vvEPA
820-R-10-016
COMMENT-RESPONSE SUMMARY REPORT
for the
PEER REVIEW
of the
FLUORIDE:
DOSE-RESPONSE ANALYSIS FOR NON-CANCER EFFECTS
DOCUMENT
November 2010
Office of Water
Office of Science and Technology
Health and Ecological Criteria Division
U.S. Environmental Protection Agency
Washington, D.C. 20004
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
TABLE OF CONTENTS
ACKNOWLEDGMENTS ii
I. INTRODUCTION 1
II. CHARGE TO THE PEER REVIEWERS 3
III. PEER REVIEW COMMENTS AND EPA RESPONSES 5
IV. RESPONSE TO DR. WHITFORD'S COMMENTS CONCERNING
FLUORIDE ANALYTICAL METHODOLOGY 30
V. EXTERNAL PEER REVIEW REPORT 35
November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
ACKNOWLEDGMENTS
This document was prepared with the technical assistance of the Toxicology and Hazard
Assessment Group of the Environmental Sciences Division of Oak Ridge National Laboratory,
Oak Ridge, Tennessee, under EPA IA No. DW-89-92209701 andDOEIAGNo. 1824-S881-A1.
Principal EPA Scientists are Joyce Morrissey Donohue, PhD, and Tina Duke, MPH, Health and
Ecological Criteria Division, Office of Science and Technology, Office of Water, U.S.
Environmental Protection Agency, Washington, DC.
The Oak Ridge National Laboratory is managed and operated by UT-Battelle, LLC., for the U.S.
Department of Energy under Contract No. DE-AC05-OOOR22725.
ii November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
I. INTRODUCTION
The United States Environmental Protection Agency (EPA), Office of Water is charged
with protecting public health and the environment from adverse exposure to chemicals and
microbials in water media, such as ambient and drinking waters, waste water/sewage sludge and
sediments. In support of this mission, the Office of Water/Office of Science and Technology
(OST) develops health standards, health criteria, health advisories, and technical guidance
documents for water and water-related media. Under this work assignment, documents prepared
by OST are to undergo external peer review.
Peer review is an important component of the scientific process. It provides a focused,
objective evaluation of a research proposal, publication, risk assessment, health advisory,
guidance or other document submitted for review. The criticisms, suggestions and new ideas
provided by the peer reviewers ensure objectivity, stimulate creative thought, strengthen the
reviewed document and confer scientific credibility on the product. Comprehensive, objective
peer review leads to good science and product acceptance within the scientific community.
The Peer Review for Fluoride: Dose-Response Analysis for Non-cancer Effects was
conducted on March 11, 2008, in Washington, DC, to allow the external peer reviewers to
discuss their evaluations of the EPA/OW document. The Peer Review was conducted under EPA
Contract Number EP-C-07-021 with ToxServices, Washington DC (Work Assignment B-02
Task 5) and managed by ICF International, Fairfax, VA.
The list of external peer reviewers and their affiliations are shown below:
Jane A. Cauley, Dr.P.H, Professor, Department of Epidemiology,
University of Pittsburgh
Pamela Den Besten, D.D.S., M.S., Professor and Chair, Division of
Pediatric Dentistry, Department of Orofacial Sciences, University of
California at San Francisco
Richard D. Jackson, D.M.D., Assistant Professor, Preventive and
Community Dentistry, School of Dentistry, Indiana, University, Oral
Health Research Institute
Gary M. Whitford, D.M.D., Ph.D., Regents' Professor, Department of Oral
Biology and Maxillofacial Pathology, Medical College of Georgia
The Charge to the Peer Reviewers is presented in Section U of this report. The complete peer
review process, including pre-peer review meeting comments, post-peer review meeting
November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
comments and a summary of the meeting comments, as prepared by ICF, are presented in Section
V. OW, with the assistance of ORNL, prepared a response to each of the general comments of the
reviewers as summarized by ICF, as well as point by point responses to the specific comments of
each of the peer reviewers; and these are presented Section HI. One reviewer (Dr. Whitford) had
major concerns about the analytical methodology that was used in the 1930's and 40's to measure
fluoride in drinking water. As a result, ORNL provided a detailed evaluation of the analytical
methodology, and this is presented in Section IV.
November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
II. CHARGE TO THE PEER REVIEWERS
1. Was the document clear and transparent in its presentation of data and explanation of the
analytical approaches used to characterize the concentration-response and dose-response
relationships for severe dental fluorosis? If not, do you have any suggestions that will
enable the authors to improve its clarity?
2. Are you aware of any significant publications related to severe dental fluorosis or the
skeletal effects of fluoride that are not included in the noncancer assessment document?
3. Are the strengths and weaknesses of Dean (1942) study (selected as critical) fully
characterized?
4. Are some teeth more susceptible to severe fluorosis than others? If so, should the OW
make modifications to the age range identified as the period of concern?
5. Are the data on cavities as collected and graphically presented by the OW consistent with
the hypothesis that there is an increased risk for cavities in susceptible individuals with
severe dental fluorosis?
6. Are there recent data that would impact the Institute of Medicine (IOM, 1997) Adequate
Intake Value of 0.05 mg/kg/day for fluoride that the OW should consider in its
assessment?
7. Can you suggest an approach to transform the water concentration data from the Dean
(1942) study to units of dose for the population susceptible to severe dental fluorosis
other than that used by the OW?
8. Can you provide input concerning the strengths and weaknesses of the approach utilized
by the OW to identify a lower bound dose for severe dental fluorosis?
9. Are you aware of dose estimates other than those from IOM (1997) and the World Health
Organization (WHO; 2002) that are appropriate critical doses for skeletal effects and/or
can you suggest a different approach that the OW might use to estimate the fluoride dose
associated with skeletal fractures using available data?
10. Are you aware of any data that can be used to demonstrate that protection of the
secondary teeth from severe dental fluorosis will also protect the primary teeth?
11. Are there factors other than those discussed in Section 3.1.4 that increase sensitivity to
fluorosis of the teeth or bones?
November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
12. Do you support the OW s conclusion that an RfD of 0.07 mg/kg/day will be protective for
severe dental fluorosis in children and skeletal effects in adults while still providing for
the beneficial effects of fluoride?
November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
III. PEER REVIEW COMMENTS AND EPA RESPONSES
1. Was the document clear and transparent in its presentation of data and
explanation of the analytical approaches used to characterize the concentration-
response and dose-response relationships for severe dental fluorosis? If not, do
you have any suggestions that will enable the authors to improve its clarity?
Reviewers' General Comments. In general, the reviewers thought that the document was clear and
transparent. In addition, the reviewers commented that the document clearly described the available
literature and presented the information in an understandable format. In particular, Dr. Jackson noted that
the individual tables summarizing the studies were very helpful.
EPA Response: No response needed.
Specific Comments:
Dr. Cauley: Recommended reorganizing the document in the format of a manuscript. She also asked for
clarification on OW and NRCs roles in the project.
EPA Response: EPA reports are not normally written in the format of a research manuscript; however,
additional information on background and rationale and purpose of the report were provided in a preface
and added to the document.
Dr. DenBesten: Page 13. State the purpose of the Secondary Maximum Contaminant Level.
EPA Response: The definition of a National Secondary Drinking Water Standard was added to the text.
Dr. Den Besten: Page 15. The statement with regards to fluoroapatite, at the end of paragraph 2.1.1 is
not accurate. The mineral formed in tooth enamel exposed to higher fluoride levels is fluoride containing
carbonated apatite. Fluoride levels in subsurface fluorotic enamel are about 200 ppm rather than the 10-
100 ppm fluoride in normal enamel, whereas fluorapatite is about 30,000 ppm fluoride. Precipitation of
fluoride mineral salts at the surface of enamel results in high surface level, though this also is not
fluoroapatite. This fluoride-substituted apatite has some increased resistance to bacterial acids that cause
tooth decay. However, the primary function of fluoride in drinking water in reducing tooth decay is
topical, primarily by the enhancement of remineralization.
EPA Response: Reference to "fluoroapatite" has been removed form the text and Dr. Den Besten's
clear explanation of the mechanism has been used in its place.
Dr. Den Besten: Page 19 section 2.2. The apatite formed in bone is also a fluoride-substituted
hydroxyapatite rather than a fluoroapatite.
EPA Response: Reference to "fluoroapatite" removed from text.
November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
Dr. Den Besten: Page 21 section 3. Change "...are preferred for evaluating the potential effects of
fluoride in drinking water", to the more accurate statement, "... are preferred for evaluating the potential
effects of ingested fluoride".
EPA Response: Text modified accordingly.
Dr. Den Besten: Page 57 paragraph 1. It should be made clear that studies on water fluoridation
conducted after 1980, are confounded by additional sources of fluoride, and changes in use of tap water.
For example, decreasing fluorosis in more recent studies may be related to reduced consumption of tap
water as use of bottled water increases. In general for all of section 3.2.2, when fluoride effects on dental
caries are discussed, the data should be divided into studies before and after 1980 when fluoride became
widely available in toothpaste, and perhaps before and after the late 1990s when bottled beverages
became widely used.
EPA Response: Statements similar to the ones requested by Dr. Den Besten are presented in the first
paragraph of Section 3.2.2 of the EPA report, although a specific time frame for the introduction of
fluoridated toothpaste and the increased use of bottled water is not given. These statements were revised
to include Dr. Den Besten's estimates.
Dr. DenBesten: Page 86; Table 3-52. Please indicate what does "complete" and "total" refer to?
EPA Response: Complete means complete fractures; total means total fractures (complete plus
incomplete, the latter of which was defined by the study authors as stress fractures observed by
roentgenography in participants reporting acute lower extremity pain syndrome.). Table 3-52 has been
modified to include the incomplete fractures, and a footnote has been added defining incomplete
fractures.
Dr. Den Besten: Page 94 section 4.4. Explain what the "NOAEL/LOAEL" approach is, or at least spell
out the acronym.
EPA Response: Acronyms were spelled out, and added to the List of Acronyms.
Dr. Den Besten: Page 97 paragraph 1. The importance of fluoride as a nutrient may need to be
reassessed, given that its primary function in caries prevention is topical. It would seem more appropriate
to focus on the upper limits for ingestion of this caries preventive agent, and leave it to future panels to
assess the relative importance of the lOM's recommended intake of fluoride and risk of severe fluorosis.
EPA Response: The issue of the importance of fluoride as a nutrient is not within the scope of the
current document. The primary task of the OW was to assess the dose-response data for severe fluorosis
as recommended by the NRC.
Dr. Den Besten: Page 98, paragraph 2, sentence 3. The statement as to the timing of secondary incisor
tooth formation is incorrect. The secondary incisors and molars begin development in utero.
Change "development" to "mineralization".
November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
EPA Response: The word "mineralization" was substituted for development; i.e., "The mineralization
of the secondary teeth begins at about 6 ± 2 months with the incisors, whereas that for the primary teeth
begins in utero (Massler and Schour, 1958)".
Dr. Den Besten: Page 101. What is the rationale for setting the BMDL at an incidence of 1% severe
fluorosis? I recommend setting the BMDL at an incidence of 99% moderate fluorosis, which would show
an intent to eliminate the adverse effect of severe fluorosis secondary to fluoride added to drinking water.
EPA Response: A BMR (Benchmark Response) of 1% is the standard level used by the EPA, as most
data sets cannot statistically support calculating a lower level because of the potential variability in
background levels. However, if the data set is large enough, a lower response level can be used. A
statistical analysis was conducted on the Dean data and it was determined that it would support a lower
BMR of 0.5% severe dental fluorosis, but not 0.01%. The BMD software was run for 99% moderate
fluorosis; the log probit model predicted a BMD of 31.6 mg/L and a BMDL of 25.6 mg/L (a poor data fit
due to the fact that the occurrence of moderate fluorosis for all study populations was less than 50%;
therefore, the use of a BMR of 99% moderate dental fluorosis is not considered appropriate.
Dr. Den Besten: Page 103. As stated above, I question a recommended fluoride intake, and feel that this
document should focus only on the dose response analysis.
EPA Response: See response above.
Dr. Den Besten: Page 104, paragraph 1. The statement that "...fluoroapatite crystals disrupt the
hydroxyapatite crystal lattice..." is incorrect and should be deleted.
EPA Response: That statement was deleted from the text.
November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
2. Are you aware of any significant publications related to severe dental fluorosis or
the skeletal effects of fluoride that are not included in the noncancer assessment
document?
Reviewers' General Comments: Three of the reviewers, Dr. Cauley, Dr. Jackson, and Dr. Whitford,
thought that there were additional publications related to severe dental fluorosis and/or skeletal effects of
fluoride that should be reviewed for possible inclusion.
EPA Response: The literature suggested by the reviewers was obtained and evaluated for inclusion in
the document.
Specific Comments:
Dr. Cauley noted that while the dental fluorosis literature review in the report covered a wide range of
fluoride levels, the skeletal/ fracture studies were limited to those with fluoride levels > 4 mg/L. She
suggested that the Office of Water consider several specific publications that discuss the occurrence of
fractures at lower levels of fluoride. Citations for the suggested references were provided.
EPA Response: A limited number of additional papers on the association of bone density and skeletal
fractures with water fluoridation levels of less than 1 mg/L were added to the report and the reader was
referred to several review papers for more detailed summaries. Because the NRC (2006) identified a
water fluoride level of 4 mg/L as being the potential threshold for skeletal effects, the EPA report
intentionally focuses on such studies rather than those that examined lower levels of fluoride in drinking
water.
November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
3. Are the strengths and weaknesses of Dean (1942) study (selected as critical) fully
characterized?
Reviewers' General Comments: In general, the reviewers believe that the strengths of the Dean (1942)
study were well characterized in the draft dose-response report.
EPA Response: No response needed.
Specific Comments:
Dr. Cauley noted that additional strengths of the study include: (1) the wide range of fluoride
concentrations, although there were fewer subjects in the high fluoride categories; (2) the dose-response
relationship illustrated by the data showing increasing risk of severe skeletal fluorosis with increasing
fluoride; and (3) the consistency of the findings across several different communities.
EPA Response: These additional points were added to the report.
Dr. Jackson doubted that anyone would disagree that the Dean (1942) study has been and will continue
to be a benchmark study in the dental literature, as well as the much broader literature related to public
health and epidemiology. He noted that Dean's study was performed when confounding fluoride sources
were not available and, thus, probably gives a very clear picture of the prevalence of the relationship of
fluoride ingestion and the subsequent development of dental fluorosis.
EPA Response: No response needed.
Dr. Cauley and Dr. Jackson noted that the Dean (1942) study does not provide information on
race/ethnicity, and given the probable characteristics of the region, the population of subjects examined
was likely not diverse in terms of racial or cultural characteristics.
EPA Response: Dean (1942: page 31) did indicate that all his study populations were Caucasian. The
dose-response report has been modified to communicate this more clearly.
Dr. Cauley noted that the data were collected the late 1930-40's. Although confounding by use of other
fluoride products would be minimal, there are many other cohort differences between children exposed to
fluoride in the late 1930-40's as compared to now. For example, dental hygiene, dietary intakes (e.g., less
water and more carbonated beverages), and body weight are very different in today's children compared
to those in the 1930s.
EPA Response: These factors were added to Section 3.1.1 in the listing of weaknesses of the study.
Dr. Cauley questioned whether puberty and/or hormonal changes may influence fluoride effects, which
may be important because age of menarche has been decreasing. Another weakness of the studies
November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
reviewed is that there appears to be no information on exposure duration (e.g., How long did these
children live in each community? Did the inclusion criteria include a minimum time of residence?).
EPA Response: The concern about hormonal changes was added as a potential weakness of the study.
Dean (1942) specifically included only children who were lifetime residents of the communities sampled.
Dr. Jackson further noted that it has been postulated that genetic factors may impact the expression of
dental fluorosis at identical levels of ingestion. The fact that data were collected in what may have been
an exclusively white population appears to limit its applicability for use as a benchmark.
EPA Response: Additional papers on genetic factors influencing susceptibility to fluoride were added to
the report (new Section 3.1.4.4). Only one study was located which specifically looked at fluorosis in
both Afro-Americans and Caucasian populations.
Dr. Whitford believes that the weaknesses of the study were not necessarily fully characterized. One
example given in his discussion of how the weaknesses of the Dean (1942) study were not fully
characterized was the lack of review of the publications presented in his response to Question #2
(Elvove, 1933). He also had the following specific comments:
A major weakness of the Dean (1942) report is the chemical method used for the determination of
fluoride concentrations in water (Elvove, 1933). The zirconium-alizarin method is rarely, or probably
never, used today because of its relative insensitivity, several interfering substances, and lack of
specificity for fluoride. In their 1952 report that described improvements to the method, Megregian and
Maier (1952) noted that Elvove's original method (1933) had several shortcomings including "non-
conformity to the color laws, limited effective fluoride range, and little color change per increment of
fluoride." It also appears that Elvove (1933) used the visual method to determine color changes in the
zirconium-alizarin reagent (since he referred to "Nessler tubes") which requires subjective judgments and
is less accurate that spectrophotometric methods.
EPA Response: For a complete response to Dr. Whitford's comment, see Section IV of this document.
A shortened version of EPA's response was added to the EPA report (see Section 3.1.1). It is
acknowledged that the method used by Elvove (1933) did not have the sensitivity or minimum level of
detection as more modern methods; however, the results are internally consistent, appear to be supported
by later studies on some of the same water sources, used mostly average values of 12 consecutive
monthly samples, therefore compensating for potential individual analysis error or seasonal variation,
and, based on water quality data from the same time period, were not likely to be compromised by high
levels of interfering substances. In addition, according to Megregian and Maier (1952) the reagent used
was sensitive to small increments of fluoride over a range of 0.0 to 3.0 ppm, the critical range for
assessing the threshold for severe fluorosis, and within this range the response approximated Beer's law.
Dr. Whitford: The study population in the Dean (1942) may not have been continuously exposed to the
community's communal water supply. Dean (1942, page 25) listed two major requisites for quantitative
evaluation of the dental effects of ingesting water containing fluoride. One of these requisites was "a
population continuously exposed throughout life to the variable under investigation (the communal water
supply)." Dr. Whitford recommended that the original papers summarized in Dean (1942) be examined
to determine the extent to which the children met the requisite cited above and that the information be
10 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
included in the Dose-Response Analysis document. If such information is not available, then the
document should note this and discuss the implications in its conclusions.
EPA Response: By listing "continuous residence" in a community as a pre-requisite for inclusion in a
drinking water study, the implication is that Dean would have used only children who were lifetime
residents of those particular communities. In support of this supposition is the statement made by Dean
(1942, pg. 29) that 289 children studied in Amarillo, TX, "used continuously throughout life the
municipal water for drinking and cooking..." In an earlier 1936 paper (Amer. J Public Health 26:567-
575), in which Dean evaluated fluorosis in children in 10 cities, it is specifically indicated (Table 11) that
the children included in the survey were those who had "used municipal water continuously". Some of
these were the same cities studied in 1942. Therefore, the conclusion is that Dean (1942) included in his
study only children who were lifetime residents of their communities.
Dr. Whitford further commented that the appropriateness of the LOAEL (2.2 ppm) and the calculated
reference dose (PvfD) (0.07 mg F/kg bw/day) reported in the OW's draft dose-response report are based
largely on the accuracy of the water fluoride concentrations shown in Dean (1942), as well as on several
other variables that may have affected the outcomes of the epidemiological studies. He indicated that his
preceding comments draw attention to several shortcomings of the chemical method used and other
limiting aspects of the studies summarized by Dean (1942) and recommended that the uncertainties
associated with these factors be discussed wherever appropriate and certainly in the "Uncertainty
Factors" section.
EPA Response: The issues raised by Dr. Whitford's are discussed in Section IV of this report. The
uncertainties associated with the analytical method used by Elove are mentioned in the EPA report.
11 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
4. Are some teeth more susceptible to severe fluorosis than others? If so, should the
OW make modifications to the age range identified as the period of concern?
Reviewers' Overall Comment: The reviewers had differing opinions on whether some (and which)
teeth are more susceptible to severe fluorosis.
EPA Response: No response needed.
Specific Comments:
Dr. Cauley thought age may be a factor in the susceptibility to severe dental fluorosis and suggested that
a more direct discussion of the age at risk is needed and that this topic should be highlighted in a separate
section. She suggested that, in particular, a table summarizing the ages of the children in each study and
the age range which appeared to be at highest risk would be helpful. She recommended that the writers
add this to Table 3-16.
EPA Response: Average ages of the study populations were added to Table 3-16. No specific
information was found in the available literature indicating that a particular age or tooth type is more
susceptible to fluorosis than any other. Because different teeth undergo mineralization at different times,
they will have different susceptibility periods; however, as noted below by Dr. Jackson, susceptibility
may be a function of the length of the mineralization period.
Dr. Den Besten thought that the report was clear in showing that posterior teeth also are susceptible, and
the case for including children up to age 14 was clear and compelling.
EPA Response: No response needed.
Neither Dr. Jackson nor Dr. Whitford has come across any literature that indicates that some teeth are
more susceptible to the development of severe dental fluorosis.
EPA Response: No response needed.
Dr. Jackson believes that the susceptibility is the same assuming that the exposure is constant and taking
into consideration how long the developing tooth is exposed to higher levels of fluoride. He stated that,
for example, maxillary third molars may present greater evidence of severe fluorosis (pitting, staining,
etc) because they take longer to develop (12-16 years-of-age) and erupt as opposed to a maxillary central
incisor (5 years-of-age).
EPA Response: No response needed. Dr. Jackson's information is supportive of the age range EPA
chose as that for greatest susceptibility to dental fluorosis
Dr. Whitford also hypothesized that the posterior teeth may be more susceptible since their development
is more protracted than that of the anterior teeth.
12 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
EPA Response: No response needed. Dr. Whitford's comment is also supportive of including the age
for the development of the third molars in the dose-response assessments. Many earlier studies have only
included the age to the development of the second molars in their assessment of fluorosis.
13 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
5. Are the data on cavities as collected and graphically presented by the OW
consistent with the hypothesis that there is an increased risk for cavities in
susceptible individuals with severe dental fluorosis?
Reviewers' General Comment: Dr. Cauley, Dr. Den Besten, and Dr. Jackson agreed that the data on
cavities presented in the report are consistent with the stated hypothesis. Dr. Whitford disagreed with
the other reviewers, noting that he felt the available data present an unclear relationship.
EPA Response: EPA has characterized the data on an increased risk for cavities in individuals with
severe dental fluorosis as suggestive yet supportive of the NRC conclusion that the thinning and pitting
of the enamel that occurs with severe dental fluorosis can lead to an increase in cavities as observed in
some but not all studies. The data also indicate that at levels beyond very mild fluorosis the reduction in
decay compared to the increase in fluorosis becomes less dramatic than that observed between no
fluorosis and mild fluorosis.
In evaluating the NRC conclusion, it is important to compare across groups that fall in the same age
range and have enough permanent dentition for a long enough period of time to support an evaluation of
differences in decayed, missing and filled teeth or decayed, missing and filled surfaces. Many of the
studies reviewed did not present the data in a manner that was suitable for evaluating the NRC
determination that there is an increased cavity risk for individuals with severe dental fluorosis compared
to those with less than severe fluorosis.
Specific Comments:
Dr. Cauley noted that the data summarized in Table 3-40 to 3-44 are convincing that dental caries are
more common in subjects with severe fluorosis. The data suggest a U-shaped relationship with fluoride: a
higher risk in those with very low intakes and intakes >2-3 mg/L.
Dr. Jackson noted that the OW may want to consider examiner bias in the methodology of these studies.
Although it would be impossible to eliminate this bias, it is something that should at least be mentioned
as a possible confounder. He also described the findings of the following study (see Appendix B for a
description of the study findings):
Jackson R, Kelly S, Katz B, Hull J, Stookey G. (1995). Dental fluorosis and caries
prevalence in children residing in communities with different levels of fluoride in the
water. J Pub I Health Dent. 88:79-84.
EPA Response: Text was added to Section 3.2 to point out the possibility of examiner bias. The
reference mentioned by Dr. Jackson was ordered and incorporated into the report.
Dr. Whitford disagreed with the other reviewers, noting that the available data present an unclear
relationship. He discussed conflicting results as follows: Driscoll's work (1983, 1986) in Illinois and
Iowa did not indicate a relationship between severe fluorosis and caries (pages 46- 47). Eklund's report
14 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
(1987) did not find a relationship when the entire dentition was considered. They found more caries in
severely fluorosed anterior teeth and premolars but not in the molars. In their Chinese study, Chen et al
(1989) found no difference in caries scores between the group without fluorosis and the group with
severe fluorosis. Warnakulasuriya et al (1992) reached a similar conclusion in their Sri Lanka study but
the validity of the conclusion was less clear because of the way they grouped the fluorosis categories. On
the other hand, Mann et al (1987, 1990) and Olsson (1979) found that DMFS scores were directly related
to the severity of fluorosis in Israel as did Wondwossen et al (2004) in Ethiopia. Ermis et al (2003)
reported a slightly higher prevalence of caries in moderate-to-severely fluorosed teeth but the relationship
was not statistically significant
Dr. Whitford noted that overall, and as summarized in Figure 3-7 on page 71, the relationship between
the severity of dental fluorosis and the risk of caries is suggestive, but not convincing. He believes that
J OO ? O
this subject requires more study with control for variables that are known risk factors for caries before a
reasonably firm conclusion can be drawn about the possibility of an association between severe dental
fluorosis and an increased risk of caries.
EPA Response: EPA agrees that the results of the various studies are not entirely consistent (as is
documented in Tables 3-16 and 3-40 in the EPA report), but notes that some of the comparisons cited by
Dr. Whitford are between caries scores in populations without fluorosis and those with severe fluorosis.
Groups without fluorosis are likely to include individuals not receiving the anticaries benefits of fluoride;
therefore, the results would be biased towards higher levels of caries in the non-fluorosis group, which
would obscure the significance of any increase in caries at the higher levels of fluorosis compared to
those at the lower levels. When these individuals are excluded from the evaluation (see results of the
Driscoll et al. 1986 study as shown in Table 3-41 in the EPA report), there is clear support for an increase
in caries with severe fluorosis. Because the Driscoll et al. study was conducted in the U.S., its results are
likely to be more relevant than studies conducted on non-U.S. populations. Nevertheless, Dr. Whitford's
suggestion that additional studies are needed that control for variables that are known risk factors for
caries is a valid one, and this suggestion has been added to the text of Section 3.2.4 of the EPA report.
15 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
6. Are there recent data that would impact the Institute of Medicine (IOM, 1997)
Adequate Intake Value of 0.05 mg/kg/day for fluoride that the OW should
consider in its assessment?
Reviewers' General Comments: Dr. Cauley, Dr. Jackson, and Dr. Whitford were unaware of any
recent data that could influence the adequate intake (AI) value.
EPA Response: No response needed.
Specific Comments:
Dr. Cauley suggested that the writers mention life stage and the upper limit of toxicity. She pointed out
that the American Dental Association website may have updated their fluoride document in 2006. In her
final comment for Question 6 (see Appendix B), she provided two tables: Criteria and Dietary Reference
Intake Values for Fluoride by Life Stage Group and Tolerable Upper Intake Levels by Life Stage Group.
EPA Response: Life stage and the upper limit of toxicity (IOM Criteria) are discussed in the EPA report.
