13 Where Bazian quotes from the York report it is inaccurate. Over unsightly dental fluorosis York did not find the difference between 1.2 ppm and lower concentrations to be "not significantly different". That was between 1.2 and 0.4 ppm: lower concentrations still (Southampton has 0.08) could well show statistical significance. PHRU made the same error (see 3.5 above).
4.2.14 It is also misleading to say that York "found no significant increase in the risk of [bone] fracture" at around 1 ppm. Finding "no increase" is different, and not just semantically, from "not finding" an increase, which was York's case with weak studies pointing both ways,: a roughly equal number of analyses found a statistically significant increase as found a decrease. Absence of evidence is not evidence of absence: the latter gives reassurance, whereas the former needs more work to be done. Bazian advocates reliance on York, however, as "directly relevant" to Southampton, but its reassurances are misleading. Because its account of York errs in favour of fluoridation one must be cautious about accepting all its criticisms of the NRC's potential risks.
4.2.15 Other points from Bazian must be omitted for lack of space. Sound criticisms are made of the NRC report and they deserve to be heeded, although no complaint has been recorded by pro-fluoridation members of the review committee about failure to include studies which would have supported their case, and the peer review procedures were thorough. But a reluctance to concede good points in the report, an apparent expectation of proof of harm before safety is taken seriously, over-reliance on a hard distinction between 1 and 4 ppm in the water, carelessness over the York report, and a drafting tone that plays down suggestions of harm, devalue its reassurances. If Prof. Newton aligns himself with the over-stringent attitude of Bazian to the evidence, it would be consistent if he showed equal approval for the approach of the York review which he characterises as unduly sceptical (see 4.2.22 below).
4.2.16 The American Dental Association took its cue from the NRC report in giving "interim guidance" that it "raised the possibility that infants could receive a greater than optimal amount of fluoride from reconstituted baby formula": "parents and caregivers should consider using water that has no or low levels of fluoride". Alarmed that opponents of fluoridation were "mischaracterizing [this] as a springboard" in their "approaches that play on the fears of the public", the ADA in 2007 toned down what was plainly a (low-level) warning to say "If parents are concerned about this possible increased risk, they may choose to use nonfluoridated water . ." (not "particularly concerned" as Prof. Newton puts it). It is significant that a long-term promotional body in the fluoridation lobby should have taken this initiative. This was not disseminated with "all the facts" by the SHA; those advocating fluoridation during the consultation seemed not to know of it; and Prof. Newton's paper plays down its significance.
4.2.17 There appears no acknowledgement of possible risk in Prof. Newton's defence of exceeding the officially safe level of fluoride in the UK: it is apparently enough for fluoridation's advocates that higher intakes are "not . . necessarily unsafe". Reassurance is unconvincing because so little work has been done to discover total fluoride intakes, and because Dr. Mansfield's cited calculations, now in draft, suggest exposure may be much higher than previously estimated, and that fluoridation of a water supply may make most of the population excessive consumers. The Committee on Toxicity in correspondence with this author some years ago seemed incapable of recognising that their figure for a safe level had been publicly revised downwards by its original author many years previously, a revision adopted in some countries but not the UK. Little reliance can therefore be placed on their reassuring conclusion quoted by Prof. Newton, nor in the pro-fluoridation Department of Health keeping the "water fluoride dose"[sic] under review. Eire, Canada and Hong Kong have reduced their levels from 1 ppm in recent years.
4.2.18 Southampton City PCT has given assurances, as Prof. Newton states, about having tried other approaches to dental health before resorting to fluoridation. But it has resisted questions of detail during the consultation into how much effort it put into the various anti-caries measures presented, and for how long, factors important to know since other places with similar socio-economic features have been able to achieve good dental health without water fluoridation. In fact proponents of fluoridation during the consultation seemed unaware of good practice with community schemes being used with apparent success in other parts of the UK, just as they were unaware of or reticent about unfavourable messages from across the Atlantic (see 4.2.1, 4.2.16 above). Prof. Newton and the SHA have apparently accepted the PCT's case at face value. The London Assembly were not so obliging when urged to fluoridate in 2003, believing that more could be attempted by other means.
4.2.19 Especially troubling is Prof. Newton's extension of the case beyond the needs of children. This ignores the status of the evidence for benefit to adults, which York and the Australian systematic review have shown to be minimal (see 2.4.8 above). It also ignores the ethical objection to people being treated against their will, which is part and parcel of the argument about population measures and cannot reasonably be side-stepped on the grounds that only scientific objections are addressed in this paper. Overstated claims of adult benefit will of course have the political advantage of cutting the ground from beneath the feet of adults who might have complained of being fluoridated to help other people's children.
4.2.20 The final five pages in Prof. Newton's evidence section revolve round the York review. Prof. Newton implicitly recognises that York is a stumbling-block, which proponents of fluoridation are seldom bold enough to declare, and he confronts the criticisms of an inadequate research base with a range of arguments designed to devalue York's conclusions.