Parts of the tables listing the IOM Criteria have been added to the EPA report (as new Table 5-1). The
ADA Report basically supports the USPHS fluoride drinking water levels.
Dr. Whitford: Although Dr. Whitford did not know of any additional data, he commented on the
interpretation of the lOM's values. He noted that the lOM's AI value represents the amount of intake of
any substance "needed to maintain a defined nutritional state or criterion of adequacy in essentially all
members of a specific healthy population." In the case of fluoride, the AI (see page 301, Dietary
Reference Intakes, 1997) "is based on estimated intakes that have been shown to reduce the occurrence
of dental caries maximally in a population without causing unwanted side effects including moderate
dental fluorosis." He noted that this does not mean that intakes somewhat higher than 0.05 mg/kg/day
increase risk of moderate dental fluorosis. In fact, the lOM's estimate for the threshold for that risk is
0.10 mg/kg/day.
EPA Response: The lOM's estimate of 0.10 mg/kg/day as the threshold for risk of moderate fluorosis is
in the EPA Report (see Section 5.1 Nutritional Guidelines). The IOM (1997) based this estimate on the
data of Dean (1942), concluding that there was less than a 5% prevalence of moderate fluorosis at 2 mg
F/L. EPA ran the Benchmark Dose software using the Dean data for moderate fluorosis (excluding the
three highest concentrations as they appeared to be outliers) and, from a log probit analysis, derived a
BMD of 2.17 mg/L and a BMDL of 2.03 mg/L for 5% moderate fluorosis. Using the same log probit
model in the BMD software, the BMD for 1% severe fluorosis is 2.31 mg/L and the BMDL is 2.13 mg/L.
Therefore, lOM's "threshold" for moderate fluorosis is at a level close to the level associated with 1%
severe fluorosis.
Dr. Den Besten disagreed with the presumption that the lOM's value is sufficient. The weight of
evidence indicates that the primary mechanism by which fluoride protects against tooth decay is a topical
16 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
effect. Therefore, she noted that the lOM's recommendation of an adequate intake value, at least relating
to tooth decay, should be reassessed.
EPA Response: If it becomes generally accepted that the beneficial anti-caries effect of fluoride is
achieved only through topical application, the status of fluoride as a "nutrient" may be subject to re-
evaluation. However, as long as an Adequate Intake for fluoride is recognized by authoritative groups
such as the IOM, EPA is obliged to treat fluoride as it does other substances which have recommended
AIs by taking those values into consideration when proposing an oral Reference Dose.
17 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
7. Can you suggest an approach to transform the water concentration data from the
Dean (1942) study to units of dose for the population susceptible to severe dental
fluorosis other than that used by the OW?
Reviewers' General Comments: The reviewers were generally accepting of the approach EPA used in
estimating doses from drinking water using the Dean (1942) data on drinking water fluoride
concentrations associated with severe dental fluorosis and estimates from Ershow and Cantor (1989) for
water intake and body weight. Several were of the opinion that EPA should present their dose estimates
without selecting a Reference Dose, leaving that decision to the users of the document.
EPA Response: Although the Office of Water might be able to use the drinking water concentration
data as the basis for their evaluation of the MCLG and MCL for fluoride, an RfD is required by other
programs within EPA. Accordingly, the post peer review version of the dose-response document still
presents an RfD based on the drinking water data and an estimate for the contributions to total exposure
from the diet at the time for the Dean (1942) study.
Specific Comments:
Dr. Cauley suggested that writers use more generalizable data from the National Health and Nutrition
Examination Survey (NHANES) for body weight. She noted that if this approach were used, the writers
could still utilize the water intake estimates from the Ershow and Cantor (1989) paper.
EPA Response: The objective was to use data that related most closely to the time period when the
Dean study was conducted (1942). The Ershow and Cantor (1989) data were from the 1977/1978
Nationwide Food Consumption Survey. NHANES I included surveys for 1971-1975. Although the
records indicate that 24-hour recall dietary information was collected for individuals ages 1 to 74 during
NHANES I (http://www.cdc.gov/nchs/nhanes/nhanesi.htm), EPA found no indication that the data had
been analyzed for age-specific water consumption as have the data from the Nationwide Food
Consumption Survey.
Dr. Den Besten and Dr. Jackson noted several difficulties in developing an approach for transforming
the water concentration data from the Dean study to units of dose for the population susceptible to severe
dental fluorosis.
EPA Response: The Dose-Response report acknowledges the limitations of the approach that was
applied.
Dr. Den Besten stated that serum fluoride levels would be the most useful measurement, but these levels
are not available. She believes that using Dean's data as a starting point to quantify total ingestion is the
best option. She recommended that future studies include random sampling of serum fluoride levels to
strengthen future decision-making relative to fluoride intake.
18 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
EPA Response: A workgroup representing several EPA program offices has submitted a list of chemical
nominations for future biomonitoring efforts at CDC through NHANES. Fluoride was among the EPA
nominations listed as a tier 1 high priority chemical for future monitoring cycles.
Dr. Jackson noted that the U.S. marketplace is constantly changing and ingestion of tap water continues
to decline. U.S. EPA's (2004) report stated that bottled water accounted for only 13% of water
consumption in the United States. More recent trade manufacturing data indicates that bottled water
consumption in the United States exceeds this percentage by a wide margin, and bottled water
consumption may surpass tap water consumption in the near future. Other published data suggest that
among Hispanic individuals, tap water is commonly perceived as "unhealthy" and again bottled water is
almost exclusively consumed. Another point that should be further explored is the possible "halo effect"
of imported foods and beverages into the United States and the fluoride content of these consumables.
EPA Response: Increases in bottled water consumption will be taken into consideration by EPA during
the preparation of the Relative Source document. Use of bottled water is assumed to have been minimal
by the children in Dean's (1942) study populations.
19 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
8. Can you provide input concerning the strengths and weaknesses of the approach
utilized by the OW to identify a lower bound dose for severe dental fluorosis?
Reviewers' General Comments: Dr. Cauley and Dr. Den Besten listed strengths and weaknesses to
the approach utilized by OW to identify a lower bound dose for severe dental fluorosis.
Strengths:
Dr. Cauley said that the main strengths of the approach were that: (1) conservative estimates
were carried out with and without the outlier community; (2) several different models were
considered; and (3) sensitivity analysis showed that even eliminating three highest data points
had little effect on the model goodness of fit or on the BMDL.
Dr. Den Besten added that: (1) the data used were limited to studies found in the Dean report
where the only exposure to fluoride was drinking water; and (2) a careful assessment was done
regarding tap water consumed and mean body weights.
Weaknesses:
Dr. Cauley said that the inherent weakness of the approach relates primarily to the weaknesses
previously cited related to the parent Dean (1942) study. The writers may consider adding a
formal test for trend in data.
EPA Response: The EPA appreciated the acknowledgement of the strengths of the approach applied. In
response to Dr. Cauley's recommendation that EPA add a test for trend in the data, the Cochran-
Armitage Test for Tend was applied to the Dean data and the results indicated that there was a positive
trend for increasing fluorosis with increasing fluoride concentration in drinking water.
Specific Comments:
Dr. Den Besten added that the assumption that 0.05 mg/kg/day is a required amount of fluoride and that
dose estimates must be above this level could be considered a weakness because the purpose of U.S.
EPA's analysis was to determine risk, not to conduct risk benefit analyses. In addition, she noted that the
assumption that the small number of children who displayed severe dental fluorosis were those who had
excess exposure to fluoride could also be considered a weakness. This assumes that these children drank
significantly more water since water was the only source of fluoride. It is more likely that genetic or
other causes are responsible for this small outlier group.
EPA Response: EPA's recognition of 0.05 mg/kg/day as a required amount of fluoride is based on the
lOM's conclusion that this dose represents an Adequate Intake (AI) which reduces the occurrence of
caries without unwanted side effects. The original choice of the 90th percentile for water consumption in
the derivation of the RfD was based on EPA policy that drinking water regulations be derived for the
protection of consumers at that level of exposure. The possibility that some children may be inherently
more susceptible to fluoride, at least above some minimal exposure level, is very plausible; however, at
the moment, there is no way to quantify the increased risk or the size of such a subpopulation. Using the
20 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
90th percentile for water consumption could, in a way, be considered a surrogate adjustment for increased
susceptibility in a segment of the entire population. However, as recommended by the peer reviewers,
the dose calculations for the mean, 75th, 90th, and 95th percentile i
that they can be considered from a risk management perspective.
the dose calculations for the mean, 75th, 90th, and 95th percentile are presented in the final document so
Dr. Whitford expressed additional concerns about the approach utilized by the OW, particularly in
reference to the determination of the LOAEL. He proposed that the variability among published studies
regarding the relationship between estimated fluoride intakes and the risk of severe dental fluorosis may
be due to the relative altitude of the cities that were studied. He provided a detailed discussion of this
issue in his final comment for Question 8. He suggested that based on the variability among published
studies regarding the relationship between estimated fluoride intakes and the risk of severe dental
fluorosis, consideration should be given to establishing a lower bound dose other than 1 percent.
Dr. Whitford had the following to say about the Clovis, New Mexico, data in Table 3-1 (page 24). "The
water fluoride concentration was listed as 2.2 ppm, the lowest concentration at which severe dental
fluorosis was recorded. The prevalence was 0.7 percent, i.e., only one subject out of the 138 examined
exhibited what was classified as having severe fluorosis. This concentration, 2.2 ppm, was selected as the
LOAEL. Unfortunately, nothing is known about the individual with severe fluorosis including whether
he/she was a permanent resident of Clovis or had lived in one or more other communities before moving
to Clovis. The prevalence of moderate dental fluorosis (Dean score 3) in Clovis was 11.0 percent. These
prevalence values for moderate and severe fluorosis are markedly higher than those for Elmhurst and
Galesburg (about 1.1 percent for moderate and an absence of severe dental fluorosis) where the water
fluoride concentrations were listed as 1.8 and 1.9 ppm, respectively, just slightly lower than the
concentration in Clovis. In view of the small differences in the water fluoride concentrations between
Clovis and the other two communities, the large differences in fluorosis prevalence values suggest that
another factor may have influenced the appearance of the teeth in Clovis. Unlike Elmhurst and
Galesburg, Clovis is located at a relatively high altitude (4300 feet). As summarized elsewhere in the
EPA document under review (pages 39-41), there is evidence from laboratory animal studies and
epidemiological studies that residence at high altitude affects amelogenesis in a way that resembles
fluorosis and that its effects may be additive to the effects of fluoride exposure. This too adds uncertainty
regarding the selection of 2.2 ppm fluoride (Clovis) as the LOAEL for severe dental fluorosis and the
appropriateness of the RfD."
He stated that a similar (but weaker) argument could be made for Colorado Springs where the average
water fluoride concentration is listed as 2.6 ppm (but with a wide range, see item 3 above). "This city is
located at an altitude of 6,035 feet. In addition to these comments and based on the variability among
published studies regarding the relationship between estimated fluoride intakes and the risk of severe
dental fluorosis, I think consideration should be given to establishing a LED other than a LBD-1%."
EPA Response: Altitude was evaluated as a contributing factor in the EPA report (Section 3.1.4.2).
Compared to non-U.S. high altitude populations (all African populations) which showed an increased
prevalence of "fluorosis", the U.S. high altitude populations included in the Dean study did not appear to
show the same level of increased "fluorosis" at the higher altitudes represented within the US datasets. It
is unknown to what extent inherent racial susceptibility to fluorosis or hypobaric hypoxia contributed to
the higher levels of severe "fluorosis" seen in the non-U.S. populations.
21 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
As noted by Dr. Whitford, two of Dean's study populations which were high altitude communities
(Clovis NM and Colorado Springs, CO) were key populations near the threshold for severe fluorosis.
Therefore, to test the weight that these two communities had in determining the threshold, the BMD
analysis was rerun allowing for the possibility that some of the cases of severe fluorosis seen in these two
communities were due to the high altitude and perhaps not fluoride-related. The prevalence of severe
fluorosis was down "adjusted" from 6/404 to 3/404 in Colorado Springs and from 1/138 to 0/138 in
Clovis. The resulting log probit BMD for 1% severe fluorosis was 2.43 mg/L and the BMDL 2.24 mg/L.
In comparison, the BMD for 1% severe fluorosis with the complete Dean data set was 2.31 mg/L and the
BMDL was 2.13 mg/L. The "altitude adjusted" values differ by only 5% from the unadjusted values.
Given the uncertainty as to whether altitude was actually a contributory factor, and given the predicted
small impact on the BMDL if it were, EPA has chosen to follow its original and slightly more
conservative approach to use the unadjusted Dean data set.
The post-peer review dose-response document followed the suggestion of Dr. Whitford and others that a
response other than 1% be used for the BMD analysis if it could be justified statistically. The post peer
review document uses a 0.5% response level.
22 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
9. Are you aware of dose estimates other than those from IOM (1997) and the World
Health Organization (WHO; 2002) that are appropriate critical doses for skeletal
effects and/or can you suggest a different approach that the OW might use to
estimate the fluoride dose associated with skeletal fractures using available data?
Reviewers' Comments:
Dr. Cauley referred to her response to Question 2 when addressing this charge question. That response
included references that would address this question. She expressed some concern about the
inconsistencies across the age groups suggesting that the significance of the differences in results across
ages could have occurred by chance. In addition, she noted that referring to the "radius" (one of the
bones of the forearm) as one unit is problematic because the distal radius (the end of the radius closest to
the wrist), is primarily trabecular bone whereas the mid radius is predominantly cortical bone. She added
that considering the radius as one unit is very problematic if the effects of fluoride differ by bone type.
Lastly, she indicated that studies that used sodium fluoride as the fluoride source may have had poor
compliance with the prescribed dosing part of the protocol because it causes gastric irritation and is
poorly tolerated. Participants in such studies may have had poor compliance with the prescribed
treatment regime.
EPA Response: EPA has incorporated the relevant references in their revisions to the peer review draft.
The distinction between trabecular bone of the distal radius and cortical bone of the mid-radius is duly
noted, and where appropriate, the text of the EPA report has been modified to clarify which portion of
the radius was being examined.
Both Dr. Den Besten and Dr. Jackson noted that they were not aware of additional dose estimates that
would be appropriate for determining critical dosage levels and their possible skeletal effects. Dr.
Jackson added that he could not think of an alternative to the approach taken to estimate fluoride intake
associated with skeletal fractures.
EPA Response: No response needed.
Dr. Whitford noted that as in the case of dental fluorosis, the critical doses for skeletal effects will be
difficult to establish with a reasonable degree of certainty. The lOM's estimate (page 307) of fluoride
exposures that may result in clinical signs of the "milder forms" of skeletal fluorosis (preclinical and
perhaps stage I) is 10 mg/day for 10 or more years. There are published exceptions suggesting that higher
exposure levels and durations are required as noted on the same page of the IOM report. Further, recent
case reports of "tea fluorosis" in the U.S. suggest that, at least for some individuals, a much higher
chronic intake is tolerated without progression to stage II skeletal fluorosis (Whyte et al, Am J Med 118:
78-82, 2005; Whyte et al, J Bone Min Res, in press). In the former report the intake was estimated at 37-
74 mg F/day from tea throughout the patient's adult life. The intake in the latter report was estimated at
more than 40 mg F/day throughout the patient's adult life. Both patients showed marked osteosclerosis
but without ligamentous calcifications which was consistent with stage I skeletal fluorosis and neither
23 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
patient experienced fractures. Hallanger-Johnson et al (Mayo Clin Proc 82: 719-724, 2007) reported four
cases with axial osteosclerosis with elevated serum fluoride levels due to chronic consumption of large
amounts of tea.
In addition to the several variables that can affect the quality and quantity of the skeleton cited in the
present document, it is of interest that much of the data relating bone fluoride concentrations to the stages
of skeletal fluorosis comes from studies of workers in aluminum processing factories. High, chronic
exposures to aluminum lead to skeletal changes that share some features in common with skeletal
fluorosis which makes it difficult to attribute the skeletal changes only to fluoride. This subject is worthy
of further exploration.
EPA Response: The lOM's estimate of fluoride exposure (at least 10 mg/day for 10 or more years) that
may result in clinical signs of the "milder forms" of skeletal fluorosis, was based on a review of the
available epidemiological studies. As with other chemicals, response to fluoride exposure is likely to
vary from individual to individual, and responses in one individual may not be typical for a population as
a whole. As noted by Dr. Whitford, IOM does point out that studies in the U.S. have not revealed many
cases of skeletal fluorosis even at fluoride drinking water levels up to 9 mg/L.
24 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
10. Are you aware of any data that can be used to demonstrate that protection of the
secondary teeth from severe dental fluorosis will also protect the primary teeth?
Reviewers' Comments:
Dr. Cauley, Dr. Jackson, and Dr. Whitford said that they are not aware of any additional data that
demonstrate that protection of the secondary teeth from severe dental fluorosis also will protect primary
teeth. Dr. Whitford added that there are a few reports suggesting that dental fluorosis in the primary
teeth may correlate with the condition in secondary teeth.
Dr. Den Besten stated that fluorosis in both primary and permanent teeth is caused by ingested fluoride,
and there are no data to suggest that primary teeth are more susceptible to fluorosis than permanent teeth.
Therefore, a measure that would protect permanent teeth would require limiting ingestion of fluoride.
These same measures would protect primary teeth.
EPA Response: Based on the reviewers' comments, it does not appear that there is any evidence that
primary teeth are more susceptible to fluoride than permanent teeth.
25 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects'
11. Are there factors other than those discussed in Section 3.1.4 that increase
sensitivity to fluorosis of the teeth or bones?
Reviewers' General Comments: All four reviewers suggested additional factors that may have some
effect on an individual's sensitivity to fluorosis of the teeth and/or bone including race/ethnicity, pubertal
stage, genetics, and amoxicillin use in children during the period of tooth formation.
EPA Response: No response needed.
Specific Comments:
Dr. Whitford noted that some researchers have suggested that African-Americans are more susceptible
to dental fluorosis but less susceptible to skeletal fluorosis.
EPA Response: EPA did not identify any data identifying a unique susceptibility of African-Americans
for either severe dental fluorosis or skeletal fluorosis.
Dr. Den Besten said that there seem to be individuals who are uniquely sensitive to the effects of
fluoride on enamel formation. These individuals may have other more "hidden" enamel defects that are
exacerbated by the effects of fluoride. Little is currently known as to why some individuals seem to be
more fluoride sensitive.
EPA Response: EPA has added information to the document on genetics as it related to dental fluorosis.
Dr. Jackson cited a study by Hong et al (2004) that appears to indicate that the use of amoxicillin could
play a contributing role in the development of primary tooth fluorosis, especially for children exposed to
lower levels of fluoride. The full citation was provided.
Hong L, Levy S, Warren J, Bergus G, Dawson D, Wefel J, Broffitt B. (2004) Primary tooth
fluorosis and amoxicillin use during infancy. J Pub I Health Dent. 64: 38-44.
EPA Response: EPA has added a discussion on the Hong et al. (2004) publication to the dose-response
document.
Dr. Jackson also indicated that there have been studies (i.e., Vieira et al., 2005; Yan et al., 2007) in mice
that suggest there may be a varying genetic response to identical levels of fluoride ingestion, and these
responses have been identified in both tooth and bone formation. He noted that Vieira et al. (2005) found
that genetic influences have a direct bearing on the biomechanical properties of the teeth. Furthermore,
Yan et al. (2007) found strain-specific effects, like increased osteoclastogenesis, when mice were
exposed to physiological level of fluoride. While he was unable to find comparable human trials, he
recommended that this area be further explored as the technological means become available. The full
citations were provided:
26 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
Vieira A, Hanocock R, Eggertsson H, Everett E, Grynpas M. (2005). Tooth quality in
dental fluorosis: genetic and environmental factors. Calcified Tissue International.
76:17-25.
Yan D, Gurumurthy A, Wright M, Pfeiler T, Loboa E, Everett E. (2007). Genetic
background influences fluoride's effects on osteoclastogenesis. Bone. 41:1036-1044.
EPA Response: EPA added data from the Viera et al. and Yan et al. studies to the dose-response
document.
Dr. Whitford cited published data that indicate that the susceptibility to dental fluorosis (Everett et al., J
Dent Res 81: 794-698, 2002) and the mechanical properties in the bone (Mousny et al,. Bone 39: 1283-
1289, 2006) are different among strains of mice. He noted that the differences presumably are due to
genetic differences among the strains.
EPA Response: EPA added data from Everett et al, as well as that from a later publication by Mousny
on the role of genetics on bone formation and mineralization to the dose-response document.
27 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
12. Do you support the OW's conclusion that an RfD of 0.07 mg/kg/day will be
protective for severe dental fluorosis in children and skeletal effects in adults
while still providing for the beneficial effects of fluoride?
[NOTE: Since completion of the Peer Review, OW has evaluated the contribution
of food to total fluoride intake, and has adjusted the oral RfD to 0.08 mg/kg/day
to account for the additional fluoride exposure from dietary intake. The dietary
contribution to total fluoride exposure at the time of the Dean (1942) study was
peer reviewed as part of the exposure and relative source contribution document
and included as an appendix in the noncancer dose-response document.
Reviewers' General Comments: In general, Dr. Cauley, Dr. Jackson, and Dr. Whitford support the
use of the 0.07 mg/kg/day RfD. However, all three of these reviewers stated that there is a considerable
degree of uncertainty regarding the estimate.
EPA Response: No response needed.
Specific Comments:
Dr. Jackson indicated that he does not think that a claim can be made that no single individual will be
completely immune from the development of severe fluorosis even at this recommended RfD. As a result
of an individual's potential genetic predisposition, he does not think that the possibility of developing
severe fluorosis can be totally ruled out. Additionally, he noted that his concern is not as much about the
amount of fluoride that is ingested through the public water supply, but the other well-known sources of
fluoride (see his final comment for Question 12 in Appendix B for other sources) that have an additive
effect to that derived from consuming fluoridated drinking water.
EPA Response: The Office of Water has developed a companion report to the dose-response document
which looks at total fluoride exposures under current conditions. In addition, the results from the Iowa
Fluoride study (Hong et al., 2006) that provide support for the conclusions reached by EPA have been
added to the report.
Dr. Whitford added that, as estimated by the IOM (1997), the RfD maybe closer to 0.10 mg/kg/day for
moderate (not severe) dental fluorosis and substantially higher than that for clinically significant skeletal
effects in the United States.
EPA Response: See response to Dr. Den Besten's comment below.
Dr. Den Besten did not agree with Dr. Whitford's recommendation because it is based on limiting severe
fluorosis to 1% of the population and the lOM's recommended adequate intake level. She suggested that
the level be lowered to eliminate severe fluorosis. She noted that one percent of the population represents
a relatively large number of individuals, and these individuals are most likely uniquely sensitive to
fluoride. She recommended that the data be analyzed without taking the lOM's recommended adequate
28 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
levels into account. Then, a secondary analysis could be conducted to include the lOM's
recommendations.
EPA Response: The Dean data were evaluated statistically to determine whether they could be used to
estimate a fluoride concentration in drinking water associated with a prevalence of severe dental fluorosis
less than 1%. It was determined that the data could support an estimate for 0.5%, but not 0.1%;
therefore, the 95% confidence limit on the lower bound for 0.5% severe dental fluorosis was used by
EPA, instead of 1%. In the development of any RfD, a level of uncertainty ("perhaps spanning an order
of magnitude") is acknowledged by EPA due to intrinsic variables which oftentimes can not be
quantified. As indicated by the differing opinions expressed by the reviewers, this uncertainty can argue
in favor of a higher or a lower RfD. Although genetic factors may make certain individuals more
susceptible to fluorosis; there is currently no way to quantify such an increase in susceptibility. As noted
by Dr. lackson, other potential sources of fluoride will contribute to dental fluorosis. The assumption
was made by EPA that drinking water was the only source of fluoride in the populations studied by Dean
(1942); however, OW later evaluated the potential contribution of dietary items to total fluoride intake
during the time of the Dean (1942) study. Using data from the studies of McClure (1943, 1949), an
adjustment was made in the derivation of the RfD to account for this additional intake of fluoride through
the diet, which was estimated to be 0.01 mg/kg/day. The resulting RfD is of 0.08 mg/kg/day.
The uncertainty associated with the RfD estimate is to some degree conservatively adjusted for by taking
the lower bound of the 95% confidence interval for the occurrence of 0.5% severe dental fluorosis (equal
to 1.87 mg/L), since it is equally likely that the value could be as high as 2.06 mg/L (upper bound on the
95% CI).
29 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
IV. RESPONSE TO DR. WHITFORD'S COMMENTS CONCERNING FLUORIDE
ANALYTICAL METHODOLOGY
Prepared by Bruce Tomkins, Annetta Watson and Dennis Opresko
ORNL
16 May 2008
Dr. Whitford(l):
(1) A major weakness of the Dean (1942) report is the chemical method used for the determination of
fluoride concentrations in water (Elvove E, 1933). The zirconiumalizarin method is rarely, or probably
never, used today because of its relative insensitivity, several interfering substances and lack of
specificity for fluoride. In their 1952 report that described improvements to the method, Megregian and
Maier (1952) noted that Elvove's original method (1933) had several shortcomings including "non-
conformity to the color laws, limited effective fluoride range, and little color change per increment of
fluoride." It also appears that Elvove (1933) used the visual method to determine color changes in the
zirconium-alizarin reagent (since he referred to "Nessler tubes") which requires subjective judgments and
is less accurate than spectrophotometric methods.
EPA Response (1): Elvove employed a reagent containing zirconium oxychloride and alizarin sodium
monosulfate. The most recent standard methods employ two reagents related to that discussed in Elvove.
One employs an acidic reagent containing zirconyl chloride and the complexing agent SPADNS [sodium
2-(parasulfophenylazo)-l, 8-dihydroxy-3, 6-naphthalene disulfonate] (APHA, AWWA and WEF 2005a).
The other is used only in an automated system to form a blue complex, and employs both alizarin and
lanthanum nitrate (APHA, AWWA and WEF 2005b). Standard Methods clearly indicates the electrode
and colorimetric methods are most satisfactory. Hence, while Elvove's reagent may not be used any
more, related procedures not only are being used, but have found their way into Standard Methods.
The Reviewer implied a lack of specificity. It is very true that more current methods for fluoride ion,
employing an ion-selective electrode (APHA, AWWA and WEF 2005c; Omega Engineering, Inc. 2008)
and ion chromatography (Hautman et al 1997) will exhibit a range, sensitivity, and specificity that are
superior to that described in Elvove. The Reviewer could also argue, correctly, that the range could be
improved by replacing the Nessler tubes, which allow color comparison with known standards, with a
standard spectrophotometer. However, when Elvove performed his measurements in 1933, the first
commercial spectrophotometer, the Beckman DU, was eight years away (Simoni et al 2003). The
fluoride-selective electrode was not available until the mid-1960's (Buck and Lindner 2001), while the
first commercial ion chromatograph was first marketed in 1981 (Evans 2004).
The only method of comparison that Elvove had at his disposal was colorimetric with Nessler tubes,
which permitted him to compare the color of his reacted unknown with that of a known standard. It
could be argued that the colorimetric difference between samples containing 0.4 and 0.7 ppm fluoride
might be difficult to distinguish; however, there should be no difficulty distinguishing 0.6 and 1.2 ppm,
for example. Analytical chemists in those days were usually trained more carefully to distinguish subtle
changes in color than chemists of today.
The Reviewer criticized the "non-conformity to the color laws", but did not state what the non-
30 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
conformity was. It is true that Elvove did not specifically state that his color reaction obeyed Beer's
Law, in which the absorbance of a solution is directly proportional to the concentration of analyte (or
derivatized analyte) present. However, such a specification is typically omitted in a professional paper
unless the reaction produces a color change that does not obey Beer's Law. In related work, Megregian
and Maier (1952) commented "This reagent [modified zirconium-alizarin reagent] is sensitive to small
increments of fluoride over a range of 0.0 to 3.0 ppm and it approximates Beer's law over this range."
The Reviewer is correct when he commented on the subjective analysis using Nessler tubes, but
comparison with Nessler tubes was clearly the standard approach when Elvove published his work.
Dr. Whitford (2\.
Megregian and Maier (1952) also reported the effects of interfering substances on the analytical results.
Sulfate at 400 ppm in the water increased the fluoride result by 0.1 ppm as did 1.1 ppm
hexametaphosphate. Chloride at 1800 ppm, bicarbonate at 400 ppm and iron at 5 ppm decreased the
fluoride result by 0.1 ppm. When the water fluoride concentration was 1.0 ppm, 1.0 ppm aluminum
reduced the result by 0.39 ppm and 3.0 ppm aluminum reduced the result by 0.63 ppm. When the fluoride
concentration was 2.0 ppm, aluminum at 1.0, 2.0 and 3.0 ppm reduced the results by 0.47, 0.86 and 1.13
ppm, respectively.
The report by Elvove (1933) that described the chemical method used to produce the water fluoride
concentrations shown in Table I of Dean's 1942 report made no mention of interfering substances. It can
be reasonably assumed that most, if not all, of the substances listed in the preceding paragraph were
present in all the water samples analyzed but at unreported or unknown concentrations. However, in his
original publication Elvove (1933) shows the concentrations of several ions in water samples obtained
from 20 different sources. With the exception of Amarillo, these sources were not those shown in the
Dean (1942) report. Among the interfering ions listed in the preceding paragraph, sulfate concentrations
were more than 400 ppm in two of the 20 water samples; bicarbonate concentrations were more than 400
ppm in five, and aluminum concentrations were more than 1.0 ppm in five. At these concentrations, each
of these ions would have affected the apparent water fluoride results.
EPA Response (2): The Reviewer's concerns regarding the presence of several interfering substances and
lack of specificity for fluoride are legitimate, and have been investigated extensively in the literature.
Elvove's reagent appears to respond primarily to fluoride ion, although the presence of certain inorganic
species will suppress fluoride response to some degree. Grutsch et al (1953) noted that modest quantities
(< 5 ppm) of either aluminum or iron will suppress, but not eliminate, a response to fluoride ion, while
the presence of sub-ppm concentrations of manganese will enhance such a response.
The critical region for community water supply characterization in the Dean (1942) study is that for
supplies with fluoride concentrations in the range of 0.9 to 4.5 ppm. USEPA has prepared the following
summary table (Table 1) compiling contemporaneous water quality values for chloride, sulfate,
aluminum, iron and bicarbonate for those communities from Dean (1942) bounded by this critical
fluoride range. In no case was chloride >1800 ppm (range of 0.5 to 689.0 ppm), sulfate or bicarbonate
>400 ppm (sulfate range of 3.4 to 308.6 ppm; bicarbonate range of 23.2 to 389.2 ppm), or aluminum >1.0
ppm (range of 0 to 0.3 ppm), or iron >5 ppm (range of 0.01 to 0.1 ppm). Instead, the values for each of
these water quality parameters falls below the concentration identified by the Reviewer as cause for
concern regarding analytical interference. Chloride ion at these concentrations appears to have no affect
upon the determination of fluoride ion.
November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
Table 1 . Water Quality Analysis forSelect Towns Used in the Dean (1942) Study
Town
Kewanee, ILa
Galesburg ILa
Clovis. NMb
Colorado Springs,
coa
Plainview, TXC
Amarillo, TXC
Conway, SC°
Lubbock, TX
Fluoride
(ppm)
0.9
1.9
2.2
2.6
2.9
3.9d
4.0d
4.4d
Chloride
(ppm)
689.0
190.5
16.5
0.5
31.8
13.5
49.5
111.8
Sulfate
(ppm)
308.6
351.7
24.2
4.9
32.5
52.0
3.4
255.0
Aluminum
(ppm)
0
0
0.2
0
0.3
0.3
0.06
0.2
Iron
(ppm)
0.01
0.1
0.02
0.04
0.1
0.08
0.06
0.1
Bicarbonate
(ppm)
300.1
295.2
234.9
23.2
331.8
389.2
256.0
342.8
aDean, 1942.
bDean, 1937.
°Dean and Elvove, 1936.
dFluoride value as reported in Dean, 1942.
Evaluation of the literature includes experimental tests by Grutsch et al (1953) with varying
concentrations of Fe (III) in solution with 0.5 and 1.00 ppm fluoride. There was no suppression of
fluoride at these concentrations when the Fe (III) concentration was 0.1 ppm. When the concentration of
Fe (III) reached 0.5 ppm, the "found" values for 0.50 and 1.00 ppm fluoride are 0.48 and 0.96 ppm,
respectively, or an approximate 4% suppression for the measured concentration of fluoride. Since the
community water supplies of interest (Table 1) all exhibit Fe concentrations < 0.1 ppm, no suppression of
actual fluoride concentrations is expected from the concentrations of Fe present.
Grutsch et al (1953) also evaluated Al (III) in water with the same two target concentrations of fluoride
and observed no suppression at either fluoride concentration when the Al (III) concentration was 0.1
ppm. When 0.3 ppm Al (III) was present, the "found" values for 0.50 and 1.00 ppm fluoride were 0.48
and 0.98 ppm, respectively, or a suppression of 4% at the low-concentration (0.5 ppm F) sample and 2%
at the 1 ppm F sample. Four out of the 8 community water supplies summarized in Table 1 exhibit Al
concentrations <0.1 ppm, while two exhibit Al concentrations of 0.2 ppm and two exhibit Al
concentrations of 0.3 ppm. No suppression of actual fluoride concentrations is expected in water
samples from communities with Al concentrations <0.1 ppm. The data of Grutsch et al (1953) indicate
that the maximal suppression of fluoride expected at 0.3 ppm Al would be 4%.
Dr. Whitford (3):
Another indication of the problem with the accuracy of the Elvove method is found in the footnote to
Table 3-1 on page 24 of the EPA Dose-Response Analysis where it is said (quoting Elvove who was the
principal chemist) that "as little as 0.01 mg F/50cc, or 0.2 ppm F, could be differentiated from the control
by application of this technique." This appears to mean that Elvove's method could differentiate between
water without fluoride and water containing 0.2 ppm fluoride. The magnitude of the error at higher
concentrations is not known to me. The scatter in the analytical results seen in the 1933/34 monthly
results for water in Colorado Springs is of particular interest (see Table 4 in Dean and Elvove, 1935).
The average of the 12 results was 2.5 ppm but the range was 1.8 to 3.0 ppm despite the fact that the water
came from a single source. While some seasonal variation in water concentrations can expected, this
32
November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
wide range (1.2 ppm) appears excessive. Further, the 12 monthly results from Monmouth ranged from
1.6 to 1.9 ppm, those from Galesburg ranged from 1.8 to 2.0 ppm, and those from Pueblo ranged from 0.3
to 0.7 ppm.
EPA Response (3): It has previously been noted by Megregian and Maier (1952)that the use of this
reagent [modified zirconium-alizarin reagent] is sensitive to small increments of fluoride over a range of
0.0 to 3.0 ppm and approximates Beer's law over this range, which includes the range of results for
Colorado Springs (fluoride content of 1.8 to 3.0 ppm from Dean and Elvove 1935). There are a number
of alternate sources of variability for natural fluoride concentrations in community water supplies of
Colorado Springs, such as snowfall, rainfall, reservoir maintenance and pumping schedules, etc., and it
should thus not be assumed that the range of 1.8 to 3.0 ppm F noted by the Reviewer in Dean and Elvove
(1935) was a consequence of some analytical difficulty. It is also noted that reported fluoride
concentrations for Colorado Springs during the 14 months observed between Jan 1940 and Feb 1941 in
Dean (1942) exhibited the narrow range of 2.4-2.8 ppm F (mean of 2.55 ppm F). It appears that the
source of variability displayed in the Dean and Elvove (1935) data set for Colorado Springs had been
resolved by the time of the Dean (1942) data collection and analysis.
The critical range for F determinations in the current assessment is <4.5 ppm F (see Table 1), a range
over which the analytical method and reagents of Elvove appear to be appropriate, adequate and
customary for the time.
REFERENCES CITED
American Public Health Association (APHA), American Water Works Association (AWWA) and Water
Environment Federation (WEF), 2005a. Method 4500-F" D. SPADNS Method. Standard Methods for
the Examination of Water & Wastewater, 21st Edition, pp 4-85 to 4-86.
American Public Health Association (APHA), American Water Works Association (AWWA) and Water
Environment Federation (WEF), 2005b. Method 4500 F- E. "Complexone Method." Standard Methods
for the Examination of Water & Wastewater, 21st Edition. Pp. 4-87.
American Public Health Association (APHA), American Water Works Association (AWWA) and Water
Environment Federation (WEF), 2005c. Method 4500-F" C. "Ion-Selective Electrode Method."
Standard Methods for the Examination of Water & Wastewater, 21st Edition. Pp 4-84 to 4-85.
Buck RP and E Lindner, 2001. Tracing the History of Selective Ion Sensors. Anal. Chem. (February 1,
2001). A 97 A, p. 88. Available on-line at http://www2.truman.edu/~blamp/chem222/manual/pdf/ise.pdf
Dean HT, 1937. Further studies on the minimal threshold of chronic endemic dental fluorosis Public
Health Reports 52 (37): 1249-1264.
Dean HT, 1942. The investigation of physiological effects by the epidemiology method. In: Fluoride
and Dental Health. Publication of AAAS, no. 19, pp. 23-31.
Dean HT and E Elvove, 1935. Studies on the minimal threshold of the dental sign of chronic endemic
fluorosis (mottled enamel). Public Health Rep. 50: 1719-1729 (Dec 6, 1935).
33 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects "
Dean HT and E Elvove, 1936. Some epidemiological aspects of chronic endemic dental fluorosis. Amer.
J. of Public Health 26: 567-575.
Elvove E, 1933. Estimation of fluorides in waters. Public Health Rep. 48 (40): 1219-1222.
Evans B, 2004. The History of Ion Chromatography: The Engineering Perspective. J. Chem. Ed. 81:
1285-1292. Available on-line at http://jchemed.chem.wisc.edu/journal/Issues/2004/Sep/absl285.html.
Grutsch JF, WH Nebergall, JC Muhler, RB Fischer, and HG Day, 1953. A Procedure for the Routine
Determination of Fluorine in Potable Waters Containing Iron, Manganese, Aluminum, and Chlorine. J.
D. Research 32 (4): 463-468.
Hautman DP, DJ Munch and JD Pfaff, 1997. Method 300.1. Determination of Inorganic Anions in
Drinking Water by Ion Chromatography, Revision 1.0. USEPA Office of Water, National Exposure
Research Laboratory, Office of R&D, USEPA, Cincinnati, OH. 45268.
http://www. epa.gov/waterscience/methods/method/file s/3 00_ 1 .pdf.
Megregian S and FJ Maier, 1952. Modified Zirconium-Alizarin Reagent for Determination of Fluoride
in Water. J. Amer. Water Works Assn. 44: 239-248.
Omega Engineering, Inc., 2008.ISE-8790 & ISE-8795, Fluoride Ion Selective Electrodes.
http://www.omega.com/manuals/manualpdf/M0780.pdf (site accessed April 2008).
Simoni RD, RL Hill, M Vaughan, and H Tabor, 2003. A Classic Instrument: The Beckman DU
Spectrophotometer and Its Inventor, Arnold O. Beckman. J. Biol. Chem. (December 5, 2003) 278 (49), 1.
Available on-line at http://www.jbc.org/cgi/content/full/278/49/el.
34 November 2010
-------
EPA Response to Peer Review of "Fluoride: Dose-Response Analysis for Non-cancer Effects'
V. EXTERNAL PEER REVIEW REPORT
35 November 2010
-------
External Peer Review of U.S. EPA's Draft Dose-Response Assessment for
Severe Dental Fluorosis and the Risk for Increased Bone Fractures Related to
Fluoride
Peer Review Report
Prepared for:
U.S. Environmental Protection Agency
Office of Water
Office of Science and Technology
Prepared by:
ICF International
9300 Lee Highway
Fairfax, VA 22031
Under Subcontract to:
ToxServices LLC
1326 18th Street, N.W. #22
Washington, DC 20036
EPA Contract No. PR-CI-06-10561
Work Assignment Number B-02
April 2009
-------
Table of Contents
Introduction 1
Peer Review Meeting 2
Comment Summary 3
Charge Question 1 3
Charge Question 2 4
Charge Question 3 6
Charge Question 4 8
Charge Question 5 9
Charge Question 6 9
Charge Question 7 10
Charge Question 8 10
Charge Question 9 11
Charge Question 10 12
Charge Question 11 12
Charge Question 12 13
Appendix A: Pre-Meeting Responses to the Charge Questions from Reviewers A-l
Appendix B: Post-Meeting Responses to the Charge Questions from Reviewers B-l
Tables
Exhibit 1: Peer Reviewer Names and Affiliations 2
Exhibit 2: Peer Reviewer Expertise Relative to the Selection Criteria 2
April 2009
-------
Introduction
On April 2, 1986, the United States Environmental Protection Agency (U.S. EPA) set a revised
Maximum Contaminant Level (MCL) for fluoride at 4 mg/L to protect against crippling skeletal
fluorosis, an adverse health effect. In addition, U.S. EPA set a non-enforceable Secondary
Maximum Contaminant Level (SMCL) at 2 mg/L to protect against objectionable dental
fluorosis. In 2003, U.S. EPA requested that the National Research Council (NRC) review the
scientific and technical basis for U.S. EPA's drinking water MCL and SMCL for fluoride as a
part of its review of all regulations promulgated prior to the 1996 reauthorization of the Safe
Drinking Water Act. NRC was asked to:
• Review new toxicology, epidemiology, and clinical data;
• Examine exposure data on orally ingested fluoride from water and other sources;
• Advise U.S. EPA on the adequacy of its MCL and SMCL to protect children and others
from adverse effects; and
• Identify data gaps and make research recommendations.
In 2006, the National Research Council (NRC) panel recommended that U.S. EPA conduct a
new quantitative risk assessment for severe dental fluorosis and the risk for increased bone
fractures related to fluoride based on the most recent dose-response data. The findings of the
2006 NRC panel diverged from those of the previous 1993 panel by categorizing severe dental
fluorosis, which results in thinning and pitting of the tooth enamel, as an adverse health effect
rather than a cosmetic effect. The 2006 panel also concluded that the present MCL of 4 mg/L
was not protective for severe dental fluorosis and may not be protective for the skeletal fracture
endpoint. The NRC committee did not evaluate health effects from lack of exposure to fluoride
or fluoride's efficacy in preventing dental cavities.
In response to the NRC 2006 report, the Office of Water (OW) has completed a dose-response
assessment based on data from studies of severe dental fluorosis and evaluated the available
dose-response data on the skeletal effects of fluoride as recommended by NRC. The report
provides a quantitative estimate of a lower bound confidence interval on the concentration in
water associated with a 1 percent risk for severe dental fluorosis derived from data collected
before fluoridation of drinking water and the introduction of fluoride into dental products. The
concentration-response data were then converted to dose using water intake and body weight
data for children during the period for formation of the secondary teeth. The relative source
contribution for drinking water compared to total fluoride exposure will be the subject of a
separate effort by the OW.
Prior to releasing the document for external peer review, it was evaluated within U.S. EPA by
representatives from the OW, Office of Children's Health, Office of Pesticide Programs, and
Office of Research and Development.
April 2009 Page 1
-------
Peer Review Meeting
On March 11, 2008, U.S. EPA convened a panel of external and independent technical experts in
Washington, D.C. to address charge questions regarding the quality, clarity, and transparency of
the draft dose-response report. The names and affiliations of the four peer reviewers for this draft
report can be found in Exhibit 1, and their areas of expertise are shown in Exhibit 2. The peer
reviewers were asked to draft initial responses to the charge questions, attend a one-day meeting
in Washington, D.C., and submit final responses to the charge questions and any additional
comments on the dose-response document following the meeting.
Exhibit 1: Peer Reviewer Names and Affiliations
Reviewer Name
Jane A.
Pamela
Richard
Gary M
Cauley, Dr.P.H
DenBesten, D.D.S., M.S.
D. Jackson, D.M.D.
. Whitford, D.M.D., Ph.D.
Reviewer Affiliation
Professor, Department of Epidemiology,
University of Pittsburgh
Professor and Chair, Division of Pediatric
Dentistry, Department of Orofacial Sciences,
University of California at San Francisco
Assistant Professor, Preventive and Community
Dentistry, School of Dentistry, Indiana University,
Oral Health Research Institute
Regents' Professor, Department of Oral Biology
and Maxillofacial Pathology, Medical College of
Georgia
Exhibit 2: Peer Reviewer Expertise Relative to the Selection Criteria
Dentist or researcher with
experience relating to
dental pitting as a result
of fluoride and the
associated potential for
tooth decay
Doctor or researcher with
knowledge of the link
between fluoride and
bone fractures
Epidemiologist with
familiarity of the health
risks from environmental
fluoride exposure
Dr. Cauley
X
X
Dr. DenBesten
X
Dr. Jackson
X
Dr. Whitford
X
X
X
April 2009
Page 2
-------
The technical experts that comprised the peer reviewer panel were selected by ICF International
based upon their independence and expertise in the subject matter. The peer reviewers'
qualifications were assessed based on the following selection criteria specified by EPA: (1)
dentists or researchers with experience relating to dental pitting as a result of fluoride and the
associated potential for tooth decay; (2) doctors or researchers with knowledge of the link
between fluoride and bone fractures; and/or (3) epidemiologists with familiarity of the health
risks from environmental fluoride exposure.
The peer reviewers' draft of initial responses to the charge questions are provided in Appendix
A, and their final responses to the charge questions and any additional comments on the dose-
response document are provided in Appendix B. (Comments are organized alphabetically by
reviewer in both appendices.) Additionally, final comments from the four peer reviewers are
organized and summarized in the main body of this document according to the charge questions.
In most cases, the comments are presented in alphabetical order according to the peer reviewer's
last name. If there is agreement among reviewers, this is stated at the beginning of the summary
for that question. Except for minor clarifications, the comment summary retains most of the
reviewers' original wording. The use of quotation marks is reserved for situations in which the
reviewer quoted a specific document or the reviewer's original wording and/or intent was
unclear.
Comment Summary
The overall response from the reviewers was positive; however, all four reviewers suggested
changes to improve the overall quality of the document. Their responses to each charge question
are summarized in the remainder of this document.
Charge Question 1. Was the document clear and transparent in its presentation of data
and explanation of the analytical approaches used to characterize the concentration-
response and dose-response relationships for severe dental fluorosis? If not, do you have
any suggestions that will enable the authors to improve its clarity?
In general, the reviewers thought that the document was clear and transparent. In addition, the
reviewers commented that the document clearly described the available literature and presented
the information in an understandable format. In particular, Dr. Jackson noted that the individual
tables summarizing the studies were very helpful.
However, the reviewers provided the following suggestions to improve the clarity of the
document:
• Dr. Cauley believes that the document could be strengthened by reorganizing it as a
manuscript. When reorganizing the document as a manuscript, the following changes
should be considered: the introduction of the document should provide background
information and a scientific rationale for this analysis; the specific objectives should be
listed; and the methods for identifying all the literature should be transparent and
consistently applied to both the fluorosis and the fracture sections.
April 2009 Page 3
-------
• Dr. Cauley also believes that it would also be helpful if the document included a
discussion on U.S. EPA OW's and NRC's roles (e.g., How will each agency find this
information helpful? What prompted this report?).
• Dr. DenBesten suggested several specific changes by page number (see Dr. DenBesten's
final comments for Question 1 in Appendix B).
• Dr. Jackson noted that he would have preferred that the studies be arranged
chronologically as opposed to alphabetically by author.
Charge Question 2. Are you aware of any significant publications related to severe dental
fluorosis or the skeletal effects of fluoride that are not included in the noncancer
assessment document?
Three of the reviewers, Dr. Cauley, Dr. Jackson, and Dr. Whitford, thought that there were
additional publications related to severe dental fluorosis and/or skeletal effects of fluoride that
should be reviewed for possible inclusion.
Dr. Cauley noted that while the dental fluorosis literature review in the report covered a wide
range of fluoride levels, the skeletal/ fracture studies were limited to those with fluoride levels
> 4 mg/L. She suggested that the Office of Water consider the following publications that discuss
the occurrence of fractures at lower levels of fluoride:
• Cauley JA, Murphy PA, Riley TJ, Buhari AM. (1995). Effects of fluoridated drinking
water on bone mass and fractures: The study of osteoporotic fractures. J Bone Miner Res.
10(7): 1076-1086.
• Phipps KR, Orwoll ES, Mason JD, Cauley JA. (2000). Community water fluoridation,
bone mineral density, and fractures: prospective study of effects in older women. BMJ
321(7265):860-864.
• Phipps KR, Burt BA. (1990). Water-borne fluoride and cortical bone mass: A comparison
of two communities. J Dent ResT. 69(6): 1256-1260. [Note: this study compares bone
mineral density in Deming, NM (0.7 mg/L) and Lordsburg, NM (3.5 mg/L)]
Dr. Cauley also expressed concern that the bone fractures section is less thorough than the
dental fluorosis section. To address this issue, she suggested the following publications be
reviewed for possible inclusion and noted that the studies by K. Phipps, C. Cooper, S. Jacobsen,
and T. Hillier are particularly noteworthy and of high quality:
• Demos LL, Kazda H, Cicuttini FM, Sinclair MI, Fairley CK. (2000). Water fluoridation,
osteoporosis, fractures—recent developments. Aust Dent J. 46(2): 80-87.
• Fabiani L, Leoni V, Vitali M. (1999). Bone-fracture incidence rate in two Italian regions
with different fluoride concentration levels in drinking water. J Trace Elem MedBiol.
13(4):232-237.
• Kleerekoper M. (1996). Fluoride and the skeleton. CritRev ClinLab Sci. 33(2): 139-161.
• Raheb J. (1995). Water fluoridation, bone density and hip fractures: a review of recent
literature. Community Dent OralEpidemiol. 23(5):309-316.
• S0gaard CH, Mosekilde L, Schwartz W, Leidig G, Minne HW, Ziegler R. (1995). Effects
of fluoride on rat vertebral body biomechanical competence and bone mass. Bone.
16(1):163-169.
April 2009 Page 4
-------
Kroger H, Alhava E, Honkanen R, Tuppurainen M, Saarikoski S. (1994). The effect of
fluoridated drinking water on axial bone mineral density—a population-based study.
Bone Miner. 27(1):33-41.
Ripa LW. (1993). A half-century of community water fluoridation in the United States:
review and commentary. JPublic Health Dent. 53(1): 17-44.
Gordon SL, Corbin SB. (1992). Summary of workshop on drinking water fluoride
influence on hip fracture on bone health. (National Institutes of Health, 10 April, 1991)
Osteoporos Int. 2(3 ): 109-117.
Colquhoun J. (1991). Water fluoride and fractures. NZMedJ. 104(917):343.
McNeill KG, Coote GE, Hitchman AJ. (1991). Uptake of fluorine in cortical and
trabecular bone. J Bone Miner Res. 6(8):859-864.
Arnala I, Alhava EM, Kivivuori R, Kauranen P. (1986). Hip fracture incidence not
affected by fluoridation. Osteofluorosis studied in Finland. Ada Orthop Scand.
57(4):344-348.
Ericsson Y, Luoma H, Ekberg O. (1986). Effects of calcium, fluoride and magnesium
supplementations on tissue mineralization in calcium- and magnesium-deficient rats.
JNutr. 116(6): 1018-1027.
Simonen O, Laitinen O. (1985). Does fluoridation of drinking-water prevent bone
fragility and osteoporosis? Lancet. 2(8452):432-434.
Madans J, Kleinman JC, Cornoni-Huntley J. (1983). The relationship between hip
fracture and water fluoridation: An analysis of national data. Am J Public Health.
73(3):296-298.
Schamschula RG, Barmes DE. (1981). Fluoride and health: dental caries, osteoporosis,
and cardiovascular disease. Annu Rev Nutr. 1:427-435.
Alhava EM, Olkkonen H, Kauranen P, Kari T. (1980). The effect of drinking water
fluoridation on the fluoride content, strength and mineral density of human bone. Ada
Orthop Scand. 51(3):413-420.
Stein ID, Granik G. (1980). Human vertebral bone: relation of strength, porosity, and
mineralization to fluoride content. Calcif Tissue Int. 32(3): 189-194.
Hegsted DM. (1967). Osteoporosis and fluoride deficiency. PostgradMed. 41(l):A49-53.
Bernstein DS, Sadowsky N, Hegsted DM, Guri CD, Stare FJ. (1966). Prevalence of
osteoporosis in high- and low-fluoride areas in North Dakota. JAMA. 198(5):499-504.
Hillier S, Cooper C, Kellingray S, Russell G, Hughes H, Coggon D. (2000). Fluoride in
drinking water and risk of hip fracture in the UK: a case-control study. Lancet.
355(9200):265-269.
Hillier S, Inskip H, Coggon D, Cooper C. (1996). Water fluoridation and osteoporotic
fracture. Community Dent Health. 13 Suppl 2:63-68.
Jacobsen SJ, Goldberg J, Cooper C, Lockwood SA. (1992). The association between
water fluoridation and hip fracture among white women and men aged 65 years and
older. A national ecologic study. Ann Epidemiol. 2(5):617-626.
Cooper C, Wickham C, Lacey RF, Barker DJ. (1990). Water fluoride concentration and
fracture of the proximal femur. J Epidemiol Community Health. 44(1): 17-19.
Phipps KR, Orwoll ES, Mason JD, Cauley JA. (2000). Community water fluoridation,
bone mineral density, and fractures: prospective study of effects in older women. BMJ.
321(7265):860-864.
April 2009 Page 5
-------
• Phipps KR, Orwoll ES, Bevan L. (1998). The association between water-borne fluoride
and bone mineral density in older adults. JDent Res. 77(9): 1739-1748.
• Phipps K. (1995). Fluoride and bone health. J Public Health Dent. 55(l):53-56.
• Phipps KR, Burt BA. (1990). Water-borne fluoride and cortical bone mass: a comparison
of two communities. J Dent Res. 69(6): 1256-1260.
Dr. Jackson noted that the document should consider the findings from a study that examined
357 school-age children in 1994 from three Indiana communities with varying levels of fluoride
in the water (0.2, 1.0, and 4.0 ppm). Severe fluorosis, as determined by the Tooth Surface Index
of Fluorosis (TSIF) (score 5, 6, or 7), was observed in 9% of the children examined in the 4.0-
ppm community. No scores of this magnitude were seen in the other two communities. A
comparison using 1992 prevalence data for the same communities was conducted, and results
showed an increase in the prevalence of fluorosis in each community, primarily for TSIF scores
1, 2, or 3. The following citation for this study was provided:
• Jackson R, Kelly S, Katz B, Brizendine E, Stookey G. (1999). Dental fluorosis in
children residing in communities with different fluoride levels in the water: 33 month
follow-up. Pediatric Dentistry. 21:248-254.