4.2.21 He is correct that many authorities and professional organisations are, or say they are, completely satisfied with fluoridation's evidence base (but see 4.2.29 below). Unless you rely on eminence-based medicine, however, you have to look deeper. Fluoridation was an attractive idea which caught on in the USA in the mid-20th century, was heavily promoted on slender evidence, and was exported to a handful of countries, mostly English-speaking. Even the Department of Health was not enthusiastic about the early evidence in written answers to this author before the York review. Premature consensus was arrived at under pressure which extended to government level, within organisations who thought they saw an easy way of alleviating dental suffering, and wanted to do good to people and be seen to do good. The Australian Dental Association has called fluoridation dentistry's "flagship". Prof. Newton now calls it, rightly, the "professional orthodoxy". Some others have called it a sacred cow, even a "religion".
4.2.22 It is easy to see how a professional bias grew up in consequence, with a (probably unconscious) urge not to look closely into safety - matched and mutually reinforced by some equivalent and opposite bias from opponents. And it is easy to see how a review like York which disturbs the orthodox view after half a century might be viewed by fluoridators as taking an "explicitly sceptical approach in order to fully test the strength of the evidence base". But Prof. Newton's description is not right. The approach by the reviewers at the NHS Centre for Reviews and Dissemination was no more sceptical than for any other review they conducted, a point that has been confirmed to this author who was part of the advisory board and has heard many objections to the way York operated from anti-fluoridators also. What York found was much genuinely weak evidence, measured by normal and appropriate standards of a high-class systematic review, which the reviewers were "surprised" by. It was indeed surprising for a review of this nature that no 'level A' study could be found in the world literature. If York had been truly sceptical it would not have dropped its standards and admitted 'level C' studies, against the original protocol, for evidence on reducing dental inequalities, in order to be able to say anything meaningful.
4.2.23 Prof. Newton's following arguments look very like special pleading. Safety studies may not be easy, but they are done in other areas of health-care and could have been done if the will was there. It could be claimed that the need for them is greater where whole populations are treated, many among whom cannot benefit or would not have consented. "Decades of surveillance", which has scarcely been whole-hearted, are no substitute; they sound good, but are almost meaningless (see 2.5.3 above). Even randomisation of some kind is not impossible (interestingly, Prof. Newton is concerned about consent here, as also about the fact that York "concluded" that fluoridation was effective, when the reviewers disclaimed such "clear confidence"). "Blinding of outcome assessment is certainly possible" (York, 9.6, in contrast to Prof. Newton), and is essential to minimise bias where professional orthodoxies are involved.
4.2.24 Prof. Newton objects to the way York addresses confounding variables in this instance (there were more that he did not mention), and argues his way through possible explanations. In doing so he takes an unusual route for a scientist in questioning standard good practice (a what-could-it-be-except-fluoridation? approach to suggested benefits). York (and other) scientists might point to the overturning of some medical beliefs, which were based on better evidence than exists for fluoridation, when high-quality studies were done, for example over HRT and cardioprotection; or to equal improvements in dental health in non-fluoridating countries. Prof. Newton's approach cannot command respect when one is looking for good science. The Australian review which he cites is only "less sceptical" than York about effectiveness when it disregards York's caveats: it found no new evidence sufficient to change York's conclusions.
4.2.25 Prof. Newton again prefers the unsystematic MRC review (see also 2.4.9 above), with its incorrect conclusion of "no evidence for any significant adverse health effects", to the systematic York review, whose senior reviewer stated in the British Medical Journal of 16th June 2001 "We have been assiduous in our paper, our full report, and our contacts with the media, not to convey a message of no evidence of harm".
4.2.26 "The important question is not whether further good quality research would help, but what does the existing research tell us we should do now?" It is usually possible to phrase a question to suit your case; but even here, were Prof. Newton hypothetically to come to the Medicines and Healthcare products Regulatory Agency with a request to give the green light to a therapeutic substance with fluoridation's levels of evidence, he would undoubtedly be told "Go away and do the kind of good-quality studies we expect before we sanction a medical intervention, and come back again when you have done so".
4.2.27 If Prof. Newton believes, however, that the quality of science should be lower for a substance given to whole populations, which will contain vulnerable sub-groups, for a lifetime, without individual consent or oversight, than for a drug given one-to-one by a doctor, this should be made clear. Lacking in this exercise, and in Prof. Newton's advice, has been any consideration of what standard of evidence is needed for this particular intervention with its unique set of features. Lacking, too, is acknowledgement that review after review has called for more research into fluoridation's effects, that this includes some fundamental as well as peripheral questions, that this has been accepted by government, and that this does not sit easily with a decision to extend fluoridation without waiting for answers. Unfortunately it is in the nature of public health professionals to want to "do something", with the risk of inappropriate or premature intervention.