Dr. Whitford suggested that the committee review a copy of the book entitled "Fluoride
Drinking Waters." He stated that this book is a compilation of many of the early papers that
address several aspects of fluoride in water. The book was published in 1962, edited by Frank J.
McClure, and is Public Health Service Publication No. 825. He highlighted the following three
references from the book that contain information pertinent to the reliability of the water fluoride
concentrations shown in the Dean (1942) study, the study selected as the "critical study for
severe dental fluorosis:"
• Elvove E. (1933). Estimation of fluorides in waters. Pub. Hlth. Rep. 48:1219-1222.
• Dean HT, Elvove E. (1935). Studies on the minimal threshold of the dental sign of
chronic endemic fluorosis (Mottled enamel). Pub. Hlth. Rep. 50:1719-1729.
• Megregian S, Maier FJ. (1952). Modified zirconium alizarin reagent for determination of
fluoride in water. J. Am. Water Works Assn. 44:239-246.
Charge Question 3. Are the strengths and weaknesses of Dean (1942) study (selected as
critical) fully characterized?
Strengths:
In general, the reviewers believe that the strengths of the Dean (1942) study were well-
characterized in the draft dose-response report.
Dr. Cauley noted that additional strengths of the study include: (1) the wide range of fluoride
concentrations, although there were fewer subjects in the high fluoride categories; (2) the dose-
response relationship illustrated by the data showing increasing risk of severe skeletal fluorosis
with increasing fluoride; and (3) the consistency of the findings across several different
communities.
April 2009 Page 6
-------
Dr. Jackson doubted that anyone would disagree that the Dean (1942) study has been and will
continue to be a benchmark study in the dental literature, as well as the much broader literature
related to public health and epidemiology. He noted that Dean's study was performed when
confounding fluoride sources were not available and, thus, probably gives a very clear picture of
the prevalence of the relationship of fluoride ingestion and the subsequent development of dental
fluorosis.
Weaknesses:
The reviewers expressed some concern about the characterization of the weaknesses of the study.
Dr. Cauley and Dr. Jackson noted that the Dean (1942) study does not provide information on
race/ethnicity, and given the probable characteristics of the region, the population of subjects
examined was likely not diverse in terms of racial or cultural characteristics.
Dr. Cauley noted that the data were collected the late 1930-40s. Although confounding by use of
other fluoride products would be minimal, there are many other cohort differences between
children exposed to fluoride in the late 1930-40s as compared to now. For example, dental
hygiene, dietary intakes (e.g., less water and more carbonated beverages), and body weight are
very different in today's children compared to those in the 1930s. Additionally, Dr. Cauley
questioned whether puberty and/or hormonal changes may influence fluoride effects, which may
be important because age of menarche has been decreasing. Another weakness of the studies
reviewed is that there appears to be no information on exposure duration (e.g., How long did
these children live in each community? Did the inclusion criteria include a minimum time of
residence?).
Dr. Jackson further noted that it has been postulated that genetic factors may impact the
expression of dental fluorosis at identical levels of ingestion. The fact that data were collected in
what may have been an exclusively white population appears to limit its applicability for use as a
benchmark.
Dr. Whitford believes that the weaknesses of the study were not necessarily fully characterized.
One example given in his discussion of how the weaknesses of the Dean (1942) study were not
fully characterized, was the lack of review of the publications presented in his response to
Question #2 (Elvove, 1933). He also had the following specific comments:
1. A major weakness of the Dean (1942) report is the chemical method used for the
determination of fluoride concentrations in water (Elvove, 1933). The zirconium-alizarin
method is rarely, or probably never, used today because of its relative insensitivity,
several interfering substances, and lack of specificity for fluoride. Please see Dr.
Whitford's final comment for Question 3 in Appendix B for a critical review of the
Elvove (1933) method.
2. The study population in the Dean (1942) may not have been continuously exposed to the
community's communal water supply. Dean (1942, page 25) listed two major requisites
for quantitative evaluation of the dental effects of ingesting water containing fluoride.
One of these requisites was "a population continuously exposed throughout life to the
variable under investigation (the communal water supply)." For a more thorough
April 2009 Page 7
-------
discussion of this issue, please see Dr. Whitford's final comment for Question 3 in
Appendix B. Dr. Whitford recommended that the original papers summarized in Dean
(1942) be examined to determine the extent to which the children met the requisite cited
above and that the information be included in the Dose-Response Analysis document. If
such information is not available, then the document should note this and discuss the
implications in its conclusions.
Dr. Whitford further commented that the appropriateness of the LOAEL (2.2 ppm) and the
calculated reference dose (RfD) (0.07 mg F/kg bw/day) reported in the OW's draft dose-response
report are based largely on the accuracy of the water fluoride concentrations shown in Dean
(1942), as well as on several other variables that may have affected the outcomes of the
epidemiological studies. He indicated that his preceding comments draw attention to several
shortcomings of the chemical method used and other limiting aspects of the studies summarized
by Dean (1942) and recommended that the uncertainties associated with these factors be
discussed wherever appropriate and certainly in the "Uncertainty Factors" section.
Charge Question 4. Are some teeth more susceptible to severe fluorosis than others? If so,
should the OW make modifications to the age range identified as the period of concern?
The reviewers had differing opinions on whether some (and which) teeth are more susceptible to
severe fluorosis.
Dr. Cauley thought age may be a factor in the susceptibility to severe fluorosis and suggested
that a more direct discussion of the age at risk is needed and that this topic should be highlighted
in a separate section. She suggested that, in particular, a table summarizing the ages of the
children in each study and the age range which appeared to be at highest risk would be helpful.
She recommended that the writers add this to Table 3-16.
Dr. DenBesten thought that the report was clear in showing that posterior teeth also are
susceptible, and the case for including children up to age 14 was clear and compelling.
Neither Dr. Jackson nor Dr. Whitford has come across any literature that indicates that some
teeth are more susceptible to the development of severe dental fluorosis. Dr. Jackson believes
that the susceptibility is the same assuming that the exposure is constant and taking into
consideration how long the developing tooth is exposed to higher levels of fluoride. He stated
that, for example, maxillary third molars may present greater evidence of severe fluorosis
(pitting, staining, etc) because they take longer to develop (12-16 years-of-age) and erupt as
opposed to a maxillary central incisor (5 years-of-age). Dr. Whitford hypothesized that the
posterior teeth may be more susceptible since their development is more protracted than that of
the anterior teeth.
April 2009 Page 8
-------
Charge Question 5. Are the data on cavities as collected and graphically presented by the
OW consistent with the hypothesis that there is an increased risk for cavities in susceptible
individuals with severe dental fluorosis?
Dr. Cauley, Dr. DenBesten, and Dr. Jackson agreed that the data on cavities presented in the
report are consistent with the stated hypothesis.
Dr. Cauley noted that the data summarized in Table 3-40 to 3-44 are convincing that dental
caries are more common in subjects with severe fluorosis. The data suggest a U-shaped
relationship with fluoride: a higher risk in those with very low intakes and intakes >2-3 mg/L.
Dr. Jackson noted that the OW may want to consider examiner bias in the methodology of these
studies. Although it would be impossible to eliminate this bias, it is something that should at
least be mentioned as a possible confounder. He also described the findings of the following
study (see Appendix B for a description of the study findings):
• Jackson R, Kelly S, Katz B, Hull J, Stookey G. (1995). Dental fluorosis and caries
prevalence in children residing in communities with different levels of fluoride in the
water. JPublHealth Dent. 88:79-84.
Dr. Whitford disagreed with the other reviewers, noting that the available data present an
unclear relationship. He discussed conflicting results from several studies in his final comments
in Appendix B. He noted that overall, and as summarized in Figure 3-7 on page 71, the
relationship between the severity of dental fluorosis and the risk of caries is suggestive, but not
convincing. He believes that this subject requires more study with control for variables that are
known risk factors for caries before a reasonably firm conclusion can be drawn about the
possibility of an association between severe dental fluorosis and an increased risk of caries.
Charge Question 6. Are there recent data that would impact the Institute of Medicine
(IOM; 1997) Adequate Intake value of 0.05 mg/kg/day for fluoride that the OW should
consider in this assessment?
Dr. Cauley, Dr. Jackson, and Dr. Whitford were unaware of any recent data that could
influence the adequate intake (AI) value.
Dr. Cauley suggested that the writers mention life stage and the upper limit of toxicity. She
pointed out that the American Dental Association website may have updated their fluoride
document in 2006. In her final comment for Question 6 (see Appendix B), she provided two
tables: Criteria and Dietary Reference Intake Values for Fluoride by Life Stage Group and
Tolerable Upper Intake Levels by Life Stage Group.
Although Dr. Whitford did not know of any additional data, he commented on the interpretation
of the lOM's values. He noted that the lOM's AI value represents the amount of intake of any
substance "needed to maintain a defined nutritional state or criterion of adequacy in essentially
all members of a specific healthy population." In the case of fluoride, the AI (see page 301,
Dietary Reference Intakes, 1997) "is based on estimated intakes that have been shown to reduce
the occurrence of dental caries maximally in a population without causing unwanted side effects
April 2009 Page 9
-------
including moderate dental fluorosis." He noted that this does not mean that intakes somewhat
higher than 0.05 mg/kg/day increase risk of moderate dental fluorosis. In fact, the lOM's
estimate for the threshold for that risk is 0.10 mg/kg/day.
Dr. DenBesten disagreed with the presumption that the lOM's value is sufficient. The weight of
evidence indicates that the primary mechanism by which fluoride protects against tooth decay is
a topical effect. Therefore, she noted that the lOM's recommendation of an adequate intake
value, at least relating to tooth decay, should be reassessed.
Charge Question 7. Can you suggest an approach to transform the water concentration
data from the Dean (1942) study to units of dose for the population susceptible to severe
dental fluorosis other than that used by the OW?
Dr. Cauley suggested that writers use more generalizable data from the National Health and
Nutrition Examination Survey (NHANES) for body weight. She noted that if this approach were
used, the writers could still utilize the water intake estimates from the Erchow and Cantor (1989)
paper.
Dr. DenBesten and Dr. Jackson noted several difficulties in developing an approach for
transforming the water concentration data from the Dean study to units of dose for the population
susceptible to severe dental fluorosis.
Dr. DenBesten stated that serum fluoride levels would be the most useful measurement, but
these levels are not available. She believes that using Dean's data as a starting point to quantify
total ingestion is the best option. She recommended that future studies include random sampling
of serum fluoride levels to strengthen future decision-making relative to fluoride intake.
Dr. Jackson noted that the U.S. marketplace is constantly changing and ingestion of tap water
continues to decline. U.S. EPA's (2004) report stated that bottled water accounted for only 13%
of water consumption in the United States. More recent trade manufacturing data indicates that
bottled water consumption in the United States exceeds this percentage by a wide margin, and
bottled water consumption may surpass tap water consumption in the near future. Other
published data suggest that among Hispanic individuals, tap water is commonly perceived as
"unhealthy" and again bottled water is almost exclusively consumed. Another point that should
be further explored is the possible "halo effect" of imported foods and beverages into the United
States and the fluoride content of these consumables.
Charge Question 8. Can you provide input concerning the strengths and weaknesses of the
approach utilized by the OW to identify a lower bound dose for severe dental fluorosis?
Dr. Cauley and Dr. DenBesten both listed strengths and weaknesses to the approach utilized by
OW to identify a lower bound dose for severe dental fluorosis.
April 2009 Page 10
-------
Strengths:
Dr. Cauley said that the main strengths of the approach were that: (1) conservative estimates
were carried out with and without the outlier community; (2) several different models were
considered; and (3) sensitivity analysis showed that even eliminating three highest data points
had little effect on the model goodness of fit or on the BMDL.
Dr. DenBesten added that: (1) the data used were limited to studies found in the Dean report
where the only exposure to fluoride was drinking water; and (2) a careful assessment was done
regarding tap water consumed and mean body weights.
Weaknesses:
Dr. Cauley said that the inherent weakness of the approach relates primarily to the weaknesses
previously cited related to the parent Dean (1942) study. The writers may consider adding a
formal test for trend in data.
Dr. DenBesten added that the assumption that 0.05 mg/kg/day is a required amount of fluoride
and that dose estimates must be above this level could be considered a weakness because the
purpose of U.S. EPA's analysis was to determine risk, not to conduct risk benefit analyses. In
addition, she noted that the assumption that the small number of children who displayed severe
fluorosis were those who had excess exposure to fluoride could also be considered a weakness.
This assumes that these children drank significantly more water since water was the only source
of fluoride. It is more likely that genetic or other causes are responsible for this small outlier
group.
Dr. Whitford expressed additional concerns about the approach utilized by the OW, particularly
in reference to the determination of the LOAEL. He proposed that the variability among
published studies regarding the relationship between estimated fluoride intakes and the risk of
severe dental fluorosis may be due to the relative altitude of the cities that were studied. He
provided a detailed discussion of this issue in his final comment for Question 8 (see Appendix
B). He suggested that based on the variability among published studies regarding the relationship
between estimated fluoride intakes and the risk of severe dental fluorosis, consideration should
be given to establishing a lower bound dose other than 1 percent.
Charge Question 9. Are you aware of dose estimates other than those from IOM (1997)
and the World Health Organization (WHO; 2002) that are appropriate critical doses for
skeletal effects and/or can you suggest a different approach that the OW might use to
estimate the fluoride dose associated with skeletal fractures using available data?
Dr. Cauley referred to her response to Question 2, which included additional references that
would address this question. She expressed some concern about the inconsistencies across the
age groups. The significance of the differences in results across age could have occurred by
chance. In addition, she noted that referring to the "radius" (one of the bones of the forearm) as a
whole is problematic because if it is the distal radius (the end of the radius closest to the wrist), it
is primarily trabecular bone mineral density (BMD), whereas the mid radius is predominantly
April 2009 Page 11
-------
cortical BMD. She added that it is very problematic if the effects of fluoride differ by bone type.
Lastly, she indicated that in the instance of randomized trials, the compliance/adherence to the
Na F is a limitation because it is poorly tolerated. She questioned whether they adhered to the
intent to treat principle.
Both Dr. DenBesten and Dr. Jackson noted that they were not aware of additional dose
estimates that would be appropriate for determining critical dosage levels and their possible
skeletal effects. Dr. Jackson added that he could not think of an alternative to the approach
taken to estimate fluoride intake associated with skeletal fractures.
Dr. Whitford noted that as in the case of dental fluorosis, the critical doses for skeletal effects
are difficult to establish with a reasonable degree of certainty. The lOM's estimate (page 307) of
fluoride exposures that may result in clinical signs of the "milder forms" of skeletal fluorosis
(preclinical and perhaps stage I) is 10 mg/day for 10 or more years. There are published
exceptions suggesting that higher exposure levels and durations are required; these are noted on
the same page of the IOM report. In addition, chronic consumption of large amounts of tea and
chronic exposure to aluminum can alter the skeletal response to fluoride. Dr. Whitford included
a discussion of recent case reports of "tea fluorosis" in the U.S. in his final comments (see
Appendix B). In addition, he stated that much of the data relating bone fluoride concentrations to
the stages of skeletal fluorosis comes from studies of workers in aluminum processing factories.
High, chronic exposures to aluminum result in skeletal changes that share some features in
common with skeletal fluorosis; this makes it difficult to attribute the skeletal changes only to
fluoride. He believes that this subject is worthy of further exploration.
Charge Question 10. Are you aware of any data that can be used to demonstrate that
protection of the secondary teeth from severe dental fluorosis will also protect the primary
teeth?
Dr. Cauley, Dr. Jackson, and Dr. Whitford said that they are not aware of any additional data
that demonstrate that protection of the secondary teeth from severe dental fluorosis also will
protect primary teeth. Dr. Whitford added that there are a few reports suggesting that dental
fluorosis in the primary teeth may correlate with the condition in secondary teeth.
Dr. DenBesten stated that fluorosis in both primary and permanent teeth is caused by ingested
fluoride, and there is no data to suggest that primary teeth are more susceptible to fluorosis than
permanent teeth. Therefore, a measure that would protect permanent teeth would require limiting
ingestion of fluoride. These same measures would protect primary teeth.
Charge Question 11. Are there factors other than those discussed in Section 3.1.4 that
increase sensitivity to fluoroisis of the teeth or bones?
All four reviewers suggested additional factors that may have some effect on an individual's
sensitivity to fluorosis of the teeth and/or bones.
April 2009 Page 12
-------
Dr. Cauley questioned whether race/ethnicity and pubertal stage are important. Likewise, Dr.
Whitford noted that some researchers have suggested that African-Americans are more
susceptible to dental fluorosis but less susceptible to skeletal fluorosis.
Dr. DenBesten said that there seem to be individuals who are uniquely sensitive to the effects of
fluoride on enamel formation. These individuals may have other more "hidden" enamel defects
that are exacerbated by the effects of fluoride. Little is currently known as to why some
individuals seem to be more fluoride sensitive.
In addition, Dr. Jackson cited a study by Hong et al (2004) that appears to indicate that the use
of amoxicillin could play a contributing role in the development of primary tooth fluorosis,
especially for children exposed to lower levels of fluoride. The full citation is:
• Hong L, Levy S, Warren J, Bergus G, Dawson D, Wefel J, Broffitt B. (2004) Primary
tooth fluorosis and amoxicillin use during infancy. JPublHealth Dent. 64: 38-44.
Dr. Jackson also indicated that there have been studies (i.e., Vieira et al., 2005; Yan et al., 2007)
in mice that suggest there may be a varying genetic response to identical levels of fluoride
ingestion, and these responses have been identified in both tooth and bone formation. He noted
that Vieira et al. (2005) found that genetic influences have a direct bearing on the biomechanical
properties of the teeth. Furthermore, Yan et al. (2007) found strain-specific effects, like increased
osteoclastogenesis, when mice were exposed to physiological level of fluoride. While he was
unable to find comparable human trials, he recommended that this area be further explored as the
technological means become available. The full citations are:
• Vieira A, Hanocock R, Eggertsson H, Everett E, Grynpas M. (2005). Tooth quality in
dental fluorosis: genetic and environmental factors. Calcified Tissue International.
76:17-25.
• Yan D, Gurumurthy A, Wright M, Pfeiler T, Loboa E, Everett E. (2007). Genetic
background influences fluoride's effects on osteoclastogenesis. Bone. 41:1036-1044.
Dr. Whitford cited published data that indicate that the susceptibility to dental fluorosis (Everett
et al., J Dent Res 81: 794-698, 2002) and the mechanical properties in the bone (Mousny et al,.
Bone 39: 1283-1289, 2006) are different among strains of mice. He noted that the differences
presumably are due to genetic differences among the strains.
Charge Question 12. Do you support the OW's conclusion that an RfD of 0.07 mg/kg/day
will be protective for severe dental fluorosis in children and skeletal effects in adults while
still providing for the beneficial effects of fluoride?
In general, Dr. Cauley, Dr. Jackson, and Dr. Whitford support the use of the 0.07 mg/kg/day
RfD. However, all three of these reviewers stated that there is a considerable degree of
uncertainty regarding the estimate.
April 2009 Page 13
-------
Dr. Jackson indicated that he does not think that a claim can be made that no single individual
will be completely immune from the development of severe fluorosis even at this recommended
RfD. As a result of an individual's potential genetic predisposition, he does not think that the
possibility of developing severe fluorosis can be totally ruled out. Additionally, he noted that his
concern is not as much about the amount of fluoride that is ingested through the public water
supply, but the other well-known sources of fluoride (see his final comment for Question 12 in
Appendix B for other sources) that have an additive effect to that derived from consuming
fluoridated drinking water.
Dr. Whitford added that, as estimated by the IOM (1997), the RfD may be closer to 0.10
mg/kg/day for moderate (not severe) dental fluorosis and substantially higher than that for
clinically significant skeletal effects in the United States.
On the other hand, Dr. DenBesten did not agree with this recommendation because it is based on
limiting severe fluorosis to 1% of the population and the lOM's recommended adequate intake
level. She suggested that the level be lowered to eliminate severe fluorosis. She noted that one
percent of the population represents a relatively large number of individuals, and these
individuals are most likely uniquely sensitive to fluoride. She recommended that the data be
analyzed without taking the lOM's recommended adequate levels into account. Then, a
secondary analysis could be conducted to include the lOM's recommendations.
April 2009 Page 14
-------
Appendix A: Pre-Meeting Responses to the Charge Questions from Reviewers
(Organized Alphabetically by Reviewer)
Page A-1
-------
Jane A. Cauley, Dr.P.H
Professor, Department of Epidemiology, University of Pittsburgh
1. Was the document clear and transparent in its presentation of data and explanation of the
analytical approaches used to characterize the concentration- response and dose-response
relationships for severe dental fluorosis? If not, do you have any suggestions that will enable the
authors to improve its clarity?
Response
The document could be strengthened by reorganizing it as a manuscript. The
Introduction should provide background and a scientific rationale for this analysis. The
specific objectives should be listed.. The methods for identifying all the literature should
be transparent and consistently applied to both the fluorosis and fracture sections. This
is typical of meta analyses which also apply certain quality scores to each paper.
I think it might also be helpful if I understand who the audience for this report is. I found
it a confusing because I do not understand the roles of the EPA, OW and the NRC : how
will each agency find this information helpful.? What prompted this report?
2. Are you aware of any significant publications related to severe dental fluorosis or the skeletal
effects of fluoride that are not included in the noncancer assessment document?
Response
The dental fluorosis literature included studies with a wide range of fluoride
levels while the skeletal/ fracture studies were limited to those with fluoride levels >
4mg/l. You might consider adding some of the literature on fractures at lower levels of
fluoride including the following:
Cauley JA, Murphy PA, Riley TJ, Buhari AM. Effects of fluoridated drinking water on
bone mass and fractures: the study of osteoporotic fractures. J Bone Miner Res. 1995
Jul; 10(7): 1076-86.
Phipps, KR, Orwoll, ES, Mason, JD, Cauley, J A. Community water fluoridation, bone
mineral density, and fractures: prospective study of effects in older women. BMJ 2000
321(7265):860-4.
In addition, Phipps KR et al published a study comparing bone mineral density in
Deming, NM (0.7mg/L) and Lordsburg, NM (3.5mg/L). The citation is: Phipps KR,
Burt BA. Water-borne fluoride and cortical bone mass: a comparison of two
communities. J Dent Res. 1990 Jun;69(6): 1256-60.
It appeared to me that this section is overall less thorough than the dental fluorosis
section. Additional Selections include the following. I would pay particular attention to
the papers by Kathy Phipps, Cyrus Cooper, S Jacobsen and T Hillier. These are high
quality in my opinion.
April 2009 Page A-2
-------
Demos LL, Kazda H, Cicuttini FM, Sinclair MI, Fairley CK.
Water fluoridation, osteoporosis, fractures—recent developments.
Aust Dent J. 2001 Jun;46(2):80-7; quiz 143.
PMID: 11491235 [PubMed - indexed for MEDLINE]
Fabiani L, Leoni V, Vital! M.
Bone-fracture incidence rate in two Italian regions with different fluoride
concentration levels in drinking water.
J Trace Elem Med Biol. 1999 Dec;13 ( 4) :232-7 .
PMID: 10707346 [PubMed - indexed for MEDLINE]
Kleerekoper M.
Fluoride and the skeleton.
Crit Rev Clin Lab Sci. 1996 Apr;33(2):139-61. Review.
PMID: 8744520 [PubMed - indexed for MEDLINE]
Raheb J.
Water fluoridation, bone density and hip fractures: a review of recent
literature.
Community Dent Oral Epidemiol. 1995 Oct;23(5):309-16. Review. No abstract
available.
PMID: 8529346 [PubMed - indexed for MEDLINE]
S0gaard CH, Mosekilde L, Schwartz W, Leidig G, Minne HW, Ziegler R.
Effects of fluoride on rat vertebral body biomechanical competence and bone
mass .
Bone. 1995 Jan;16(1):163-9.
PMID: 7742076 [PubMed - indexed for MEDLINE]
Kroger H, Alhava E, Honkanen R, Tuppurainen M, Saarikoski S.
The effect of fluoridated drinking water on axial bone mineral density—a
population-based study.
Bone Miner. 1994 Oct;27(1):33-41.
PMID: 7849544 [PubMed - indexed for MEDLINE]
Ripa LW.
A half-century of community water fluoridation in the United States: review
and
commentary.
J Public Health Dent. 1993 Winter;53(1):17-44 . Review.
PMID: 8474047 [PubMed - indexed for MEDLINE]
Gordon SL, Corbin SB.
Summary of workshop on drinking water fluoride influence on hip fracture on
bone
health. (National Institutes of Health, 10 April, 1991)
Osteoporos Int. 1992 May;2(3) :109-17 . No abstract available.
PMID: 1627897 [PubMed - indexed for MEDLINE]
Colquhoun J.
Water fluoride and fractures.
N Z Med J. 1991 Aug 14;104(917):343. No abstract available.
April 2009 Page A-3
-------
PMID: 1876343 [PubMed - indexed for MEDLINE]
McNeill KG, Coote GE, Hitchman AJ.
Uptake of fluorine in cortical and trabecular bone.
J Bone Miner Res. 1991 Aug;6(8):859-64.
PMID: 1785375 [PubMed - indexed for MEDLINE]
Arnala I, Alhava EM, Kivivuori R, Kauranen P.
Hip fracture incidence not affected by fluoridation. Osteofluorosis studied
in
Finland.
Acta Orthop Scand. 1986 Aug;57(4):344-8.
PMID: 3788501 [PubMed - indexed for MEDLINE]
Ericsson Y, Luoma H, Ekberg 0.
Effects of calcium, fluoride and magnesium supplementations on tissue
mineralization in calcium- and magnesium-deficient rats.
J Nutr. 1986 Jun;116(6):1018-27.
PMID: 3723198 [PubMed - indexed for MEDLINE]
Simonen 0, Laitinen 0.
Does fluoridation of drinking-water prevent bone fragility and osteoporosis?
Lancet. 1985 Aug 24;2 ( 8452) :432-4.
PMID: 2863455 [PubMed - indexed for MEDLINE]
Madans J, Kleinman JC, Cornoni-Huntley J.
The relationship between hip fracture and water fluoridation: an analysis of
national data.
Am J Public Health. 1983 Mar;73(3):296-8.
PMID: 6824115 [PubMed - indexed for MEDLINE]
Schamschula RG, Barmes DE.
Fluoride and health: dental caries, osteoporosis, and cardiovascular
disease.
Annu Rev Nutr. 1981;1:427-35. Review.
PMID: 6764723 [PubMed - indexed for MEDLINE]
Alhava EM, Olkkonen H, Kauranen P, Kari T.
The effect of drinking water fluoridation on the fluoride content, strength
and
mineral density of human bone.
Acta Orthop Scand. 1980 Jun;51(3):413-20.
PMID: 7446020 [PubMed - indexed for MEDLINE]
Stein ID, Granik G.
Human vertebral bone: relation of strength, porosity, and mineralization to
fluoride content.