4.2.28 The precautionary principle as set out by the European Commission and quoted in Prof. Newton's paper gives scope for debate over uncertainty of evidence and reasonableness of grounds for concern. On the one side over fluoridation is a professional consensus, as he says, which has existed for decades; on the other are the leading participants in the only truly scientifically reputable review, some of whom have been trying for eight years to make an unwilling community aware of the true state of the evidence, and an unknown number of others. Prof. Newton reverts here to pre-York claims that "water fluoridation is in fact safe and effective". It bears repeating that the chair of the York advisory board wrote "The review did not show water fluoridation to be safe", in an open letter in 2001 which took the medical and dental professions to task for their misleading statements; and that the York reviewers wrote that "To have clear confidence in the ability to answer the question [about efficacy], the quality of the evidence would need to be higher." The sharpness of the contrast has been fudged by most professional bodies since 2000, and by Prof. Newton in his report.
4.2.29 Consensus is not unanimity. Nobel prize winners may speak and have spoken against fluoridation. But the average British Medical or Dental Association member or NHS employee in a hierarchical profession does not oppose the party line in public, for good career reasons; nor does s/he read the original studies or reviews and form an independent opinion - only a tiny minority does this, in any profession. In the Southampton case there has already been a formal complaint against a health services research academic who did not keep his views against NHS fluoridation policy to himself. Opposition has its disincentives, and consensus is not always what it sounds.
5.1 Prof. Newton concludes with a short summary and a statement that the concerns raised in section 4 above are not cogent arguments that outweigh the health arguments for fluoridation. If cogency is in the eye of the beholder, then the professional orthodoxy will win every time under this system. The SHA should have been well aware of key statements about the evidence coming from leading scientists from the York review, who can be expected to know better than anyone about fluoridation's disputed benefits and harms; if they did not read these - and they were submitted as written with the response on 17th December by the All-Party Group which the present author co-chairs - there has been neglect in some quarters. Ethics do not form part of Prof. Newton's case, nor it seems did the board have before it a balanced dissection of the arguments for and against coercion as they apply to an intervention making medical claims (Paper HA09/ 019 dealt somewhat inaccurately with the issue). By the Regulations these need to be put into the scales as well, together with environmental arguments which have not been addressed at all in this consultation process, despite the SHA's claim in its consultation document that it had done so.
5.2 It was good practice to put before the SHA board, as Prof. Newton did, the 'official' summaries of the two systematic reviews which constituted best evidence on the science. The board, faced with much paperwork and with limited expertise in interpreting disputed scientific evidence, will have relied on their Regional Director of Public Health's guidance; to go against this in a matter of such import would have been difficult. When he writes (53) that "The balance of evidence in the systematic reviews . . is in favour of the effectiveness and safety of fluoridation" it is not clear what he is claiming. If he advises that fluoridation be initiated on (say) a 55% probability of effectiveness and safety, he is departing a long way from the standards of good science that require at least 95% certainty, the traditional figure for statistical significance. (Any figure is arbitrary, but 95+% is the convention.) If on the other hand he means that on balance he judges the evidence to have reached this conventionally accepted level of proof, he is out of step with the scientists from York among others. (See 2.4.5 above for overstatements by the later Australian review.)
6.1 Prof. Newton's report resembles others from the professional orthodoxy that have been written since 2000. There is acknowledgement of the primacy of the York review, which not all recent reports have recognised. Yet there is a pick-and-mix over what York said, with its findings overlooked or watered down where they are inconvenient for the case being made. The selectiveness begins with the age-group of children which underpins the SHA's case. There are errors and overstatements, in each case in support of fluoridation. There is an unspoken assumption in the Bazian document that adverse effects must be proven before they are factored in, which is in conflict with the precautionary principle cited, which in turn is inappropriately discarded as not applicable to Southampton because of a pre-existing "consensus" view about safety.
6.2 Few if any drawbacks to fluoridation are countenanced, and there is a demonstrable professional failure to keep up with developments in the field that might cast doubt on a fluoridation policy. The appeal to 50 years of history is not scientific, and if monitoring had been effectively done, as is suggested, York would have included it. The narrow therapeutic margin is not mentioned. Crucially, there is a failure to acknowledge the lack of any good evidence for reducing inequalities in dental health, which goes to the heart of what fluoridation is about in public health terms. Nor is there reference to historical perspective in the case of a short-term unexplained rise in caries followed by a more recent decline, in one age-group of children only.
6.2 Our society has suffered before because of the deletion of cautions from expert bodies, notably about destructive weapons in the Middle East. It is a temptation to public health scientists to take action, and with political pressure coming from the Department of Health questions such as the possibility of adverse effects, the paucity of safety (and indeed many other) data, normal ethical standards on consent, and the views of the population affected, can appear minor obstacles in the path of doing what seems to be good. In addition, with caries at historically low levels regardless of fluoridation, and the view of at least one experienced community dental health expert that no gains can be expected from fluoridation at current levels of dental decay, the present exercise has been driven more by faith and the urge to do good while implementing departmental policy than by good science. Prof. Newton's paper, a product of the long-standing pro-fluoridation culture of the "professional orthodoxy", does not provide a reliable guide to the state of the evidence.
29th May 2009.