Calcif Tissue Int. 1980;32 ( 3):189-94 .
PMID: 6775787 [PubMed - indexed for MEDLINE]
Hegsted DM.
Osteoporosis and fluoride deficiency.
Postgrad Med. 1967 Jan;41(1):A49-53. No abstract available.
PMID: 6036214 [PubMed - indexed for MEDLINE]
Bernstein DS, Sadowsky N, Hegsted DM, Guri CD, Stare FJ.
Prevalence of osteoporosis in high- and low-fluoride areas in North Dakota.
April 2009 Page A-4
-------
JAMA. 1966 Oct 31;198(5):499-504. No abstract available.
PMID: 5953273 [PubMed - indexed for MEDLINE]
Hillier S, Cooper C, Kellingray S, Russell G, Hughes H, Coggon D.
Fluoride in drinking water and risk of hip fracture in the UK: a case-
control
study.
Lancet. 2000 Jan 22;355(9200):265-9.
PMID: 10675073 [PubMed - indexed for MEDLINE]
Hillier S, Inskip H, Coggon D, Cooper C.
Water fluoridation and osteoporotic fracture.
Community Dent Health. 1996 Sep;13 Suppl 2:63-8. Review.
PMID: 8897754 [PubMed - indexed for MEDLINE]
Jacobsen SJ, Goldberg J, Cooper C, Lockwood SA.
The association between water fluoridation and hip fracture among white
women and
men aged 65 years and older. A national ecologic study.
Ann Epidemiol. 1992 Sep;2(5):617-26.
PMID: 1342313 [PubMed - indexed for MEDLINE]
Cooper C, Wickham C, Lacey RF, Barker DJ.
Water fluoride concentration and fracture of the proximal femur.
J Epidemiol Community Health. 1990 Mar;44(1):17-9.
PMID: 2348142 [PubMed - indexed for MEDLINE]
1: Phipps KR, Orwoll ES, Mason JD, Cauley JA.
Community water fluoridation, bone mineral density, and fractures:
prospective
study of effects in older women.
BMJ. 2000 Oct 7;321 (7265) :860-4.
PMID: 11021862 [PubMed - indexed for MEDLINE]
Phipps KR, Orwoll ES, Bevan L.
The association between water-borne fluoride and bone mineral density in
older
adults.
J Dent Res. 1998 Sep;77 ( 9):1739-48 .
PMID: 9759671 [PubMed - indexed for MEDLINE]
Phipps K.
Fluoride and bone health.
J Public Health Dent. 1995 Winter;55(1):53-6. Review.
PMID: 7776293 [PubMed - indexed for MEDLINE]
Phipps KR, Burt BA.
Water-borne fluoride and cortical bone mass: a comparison of two
communities.
J Dent Res. 1990 Jun;69(6):1256-60.
PMID: 2355118 [PubMed - indexed for MEDLINE]
3. Are the strengths and weaknesses of Dean (1942) study (selected as critical) fully
characterized?
April 2009 Page A-5
-------
Response:
Additional strengths include the wide range of fluoride concentrations, although there
were fewer subjects in the high fluoride categories. The dose response in the data showing
increasing risk of severe skeletal fluorosis with increasing fluoride is also a strength. Finally the
consistency of the findings across several different communities.
Other weaknesses: No information is provided on race/ethnicity. The data were
collected the late 1930-40s. Although confounding by use of other fluoride products would be
minimal, there are many other cohort differences between children exposed to fluoride in the late
1930-40s compared to now. For example, dental hygiene, dietary intakes (less water... more
carbonated beverages), body weight are very different in today's children compared to children
in the 30s.. Do puberty/ hormonal changes influence fluoride effects? This may be important
because age of menarche has been decreasing. Hormones have been shown to influence dental
health. Another weakness of the studies reviewed is that there appears to be no information on
duration of exposure. How long did these children live in each community? Did the inclusion
criteria include a minimum time of residence? If so this point needs to be made in the document.
4. Are some teeth more susceptible to sever fluorosis than others? If so, should the OW make
modifications to the age range identified as the period of concern?
Response
This should be highlighted in a separate section. A more direct discussion of the
age at risk is needed. A table summarizing the ages of the children in each study and the
age range which appeared to be at highest risk would be helpful. You could add this to
Table 3-16.
5. Are the data on cavities as collected and graphically presented by the OW consistent with the
hypothesis that there is an increased risk for cavities in susceptible individuals with severe dental
fluorosis?
Response
The data summarized in Table 3-40 to 3-44 are convincing that dental caries are
more common in subjects with severe fluorosis. The data suggest a U shaped
relationship with fluoride: a higher risk in those with very low intakes and intakes
>2-3mg/L.
6. Are there recent data that would impact the IOM (1997) Adequate Intake value of 0.05
mg/kg/day for fluoride that the OW should consider in this assessment?
Response
I do not know of any more recent data. Perhaps you want to mention life stage
and Upper limit of toxicity. The ADA website looks like they updated their
Fluoride document in 2006.
April 2009 Page A-6
-------
TABLE S-5
Criteria and Dietary Reference Intake Values for Fluoride by Life Stage Group
LifeStagie
10 through 6
imonths
|7 through 12
imonths
11 through 3
[years
[4 through 8
iyears
19 through 13
iyears
i 14 through 18
iyears
119 through 30
[years
[31 through 50
iyears
151 through 70
[years
> i70 years
Criterion
A! (rng/day)
Male! Female
Pregnancy
<= 118 years
119 through 50
[years
Lactation
<= 118 years
119 through 50
[years
[Human milk * 0.01
[ content [
[Caries j 0.5
[prevention [
[Caries 1 0.7
[prevention [
[Caries 1 1
{prevention j
[Caries I 2
[prevention 1
^Caries I 3
[prevention 1
[Caries I 4
[prevention [
[Caries [ 4
[prevention \
[Caries \ 4
[prevention [
[Caries \ 4
[prevention |
[Caries \
[prevention \
[Caries \
[prevention |
[Caries i
[prevention [
[Caries [
^prevention |
~ooT
0.5
O7
1
2
3
3
3
3
3
3
3
3
o
5
"AI = Adequate Intake. For healthy infants fed human milk, AI is the mean intake. The observed estimate
of nutrient intake that reduces the incidence of dental caries maximally in a group of healthy people. The
AI is used if the scientific evidence is not available to derive an EAR. The AI is believed to cover their
April 2009
Page A-7
-------
needs, but lack of data or uncertainty in the data prevent being able to specify with confidence the
percentage of individuals covered by this intake.
TABLE S-6
Tolerable Upper Intake Levels (ULa), by Life Stage Group
PhOi^
(g/dayy
through 6 months
\1 through 12 months
\ 1 through 3 years
\4 through 8 years
19 through 18 years
119 through 70 years
> §70 years
Pregnancy
<= * 18 years
>19 through 50 years
Lactation
years
119 through 50 years
Mr!
ND!
2.5!
2l!
2.5
ND
2.5
2.5
2.5
___
2.5
ND
3
3"
4
4
3
3.5
3.5
4
4
ND!
ND!
65!
no!
350 I
350!
350!
350!
350 |
350!
350 I
25
25
50
50
__
50
50
50
50
50
50
0.7!
0.9!
1.3 |
___^
10]
10 1
10 1
10
10
10 1
"UL = the maximum level of daily nutrient intake that is likely to pose no risk of adverse effects to members
of the healthy general population. Unless specified otherwise, the UL represents total intake from food,
water, and supplements.
*The UL for magnesium represents intake from pharmacological agents only and does not include intake
from food and water.
cAs cholecalciferol. 1 ug cholecalciferol = 40 IU vitamin D.
''ND. Not determinable due to lack of data of adverse effects in this age group and concern with regard to
lack of ability to handle excess amounts. Source of intake should be from food only in order to prevent high
April 2009
Page A-8
-------
levels of intake.
7. Can you suggest an approach to transform the water concentration data from the Dean (1942)
study to units of dose for the population susceptible to severe dental fluorosis other than that
used by the OW?
Response
To estimate the RFD of 0.07 mg/Kg/day, the OW attempted to obtain detailed
analysis of average body weights and water intakes during the time the Dean data was collected.
Data from the Erchow and Cantor were used (1989). Why not use more generalizable data from
NHANES at least for body weight? You could still use the water intake estimates from the
Cantor paper.
8. Can you provide input concerning the strengths and weaknesses of the approach utilized by
the OW to identify a lower bound dose for severe dental fluorosis?
Response
Strengths: Conservative estimates carried out with and without the outlier
community; several different models were considered, sensitivity analysis showed
that even eliminating three highest data points had little effect on the model
goodness of fit or on the BMDL.
Weakness: Inherent weakness of the approach relate primarily to the weaknesses
cited related to the parent Dean Study. You might consider adding a formal test
for trend in data.
9. Are you aware of dose estimates other than those from IOM (1997) and WHO (2002) that are
appropriate critical doses for skeletal effects and/or can you suggest a different approach that the
OW might use to estimate the fluoride dose associated with skeletal fractures using available
data?
Response
See answer to #2 for additional references.
Some concern about the inconsistencies across the age groups. The significance of the
differences in results across age could have occurred by chance alone.
Referring to the" radius" is problematic because if it s the distal radius, it is primarily
trabecular BMD and mid radius is predominantly cortical BMD. If the effects of
fluoride differs by type of bone, this is very problematic.
April 2009 Page A-9
-------
Randomized trials: please note the compliance/adherence to the Na F because it is poorly
tolerated. This is a limitation to theses studies. Did they adhere to the intent to treat
principle?
10. Are you aware of any data that can be used to demonstrate that protection of the secondary
teeth from severe dental fluorosis will also protect the primary teeth?
Response
No, I am not aware of any additional data.
11. Are there factors other than those discussed in Section 3.1.4 that increase sensitivity to
fluoroisis of the teeth or bones?
Response
I am not sure but I question whether race/ ethnicity and pubertal stage is
important.
12. Do you support the OW's conclusion that an RfD of 0.07 mg/kg/day will be protective for
severe dental fluorosis in children and skeletal effects in adults while still providing for the
beneficial effects of fluoride?
Response
Yes, I am in support of the 0.07 mg/kg/day but I would admit the uncertainty that surrounds this
estimate..
April 2009 Page A-10
-------
Pamela Denbesten, D.D.S., M.S.
Professor and Chair, Division of Pediatric Dentistry, Department of Orofacial Sciences,
University of California at San Francisco
1. Was the document clear and transparent in its presentation of data and explanation of the
analytical approaches used to characterize the concentration- response and dose-response
relationships for severe dental fluorosis? If not, do you have any suggestions that will enable the
authors to improve its clarity?
Response:
The document was for the most part clear and transparent in its presentation. To
improve clarity, I suggest the following:
Page 13. Please indicate the purpose of the Secondary Maximum Contaminant
Level, and what this was used for.
Page 15, section 2.1.1. The statement with regards to fluoroapatite, at the end of
paragraph 2.1.1 is not accurate. The mineral formed in tooth enamel exposed to higher
fluoride levels is fluoride-containing carbonated apatite, rather than fluoroapatite. In
fluorotic enamel, fluoride levels are increased to about 200 ppm rather than the 10-100
ppm fluoride in normal enamel below the surface, while fluorapatite is about 30,000 ppm
fluoride. The fluoride-substituted apatite, formed with increased levels of ingested
fluoride, has some increased resistance to bacterial acids that cause tooth decay.
However, the primary function of fluoride in drinking water to reduce tooth decay is
topical, primarily by the enhancement of remineralization.
References:
1. Featherstone JOB. Prevention and reversal of dental caries: role of low level fluoride.
Community Dent and Oral Epidemiol 1999;27:31-40.
2. Featherstone JOB. The science and practice of caries prevention. JADA 2000; 131:887-99.
3. Fejerskov O, Thylstrup A, Larsen MJ. Rational use of fluorides in caries prevention. Acta
Odont. Scand. 1981;39:241-49.
4. LeGeros RZ. Calcium Phosphates in Enamel, Dentin and Bone. In: Myers HM, editor.
Calcium Phosphates in Oral Biology and Medicine. Basel: Karger; 1991. p. 108-29.
5. Robinson C, Kirkham J, Weatherell JA. Fluoride in teeth and bone. In: Fejerskov O, Ekstrand
J, Burt BA, editors. Fluoride in Dentistry. 2nd ed. Copenhagen: Munksgaard; 1996. p. 69-87.
6.Ten Gate JM, Featherstone JOB. Mechanistic aspects of the interactions between fluoride and
dental enamel. CRC Critical Reviews in Oral Biology 1991;2:283-96.
Page 19 section 2.2. Determine whether in fact the apatite formed in bone is
fluoroapatite or like in tooth, is a fluoride-substituted hydroxyapatite.
Page 21 section 3. Change "...are preferred for evaluating the potential effects
of fluoride in drinking water", to the more accurate statement, "...are preferred for
evaluating the potential effects of ingested fluoride".
Page 57; paragraph 1. Statements relating fluoride in the drinking water to anti-
caries benefit from studies conducted after 1980, should be qualified in that other
sources of fluoride may be confounding factors. For example, fluorosis incidence in
more recent studies may decrease as consumption of tap water is altered by
consumption of bottled beverages.
April 2009 Page A-11
-------
In general for all of section 3.2.2, when fluoride effects on dental caries are
discussed, the data should be divided into studies before and after 1980 when fluoride
was widely available in toothpaste, and perhaps before and after the late 1990s when
bottled beverages became widely used.
Page 86; Table 3-52. What does "complete" and "total" refer to?
Page 94; section 4.4. Explain with the "NOAEL/LOAEL" approach is, or at least
spell out the acronmym.
Page 97 paragraph 1. The importance of fluoride as a nutrient may need to be
reassessed, given that its primary function in caries prevention is topical. It would seem
more appropriate focus on the upper limits for ingestion of this caries preventive agent.
Page 98. The statement describing the timing incisor tooth formation is
incorrect. The secondary incisors and molars begin development in utero, and continue
with crown formation complete at about 3 years of age.
Page 101. Please justify the rationale for setting the BMDL at an incidence of 1 %
severe fluorosis, rather than 0% severe fluorosis.
Page 103. As stated above, I question the concept of a recommended fluoride
intake.
Page 104, paragraph 1. The statement that "...fluoroapatite crystals disrupt the
hydroxyapatite crysal lattice..." is incorrect and should be deleted.
2. Are you aware of any significant publications related to severe dental fluorosis or the skeletal
effects of fluoride that are not included in the noncancer assessment document?
Response:
No I am not
3. Are the strengths and weaknesses of Dean (1942) study (selected as critical) fully
characterized?
Response:
Yes, to me they are.
4. Are some teeth more susceptible to severe fluorosis than others? If so, should the OW make
modifications to the age range identified as the period of concern?
Response:
The OW report was clear in showing that posterior teeth are also susceptible, and the
case for including children up to age 14 was clear and compelling.
5. Are the data on cavities as collected and graphically presented by the OW consistent with the
hypothesis that there is an increased risk for cavities in susceptible individuals with severe dental
fluorosis?
Response
Yes
April 2009 Page A-12
-------
6. Are there recent data that would impact the IOM (1997) Adequate Intake value of 0.05
mg/kg/day for fluoride that the OW should consider in this assessment?
Response.
I do not know on what basis the IOM recommended that fluoride levels at 0.05
mg/kg/day be ingested. However the weight of evidence is that the primary mechanism
by which fluoride protects against tooth decay is topical.
I would agree with the recent NRC report that states; "The recommended optimal
fluoride intake for children to maximize caries prevention and minimize the occurrence
of dental fluorosis is often stated as being 0.05-0.07 mg/kg/day (Levy1994; Heller et al.
1999, 2000). Burt (1992) attempted to track down the origin of the estimate of 0.05-0.07
mg/kg/day as an optimum intake of fluoride but was unable to find it. He interpreted the
available evidence as suggesting that 0.05-0.07 mg/kg/day (from all sources) remains a
useful upper limit for fluoride intake in children".
I agree with Burt's interpretation, and do not believe that the EPA should consider that
there is a particular desired fluoride intake. For further discussion on the relative
importance of fluoride ingestion in caries prevention, please see the response to
question 1.
7. Can you suggest an approach to transform the water concentration data from the Dean (1942)
study to units of dose for the population susceptible to severe dental fluorosis other than that
used by the OW?
Response:
No I can not. Of course serum fluoride levels would be the most useful measurement,
but these levels are not available. I believe the using Dean's data as a starting point to
quantify total ingestion is the best that we can do. I suggest that the suggestion be
given that future studies include random sampling of serum fluoride levels to strengthen
future decision making relative to fluoride intake.
8. Can you provide input concerning the strengths and weaknesses of the approach utilized by
the OW to identify a lower bound dose for severe dental fluorosis?
Response:
9. Are you aware of dose estimates other than those from IOM (1997) and WHO (2002) that are
appropriate critical doses for skeletal effects and/or can you suggest a different approach that the
OW might use to estimate the fluoride dose associated with skeletal fractures using available
data?
Response: No I can not.
10. Are you aware of any data that can be used to demonstrate that protection of the secondary
teeth from severe dental fluorosis will also protect the primary teeth?
April 2009 Page A-13
-------
Response:
Primary teeth are known to be less affected by fluoride ingestion, most likely because
they are formed in utero with a dilution of maternal serum fluoride levels as fluoride in
the placenta. This suggests that minimizing the risk of severe dental fluorosis in
secondary teeth will also protect the primary teeth
11. Are there factors other than those discussed in Section 3.1.4 that increase sensitivity to
fluoroisis of the teeth or bones?
Response:
There seem to be individuals who are uniquely sensitive to the effects of fluoride on
enamel formation. These individuals may have other more "hidden" enamel defects that
are exacerbated by the effects of fluoride. Little is currently know as to why some
individuals seem to be more fluoride sensitive.
12. Do you support the OW's conclusion that an RfD of 0.07 mg/kg/day will be protective for
severe dental fluorosis in children and skeletal effects in adults while still providing for the
beneficial effects of fluoride?
Response: I do not agree with this recommendation as it is based on limiting severe
fluorosis to 1 % of the population. I suggest that the level be lowered to eliminate severe
fluorosis. One percent of the population represents a relatively large number of
individuals, and given the data showing the primarily topical effects of fluoride I do not
see a rationale for acceptance of the 1 % severe fluorosis
April 2009 Page A-14
-------
Richard D. Jackson, D.M.D.
Assistant Professor, Preventive and Community Dentistry, School of Dentistry, Indiana
University, Oral Health Research Institute
1. Was the document clear and transparent in its presentation of data and explanation of the
analytical approaches used to characterize the concentration- response and dose-response
relationships for severe dental fluorosis? If not, do you have any suggestions that will enable the
authors to improve its clarity?
I felt that the document clearly described what is available in the literature and presented the
information in an understandable format. I am glad individual tables summarizing some of the
studies were included for reference. As an aside, I would have preferred that the studies be
described chronologically as opposed to alphabetically by author.
2. Are you aware of any significant publications related to severe dental fluorosis or the skeletal
effects of fluoride that are not included in the noncancer assessment document?
Jackson R, Kelly S, Katz B, Brizendine E, Stookey G. Dental fluorosis in children residing in
communities with different fluoride levels in the water: 33 month follow-up. Pediatric Dentistry
21:248-254.
Examination of 357 school-age children in 1994 from three Indiana communities with varying
levels of fluoride in the water (0.2, 1.0 and 4.0 ppm). Severe fluorosis using the TSIF (score 5, 6
or 7) was noted in 9% of the children examined in the 4.0 ppm community. No scores of this
magnitude were seen in the other two communities. Based on comparison to previous prevalence
data in the same communities in 1992, the prevalence of fluorosis had increased in each
community but primarily in the categories rated as a 1, 2 or 3 using the TSIF.
3. Are the strengths and weaknesses of Dean (1942) study (selected as critical) fully
characterized?
I think the strengths of the investigation conducted by Dean (1942) are fully characterized. I
doubt that anyone would disagree that Dean (1942) has been and will continue to be a
benchmark study in the dental literature as well as the much broader literature related to public
health and epidemiology. It is true that Dean's study was performed when confounding fluoride
sources were not available and thus in all probability gives a very clear picture of the prevalence
of the relationship of fluoride ingestion and the subsequent development of dental fluorosis.
However, one point of weakness that could be argued is that the population of subjects examined
was not diverse in terms of racial or cultural characteristics. As it has been postulated that
genetic factors may impact on the expression of dental fluorosis at identical levels of ingestion,
the fact that the data was collected in an exclusively white population appears to limit its
applicability for use as a benchmark.
4. Are some teeth more susceptible to severe fluorosis than others? If so, should the OW make
modifications to the age range identified as the period of concern?
April 2009 Page A-15
-------
I was a little unclear as to the intent of the question. My initial reaction was to answer "No". I
have not come across any literature that indicates that some teeth are more susceptible to the
development of severe dental fluorosis assuming that the exposure is constant and taking into
consideration how long the developing tooth is exposed to higher levels of fluoride. For example,
maxillary third molars may present with greater evidence of severe fluorosis (pitting, staining,
etc) based on the fact that they take longer to develop (Cc 12-16 years-of-age) and erupt as
opposed to a maxillary central incisor (Cc 5 years-of-age).
5. Are the data on cavities as collected and graphically presented by the OW consistent with the
hypothesis that there is an increased risk for cavities in susceptible individuals with severe dental
fluorosis?
It appears that the data are consistent with and support this hypothesis. A point of discussion may
be made concerning examiner bias in the methodology of these studies. It is obvious that it
would be impossible to eliminate bias as the most severely affected teeth would be evident by the
caries examiner when determining the DMFT/DMFS of the subject. I don't see any solution for
this dilemma with the use of a single examiner or multiple examiners but it is something that
should at least be mentioned as a possible confounder.
Jackson et al (1995) also examined children residing in communities with varying levels of
fluoride in their public water supplies. No child in the negligibly-fluoridated community or the
optimally-fluoridated community was found to have TSIF scores greater than 2. However, the
DMFT and DMFS in the 7-10 year-olds residing in the 4X optimum community were not
significantly different from those residing in the 1.0 ppm community. For 11-14 year-olds DMFS
was significantly lower than in the negligibly fluoridated community.
6. Are there recent data that would impact the IOM (1997) Adequate Intake value of 0.05
mg/kg/day for fluoride that the OW should consider in this assessment?
I am not aware of any new data that could influence the value of 0.5 mg/kg/day as the AI value.
7. Can you suggest an approach to transform the water concentration data from the Dean (1942)
study to units of dose for the population susceptible to severe dental fluorosis other than that
used by the OW?
No, using the data available the approach seems to be logical. The problem, as I see it. is the US
marketplace is constantly changing and ingestion of tap water continue to decline. The EPA
(2004) report stated that bottled water accounted for only 13% of water consumption in the
United States. More recent trade manufacturing data indicates that bottled water consumption in
the United States exceeds this percentage by a wide margin and bottled water consumption may
surpass tapwater consumption in the near future. Other published data suggest that among
Hispanic individuals, tap water is commonly perceived as "unhealthy" and again bottled water is
consumed almost exclusively. Another point that should be further explored is the possible "halo
effect" of imported foods and beverages into the United States and the fluoride content of these
consumables.
April 2009 Page A-16
-------
8. Can you provide input concerning the strengths and weaknesses of the approach utilized by
the OW to identify a lower bound dose for severe dental fluorosis?
9. Are you aware of dose estimates other than those from IOM (1997) and WHO (2002) that are
appropriate critical doses for skeletal effects and/or can you suggest a different approach that the
OW might use to estimate the fluoride dose associated with skeletal fractures using available
data?
I am not aware of additional dose estimates that would be appropriate for determining critical
dosage levels and their possible skeletal effects.
IOM.
Recommended total dietary fluoride intake
Age Reference weight* Adequate intakef Tolerable upper intake§
kg Ib mg/day mg/day
0-6 months 7 16 0.01 0.7
6-12 months 9 20 0.5 0.9
1-3 years 13 29 0.7 1.3
4-8 years 22 48 1.1 2.2
>9 years 40/76 88/166 2.0/3.8 10.0
* Values based on data collected during 1988-1994 as part of the third National Health and
Nutrition Examination Survey.
f Intake that maximally reduces occurrence of dental caries without causing unwanted side
effects, including moderate enamel fluorosis.
§ Highest level of nutrient intake that is likely to pose no risks for adverse health effects in
almost all persons.
Source: Adapted from Institute of Medicine. Fluoride. In: Dietary reference intakes for
calcium,phosphorus, magnesium, vitamin D, and fluoride. Washington, DC: National Academy
Press,1997:288-313.
BONE FRACTURE AND BONE DEVELOPMENT PROBLEMS
There were 29 studies included on the association between bone fracture and bone
development problems and water fluoridation. Other than fluorosis, bone effects (not
including bone cancers) were the most studied potential adverse effect. These studies had a
mean validity score of 3.4 out of 8. All but one study were of evidence level C. These studies
included both cohort and ecological designs, some of which included analyses controlling for
potential confounding factors. Observer bias could potentially play a role in bone fracture
studies, depending on how the study is conducted.
The evidence on bone fracture can be classified into hip fracture and other sites because
there are more studies on hip fracture than any other site. Using a qualitative method of
analysis (Figure 8.1), there is no clear association of hip fracture with water fluoridation.
Theevidence on other fractures is similar. Overall, the findings of studies of bone fracture effects
showed small variations around the 'no effect' mark. A meta-regression of bone fracture studies
also found no association with water fluoridation.
10. Are you aware of any data that can be used to demonstrate that protection of the secondary
teeth from severe dental fluorosis will also protect the primary teeth?
I am not aware of any data that supports this possibility.
April 2009 Page A-17
-------
11. Are there factors other than those discussed in Section 3.1.4 that increase sensitivity to
fluorosis of the teeth or bones?
Section 3.1.4. states that climate and altitude of place of residence; dietary habits, nutritional
status, physiological state, and certain pathological conditions may affect the occurrence and
severity of fluoride-induced dental fluorosis. The section further mentions that certain genetic
conditions such as amelogenesis imperfecta may lead to defective development of enamel.
There have been studies (Vieira et al, 2005; Yan et al, 2007) that suggest, in mice, that there
may be a varying genetic response to identical levels of fluoride ingestion. These changes have
been identified in both tooth and bone formation. Vieira et al found that genetic influences have
a direct bearing on the biomechanical properties of the teeth. Yan et al. found strain-specific
effects of physiological level of fluoride with increased osetoclastogenesis in some mouse
strains. While I was unable to find comparable human trials, this area should be further explored
as the technological means become available.
A study by Hong et al (2004) appears to indicate that the use of amoxicillin could play a
contributing role in the development of primary tooth fluorosis, especially for children exposed
to lower levels of fluoride.
12. Do you support the OW's conclusion that an RfD of 0.07 mg/kg/day will be protective for
severe dental fluorosis in children and skeletal effects in adults while still providing for the
beneficial effects of fluoride?
References
Hong L, Levy S, Warren J, Bergus G, Dawson D, Wefel J, Broffitt B. Primary tooth fluorosis
and amoxicillin use during infancy. Journal of Public Health Dentistry 64: 38-44 (2004).
Vieira A, Hanocock R, Eggertsson H, Everett E, Grynpas M. Tooth quality in dental fluorosis:
genetic and environmental factors. Calcified Tissue International 76:17-25 (2005).
Yan D, Gurumurthy A, Wright M, Pfeiler T, Loboa E, Everett E. Genetic background influences
fluoride's effects on osteoclstogenesis. Bone 41:1036-1044 (2007).
Jackson R, Kelly S, Katz B, Hull J, Stookey G. Dental fluorosis and caries prevalence in children
residing in communities with different levels of fluoride in the water. Journal of Public Health
Dentistry 88: 79-84 (1995).
A systematic review of public water fluoridation
NHS Centre for Review and Dissemination
University of York 2000
EXECUTIVE SUMMARY
This systematic review has been commissioned by the Chief Medical Officer of the
Department of Health to 'carry out an up to date expert scientific review of fluoride and health'
(Paragraph 9.20, Our Healthier Nation).
Overall, the aim has been to assess the evidence on the positive and negative effects of
population wide drinking water fluoridation strategies to prevent caries. To achieve this aim
five objectives were identified:
April 2009 Page A-18
-------
Objective 1: What are the effects of fluoridation of drinking water supplies on the incidence of
caries?
Objective 2: If water fluoridation is shown to have beneficial effects, what is the effect over
and above that offered by the use of alternative interventions and strategies?
Objective 3: Does water fluoridation result in a reduction of caries across social groups and
between geographical locations, bringing equity?
Objective 4: Does water fluoridation have negative effects?
Objective 5: Are there differences in the effects of natural and artificial water fluoridation?
Methods
A search of 25 electronic databases (with no language restrictions) and the wo rid-wide-web
was undertaken. Relevant journals and indices were hand searched and attempts were
made to contact authors for further information.
Quality inclusion criteria were based on a pre-defined hierarchy of evidence (A, B, and C).
Studies of efficacy were included if they were of evidence level A or B. In order to allow the
broadest search for evidence on potential adverse effects, studies of all levels of evidence
were included. Objective specific inclusion criteria, based on selection of participants,
intervention, outcomes assessed, and study design appropriate for a given objective were
then applied. Study validity was formally assessed using a published checklist modified for
this review (CRD Report 4, 1996).
Inclusion criteria were assessed independently by at least two reviewers. Extraction of data
from, and validity assessment of, included studies was independently performed by two
reviewers, and checked by a third reviewer. Disagreements were resolved through
consensus.
Where the data were in a suitable format, measures of effect and 95% confidence intervals
(Cl) were plotted. Heterogeneity was investigated by visual examination and statistically
using the Q-statistic. Where no evidence of heterogeneity was found a meta-analysis was
conducted to produce a pooled estimate of the measure of effect. Statistically significant
heterogeneity was investigated using meta-regression. Multiple regression analysis was used
to explore the relationship between fluoridation and fluorosis.
Results
214 studies met full inclusion criteria for one or more of the objectives. No randomised
controlled trials of the effects of water fluoridation were found. The study designs used
included 45 'before and after' studies, 102 cross-sectional studies, 47 ecological studies, 13
cohort (prospective or retrospective) studies and 7 case-control studies. Several studies were
reported in multiple papers over a number of years.
Results by Objective
Objective 1
A total of 26 studies of the effect of water fluoridation on dental caries were found. For this
objective, the quality of studies found was moderate (no level A studies). A large number of
studies were excluded because they were cross-sectional studies and therefore did not meet
the inclusion criteria of being evidence level B or above. All but three of the studies included
were before-after studies, two included studies used prospective cohort designs, and one
used a retrospective cohort design. All before-after studies located by the search were
included. The most serious defect of these studies was the lack of appropriate analysis.
Many studies did not present an analysis at all, while others only did simple analyses without
attempting to control for potentially confounding factors. While some of these studies were
conducted in the 1940's and 50's, prior to the common use of such analyses, studies
conducted much later also failed to use methods that were commonplace at the time of the
study.
Another defect of many studies was the lack of any measure of variance for the estimates of
decay presented. While most studies that presented the proportion of caries-free children
contained sufficient data to calculate standard errors, this was not possible for the studies that
presented dmft/DMFT scores. Only four of the eight studies using these data provided
estimates of variance.
The best available evidence suggests that fluoridation of drinking water supplies does reduce
caries prevalence, both as measured by the proportion of children who are caries free and by
April 2009 Page A-19
-------
the mean change in dmft/DMFT score. The studies were of moderate quality (level B), but of
limited quantity. The degree to which caries is reduced, however, is not clear from the data
available. The range of the mean difference in the proportion (%) of caries-free children is -5.0
to 64%, with a median of 14.6% (interquartile range 5.05, 22.1%). The range of mean change
in dmft/DMFT score was from 0.5 to 4.4, median 2.25 teeth (interquartile range 1.28, 3.63
teeth). It is estimated that a median of six people need to receive fluoridated water for one
extra person to be caries-free (interquartile range of study NNTs 4, 9). The best available
evidence from studies following withdrawal of water fluoridation indicates that caries
prevalence increases, approaching the level of the low fluoride group. Again, however, the
studies were of moderate quality (level B), and limited quantity. The estimates of effect could
be biased due to poor adjustment for the effects of potential confounding factors.
Objective 2
To address this objective, studies conducted after 1974 were examined. While only nine
studies were included for Objective 2, these would have been enough to provide a confident
answer to the objective's question if the studies had been of sufficient quality. Since these
studies were completed after 1974, one might expect that the validity assessments would be
higher than the earlier studies following the introduction of more rigorous study methodology
and analytic techniques. However, the average validity checklist score and level of evidence
was essentially the same for studies after 1974 as those conducted prior to 1974. Hence, the
ability to answer this objective is similar to that in Objective 1.
In those studies completed after 1974, a beneficial effect of water fluoridation was still evident
in spite of the assumed exposure to non-water fluoride in the populations studied. The metaregression
conducted for Objective 1 confirmed this finding.
Objective 3
No level A or B studies examining the effect of water fluoridation on the inequalities of dental
health between social classes were identified. However, because of the importance of this
objective, level C studies conducted in England were included. A total of 15 studies
investigating the association of water fluoridation, dental caries and social class in England
were identified. The quality of the evidence of the studies was low, and the measures of
social class that were used varied. Variance data were not reported in most of these studies,
so a statistical analysis was not undertaken.
There appears to be some evidence that water fluoridation reduces the inequalities in dental
health across social classes in 5 and 12 year-olds, using the dmft/DMFT measure. This effect
was not seen in the proportion of caries-free children among 5 year-olds. The data for the
effects in children of other ages did not show an effect. The small quantity of studies,
differences between these studies, and their low quality rating, suggest caution in interpreting
these results.
Objective 4
DENTAL FLUOROSIS
Dental fluorosis was the most widely and frequently studied of all negative effects. The
fluorosis studies were largely cross-sectional designs, with only four before-after designs.
Although 88 studies of fluorosis were included, they were of low quality. The mean validity
score for fluorosis was only 2.8 out of 8. All, but one, of the studies were of evidence level C.
Observer bias may be of particular importance in studies assessing fluorosis. Efforts to
control for the effects of potential confounding factors, or reducing potential observer bias
were uncommon.
As there may be some debate about the significance of a fluorosis score at the lowest level of
each index being used to define a person as 'fluorosed', a second method of determining the
proportion 'fluorosed' was selected. This method describes the number of children having
dental fluorosis that may cause 'aesthetic concern'.
With both methods of identifying the prevalence of fluorosis, a significant dose-response
relationship was identified through a regression analysis. The prevalence of fluorosis at a
water fluoride level of 1.0 ppm was estimated to be 48% (95% Cl 40 to 57) and for fluorosis of
aesthetic concern it was predicted to be 12.5% (95% Cl 7.0 to 21.5). A very rough estimate
of the number of people who would have to be exposed to water fluoride levels of 1.0 ppm for
one additional person to develop fluorosis of any level is 6 (95% Cl 4 to 21), when compared
April 2009 Page A-20
-------
with a theoretical low fluoride level of 0.4 ppm. Of these approximately one quarter will have
fluorosis of aesthetic concern, but the precision of these rough estimates is low. These
estimates only apply to the comparison of 1.0 ppm to 0.4 ppm, and would be different if other
levels were compared.
BONE FRACTURE AND BONE DEVELOPMENT PROBLEMS
There were 29 studies included on the association between bone fracture and bone
development problems and water fluoridation. Other than fluorosis, bone effects (not
including bone cancers) were the most studied potential adverse effect. These studies had a
mean validity score of 3.4 out of 8. All but one study were of evidence level C. These studies
included both cohort and ecological designs, some of which included analyses controlling for
potential confounding factors. Observer bias could potentially play a role in bone fracture
studies, depending on how the study is conducted.
The evidence on bone fracture can be classified into hip fracture and other sites because
there are more studies on hip fracture than any other site. Using a qualitative method of
analysis (Figure 8.1), there is no clear association of hip fracture with water fluoridation. The
evidence on other fractures is similar. Overall, the findings of studies of bone fracture effects
showed small variations around the 'no effect' mark. A meta-regression of bone fracture
studies also found no association with water fluoridation.
CANCER STUDIES
There were 26 studies of the association of water fluoridation and cancer included. Eighteen
of these studies are from the lowest level of evidence (level C) with the highest risk of bias.
There is no clear association between water fluoridation and overall cancer incidence and
mortality. This was also true for osteosarcoma and bone/joint cancers. Only two studies
considered thyroid cancer and neither found a statistically significant association with water
fluoridation.
Overall, no clear association between water fluoridation and incidence or mortality of bone
cancers, thyroid cancer or all cancers was found.
OTHER POSSIBLE NEGATIVE EFFECTS
A total of 33 studies of the association of water fluoridation with other possible negative
effects were included in the review. Interpreting the results of studies of other possible
negative effects is very difficult because of the small numbers of studies that met inclusion
criteria on each specific outcome, and poor study quality. A major weakness of these studies
generally was failure to control for any confounding factors.
Overall, the studies examining other possible negative effects provide insufficient evidence on
any particular outcome to permit confident conclusions. Further research in these areas
needs to be of a much higher quality and should address and use appropriate methods to
control for confounding factors.
Objective 5:
The assessment of natural versus artificial water fluoridation effects is greatly limited due to
the lack of studies making this comparison. Very few studies included both natural and
artificially fluoridated areas, and direct comparisons were not possible for most outcomes. No
major differences were apparent in this review, however, the evidence is not adequate to
make a conclusion regarding this objective.
Conclusions
This review presents a summary of the best available and most reliable evidence on the
safety and efficacy of water fluoridation.
Given the level of interest surrounding the issue of public water fluoridation, it is surprising to
find that little high quality research has been undertaken. As such, this review should provide
both researchers and commissioners of research with an overview of the methodological
limitations of previous research conducted in this area.
The evidence of a benefit of a reduction in caries should be considered together with the
increased prevalence of dental fluorosis. The research evidence is of insufficient quality to
allow confident statements about other potential harms or whether there is an impact on
social inequalities. This evidence on benefits and harms needs to be considered along with
the ethical, environmental, ecological, costs and legal issues that surround any decisions
about water fluoridation. All of these issues fell outside the scope of this review.
April 2009 PageA-21
-------
Any future research into the safety and efficacy of water fluoridation should be carried out with
appropriate methodology to improve the quality of the existing evidence base.
April 2009 Page A-22
-------
Gary M. Whitford, D.M.D., Ph.D.
Regents' Professor, Department of Oral Biology and Maxillofacial Pathology, Medical
College of Georgia
1. Was the document clear and transparent in its presentation of data and explanation of the
analytical approaches used to characterize the concentration- response and dose-response
relationships for severe dental fluorosis? If not, do you have any suggestions that will enable the
authors to improve its clarity?
Answer:
Yes, the document was clear and transparent. No, I have no suggestions to improve its
clarity.
2. Are you aware of any significant publications related to severe dental fluorosis or the skeletal
effects of fluoride that are not included in the noncancer assessment document?
Answer:
The report by Dean and Elvove is relevant to question 3. The reference is:
Dean HT, Elvove E. Studies on the minimal threshold of the dental sign of chronic endemic
fluorosis (Mottled enamel). Pub Hlth Rep 50: 1719-1729, 1935.
3. Are the strengths and weaknesses of Dean (1942) study (selected as critical) fully
characterized?
Answer:
Not entirely. One weakness of the Dean (1942) report was the chemical method used for
the determination of fluoride concentrations in water. The zirconium-alizarin method is rarely
used today because of its relative insensitivity, several interfering substances and lack of
specificity for fluoride. For the SPANDS colorimetric method, interfering substances include Al
(0.1 ppm) and Fe (10 ppm) which reduce absorbance and phosphorus (16 ppm) and sulfate (200
ppm) which increase absorbance. I'm not sure how these ions affect the zirconium-alizarin
colorimetric method but Elvove's earlier papers should provide the information.
An indication of the problem is found in the footnote to Table 3-1 on page 24 where it is
said (quoting Elvove who was the principal chemist) that "as little as 0.01 mg F/50cc, or 0.2 ppm
F, could be differentiated from the control by application of this technique." This appears to
mean that Elvove's method could differentiate between water without fluoride and water
containing 0.2 ppm fluoride. The magnitude of the error at higher concentrations is not known to
me but could be similar. Again, Elvove's earlier papers should provide the information.
One indication of the scatter in analytical results can be seen in the 1933/34 monthly
results for water in Colorado Springs (see Table 4 in Dean and Elvove, Studies on the minimal
threshold of the dental sign of chronic endemic fluorosis (mottled enamel), Public Hlth Rep 50:
1719-1729, 1935). The average of the 12 results was 2.5 ppm but the range was 1.8 to 3.0 ppm
despite the fact that the water came from a single source. While some seasonal variation in water
concentrations can expected, this wide range (1.2 ppm) appears excessive. The 12 monthly
results from Monmouth ranged from 1.6 to 1.9 ppm, those from Galesburg ranged from 1.8 to
2.0 ppm, and those from Pueblo ranged from 0.3 to 0.7 ppm.
4. Are some teeth more susceptible to sever fluorosis than others? If so, should the OW make
modifications to the age range identified as the period of concern?
April 2009 Page A-23
-------
Answer:
I am not aware of data showing that some teeth more susceptible to severe fluorosis than
others. I would guess that the posterior teeth may be more susceptible since their development is
more protracted than that of the anterior teeth.
5. Are the data on cavities as collected and graphically presented by the OW consistent with the
hypothesis that there is an increased risk for cavities in susceptible individuals with severe dental
fluorosis?
Answer:
The available data present an unclear relationship. Driscoll's work (1983, 1986) in
Illinois and Iowa did not indicate a relationship between severe fluorosis and caries (pages 46-
47). Eklund's report (1987) did not find a relationship when the entire dentition was considered.
They found more caries in severely fluorosed anterior teeth and premolars but not in the molars.
In their Chinese study, Chen et al (1989) found no difference in caries scores between the group
without fluorosis and the group with severe fluorosis. Warnakulasuriya et al (1992) reached a
similar conclusion in their Sri Lanka study but the validity of the conclusion was less clear
because of the way they grouped the fluorosis categories. On the other hand, Mann et al (1987,
1990) and Olsson (1979) found that DMFS scores were directly related to the severity of
fluorosis in Israel as did Wondwossen et al (2004) in Ethiopia. Ermis et al (2003) reported a
slightly higher prevalence of caries in moderate-to-severely fluorosed teeth but the relationship
was not significant.
Overall and as summarized in Figure 3-7 on page 71, the relationship between the
severity of dental fluorosis and the risk of caries is suggestive but not convincing. I think this
subject requires more study with control for variables that are known risk factors for caries
before a reasonably firm conclusion can be drawn about an increased risk of caries associated
with severe dental fluorosis.
6. Are there recent data that would impact the IOM (1997) Adequate Intake value of 0.05
mg/kg/day for fluoride that the OW should consider in this assessment?
Answer:
The lOM's Adequate Intake (Al) represents the amount of intake of any substance
"needed to maintain a defined nutritional state or criterion of adequacy in essentially all members
of a specific healthy population." In the case of fluoride, the Al "is based on estimated intakes
that have been shown to reduce the occurrence of dental caries maximally in a population
without causing unwanted side effects including moderate dental fluorosis (page 301, Dietary
Reference Intakes, 1997). This does not mean that intakes somewhat higher than 0.05 mg/kg/day
increase risk of moderate dental fluorosis. In fact, the lOM's estimate for the threshold for that
risk is 0.10 mg/kg/day. I am not aware of data that would call for a change in the Al of 0.05
mg/kg F/day.
7. Can you suggest an approach to transform the water concentration data from the Dean (1942)
study to units of dose for the population susceptible to severe dental fluorosis other than that
used by the OW?
Answer: Not at this time.
April 2009 Page A-24
-------
8. Can you provide input concerning the strengths and weaknesses of the approach utilized by
the OW to identify a lower bound dose for severe dental fluorosis?
Answer:
I am looking at the Clovis, New Mexico, data in Table 3-1 (page 24). The water
fluoride concentration was listed as 2.2 ppm, the lowest concentration at which severe dental
fluorosis was recorded. The prevalence was 0.7 percent. This concentration, therefore, was
selected as the LOAEL. The prevalence of moderate dental fluorosis (Dean score 3) was 11.0
percent. These prevalence values are markedly higher than those for Elmhurst and Galesburg,
Illinois, (about 1.1 percent for moderate and an absence of severe dental fluorosis) where the
water fluoride concentrations were listed as 1.8 and 1.9 ppm, respectively, just slightly lower
than the concentration in Clovis.
In view of the small differences in the water fluoride concentrations between Clovis and
the other two communities, the large differences in fluorosis prevalence values suggest that
another factor influenced the appearance of the teeth in Clovis. Unlike Elmhurst and Galesburg,
Clovis is located at a relatively high altitude (4300 feet). As summarized elsewhere in the
document (pages 39-41), there is evidence from laboratory animal studies and epidemiological
studies that residence at high altitude affects amelogenesis in a way that resembles fluorosis and
that its effects may be additive to the effects of fluoride exposure. This adds uncertainty
regarding the selection of 2.2 ppm fluoride (Clovis) as the LOAEL for severe dental fluorosis.
A similar (but weaker) argument can be made for Colorado Springs where the average
water fluoride concentration is listed as 2.6 ppm (but with a wide range, see item 3 above). This
city is located at an altitude of 6,035 feet.
In addition to these comments and based on the variability among published studies
regarding the relationship between estimated fluoride intakes and the risk of severe dental
fluorosis, I think consideration should be given to establishing a LBD-5% rather than a LBD-1%.
9. Are you aware of dose estimates other than those from IOM (1997) and WHO (2002) that are
appropriate critical doses for skeletal effects and/or can you suggest a different approach that the
OW might use to estimate the fluoride dose associated with skeletal fractures using available
data?
Answer:
The IOM (page 307) estimate of fluoride exposures that may result in clinical signs of the
milder forms skeletal fluorosis (presumably preclinical and stage I) is 10 mg/day for 10 or more
years. There are published exceptions as noted on the same page of the IOM report. Recent case
reports of "tea fluorosis" in the U.S. suggest that, at least for some individuals, a much higher
chronic intake is tolerated without progression to stage II skeletal fluorosis (Whyte et al, Am J
Med 118: 78-82, 2005; Whyte et al, J Bone Min Res, in press). In the former report the intake
was estimated at 37-74 mg F/day from tea throughout the patient's adult life. The intake in the
latter report was estimated at more than 40 mg F/day throughout the patient's adult life. Both
patients showed marked osteosclerosis without ligamentous calcifications consistent with stage I
skeletal fluorosis. Hallanger-Johnson et al (Mayo Clin Proc 82: 719-724, 2007) reported four
cases with axial osteosclerosis with elevated serum fluoride levels due to chronic consumption of
large amounts of tea.
In addition to the several variables that can affect the quality and quantity of the skeleton
cited in the present document, it is of interest that much of the data relating bone fluoride
concentrations to the stages of skeletal fluorosis comes from studies of workers in aluminum
April 2009 Page A-25
-------
processing factories. High, chronic exposures to aluminum lead to skeletal changes that share
some features in common with skeletal fluorosis which makes it difficult to attribute the skeletal
changes only to fluoride. I think this subject is worthy of further exploration.
10. Are you aware of any data that can be used to demonstrate that protection of the secondary
teeth from severe dental fluorosis will also protect the primary teeth?
Answer:
There are a few reports that suggest that dental fluorosis in the primary teeth may
correlate with the condition in secondary teeth. I am not aware of reports suggesting that
protection of the secondary teeth from severe dental fluorosis will also protect the primary teeth.
11. Are there factors other than those discussed in Section 3.1.4 that increase sensitivity to
fluoroisis of the teeth or bones?
Answer:
As I recall, there are suggestions that African-Americans are more susceptible to dental
fluorosis but less susceptible to skeletal fluorosis. Also, published data indicate that there are
differences among strains of mice regarding susceptibility to dental fluorosis. It is assumed that
the differences are due to genetics. NIDCR has requested applications for further
pharmacogenetic studies.
12. Do you support the OW's conclusion that an RfD of 0.07 mg/kg/day will be protective for
severe dental fluorosis in children and skeletal effects in adults while still providing for the
beneficial effects of fluoride?
Answer:
Yes, 0.07 mg/kg/day will be protective. For reasons listed above, however, I think there
is a considerable degree of uncertainty regarding the accuracy of the estimate. It may be closer to
0.10 mg/kg/day for moderate/severe dental fluorosis and substantially higher than that for
clinically significant skeletal fluorosis in the United States.
April 2009 Page A-26
-------
Appendix B: Post-Meeting Responses to the Charge Questions from
Reviewers (Organized Alphabetically by Reviewer)
PageB-l
-------
Jane A. Cauley, Dr.P.H
Professor, Department of Epidemiology, University of Pittsburgh
1. Was the document clear and transparent in its presentation of data and explanation of the
analytical approaches used to characterize the concentration- response and dose-response
relationships for severe dental fluorosis? If not, do you have any suggestions that will enable the
authors to improve its clarity?
Response
The document could be strengthened by reorganizing it as a manuscript. The
Introduction should provide background and a scientific rationale for this analysis. The
specific objectives should be listed.. The methods for identifying all the literature should
be transparent and consistently applied to both the fluorosis and fracture sections. This
is typical of meta analyses which also apply certain quality scores to each paper.
I think it might also be helpful if I understand who the audience for this report isd . I
found it a confusing because I do not understand the roles of the EPA, OW and the NRC :
how will each agency find this information helpful.? What prompted this report?
2. Are you aware of any significant publications related to severe dental fluorosis or the skeletal
effects of fluoride that are not included in the noncancer assessment document?
Response
The dental fluorosis literature included studies with a wide range of fluoride
levels while the skeletal/ fracture studies were limited to those with fluoride levels >
4mg/l. You might consider adding some of the literature on fractures at lower levels of
fluoride including the following:
Cauley JA, Murphy PA, Riley TJ, Buhari AM. Effects of fluoridated drinking water on
bone mass and fractures: the study of osteoporotic fractures. J Bone Miner Res. 1995
Jul; 10(7): 1076-86.
Phipps, KR, Orwoll, ES, Mason, JD, Cauley, J A. Community water fluoridation, bone
mineral density, and fractures: prospective study of effects in older women. BMJ 2000
321(7265):860-4.
In addition, Phipps KR et al published a study comparing bone mineral density in
Deming, NM (0.7mg/L) and Lordsburg, NM (3.5mg/L). The citation is: Phipps KR,
Burt BA. Water-borne fluoride and cortical bone mass: a comparison of two
communities. J Dent Res. 1990 Jun;69(6): 1256-60.
It appeared to me that this section is overall less thorough than the dental flourosis
section. Additional Selections include the following. I would pay particular attention to
the papers by Kathy Phipps, Cyrus Cooper, S Jacobsen and T Hillier. These are high
quality in my opinion.
April 2009 Page B-2
-------
Demos LL, Kazda H, Cicuttini FM, Sinclair MI, Fairley CK.
Water fluoridation, osteoporosis, fractures--recent developments.
Aust Dent J. 2001 Jun;46(2):80-7; quiz 143.
PMID: 11491235 [PubMed - indexed for MEDLINE]
Fabiani L, Leoni V, Vital! M.
Bone-fracture incidence rate in two Italian regions with different fluoride
concentration levels in drinking water.
J Trace Elem Med Biol. 1999 Dec;13 ( 4) :232-7 .
PMID: 10707346 [PubMed - indexed for MEDLINE]
Kleerekoper M.
Fluoride and the skeleton.
Crit Rev Clin Lab Sci. 1996 Apr;33(2):139-61. Review.
PMID: 8744520 [PubMed - indexed for MEDLINE]
Raheb J.
Water fluoridation, bone density and hip fractures: a review of recent
literature.
Community Dent Oral Epidemiol. 1995 Oct;23(5):309-16. Review. No abstract
available.
PMID: 8529346 [PubMed - indexed for MEDLINE]
S0gaard CH, Mosekilde L, Schwartz W, Leidig G, Minne HW, Ziegler R.
Effects of fluoride on rat vertebral body biomechanical competence and bone
mass .
Bone. 1995 Jan;16(1):163-9.
PMID: 7742076 [PubMed - indexed for MEDLINE]
Kroger H, Alhava E, Honkanen R, Tuppurainen M, Saarikoski S.
The effect of fluoridated drinking water on axial bone mineral density—a
population-based study.
Bone Miner. 1994 Oct;27(1):33-41.
PMID: 7849544 [PubMed - indexed for MEDLINE]
Ripa LW.
A half-century of community water fluoridation in the United States: review
and
commentary.
J Public Health Dent. 1993 Winter;53(1):17-44 . Review.
PMID: 8474047 [PubMed - indexed for MEDLINE]
Gordon SL, Corbin SB.
Summary of workshop on drinking water fluoride influence on hip fracture on
bone
health. (National Institutes of Health, 10 April, 1991)
Osteoporos Int. 1992 May;2(3) :109-17 . No abstract available.
PMID: 1627897 [PubMed - indexed for MEDLINE]
Colquhoun J.
Water fluoride and fractures.
N Z Med J. 1991 Aug 14;104(917):343. No abstract available.
PMID: 1876343 [PubMed - indexed for MEDLINE]
McNeill KG, Coote GE, Hitchman AJ.
Uptake of fluorine in cortical and trabecular bone.
April 2009 Page B-3
-------
J Bone Miner Res. 1991 Aug;6(8):859-64.
PMID: 1785375 [PubMed - indexed for MEDLINE]
Arnala I, Alhava EM, Kivivuori R, Kauranen P.
Hip fracture incidence not affected by fluoridation. Osteofluorosis studied
in
Finland.
Acta Orthop Scand. 1986 Aug;57(4):344-8.
PMID: 3788501 [PubMed - indexed for MEDLINE]
Ericsson Y, Luoma H, Ekberg 0.
Effects of calcium, fluoride and magnesium supplementations on tissue
mineralization in calcium- and magnesium-deficient rats.
J Nutr. 1986 Jun;116(6):1018-27.
PMID: 3723198 [PubMed - indexed for MEDLINE]
Simonen 0, Laitinen 0.
Does fluoridation of drinking-water prevent bone fragility and osteoporosis?
Lancet. 1985 Aug 24;2 ( 8452) :432-4.
PMID: 2863455 [PubMed - indexed for MEDLINE]
Madans J, Kleinman JC, Cornoni-Huntley J.
The relationship between hip fracture and water fluoridation: an analysis of
national data.
Am J Public Health. 1983 Mar;73(3):296-8.
PMID: 6824115 [PubMed - indexed for MEDLINE]
Schamschula RG, Barmes DE.
Fluoride and health: dental caries, osteoporosis, and cardiovascular
disease.
Annu Rev Nutr. 1981;1:427-35. Review.
PMID: 6764723 [PubMed - indexed for MEDLINE]
Alhava EM, Olkkonen H, Kauranen P, Kari T.
The effect of drinking water fluoridation on the fluoride content, strength
and
mineral density of human bone.
Acta Orthop Scand. 1980 Jun;51(3):413-20.
PMID: 7446020 [PubMed - indexed for MEDLINE]
Stein ID, Granik G.
Human vertebral bone: relation of strength, porosity, and mineralization to
fluoride content.
Calcif Tissue Int. 1980;32 ( 3):189-94 .
PMID: 6775787 [PubMed - indexed for MEDLINE]
Hegsted DM.
Osteoporosis and fluoride deficiency.
Postgrad Med. 1967 Jan;41(1):A49-53. No abstract available.
PMID: 6036214 [PubMed - indexed for MEDLINE]
Bernstein DS, Sadowsky N, Hegsted DM, Guri CD, Stare FJ.
Prevalence of osteoporosis in high- and low-fluoride areas in North Dakota.
JAMA. 1966 Oct 31;198(5):499-504. No abstract available.
PMID: 5953273 [PubMed - indexed for MEDLINE]
April 2009 Page B-4
-------
Hillier S, Cooper C, Kellingray S, Russell G, Hughes H, Coggon D.
Fluoride in drinking water and risk of hip fracture in the UK: a case-
control
study.
Lancet. 2000 Jan 22;355(9200):265-9.
PMID: 10675073 [PubMed - indexed for MEDLINE]
Hillier S, Inskip H, Coggon D, Cooper C.
Water fluoridation and osteoporotic fracture.
Community Dent Health. 1996 Sep;13 Suppl 2:63-8. Review.
PMID: 8897754 [PubMed - indexed for MEDLINE]
Jacobsen SJ, Goldberg J, Cooper C, Lockwood SA.
The association between water fluoridation and hip fracture among white
women and
men aged 65 years and older. A national ecologic study.
Ann Epidemiol. 1992 Sep;2(5):617-26.
PMID: 1342313 [PubMed - indexed for MEDLINE]
Cooper C, Wickham C, Lacey RF, Barker DJ.
Water fluoride concentration and fracture of the proximal femur.
J Epidemiol Community Health. 1990 Mar;44(1):17-9.
PMID: 2348142 [PubMed - indexed for MEDLINE]
1: Phipps KR, Orwoll ES, Mason JD, Cauley JA.
Community water fluoridation, bone mineral density, and fractures:
prospective
study of effects in older women.
BMJ. 2000 Oct 7;321(7265):860-4.
PMID: 11021862 [PubMed - indexed for MEDLINE]
Phipps KR, Orwoll ES, Bevan L.
The association between water-borne fluoride and bone mineral density in
older
adults.
J Dent Res. 1998 Sep;77 ( 9):1739-48 .
PMID: 9759671 [PubMed - indexed for MEDLINE]
Phipps K.
Fluoride and bone health.
J Public Health Dent. 1995 Winter;55(1):53-6. Review.
PMID: 7776293 [PubMed - indexed for MEDLINE]
Phipps KR, Burt BA.
Water-borne fluoride and cortical bone mass: a comparison of two
communities.
J Dent Res. 1990 Jun;69(6):1256-60.
PMID: 2355118 [PubMed - indexed for MEDLINE]
3. Are the strengths and weaknesses of Dean (1942) study (selected as critical) fully
characterized?
Response:
Additional strengths include the wide range of fluoride concentrations, although there
were fewer subjects in the high fluoride categories. The dose response in the data showing
April 2009 Page B-5
-------
increasing risk of severe skeletal fluorosis with increasing fluoride is also a strength. Finally the
consistency of the findings across several different communities.
Other weaknesses: No information is provided on race/ethnicity. The data were
collected the late 1930-40s. Although confounding by use of other fluoride products would be
minimal, there are many other cohort differences between children exposed to fluoride in the late
1930-40s compared to now. For example, dental hygiene, dietary intakes (less water... more
carbonated beverages), body weight are very different in today's children compared to children
in the 30s.. Do puberty/ hormonal changes influence fluoride effects? This may be important
because age of menarche has been decreasing. Hormones have been shown to influence dental
health. Another weakness of the studies reviewed is that there appears to be no information on
duration of exposure. How long did these children live in each community? Did the inclusion
criteria include a minimum time of residence? If so this point needs to be made in the document.
4. Are some teeth more susceptible to sever fluorosis than others? If so, should the OW make
modifications to the age range identified as the period of concern?
Response
This should be highlighted in a separate section. A more direct discussion of the
age at risk is needed. A table summarizing the ages of the children in each study and the
age range which appeared to be at highest risk would be helpful. You could add this to
Table 3-16.
5. Are the data on cavities as collected and graphically presented by the OW consistent with the
hypothesis that there is an increased risk for cavities in susceptible individuals with severe dental
fluorosis?
Response
The data summarized in Table 3-40 to 3-44 are convincing that dental caries are
more common in subjects with severe fluorosis. The data suggest a U shaped
relationship with fluoride: a higher risk in those with very low intakes and intakes
>2-3mg/L.
6. Are there recent data that would impact the IOM (1997) Adequate Intake value of 0.05
mg/kg/day for fluoride that the OW should consider in this assessment?
Response
I do not know of any more recent data. Perhaps you want to mention life stage
and Upper limit of toxicity. The ADA website looks like they updated their
Fluoride document in 2006.
TABLE S-5
Criteria and Dietary Reference Intake Values for Fluoride by Life Stage Group
April 2009 Page B-6
-------
Al (mg/day)
MatelPemale
[0 through 6
[months
[7 through 12
[months
! 1 through 3
iyears
|4 through 8
iyears
[9 through 13
Iyears
[14 through 18
iyears
1 19 through 30
iyears
131 through 50
[years
[51 through 70
Iyears
> [70 years
Pregnancy
<= [ 1 8 years
i 19 through 50
iyears
Lactation
<= 1 1 8 years
i 19 through 50
iyears
(Human milk « 0.01
bontent [
jCaries \ 0.5
^prevention 1
SCaries 1 0.7
^prevention j
jCaries j 1
prevention \
jCaries j 2
^prevention I
[Caries [ 3
jprevention j
[Caries I 4
Jprevention |
(Caries ! 4
prevention
jCaries « 4
^prevention I
jCaries | 4
jprevention 1
SCaries I
^prevention |
JCaries |
^prevention 1
JCaries >
(prevention |
jCaries j
^prevention |
0.01
0.5
0.7
r
0
3
3
3
3'
3
3'
'-y
3
'-y
"AI = Adequate Intake. For healthy infants fed human milk, AI is the mean intake. The observed estimate
of nutrient intake that reduces the incidence of dental caries maximally in a group of healthy people. The
AI is used if the scientific evidence is not available to derive an EAR. The AI is believed to cover their
needs, but lack of data or uncertainty in the data prevent being able to specify with confidence the
percentage of individuals covered by this intake.
April 2009
Page B-7
-------
TABLE S-6
Tolerable Upper Intake Levels (ULa), by Life Stage Group
Life Stage Grouo
through 6 months
17 through 12 months
PhOi^
(g/dayj
11 through 3 years
54 through 8 years
§9 through 18 years
j!9 through 70 years
> 170 years
Pregnancy
<= i 18 years
19 through 50 years
Lactation
<= 118 years
119 through 50 years
d\
ND
ND
2.5
2.5
25
25
25
2.5
2.5
2.5
25
ND
ND
3
3
4
3.5
3.5
NDS
NCH
65 t
25 |
25!
50 I
110J
___
50!
50!
3505
50!
350 i
350 |
350 |
350 j
"350T
50|
50!
50 |
50!
50!
0.7\
0.9\
1.3 |
2.21
10 [
o
10]
10!
10J
To]
"UL = the maximum level of daily nutrient intake that is likely to pose no risk of adverse effects to members
of the healthy general population. Unless specified otherwise, the UL represents total intake from food,
water, and supplements.
6The UL for magnesium represents intake from pharmacological agents only and does not include intake
from food and water.
cAs cholecalciferol. 1 ug cholecalciferol = 40 IU vitamin D.
''ND. Not determinable due to lack of data of adverse effects in this age group and concern with regard to
lack of ability to handle excess amounts. Source of intake should be from food only in order to prevent high
levels of intake.
April 2009
Page B-8
-------
7. Can you suggest an approach to transform the water concentration data from the Dean (1942)
study to units of dose for the population susceptible to severe dental fluorosis other than that
used by the OW?
Response
To estimate the RFD of 0.07 mg/Kg/day, the OW attempted to obtain detailed
analysis of average body weights and water intakes during the time the Dean data was collected.
Data from the Erchow and Cantor were used (1989). Why not use more generalizable data from
NHANES at least for body weight? You could still use the water intake estimates from the
Cantor paper.
8. Can you provide input concerning the strengths and weaknesses of the approach utilized by
the OW to identify a lower bound dose for severe dental fluorosis?
Response
Strengths: Conservative estimates carried out with and without the outlier
community; several different models were considered, sensitivity analysis showed
that even eliminating three highest data points had little effect on the model
goodness of fit or on the BMDL.
Weakness: Inherent weakness of the approach relate primarily to the weaknesses
cited related to the parent Dean Study. You might consider adding a formal test
for trend in data.
9. Are you aware of dose estimates other than those from IOM (1997) and WHO (2002) that are
appropriate critical doses for skeletal effects and/or can you suggest a different approach that the
OW might use to estimate the fluoride dose associated with skeletal fractures using available
data?
Response
See answer to #2 for additional references.
Some concern about the inconsistencies across the age groups. The significance of the
differences in results across age could have occurred by chance alone.
Referring to the" radius" is problematic because if it s the distal radius, it is primarily
trabecular BMD and mid radius is predominantly cortical BMD. If the effects of
fluoride differs by type of bone, this is very problematic.
Randomized trials: please note the compliance/adherence to the Na F because it is poorly
tolerated. This is a limitation to theses studies. Did they adhere to the intent to treat
principle?
April 2009 Page B-9
-------
10. Are you aware of any data that can be used to demonstrate that protection of the secondary
teeth from severe dental fluorosis will also protect the primary teeth?
Response
No, I am not aware of any additional data.
11. Are there factors other than those discussed in Section 3.1.4 that increase sensitivity to
fluoroisis of the teeth or bones?
Response
I am not sure but I question whether race/ ethnicity and pubertal stage is
important.
13. Do you support the OW's conclusion that an RfD of 0.07 mg/kg/day will be protective for
severe dental fluorosis in children and skeletal effects in adults while still providing for the
beneficial effects of fluoride?
Response
Yes, I am in support of the 0.07 mg/kg/day but I would admit the uncertainty that surrounds this
estimate..
April 2009 Page B-10
-------
Pamela Denbesten, D.D.S., M.S.
Professor and Chair, Division of Pediatric Dentistry, Department of Orofacial Sciences,
University of California at San Francisco
1. Was the document clear and transparent in its presentation of data and explanation of the
analytical approaches used to characterize the concentration- response and dose-response
relationships for severe dental fluorosis? If not, do you have any suggestions that will enable the
authors to improve its clarity?
Response:
The document was for the most part clear and transparent in its presentation. To improve clarity,
I suggest the following.
Page 13. State the purpose of the Secondary Maximum Contaminant Level
Page 15. The statement with regards to fluoroapatite, at the end of paragraph 2.1.1 is not
accurate. The mineral formed in tooth enamel exposed to higher fluoride levels is fluoride
containing carbonated apatite. Fluoride levels in subsurface fluorotic enamel are about 200
ppm rather than the 10-100 ppm fluoride in normal enamel, whereas fluorapatite is about
30,000 ppm fluoride. Precipitation of fluoride mineral salts at the surface of enamel results
in high surface level, though this also is not fluoroapatite. This fluoride-substituted apatite
has some increased resistance to bacterial acids that cause tooth decay. However, the
primary function of fluoride in drinking water in reducing tooth decay is topical, primarily
by the enhancement of remineralization.
Page 19 section 2.2. The apatite formed in bone is also a fluoride-substituted hydroxyapatite
rather than a fluoroapatite.
Page 21 section 3. Change ".. .are preferred for evaluating the potential effects of fluoride in
drinking water", to the more accurate statement, ".. .are preferred for evaluating the
potential effects of ingested fluoride".
Page 57 paragraph 1. It should be made clear that studies on water fluoridation conducted after
1980, are confounded by additional sources of fluoride, and changes in use of tap water.
For example, decreasing fluorosis in more recent studies may be related to reduced
consumption of tap water as use of bottled water increases. In general for all of section
3.2.2, when fluoride effects on dental caries are discussed, the data should be divided into
studies before and after 1980 when fluoride became widely available in toothpaste, and
perhaps before and after the late 1990s when bottled beverages became widely used.
Page 86; Table 3-52. Please indicate what does "complete" and "total" refer to?
Page 94 section 4.4. Explain with the "NOAEL/LOAEL" approach is, or at least spell out the
acronym.
Page 97 paragraph 1. The importance of fluoride as a nutrient may need to be reassessed, given
that its primary function in caries prevention is topical. It would seem more appropriate
April 2009 Page B-11
-------
focus on the upper limits for ingestion of this caries preventive agent, and leave it to
future panels to assess the relative importance of the lOMs recommended intake of
fluoride and risk of severe fluorosis.
Page 98, paragraph 2, sentence 3. The statement as to the timing secondary incisor tooth
formation is incorrect. The secondary incisors and molars begin development in utero.
Change "development" to "mineralization".
Page 101. What is the rationale for setting the BMDL at an incidence of 1% severe fluorosis? I
recommend setting the BMDL at an incidence of 99% moderate fluorosis, which would
show an intent to eliminate the adverse effect of severe fluorosis secondary to fluoride
added to drinking water.
Page 103. As stated above, I question a recommended fluoride intake, and feel that this
document should focus only on the dose response analysis.
Page 104, paragraph 1. The statement that"... fluoroapatite crystals disrupt the hydroxyapatite
crysal lattice..." is incorrect and should be deleted.
2. Are you aware of any significant publications related to severe dental fluorosis or the skeletal
effects of fluoride that are not included in the noncancer assessment document?
Response:
No I am not
3. Are the strengths and weaknesses of Dean (1942) study (selected as critical) fully
characterized?
Response:
Yes, to me they are.
4. Are some teeth more susceptible to severe fluorosis than others? If so, should the OW make
modifications to the age range identified as the period of concern?
Response:
The OW report was clear in showing that posterior teeth are also susceptible, and the case for
including children up to age 14 was clear and compelling.
5. Are the data on cavities as collected and graphically presented by the OW consistent with the
hypothesis that there is an increased risk for cavities in susceptible individuals with severe dental
fluorosis?
Response:
Yes
6. Are there recent data that would impact the IOM (1997) Adequate Intake value of 0.05
mg/kg/day for fluoride that the OW should consider in this assessment?
April 2009 Page B-12
-------
Response.
The weight of evidence is that the primary mechanism by which fluoride protects against tooth
decay is a topical effect. Therefore, the recommendation by the IOM that there is an adequate
intake value, at least relating to tooth decay should be reassessed.
7. Can you suggest an approach to transform the water concentration data from the Dean (1942)
study to units of dose for the population susceptible to severe dental fluorosis other than that
used by the OW?
Response:
No I can not. Of course serum fluoride levels would be the most useful measurement, but these
levels are not available. I believe the using Dean's data as a starting point to quantify total
ingestion is the best that we can do. I suggest a recommendation that future studies include
random sampling of serum fluoride levels to strengthen future decision-making relative to
fluoride intake.
8. Can you provide input concerning the strengths and weaknesses of the approach utilized by
the OW to identify a lower bound dose for severe dental fluorosis?
Response:
Strengths.
1) The data was used limited to Dean's studies where fluoride was only in drinking water.
2) A careful assessment was done of the tap water consumed and mean body weights
Weaknesses
1) The assumption that 0.05 mg/kg/day is a required amount of fluoride and that dose estimates
must be above this level. The purpose of the EPA's analysis was to determine risk, not risk
benefit analyses.
2) The assumption that the small number of children who displayed severe fluorosis were those
who had excess exposure to fluoride. Given the fact that water was the only source of fluoride,
this would assume that these children drank significantly more water. It is more likely that
genetic or other causes are responsible for this small outlier group.
9. Are you aware of dose estimates other than those from IOM (1997) and WHO (2002) that are
appropriate critical doses for skeletal effects and/or can you suggest a different approach that the
OW might use to estimate the fluoride dose associated with skeletal fractures using available
data?
Response:
No I can not.
10. Are you aware of any data that can be used to demonstrate that protection of the secondary
teeth from severe dental fluorosis will also protect the primary teeth?
April 2009 Page B-13
-------
Response:
Fluorosis in both primary and permanent teeth is caused by ingested fluoride, and there is no data
to suggest that primary teeth are more susceptible to fluorosis than permanent teeth. Therefore,
measure that would protect permanent teeth would require limiting ingestion of fluoride. These
same measures would protect primary teeth.
11. Are there factors other than those discussed in Section 3.1.4 that increase sensitivity to
fluoroisis of the teeth or bones?
Response:
There seem to be individuals who are uniquely sensitive to the effects of fluoride on enamel
formation. These individuals may have other more "hidden" enamel defects that are exacerbated
by the effects of fluoride. Little is currently know as to why some individuals seem to be more
fluoride sensitive.
12. Do you support the OW's conclusion that an RfD of 0.07 mg/kg/day will be protective for
severe dental fluorosis in children and skeletal effects in adults while still providing for the
beneficial effects of fluoride?
Response:
I do not agree with this recommendation as it is based on limiting severe fluorosis to 1% of the
population, and the lOM's recommended adequate intake level. I suggest that the level be
lowered to eliminate severe fluorosis. One percent of the population represents a relatively large
number of individuals. These are the individuals who are most likely uniquely sensitive to
fluoride. The data should analyzed without taking the IOM recommended adequate levels into
account. A secondary analysis could then be done to include the IOM recommendations.
April 2009 Page B-14
-------
Richard D. Jackson, D.M.D.
Assistant Professor, Preventive and Community Dentistry, School of Dentistry, Indiana
University, Oral Health Research Institute
1. Was the document clear and transparent in its presentation of data and explanation of the
analytical approaches used to characterize the concentration- response and dose-response
relationships for severe dental fluorosis? If not, do you have any suggestions that will enable the
authors to improve its clarity?
I felt that the document clearly described what is available in the literature and presented the
information in an understandable format. I am glad individual tables summarizing some of the
studies were included for reference. As an aside, I would have preferred that the studies be
described chronologically as opposed to alphabetically by author.
2. Are you aware of any significant publications related to severe dental fluorosis or the skeletal
effects of fluoride that are not included in the noncancer assessment document?
Jackson R, Kelly S, Katz B, Brizendine E, Stookey G. Dental fluorosis in children residing in
communities with different fluoride levels in the water: 33 month follow-up. Pediatric Dentistry
21:248-254.
Examination of 357 school-age children in 1994 from three Indiana communities with varying
levels of fluoride in the water (0.2, 1.0 and 4.0 ppm). Severe fluorosis using the TSIF (score 5, 6
or 7) was noted in 9% of the children examined in the 4.0 ppm community. No scores of this
magnitude were seen in the other two communities. Based on comparison to previous prevalence
data in the same communities in 1992, the prevalence of fluorosis had increased in each
community but primarily in the categories rated as a 1, 2 or 3 using the TSIF.
I believe I included this paper in the packet that I sent to Lisa shortly after the meeting was
conducted.
3. Are the strengths and weaknesses of Dean (1942) study (selected as critical) fully
characterized?
I think the strengths of the investigation conducted by Dean (1942) are fully characterized. I
doubt that anyone would disagree that Dean (1942) has been and will continue to be a
benchmark study in the dental literature as well as the much broader literature related to public
health and epidemiology. It is true that Dean's study was performed when confounding fluoride
sources were not available and thus in all probability gives a very clear picture of the prevalence
of the relationship of fluoride ingestion and the subsequent development of dental fluorosis.
However, one point of weakness that could be argued is that the population of subjects examined
was not diverse in terms of racial or cultural characteristics. As it has been postulated that
genetic factors may impact on the expression of dental fluorosis at identical levels of ingestion,
the fact that the data was collected in what may have been an exclusively white population
appears to limit its applicability for use as a benchmark. I was unable to find any information
concerning the racial/ethnic composition of the children involved in this investigation.
April 2009 Page B-15
-------
4. Are some teeth more susceptible to severe fluorosis than others? If so, should the OW make
modifications to the age range identified as the period of concern?
I was a little unclear as to the intent of the question. My initial reaction was to answer "No". I
have not come across any literature that indicates that some teeth are more susceptible to the
development of severe dental fluorosis assuming that the exposure is constant and taking into
consideration how long the developing tooth is exposed to higher levels of fluoride. For example,
maxillary third molars may present with greater evidence of severe fluorosis (pitting, staining,
etc) based on the fact that they take longer to develop (Cc 12-16 years-of-age) and erupt as
opposed to a maxillary central incisor (Cc 5 years-of-age).
5. Are the data on cavities as collected and graphically presented by the OW consistent with the
hypothesis that there is an increased risk for cavities in susceptible individuals with severe dental
fluorosis?
It appears that the data are consistent with and support this hypothesis. A point of discussion may
be made concerning examiner bias in the methodology of these studies. It is obvious that it
would be impossible to eliminate bias as the most severely affected teeth would be evident by the
caries examiner when determining the DMFT/DMFS of the subject. I don't see any solution for
this dilemma with the use of a single examiner or multiple examiners but it is something that
should at least be mentioned as a possible confounder.
Jackson et al (1995) also examined children residing in communities with varying levels of
fluoride in their public water supplies. No child in the negligibly-fluoridated community or the
optimally-fluoridated community was found to have TSIF scores greater than 2. However, the
DMFT and DMFS in the 7-10 year-olds residing in the 4X optimum community were not
significantly different from those residing in the 1.0 ppm community. For 11-14 year-olds DMFS
was significantly lower than in the negligibly fluoridated community.
6. Are there recent data that would impact the IOM (1997) Adequate Intake value of 0.05
mg/kg/day for fluoride that the OW should consider in this assessment?
I am not aware of any new data that could influence the value of 0.5 mg/kg/day as the AI value.
7. Can you suggest an approach to transform the water concentration data from the Dean (1942)
study to units of dose for the population susceptible to severe dental fluorosis other than that
used by the OW?
No, using the data available the approach seems to be logical. The problem, as I see it. is the US
marketplace is constantly changing and ingestion of tap water continue to decline. The EPA
(2004) report stated that bottled water accounted for only 13% of water consumption in the
United States. More recent trade manufacturing data indicates that bottled water consumption in
the United States exceeds this percentage by a wide margin and bottled water consumption may
surpass tapwater consumption in the near future. Other published data suggest that among
Hispanic individuals, tap water is commonly perceived as "unhealthy" and again bottled water is
April 2009 Page B-16
-------
consumed almost exclusively. Another point that should be further explored is the possible "halo
effect" of imported foods and beverages into the United States and the fluoride content of these
consumables.
8. Can you provide input concerning the strengths and weaknesses of the approach utilized by
the OW to identify a lower bound dose for severe dental fluorosis?
I believe that the ideas brought up during the meeting may clarify things somewhat. I really
didn't have any additional comments that I thought would benefit how to improve on the
approach taken by the OW.
9. Are you aware of dose estimates other than those from IOM (1997) and WHO (2002) that are
appropriate critical doses for skeletal effects and/or can you suggest a different approach that the
OW might use to estimate the fluoride dose associated with skeletal fractures using available
data?
I am not aware of additional dose estimates that would be appropriate for determining critical
dosage levels and their possible skeletal effects. From my review of the literature and the
discussion the day of the meeting I cant think of an alternative to the approach taken to estimate
fluoride intake associated with skeletal fractures.
IOM.
Recommended total dietary fluoride intake
Age Reference weight* Adequate intakef Tolerable upper intake§
kg Ib mg/day mg/day
0-6 months 7 16 0.01 0.7
6-12 months 9 20 0.5 0.9
1-3 years 13 29 0.7 1.3
4-8 years 22 48 1.1 2.2
>9 years 40/76 88/166 2.0/3.8 10.0
* Values based on data collected during 1988-1994 as part of the third National Health and
Nutrition Examination Survey.
f Intake that maximally reduces occurrence of dental caries without causing unwanted side
effects, including moderate enamel fluorosis.
§ Highest level of nutrient intake that is likely to pose no risks for adverse health effects in
almost all persons.
Source: Adapted from Institute of Medicine. Fluoride. In: Dietary reference intakes for
calcium,phosphorus, magnesium, vitamin D, and fluoride. Washington, DC: National Academy
Press,1997:288-313.
BONE FRACTURE AND BONE DEVELOPMENT PROBLEMS
There were 29 studies included on the association between bone fracture and bone
development problems and water fluoridation. Other than fluorosis, bone effects (not
including bone cancers) were the most studied potential adverse effect. These studies had a
mean validity score of 3.4 out of 8. All but one study were of evidence level C. These studies
included both cohort and ecological designs, some of which included analyses controlling for
potential confounding factors. Observer bias could potentially play a role in bone fracture
studies, depending on how the study is conducted.
The evidence on bone fracture can be classified into hip fracture and other sites because
there are more studies on hip fracture than any other site. Using a qualitative method of
April 2009 Page B-17
-------
analysis (Figure 8.1), there is no clear association of hip fracture with water fluoridation.
Theevidence on other fractures is similar. Overall, the findings of studies of bone fracture effects
showed small variations around the 'no effect' mark. A meta-regression of bone fracture studies
also found no association with water fluoridation.
10. Are you aware of any data that can be used to demonstrate that protection of the secondary
teeth from severe dental fluorosis will also protect the primary teeth?
I am not aware of any data that supports this possibility.
11. Are there factors other than those discussed in Section 3.1.4 that increase sensitivity to
fluorosis of the teeth or bones?
Section 3.1.4. states that climate and altitude of place of residence; dietary habits, nutritional
status, physiological state, and certain pathological conditions may affect the occurrence and
severity of fluoride-induced dental fluorosis. The section further mentions that certain genetic
conditions such as amelogenesis imperfecta may lead to defective development of enamel.
There have been studies (Vieira et al, 2005; Yan et al, 2007) that suggest, in mice, that there
may be a varying genetic response to identical levels of fluoride ingestion. These changes have
been identified in both tooth and bone formation. Vieira et al found that genetic influences have
a direct bearing on the biomechanical properties of the teeth. Yan et al. found strain-specific
effects of physiological level of fluoride with increased osetoclastogenesis in some mouse
strains. While I was unable to find comparable human trials, this area should be further explored
as the technological means become available.
A study by Hong et al (2004) appears to indicate that the use of amoxicillin could play a
contributing role in the development of primary tooth fluorosis, especially for children exposed
to lower levels of fluoride. I believe I sent a copy of this study to Lisa in the packet following
the meeting.
12. Do you support the OW's conclusion that an RfD of 0.07 mg/kg/day will be protective for
severe dental fluorosis in children and skeletal effects in adults while still providing for the
beneficial effects of fluoride?
Based on the data that was presented and the rationales provided I think the OW does make a
very convincing argument that 0.07mg/kg/day will be protective against the more severe forms
of fluorosis in children and the skeletal effects in adults in a vast majority of the population. I
don't think that a claim can be made that no single individual will be completely immune from
the development of severe fluorosis even at this recommended RfD. Based on possible a possible
genetic predisposition, I don't think that the possibility can be totally ruled out. As always my
concern is not so much the amount of fluoride that is ingested through the public water supply
but the other well-known sources of fluoride that have an additive effect to that derived from
consuming fluoridated drinking water. Dentifrice/fluoridated mouthrinse ingestion by young
children, bottled waters which are not tested for their fluoride content, possible inappropriate use
of fluoride supplements, ingestion of foods especially imported from other countries are in my
mind more likely to be the possible causative factors if severe dental fluorosis occurs. Because
the literature contains so few cases of severe skeletal fluorosis occurring in the United States I
April 2009 Page B-18
-------
am more confident that the RfD of 0.7 mg/kg/day is well within an acceptable safety margin for
minimizing the possible development of skeletal fluorosis.
References
Hong L, Levy S, Warren J, Bergus G, Dawson D, Wefel J, Broffitt B. Primary tooth fluorosis
and amoxicillin use during infancy. Journal of Public Health Dentistry 64: 38-44 (2004).
Vieira A, Hanocock R, Eggertsson H, Everett E, Grynpas M. Tooth quality in dental fluorosis:
genetic and environmental factors. Calcified Tissue International 76:17-25 (2005).
Yan D, Gurumurthy A, Wright M, Pfeiler T, Loboa E, Everett E. Genetic background influences
fluoride's effects on osteoclstogenesis. Bone 41:1036-1044 (2007).
Jackson R, Kelly S, Katz B, Hull J, Stookey G. Dental fluorosis and caries prevalence in children
residing in communities with different levels of fluoride in the water. Journal of Public Health
Dentistry 88: 79-84 (1995).
A systematic review of public water fluoridation
NHS Centre for Review and Dissemination
University of York 2000
EXECUTIVE SUMMARY
This systematic review has been commissioned by the Chief Medical Officer of the
Department of Health to 'carry out an up to date expert scientific review of fluoride and health'
(Paragraph 9.20, Our Healthier Nation).
Overall, the aim has been to assess the evidence on the positive and negative effects of
population wide drinking water fluoridation strategies to prevent caries. To achieve this aim
five objectives were identified:
Objective 1: What are the effects of fluoridation of drinking water supplies on the incidence of
caries?
Objective 2: If water fluoridation is shown to have beneficial effects, what is the effect over
and above that offered by the use of alternative interventions and strategies?
Objective 3: Does water fluoridation result in a reduction of caries across social groups and
between geographical locations, bringing equity?
Objective 4: Does water fluoridation have negative effects?
Objective 5: Are there differences in the effects of natural and artificial water fluoridation?
Methods
A search of 25 electronic databases (with no language restrictions) and the wo rid-wide-web
was undertaken. Relevant journals and indices were hand searched and attempts were
made to contact authors for further information.
Quality inclusion criteria were based on a pre-defined hierarchy of evidence (A, B, and C).
Studies of efficacy were included if they were of evidence level A or B. In order to allow the
broadest search for evidence on potential adverse effects, studies of all levels of evidence
were included. Objective specific inclusion criteria, based on selection of participants,
intervention, outcomes assessed, and study design appropriate for a given objective were
then applied. Study validity was formally assessed using a published checklist modified for
this review (CRD Report 4, 1996).
Inclusion criteria were assessed independently by at least two reviewers. Extraction of data
from, and validity assessment of, included studies was independently performed by two
reviewers, and checked by a third reviewer. Disagreements were resolved through
consensus.
Where the data were in a suitable format, measures of effect and 95% confidence intervals
April 2009 Page B-19
-------
(Cl) were plotted. Heterogeneity was investigated by visual examination and statistically
using the Q-statistic. Where no evidence of heterogeneity was found a meta-analysis was
conducted to produce a pooled estimate of the measure of effect. Statistically significant
heterogeneity was investigated using meta-regression. Multiple regression analysis was used
to explore the relationship between fluoridation and fluorosis.
Results
214 studies met full inclusion criteria for one or more of the objectives. No randomised
controlled trials of the effects of water fluoridation were found. The study designs used
included 45 'before and after' studies, 102 cross-sectional studies, 47 ecological studies, 13
cohort (prospective or retrospective) studies and 7 case-control studies. Several studies were
reported in multiple papers over a number of years.
Results by Objective
Objective 1
A total of 26 studies of the effect of water fluoridation on dental caries were found. For this
objective, the quality of studies found was moderate (no level A studies). A large number of
studies were excluded because they were cross-sectional studies and therefore did not meet
the inclusion criteria of being evidence level B or above. All but three of the studies included
were before-after studies, two included studies used prospective cohort designs, and one
used a retrospective cohort design. All before-after studies located by the search were
included. The most serious defect of these studies was the lack of appropriate analysis.
Many studies did not present an analysis at all, while others only did simple analyses without
attempting to control for potentially confounding factors. While some of these studies were
conducted in the 1940's and 50's, prior to the common use of such analyses, studies
conducted much later also failed to use methods that were commonplace at the time of the
study.
Another defect of many studies was the lack of any measure of variance for the estimates of
decay presented. While most studies that presented the proportion of caries-free children
contained sufficient data to calculate standard errors, this was not possible for the studies that
presented dmft/DMFT scores. Only four of the eight studies using these data provided
estimates of variance.
The best available evidence suggests that fluoridation of drinking water supplies does reduce
caries prevalence, both as measured by the proportion of children who are caries free and by
the mean change in dmft/DMFT score. The studies were of moderate quality (level B), but of
limited quantity. The degree to which caries is reduced, however, is not clear from the data
available. The range of the mean difference in the proportion (%) of caries-free children is -5.0
to 64%, with a median of 14.6% (interquartile range 5.05, 22.1%). The range of mean change
in dmft/DMFT score was from 0.5 to 4.4, median 2.25 teeth (interquartile range 1.28, 3.63
teeth). It is estimated that a median of six people need to receive fluoridated water for one
extra person to be caries-free (interquartile range of study NNTs 4, 9). The best available
evidence from studies following withdrawal of water fluoridation indicates that caries
prevalence increases, approaching the level of the low fluoride group. Again, however, the
studies were of moderate quality (level B), and limited quantity. The estimates of effect could
be biased due to poor adjustment for the effects of potential confounding factors.
Objective 2
To address this objective, studies conducted after 1974 were examined. While only nine
studies were included for Objective 2, these would have been enough to provide a confident
answer to the objective's question if the studies had been of sufficient quality. Since these
studies were completed after 1974, one might expect that the validity assessments would be
higher than the earlier studies following the introduction of more rigorous study methodology
and analytic techniques. However, the average validity checklist score and level of evidence
was essentially the same for studies after 1974 as those conducted prior to 1974. Hence, the
ability to answer this objective is similar to that in Objective 1.
In those studies completed after 1974, a beneficial effect of water fluoridation was still evident
in spite of the assumed exposure to non-water fluoride in the populations studied. The metaregression
conducted for Objective 1 confirmed this finding.
Objective 3
April 2009 Page B-20
-------
No level A or B studies examining the effect of water fluoridation on the inequalities of dental
health between social classes were identified. However, because of the importance of this
objective, level C studies conducted in England were included. A total of 15 studies
investigating the association of water fluoridation, dental caries and social class in England
were identified. The quality of the evidence of the studies was low, and the measures of
social class that were used varied. Variance data were not reported in most of these studies,
so a statistical analysis was not undertaken.
There appears to be some evidence that water fluoridation reduces the inequalities in dental
health across social classes in 5 and 12 year-olds, using the dmft/DMFT measure. This effect
was not seen in the proportion of caries-free children among 5 year-olds. The data for the
effects in children of other ages did not show an effect. The small quantity of studies,
differences between these studies, and their low quality rating, suggest caution in interpreting
these results.
Objective 4
DENTAL FLUOROSIS
Dental fluorosis was the most widely and frequently studied of all negative effects. The
fluorosis studies were largely cross-sectional designs, with only four before-after designs.
Although 88 studies of fluorosis were included, they were of low quality. The mean validity
score for fluorosis was only 2.8 out of 8. All, but one, of the studies were of evidence level C.
Observer bias may be of particular importance in studies assessing fluorosis. Efforts to
control for the effects of potential confounding factors, or reducing potential observer bias
were uncommon.
As there may be some debate about the significance of a fluorosis score at the lowest level of
each index being used to define a person as 'fluorosed', a second method of determining the
proportion 'fluorosed' was selected. This method describes the number of children having
dental fluorosis that may cause 'aesthetic concern'.
With both methods of identifying the prevalence of fluorosis, a significant dose-response
relationship was identified through a regression analysis. The prevalence of fluorosis at a
water fluoride level of 1.0 ppm was estimated to be 48% (95% Cl 40 to 57) and for fluorosis of
aesthetic concern it was predicted to be 12.5% (95% Cl 7.0 to 21.5). A very rough estimate
of the number of people who would have to be exposed to water fluoride levels of 1.0 ppm for
one additional person to develop fluorosis of any level is 6 (95% Cl 4 to 21), when compared
with a theoretical low fluoride level of 0.4 ppm. Of these approximately one quarter will have
fluorosis of aesthetic concern, but the precision of these rough estimates is low. These
estimates only apply to the comparison of 1.0 ppm to 0.4 ppm, and would be different if other
levels were compared.
BONE FRACTURE AND BONE DEVELOPMENT PROBLEMS
There were 29 studies included on the association between bone fracture and bone
development problems and water fluoridation. Other than fluorosis, bone effects (not
including bone cancers) were the most studied potential adverse effect. These studies had a
mean validity score of 3.4 out of 8. All but one study were of evidence level C. These studies
included both cohort and ecological designs, some of which included analyses controlling for
potential confounding factors. Observer bias could potentially play a role in bone fracture
studies, depending on how the study is conducted.
The evidence on bone fracture can be classified into hip fracture and other sites because
there are more studies on hip fracture than any other site. Using a qualitative method of
analysis (Figure 8.1), there is no clear association of hip fracture with water fluoridation. The
evidence on other fractures is similar. Overall, the findings of studies of bone fracture effects
showed small variations around the 'no effect' mark. A meta-regression of bone fracture
studies also found no association with water fluoridation.
CANCER STUDIES
There were 26 studies of the association of water fluoridation and cancer included. Eighteen
of these studies are from the lowest level of evidence (level C) with the highest risk of bias.
There is no clear association between water fluoridation and overall cancer incidence and
mortality. This was also true for osteosarcoma and bone/joint cancers. Only two studies
considered thyroid cancer and neither found a statistically significant association with water
April 2009 PageB-21
-------
fluoridation.
Overall, no clear association between water fluoridation and incidence or mortality of bone
cancers, thyroid cancer or all cancers was found.
OTHER POSSIBLE NEGATIVE EFFECTS
A total of 33 studies of the association of water fluoridation with other possible negative
effects were included in the review. Interpreting the results of studies of other possible
negative effects is very difficult because of the small numbers of studies that met inclusion
criteria on each specific outcome, and poor study quality. A major weakness of these studies
generally was failure to control for any confounding factors.
Overall, the studies examining other possible negative effects provide insufficient evidence on
any particular outcome to permit confident conclusions. Further research in these areas
needs to be of a much higher quality and should address and use appropriate methods to
control for confounding factors.
Objective 5:
The assessment of natural versus artificial water fluoridation effects is greatly limited due to
the lack of studies making this comparison. Very few studies included both natural and
artificially fluoridated areas, and direct comparisons were not possible for most outcomes. No
major differences were apparent in this review, however, the evidence is not adequate to
make a conclusion regarding this objective.
Conclusions
This review presents a summary of the best available and most reliable evidence on the
safety and efficacy of water fluoridation.
Given the level of interest surrounding the issue of public water fluoridation, it is surprising to
find that little high quality research has been undertaken. As such, this review should provide
both researchers and commissioners of research with an overview of the methodological
limitations of previous research conducted in this area.
The evidence of a benefit of a reduction in caries should be considered together with the
increased prevalence of dental fluorosis. The research evidence is of insufficient quality to
allow confident statements about other potential harms or whether there is an impact on
social inequalities. This evidence on benefits and harms needs to be considered along with
the ethical, environmental, ecological, costs and legal issues that surround any decisions
about water fluoridation. All of these issues fell outside the scope of this review.
Any future research into the safety and efficacy of water fluoridation should be carried out with
appropriate methodology to improve the quality of the existing evidence base.
April 2009 Page B-22
-------
Gary M. Whitford, D.M.D., Ph.D.
Regents' Professor, Department of Oral Biology and Maxillofacial Pathology, Medical
College of Georgia
1. Was the document clear and transparent in its presentation of data and explanation of the
analytical approaches used to characterize the concentration- response and dose-response
relationships for severe dental fluorosis? If not, do you have any suggestions that will enable the
authors to improve its clarity?
Answer:
Yes, the document was clear and transparent.
2. Are you aware of any significant publications related to severe dental fluorosis or the skeletal
effects of fluoride that are not included in the noncancer assessment document?
Answer:
Each of the following publications contains information pertinent to the reliability of the
water fluoride concentrations shown in Dean (1942), the report selected as the "critical study for
severe dental fluorosis."
1. Elvove E. Estimation of fluorides in waters. Pub Hlth Rep 48: 1219-1222, 1933.
2. Dean HT, Elvove E. Studies on the minimal threshold of the dental sign of chronic
endemic fluorosis (Mottled enamel). Pub Hlth Rep 50: 1719-1729, 1935.
3. Megregian S, Maier FJ. Modified zirconium alizarin reagent for determination of
fluoride in water. J Am Water Works Assn 44: 239-246, 1952.
I also recommend that a copy of the book entitled "Fluoride Drinking Waters" be
consulted. It is a compilation of many of the early papers, including the three cited above, that
dealt with several aspects of fluoride in water. The book is Public Health Service Publication No.
825. It was edited by Frank J. McClure and published in 1962.1 am willing to loan my copy if
the committee cannot locate the book.
3. Are the strengths and weaknesses of Dean (1942) study (selected as critical) fully
characterized?
Answer:
Not entirely. A major weakness of the Dean (1942) report is the chemical method used
for the determination of fluoride concentrations in water (Elvove E, 1933). The zirconium-
alizarin method is rarely, or probably never, used today because of its relative insensitivity,
several interfering substances and lack of specificity for fluoride. In their 1952 report that
described improvements to the method, Megregian and Maier (1952) noted that Elvove's
original method (1933) had several shortcomings including "non-conformity to the color laws,
limited effective fluoride range, and little color change per increment of fluoride." It also appears
that Elvove (1933) used the visual method to determine color changes in the zirconium-alizarin
reagent (since he referred to "Nessler tubes") which requires subjective judgments and is less
accurate that spectrophotometric methods.
Megregian and Maier (1952) also reported the effects of interfering substances on the
analytical results. Sulfate at 400 ppm in the water increased the fluoride result by 0.1 ppm as did
1.1 ppm hexametaphosphate. Chloride at 1800 ppm, bicarbonate at 400 ppm and iron at 5 ppm
decreased the fluoride result by 0.1 ppm. When the water fluoride concentration was 1.0 ppm,
1.0 ppm aluminum reduced the result by 0.39 ppm and 3.0 ppm aluminum reduced the result by
April 2009 Page B-23
-------
0.63 ppm. When the fluoride concentration was 2.0 ppm, aluminum at 1.0, 2.0 and 3.0 ppm
reduced the results by 0.47, 0.86 and 1.13 ppm, respectively.
The report by Elvove (1933) that described the chemical method used to produce the
water fluoride concentrations shown in Table I of Dean's 1942 report made no mention of
interfering substances. It can be reasonably assumed that most, if not all, of the substances listed
in the preceding paragraph were present in all the water samples analyzed but at unreported or
unknown concentrations. However, in his original publication Elvove (1933) shows the
concentrations of several ions in water samples obtained from 20 different sources. With the
exception of Amarillo, these sources were not those shown in the Dean (1942) report. Among the
interfering ions listed in the preceding paragraph, sulfate concentrations were more than 400 ppm
in two of the 20 water samples, bicarbonate concentrations were more than 400 ppm in five, and
aluminum concentrations were more than 1.0 ppm in five. At these concentrations, each of these
ions would have affected the apparent water fluoride results.
It may be possible to access 1930-1940 analytical records from the communities shown
in Table I (Dean, 1942) and make a judgment concerning the possible effects of the interfering
substances of the reported water fluoride concentrations.
Another indication of the problem with the accuracy of the Elvove method is found in the
footnote to Table 3-1 on page 24 of the EPA Dose-Response Analysis where it is said (quoting
Elvove who was the principal chemist) that "as little as 0.01 mg F/50cc, or 0.2 ppm F, could be
differentiated from the control by application of this technique." This appears to mean that
Elvove's method could differentiate between water without fluoride and water containing 0.2
ppm fluoride. The magnitude of the error at higher concentrations is not known to me.
The scatter in the analytical results seen in the 1933/34 monthly results for water in
Colorado Springs is of particular interest (see Table 4 in Dean and Elvove, 1935). The average of
the 12 results was 2.5 ppm but the range was 1.8 to 3.0 ppm despite the fact that the water came
from a single source. While some seasonal variation in water concentrations can expected, this
wide range (1.2 ppm) appears excessive. Further, the 12 monthly results from Monmouth ranged
from 1.6 to 1.9 ppm, those from Galesburg ranged from 1.8 to 2.0 ppm, and those from Pueblo
ranged from 0.3 to 0.7 ppm.
Dean (1942, page 25) listed two major requisites for quantitative evaluation of the dental
effects of ingesting water containing fluoride. One was "a population continuously exposed
throughout life to the variable under investigation (the communal water supply)." In the Dean
and Elvove (1935) paper, the first paragraph in the Discussion contains this sentence: "As a
result of checking the water histories as given by the child by a followup recheck with the
parents, only about 20 percent of the children in the age group studied and present in the school
on the day of the examination were found to have had an unbroken history of residence and
constant use of the city water supply." More details, such as how long and/or how frequently the
children were away from the home water supply, were not given.
The studies summarized by Dean (1942) were done during or around the time of the
Great Depression when large numbers of families were resettling in other locations in search of
employment and other necessities. I recommend that the original papers summarized in Dean
(1942) be examined to determine the extent to which the children met the requisite cited above
and that the information be included in the Dose-Response Analysis document. If such
information is not available, then the document should say so and discuss the implications for its
conclusions.
April 2009 Page B-24
-------
Finally, I note that the appropriateness of the LOAEL (2.2 ppm) and the calculated RfD
(0.07 mg F/kg bw/day) are based largely on the accuracy of the water fluoride concentrations
shown in Dean (1942) but also on several other variables that may have affected the outcomes of
the epidemiological studies. The preceding comments draw attention to several shortcomings of
the chemical method used and other limiting aspects of the studies summarized by Dean (1942).
The uncertainties associated with these factors should be discussed wherever appropriate and
certainly in the "Uncertainty factors" section.
4. Are some teeth more susceptible to sever fluorosis than others? If so, should the OW make
modifications to the age range identified as the period of concern?
Answer:
I am not aware of data showing that some teeth more susceptible to severe fluorosis than
others. I would guess that the posterior teeth may be more susceptible since their development is
more protracted than that of the anterior teeth.
5. Are the data on cavities as collected and graphically presented by the OW consistent with the
hypothesis that there is an increased risk for cavities in susceptible individuals with severe dental
fluorosis?
Answer:
The available data present an unclear relationship. Driscoll's work (1983, 1986) in
Illinois and Iowa did not indicate a relationship between severe fluorosis and caries (pages 46-
47). Eklund's report (1987) did not find a relationship when the entire dentition was considered.
They found more caries in severely fluorosed anterior teeth and premolars but not in the molars.
In their Chinese study, Chen et al (1989) found no difference in caries scores between the group
without fluorosis and the group with severe fluorosis. Warnakulasuriya et al (1992) reached a
similar conclusion in their Sri Lanka study but the validity of the conclusion was less clear
because of the way they grouped the fluorosis categories. On the other hand, Mann et al (1987,
1990) and Olsson (1979) found that DMFS scores were directly related to the severity of
fluorosis in Israel as did Wondwossen et al (2004) in Ethiopia. Ermis et al (2003) reported a
slightly higher prevalence of caries in moderate-to-severely fluorosed teeth but the relationship
was not statistically significant.
Overall and as summarized in Figure 3-7 on page 71, the relationship between the
severity of dental fluorosis and the risk of caries is suggestive but not convincing. I think this
subject requires more study with control for variables that are known risk factors for caries
before a reasonably firm conclusion can be drawn about the possibility of an association between
severe dental fluorosis and an increased risk of caries.
6. Are there recent data that would impact the IOM (1997) Adequate Intake value of 0.05
mg/kg/day for fluoride that the OW should consider in this assessment?
Answer:
The lOM's Adequate Intake (Al) represents the amount of intake of any substance
"needed to maintain a defined nutritional state or criterion of adequacy in essentially all members
of a specific healthy population." In the case of fluoride, the Al (see page 301, Dietary Reference
Intakes, 1997) "is based on estimated intakes that have been shown to reduce the occurrence of
dental caries maximally in a population without causing unwanted side effects including
moderate dental fluorosis. This does not mean that intakes somewhat higher than 0.05 mg/kg/day
April 2009 Page B-25
-------
increase risk of moderate dental fluorosis. In fact, the lOM's estimate for the threshold for that
risk is 0.10 mg/kg/day. I am not aware of data that would call for a change in the AI of 0.05
mg/kg F/day.
7. Can you suggest an approach to transform the water concentration data from the Dean (1942)
study to units of dose for the population susceptible to severe dental fluorosis other than that
used by the OW?
Answer: Not at this time.
8. Can you provide input concerning the strengths and weaknesses of the approach utilized by
the OW to identify a lower bound dose for severe dental fluorosis?
Answer:
I am looking at the Clovis, New Mexico, data in Table 3-1 (page 24). The water
fluoride concentration was listed as 2.2 ppm, the lowest concentration at which severe dental
fluorosis was recorded. The prevalence was 0.7 percent, ie, only one subject out of the 138
examined exhibited what was classified as having severe fluorosis. This concentration, 2.2 ppm,
was selected as the LOAEL. Unfortunately, nothing is known about the individual with severe
fluorosis including whether he/she was a permanent resident of Clovis or had lived in one or
more other communities before moving to Clovis. The prevalence of moderate dental fluorosis
(Dean score 3) in Clovis was 11.0 percent. These prevalence values for moderate and severe
fluorosis are markedly higher than those for Elmhurst and Galesburg (about 1.1 percent for
moderate and an absence of severe dental fluorosis) where the water fluoride concentrations
were listed as 1.8 and 1.9 ppm, respectively, just slightly lower than the concentration in Clovis.
In view of the small differences in the water fluoride concentrations between Clovis and
the other two communities, the large differences in fluorosis prevalence values suggest that
another factor may have influenced the appearance of the teeth in Clovis. Unlike Elmhurst and
Galesburg, Clovis is located at a relatively high altitude (4300 feet). As summarized elsewhere in
the EPA document under review (pages 39-41), there is evidence from laboratory animal studies
and epidemiological studies that residence at high altitude affects amelogenesis in a way that
resembles fluorosis and that its effects may be additive to the effects of fluoride exposure. This
too adds uncertainty regarding the selection of 2.2 ppm fluoride (Clovis) as the LOAEL for
severe dental fluorosis and the appropriateness of the RfD.
A similar (but weaker) argument can be made for Colorado Springs where the average
water fluoride concentration is listed as 2.6 ppm (but with a wide range, see item 3 above). This
city is located at an altitude of 6,035 feet.
In addition to these comments and based on the variability among published studies
regarding the relationship between estimated fluoride intakes and the risk of severe dental
fluorosis, I think consideration should be given to establishing a LED other than a LBD-1%.
9. Are you aware of dose estimates other than those from IOM (1997) and WHO (2002) that are
appropriate critical doses for skeletal effects and/or can you suggest a different approach that the
OW might use to estimate the fluoride dose associated with skeletal fractures using available
data?
Answer:
As in the case of dental fluorosis, the critical doses for skeletal effects will be difficult to
establish with a reasonable degree of certainty. The lOM's estimate (page 307) of fluoride
April 2009 Page B-26
-------
exposures that may result in clinical signs of the "milder forms" of skeletal fluorosis (preclinical
and perhaps stage I) is 10 mg/day for 10 or more years. There are published exceptions
suggesting that higher exposure levels and durations are required as noted on the same page of
the IOM report. Further, recent case reports of "tea fluorosis" in the U.S. suggest that, at least for
some individuals, a much higher chronic intake is tolerated without progression to stage II
skeletal fluorosis (Whyte et al, Am J Med 118: 78-82, 2005; Whyte et al, J Bone Min Res, in
press). In the former report the intake was estimated at 37-74 mg F/day from tea throughout the
patient's adult life. The intake in the latter report was estimated at more than 40 mg F/day
throughout the patient's adult life. Both patients showed marked osteosclerosis but without
ligamentous calcifications which was consistent with stage I skeletal fluorosis and neither patient
experienced fractures. Hallanger-Johnson et al (Mayo Clin Proc 82: 719-724, 2007) reported four
cases with axial osteosclerosis with elevated serum fluoride levels due to chronic consumption of
large amounts of tea.
In addition to the several variables that can affect the quality and quantity of the skeleton
cited in the present document, it is of interest that much of the data relating bone fluoride
concentrations to the stages of skeletal fluorosis comes from studies of workers in aluminum
processing factories. High, chronic exposures to aluminum lead to skeletal changes that share
some features in common with skeletal fluorosis which makes it difficult to attribute the skeletal
changes only to fluoride. This subject is worthy of further exploration.
10. Are you aware of any data that can be used to demonstrate that protection of the secondary
teeth from severe dental fluorosis will also protect the primary teeth?
Answer:
There are a few reports that suggest that dental fluorosis in the primary teeth may
correlate with the condition in secondary teeth. I am not aware of reports suggesting that
protection of the secondary teeth from severe dental fluorosis will also protect the primary teeth.
11. Are there factors other than those discussed in Section 3.1.4 that increase sensitivity to
fluorosis of the teeth or bones?
Answer:
As I recall, there are suggestions that African-Americans are more susceptible to dental
fluorosis but less susceptible to skeletal fluorosis. Published data indicate that there are
differences among strains of mice regarding susceptibility to dental fluorosis (Everett et al. J
Dent Res 81: 794-698, 2002) and mechanical properties of bone (Mousny et al. Bone 39: 1283-
1289, 2006). It is assumed that the differences are due to genetic differences among the strains.
NIDCR has requested applications for further pharmacogenetic studies.
12. Do you support the OW's conclusion that an RfD of 0.07 mg/kg/day will be protective for
severe dental fluorosis in children and skeletal effects in adults while still providing for the
beneficial effects of fluoride?
Answer:
Yes. For reasons listed above, however, there is a considerable degree of uncertainty
regarding the appropriateness of the estimate. As estimated by the IOM (1997), the RfD may be
closer to 0.10 mg/kg/day for moderate (not severe) dental fluorosis and substantially higher than
that for clinically significant skeletal effects in the United States.
April 2009 Page B-27
------- |